Merge branch 'develop' into fix/json_pointer_create_object_5357

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

tests/CMakeLists.txt

@@ -2,6 +2,9 @@ cmake_minimum_required(VERSION 3.13...4.0)
 
 option(JSON_Valgrind    "Execute test suite with Valgrind." OFF)
 option(JSON_FastTests   "Skip expensive/slow tests." OFF)
+option(JSON_TestSimdutf "Build the unit tests against the simdutf UTF-8 validation backend." OFF)
+
+set(JSON_SIMDUTF_VERSION 9.1.0 CACHE STRING "The simdutf version used by JSON_TestSimdutf.")
 
 set(JSON_32bitTest     AUTO CACHE STRING "Enable the 32bit unit test (ON/OFF/AUTO/ONLY).")
 set(JSON_TestStandards "" CACHE STRING "The list of standards to test explicitly.")
@@ -72,7 +75,15 @@ target_compile_options(test_main PUBLIC
     #                              is annotated JSON_HEDLEY_NO_RETURN (it always throws), which
     #                              makes MSVC flag the code following its call in binary_reader.hpp
     #                              as unreachable for that instantiation, in both Debug and Release
-    $<$<CXX_COMPILER_ID:MSVC>:/W4;/wd4566;/wd4996;/wd4702>
+    #       Disable warning C4503: decorated name length exceeded, name was truncated; the deep
+    #                              copy support added for #5387 pushes the mangled name of
+    #                              std::allocator_traits<...>::construct for the custom-base-class
+    #                              test's map type past VS2015's limit. The name is only used for
+    #                              debug info, so truncation does not affect the build.
+    #       Disable warning C5285: cannot declare a specialization for 'std::tuple'; MSVC 19.51
+    #                              reports the forward declarations of standard library
+    #                              templates in the vendored doctest.h
+    $<$<CXX_COMPILER_ID:MSVC>:/W4;/wd4566;/wd4996;/wd4702;/wd4503;/wd5285>
     # https://github.com/nlohmann/json/issues/1114
     $<$<CXX_COMPILER_ID:MSVC>:/bigobj> $<$<BOOL:${MINGW}>:-Wa,-mbig-obj>
 
@@ -101,7 +112,7 @@ endif()
 
 if (CMAKE_CXX_COMPILER_ID STREQUAL "MSVC")
     # avoid stack overflow, see https://github.com/nlohmann/json/issues/2955
-    json_test_set_test_options("test-cbor;test-msgpack;test-ubjson;test-bjdata;test-binary_formats" LINK_OPTIONS /STACK:4000000)
+    json_test_set_test_options("test-bon8;test-cbor;test-msgpack;test-ubjson;test-bjdata;test-binary_formats" LINK_OPTIONS /STACK:4000000)
 endif()
 
 # disable exceptions for test-disabled_exceptions
@@ -125,6 +136,51 @@ json_test_set_test_options(test-unicode4 TEST_PROPERTIES TIMEOUT 3000)
 # add unit tests
 #############################################################################
 
+# Generate the leak checks for every JSON_HEDLEY_* macro defined in
+# hedley.hpp; tests/src/unit-no-macro-leak.cpp #include-s the result after
+# nlohmann/json.hpp (see issue #5408). Using the shared
+# cmake/scripts/gen_hedley_undef_check.cmake script (also used by `make
+# update_hedley_undef`) instead of a hand-maintained list of macro names
+# means this test can never go stale after a future `make update_hedley`.
+set(hedley_hpp "${PROJECT_SOURCE_DIR}/include/nlohmann/thirdparty/hedley/hedley.hpp")
+set(hedley_undef_check_script "${PROJECT_SOURCE_DIR}/cmake/scripts/gen_hedley_undef_check.cmake")
+set(hedley_undef_checks "${PROJECT_BINARY_DIR}/include/hedley_undef_checks.inc")
+
+# Reconfigure whenever the vendored header or the generator script changes,
+# so a `cmake --build` after `make update_hedley` does not silently keep a
+# stale generated file around.
+set_property(DIRECTORY APPEND PROPERTY CMAKE_CONFIGURE_DEPENDS
+    "${hedley_hpp}"
+    "${hedley_undef_check_script}")
+
+# Generate once at configure time, so the very first build (before any
+# custom-command build step has run) already has an up-to-date file.
+execute_process(
+    COMMAND ${CMAKE_COMMAND}
+        "-DHEDLEY_HPP=${hedley_hpp}"
+        "-DOUTPUT=${hedley_undef_checks}"
+        -DMODE=checks
+        -P "${hedley_undef_check_script}"
+    RESULT_VARIABLE hedley_undef_check_result
+)
+if(NOT hedley_undef_check_result EQUAL 0)
+    message(FATAL_ERROR "Failed to generate ${hedley_undef_checks}")
+endif()
+
+# Also (re)generate as a build step, so an incremental build after editing
+# hedley.hpp without a full reconfigure still picks up the change.
+add_custom_command(
+    OUTPUT "${hedley_undef_checks}"
+    COMMAND ${CMAKE_COMMAND}
+        "-DHEDLEY_HPP=${hedley_hpp}"
+        "-DOUTPUT=${hedley_undef_checks}"
+        -DMODE=checks
+        -P "${hedley_undef_check_script}"
+    DEPENDS "${hedley_hpp}" "${hedley_undef_check_script}"
+    COMMENT "Generating Hedley undef leak checks"
+    VERBATIM)
+add_custom_target(generate_hedley_undef_checks DEPENDS "${hedley_undef_checks}")
+
 if("${JSON_TestStandards}" STREQUAL "")
     set(test_cxx_standards 11 14 17 20 23)
     unset(test_force)
@@ -149,6 +205,71 @@ if(test_force)
 endif()
 message(STATUS "${msg}")
 
+#############################################################################
+# optionally validate UTF-8 with simdutf (JSON_USE_SIMDUTF)
+#############################################################################
+
+# The simdutf backend is opt-in and not vendored, so it is fetched here rather
+# than being a checked-in dependency. Everything below hangs off test_main,
+# whose usage requirements every test target inherits; the library target and
+# the installed CMake package are deliberately left untouched.
+if (JSON_TestSimdutf)
+    # simdutf requires C++17, both to compile itself and to be reachable from
+    # the library, which keeps its scalar validator below that. Find a tested
+    # standard that satisfies it.
+    set(simdutf_standard "")
+    foreach(cxx_standard ${test_cxx_standards})
+        if(NOT cxx_standard LESS 17 AND compiler_supports_cpp_${cxx_standard})
+            set(simdutf_standard ${cxx_standard})
+            break()
+        endif()
+    endforeach()
+
+    if("${simdutf_standard}" STREQUAL "")
+        # Building simdutf would fail outright without a C++17 compiler, and
+        # even with one it would go unused if no C++17-or-later standard is
+        # tested. Say so and fall back to the scalar validator rather than
+        # failing the build.
+        if(NOT compiler_supports_cpp_17)
+            set(simdutf_reason "the compiler does not support C++17")
+        else()
+            set(simdutf_reason "no tested standard is C++17 or later (testing ${msg_standards})")
+        endif()
+        message(WARNING
+            "JSON_TestSimdutf is enabled, but ${simdutf_reason}. simdutf requires C++17, so it "
+            "is not fetched and JSON_USE_SIMDUTF is not defined: the tests run against the "
+            "built-in scalar UTF-8 validator instead. Set JSON_TestStandards to include 17 or "
+            "later, or build with a compiler that supports C++17.")
+    else()
+        if (CMAKE_VERSION VERSION_LESS 3.18)
+            message(FATAL_ERROR "JSON_TestSimdutf requires CMake 3.18 or later (simdutf's minimum).")
+        endif()
+
+        include(FetchContent)
+
+        # simdutf builds its tests and tools by default, and its tests pull
+        # further dependencies of their own; only the library is needed here
+        set(SIMDUTF_TESTS OFF CACHE BOOL "" FORCE)
+        set(SIMDUTF_TOOLS OFF CACHE BOOL "" FORCE)
+        set(SIMDUTF_BENCHMARKS OFF CACHE BOOL "" FORCE)
+        set(SIMDUTF_ICONV OFF CACHE BOOL "" FORCE)
+
+        FetchContent_Declare(simdutf
+            URL https://github.com/simdutf/simdutf/archive/refs/tags/v${JSON_SIMDUTF_VERSION}.tar.gz
+            DOWNLOAD_EXTRACT_TIMESTAMP TRUE
+        )
+        FetchContent_MakeAvailable(simdutf)
+
+        target_compile_definitions(test_main PUBLIC JSON_USE_SIMDUTF)
+        target_link_libraries(test_main PUBLIC simdutf::simdutf)
+
+        # simdutf.h requires C++17; below that the library keeps its scalar
+        # validator, so any C++11/14 test targets exercise the fallback and the
+        # C++17-and-later ones exercise simdutf. Both must agree.
+        message(STATUS "UTF-8 validation delegated to simdutf ${JSON_SIMDUTF_VERSION} for C++17 and later (JSON_USE_SIMDUTF)")
+    endif()
+endif()
+
 # *DO* use json_test_set_test_options() above this line
 
 json_test_should_build_32bit_test(json_32bit_test json_32bit_test_only "${JSON_32bitTest}")
@@ -163,6 +284,14 @@ foreach(file ${files})
     json_test_add_test_for(${file} MAIN test_main CXX_STANDARDS ${test_cxx_standards} ${test_force})
 endforeach()
 
+# tests/src/unit-no-macro-leak.cpp #include-s the generated leak-check file,
+# so its test targets must be built after generate_hedley_undef_checks.
+foreach(cxx_standard ${test_cxx_standards})
+    if(TARGET test-no-macro-leak_cpp${cxx_standard})
+        add_dependencies(test-no-macro-leak_cpp${cxx_standard} generate_hedley_undef_checks)
+    endif()
+endforeach()
+
 if(json_32bit_test_only)
     # Skip all other tests in this file
     return()
@@ -177,6 +306,24 @@ json_test_add_test_for(src/unit-comparison.cpp
     MAIN test_main CXX_STANDARDS ${test_cxx_standards} ${test_force}
 )
 
+# test the parser again with JSON_DIAGNOSTIC_POSITIONS enabled
+json_test_set_test_options(test-class_parser_diagnostic_positions
+    COMPILE_DEFINITIONS JSON_DIAGNOSTIC_POSITIONS=1
+)
+json_test_add_test_for(src/unit-class_parser.cpp
+    NAME test-class_parser_diagnostic_positions
+    MAIN test_main CXX_STANDARDS ${test_cxx_standards} ${test_force}
+)
+
+# test diagnostic positions again without regular diagnostics (JSON pointer paths)
+json_test_set_test_options(test-diagnostic-positions_only
+    COMPILE_DEFINITIONS JSON_DIAGNOSTICS=0
+)
+json_test_add_test_for(src/unit-diagnostic-positions.cpp
+    NAME test-diagnostic-positions_only
+    MAIN test_main CXX_STANDARDS ${test_cxx_standards} ${test_force}
+)
+
 # *DO NOT* use json_test_set_test_options() below this line
 
 #############################################################################

tests/Makefile

@@ -10,7 +10,7 @@ CXXFLAGS += -std=c++11
 CPPFLAGS += -I ../single_include
 
 FUZZER_ENGINE = src/fuzzer-driver_afl.cpp
-FUZZERS = parse_afl_fuzzer parse_bson_fuzzer parse_cbor_fuzzer parse_msgpack_fuzzer parse_ubjson_fuzzer parse_bjdata_fuzzer
+FUZZERS = parse_afl_fuzzer parse_bson_fuzzer parse_cbor_fuzzer parse_msgpack_fuzzer parse_ubjson_fuzzer parse_bjdata_fuzzer parse_bon8_fuzzer
 fuzzers: $(FUZZERS)
 
 parse_afl_fuzzer:
@@ -30,3 +30,6 @@ parse_ubjson_fuzzer:
 
 parse_bjdata_fuzzer:
 	$(CXX) $(CXXFLAGS) $(CPPFLAGS) $(FUZZER_ENGINE) src/fuzzer-parse_bjdata.cpp -o $@
+
+parse_bon8_fuzzer:
+	$(CXX) $(CXXFLAGS) $(CPPFLAGS) $(FUZZER_ENGINE) src/fuzzer-parse_bon8.cpp -o $@

tests/abi/config/CMakeLists.txt

@@ -14,6 +14,20 @@ add_test(
     NAME test-abi_config_noversion
     COMMAND abi_config_noversion ${DOCTEST_TEST_FILTER})
 
+# test default and no version namespace with all ABI tags enabled, so the
+# expected tag order is checked regardless of the JSON_* CMake options
+foreach(test default noversion)
+    add_executable(abi_config_${test}_all_tags ${test}.cpp)
+    target_compile_definitions(abi_config_${test}_all_tags PRIVATE
+        JSON_DIAGNOSTICS=1
+        JSON_DIAGNOSTIC_POSITIONS=1
+        JSON_USE_LEGACY_DISCARDED_VALUE_COMPARISON=1)
+    target_link_libraries(abi_config_${test}_all_tags PRIVATE abi_compat_main)
+    add_test(
+        NAME test-abi_config_${test}_all_tags
+        COMMAND abi_config_${test}_all_tags ${DOCTEST_TEST_FILTER})
+endforeach()
+
 # test custom namespace
 add_executable(abi_config_custom custom.cpp)
 target_link_libraries(abi_config_custom PRIVATE abi_compat_main)

tests/abi/config/default.cpp

@@ -24,12 +24,24 @@ TEST_CASE("default namespace")
         expected += "_diag";
 #endif
 
+#if JSON_USE_LEGACY_DISCARDED_VALUE_COMPARISON
+        expected += "_ldvcmp";
+#endif
+
 #if JSON_DIAGNOSTIC_POSITIONS
         expected += "_dp";
 #endif
 
-#if JSON_USE_LEGACY_DISCARDED_VALUE_COMPARISON
-        expected += "_ldvcmp";
+#if JSON_BRACE_INIT_COPY_SEMANTICS
+        expected += "_bics";
+#endif
+
+#if JSON_PRECISE_STREAM_POSITION
+        expected += "_psp";
+#endif
+
+#if JSON_STRICT_NUL_HANDLING
+        expected += "_snul";
 #endif
 
         expected += "_v" STRINGIZE(NLOHMANN_JSON_VERSION_MAJOR);

tests/abi/config/noversion.cpp

@@ -25,12 +25,24 @@ TEST_CASE("default namespace without version component")
         expected += "_diag";
 #endif
 
+#if JSON_USE_LEGACY_DISCARDED_VALUE_COMPARISON
+        expected += "_ldvcmp";
+#endif
+
 #if JSON_DIAGNOSTIC_POSITIONS
         expected += "_dp";
 #endif
 
-#if JSON_USE_LEGACY_DISCARDED_VALUE_COMPARISON
-        expected += "_ldvcmp";
+#if JSON_BRACE_INIT_COPY_SEMANTICS
+        expected += "_bics";
+#endif
+
+#if JSON_PRECISE_STREAM_POSITION
+        expected += "_psp";
+#endif
+
+#if JSON_STRICT_NUL_HANDLING
+        expected += "_snul";
 #endif
 
         expected += "::basic_json";

tests/benchmarks/CMakeLists.txt

@@ -1,4 +1,4 @@
-cmake_minimum_required(VERSION 3.11...3.14)
+cmake_minimum_required(VERSION 3.14)
 project(JSON_Benchmarks LANGUAGES CXX)
 
 # set compiler flags
@@ -6,29 +6,35 @@ if((CMAKE_CXX_COMPILER_ID MATCHES GNU) OR (CMAKE_CXX_COMPILER_ID MATCHES Clang))
     set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -flto -DNDEBUG -O3")
 endif()
 
-# configure Google Benchmarks
+# configure Google Benchmark; a fixed release, so that results stay comparable
+set(JSON_GOOGLE_BENCHMARK_VERSION 1.9.5)
 include(FetchContent)
-FetchContent_Declare(
-    benchmark
-    GIT_REPOSITORY https://github.com/google/benchmark.git
-    GIT_TAG origin/main
-    GIT_SHALLOW TRUE
-)
 
-FetchContent_GetProperties(benchmark)
-if(NOT benchmark_POPULATED)
-    FetchContent_Populate(benchmark)
-    set(BENCHMARK_ENABLE_TESTING OFF CACHE INTERNAL "" FORCE)
-    add_subdirectory(${benchmark_SOURCE_DIR} ${benchmark_BINARY_DIR})
-endif()
+# only the library is needed; -Werror would break the pinned release as soon as
+# a newer compiler adds a warning
+set(BENCHMARK_ENABLE_TESTING OFF CACHE BOOL "" FORCE)
+set(BENCHMARK_ENABLE_INSTALL OFF CACHE BOOL "" FORCE)
+set(BENCHMARK_ENABLE_WERROR OFF CACHE BOOL "" FORCE)
+
+FetchContent_Declare(benchmark
+    URL https://github.com/google/benchmark/archive/refs/tags/v${JSON_GOOGLE_BENCHMARK_VERSION}.tar.gz
+    URL_HASH SHA256=9631341c82bac4a288bef951f8b26b41f69021794184ece969f8473977eaa340
+    DOWNLOAD_EXTRACT_TIMESTAMP TRUE
+)
+FetchContent_MakeAvailable(benchmark)
 
 # download test data
 set(CMAKE_MODULE_PATH ${CMAKE_CURRENT_SOURCE_DIR}/../../cmake ${CMAKE_MODULE_PATH})
 include(download_test_data)
 
+# the header to benchmark; point this at a directory holding another version's
+# nlohmann/json.hpp to compare versions (see README.md)
+set(JSON_BENCHMARK_INCLUDE_DIR "${CMAKE_CURRENT_SOURCE_DIR}/../../single_include" CACHE PATH
+    "directory containing the nlohmann/json.hpp to benchmark")
+
 # benchmark binary
 add_executable(json_benchmarks src/benchmarks.cpp)
 target_compile_features(json_benchmarks PRIVATE cxx_std_11)
 target_link_libraries(json_benchmarks benchmark ${CMAKE_THREAD_LIBS_INIT})
 add_dependencies(json_benchmarks download_test_data)
-target_include_directories(json_benchmarks PRIVATE ${CMAKE_SOURCE_DIR}/../../single_include ${CMAKE_BINARY_DIR}/include)
+target_include_directories(json_benchmarks PRIVATE ${JSON_BENCHMARK_INCLUDE_DIR} ${CMAKE_BINARY_DIR}/include)

tests/benchmarks/README.md

@@ -0,0 +1,130 @@
+# Benchmarks
+
+Micro-benchmarks for parsing, serialization and the binary formats, written with
+[Google Benchmark](https://github.com/google/benchmark). They are not run by CI; see
+[When to run them](#when-to-run-them).
+
+## What is measured
+
+| benchmark | what it does |
+|---|---|
+| `ParseFile`, `ParseString` | parse JSON from a file stream or a string |
+| `ParseIndented` | parse the large files re-indented by 4 spaces, for the lexer's whitespace handling |
+| `Dump` | serialize, compact (`-`) and indented (`4`) |
+| `ToCbor`, `BinaryToCbor` | write CBOR; `BinaryToCbor` writes binary values of growing size |
+| `FromMsgpack` | read MessagePack; unchanged over the years, so its numbers stay comparable across releases |
+| `FromBinaryBuffer`, `FromBinaryFile` | read CBOR, MessagePack, UBJSON, BJData and BSON from a buffer or a `FILE*` |
+| `FromBinaryShape` | read deeply nested, container-heavy and scalar-heavy documents in every binary format |
+| `FromCborChunkedString` | read CBOR strings split into indefinite-length chunks |
+
+The input files are those of [nativejson-benchmark](https://github.com/miloyip/nativejson-benchmark) (`canada`,
+`citm_catalog`, `twitter`), a large `jeopardy` file, and number-heavy files (`floats`, `signed_ints`, ...).
+`bytes_per_second` counts the bytes read or written: the JSON text when parsing, the output when serializing.
+
+## Requirements
+
+- CMake 3.14 or later, a C++11 compiler, and Ninja for the `make` target.
+- Network access on the first configure: CMake downloads Google Benchmark and the
+  [test data](https://github.com/nlohmann/json_test_data) into the build directory. To reuse a download of the test
+  data, pass `-DJSON_TestDataDirectory=<build directory>/test_files`.
+- Google Benchmark is pinned to a release (1.9.5), so that results from different days stay comparable. To update it,
+  change `JSON_GOOGLE_BENCHMARK_VERSION` and the archive's `URL_HASH` in `CMakeLists.txt` together.
+- The benchmarks include `single_include/nlohmann/json.hpp`, so run `make amalgamate` after changing anything in
+  `include/`.
+
+GCC and Clang builds use `-O3 -flto -DNDEBUG`.
+
+## Running them
+
+From the repository root, this builds everything from scratch in `cmake-build-benchmarks` and runs all benchmarks:
+
+```sh
+make run_benchmarks
+```
+
+To build once and run selectively:
+
+```sh
+cmake -S tests/benchmarks -B build-benchmarks -G Ninja -DCMAKE_BUILD_TYPE=Release
+cmake --build build-benchmarks
+build-benchmarks/json_benchmarks --benchmark_filter='ParseString|Dump'
+```
+
+Useful options of `json_benchmarks`:
+
+| option | effect |
+|---|---|
+| `--benchmark_list_tests` | list the benchmarks instead of running them |
+| `--benchmark_filter=<regex>` | run only the benchmarks whose names match |
+| `--benchmark_repetitions=<n>` | run every benchmark `n` times and add mean, median, standard deviation and coefficient of variation |
+| `--benchmark_enable_random_interleaving=true` | run the repetitions in random order, which spreads out drifts such as thermal throttling |
+| `--benchmark_min_time=<seconds>s` | run each benchmark at least this long (e.g. `2s`) |
+| `--benchmark_out=<file> --benchmark_out_format=json` | also write the results to a file, e.g. for `compare.py` |
+
+## Reading the output
+
+Each line shows the wall-clock `Time` and the `CPU` time per iteration, the number of `Iterations` Google Benchmark
+chose, and the throughput in `bytes_per_second`. With repetitions, the lines ending in `_median` are the ones to
+compare. A `_cv` (coefficient of variation) above a few percent means the machine was too noisy for small
+differences to mean anything.
+
+## Comparing two versions
+
+To see what a change or a release did, build the same benchmarks twice: once against the header of the version to
+compare with, and once against the current one. `JSON_BENCHMARK_INCLUDE_DIR` names the directory holding the
+`nlohmann/json.hpp` to benchmark. For example, to compare the current checkout with 3.12.0:
+
+```sh
+# the header of the version to compare with
+mkdir -p build-baseline-header/nlohmann
+git show v3.12.0:single_include/nlohmann/json.hpp > build-baseline-header/nlohmann/json.hpp
+
+# the same benchmarks, built against either header
+cmake -S tests/benchmarks -B build-baseline -G Ninja -DCMAKE_BUILD_TYPE=Release \
+      -DJSON_BENCHMARK_INCLUDE_DIR="$PWD/build-baseline-header"
+cmake -S tests/benchmarks -B build-current -G Ninja -DCMAKE_BUILD_TYPE=Release
+cmake --build build-baseline
+cmake --build build-current
+
+# run both, back to back
+build-baseline/json_benchmarks --benchmark_repetitions=10 --benchmark_enable_random_interleaving=true \
+                               --benchmark_out=build-baseline/results.json --benchmark_out_format=json
+build-current/json_benchmarks --benchmark_repetitions=10 --benchmark_enable_random_interleaving=true \
+                              --benchmark_out=build-current/results.json --benchmark_out_format=json
+```
+
+Google Benchmark ships a tool to compare the two result files. It needs NumPy and SciPy:
+
+```sh
+python3 -m venv build-venv
+build-venv/bin/pip install numpy scipy
+build-venv/bin/python build-current/_deps/benchmark-src/tools/compare.py -a benchmarks build-baseline/results.json build-current/results.json
+```
+
+The tool's own `tools/requirements.txt` pins NumPy and SciPy versions that need Python 3.11 or later; with an older
+Python, unpinned versions work as well. In its output:
+
+- the `Time` and `CPU` columns are relative changes: `-0.35` means 35% faster, `+0.10` means 10% slower;
+- `_pvalue` lines report a Mann-Whitney U test of whether the two versions differ. It needs at least 9
+  repetitions, and a p-value below 0.05 means the difference is unlikely to be noise;
+- `OVERALL_GEOMEAN` summarizes all benchmarks;
+- `-a` shows only the aggregates, not every repetition.
+
+The header you compare with must support everything the benchmarks use. The current benchmarks build against 3.12.0.
+Only benchmarks present in both result files are compared, so for older releases, either filter the benchmarks or
+build that release's own `tests/benchmarks` against its own header.
+
+## Getting stable numbers
+
+- Build and run both versions on the same machine, one right after the other.
+- Keep the machine otherwise idle: no builds, no browser, and a laptop plugged in.
+- On Linux, set the CPU frequency governor to `performance`, e.g. `sudo cpupower frequency-set --governor performance`.
+  Google Benchmark prints a warning when frequency scaling is enabled. Pinning the process to a core
+  (`taskset -c 2 ...`) helps as well.
+- Use 10 or more repetitions with random interleaving, compare medians, and treat changes within the `_cv` as noise.
+
+## When to run them
+
+They are a manual step, not part of CI: shared CI runners vary more between runs than most of the effects measured.
+Run the comparison above before a release, comparing the previous release tag with `develop`, and for pull requests
+that claim to change performance.

tests/benchmarks/src/benchmarks.cpp

@@ -81,6 +81,44 @@ BENCHMARK_CAPTURE(ParseString, signed_ints,       TEST_DATA_DIRECTORY "/regressi
 BENCHMARK_CAPTURE(ParseString, unsigned_ints,     TEST_DATA_DIRECTORY "/regression/unsigned_ints.json");
 BENCHMARK_CAPTURE(ParseString, small_signed_ints, TEST_DATA_DIRECTORY "/regression/small_signed_ints.json");
 
+//////////////////////////////////////////////////////////////////////////////
+// parse pretty-printed JSON from string
+//
+// Every file in the corpus above is minified or only lightly spaced, so none of
+// them exercise the lexer's whitespace handling. Real-world JSON is frequently
+// indented - configuration files, pretty-printed API responses, anything kept
+// under version control - where insignificant whitespace can outweigh the data.
+// Re-serializing a document with an indentation and parsing that keeps the
+// content identical to the ParseString row above, so the pair isolates the cost
+// of the whitespace alone.
+//////////////////////////////////////////////////////////////////////////////
+
+static void ParseIndented(benchmark::State& state, const char* filename, int indent)
+{
+    std::ifstream f(filename);
+    std::string str((std::istreambuf_iterator<char>(f)), std::istreambuf_iterator<char>());
+    const std::string indented = json::parse(str).dump(indent);
+
+    while (state.KeepRunning())
+    {
+        state.PauseTiming();
+        auto* j = new json();
+        state.ResumeTiming();
+
+        *j = json::parse(indented);
+
+        state.PauseTiming();
+        delete j;
+        state.ResumeTiming();
+    }
+
+    state.SetBytesProcessed(state.iterations() * indented.size());
+}
+BENCHMARK_CAPTURE(ParseIndented, jeopardy / 4,     TEST_DATA_DIRECTORY "/jeopardy/jeopardy.json",                 4);
+BENCHMARK_CAPTURE(ParseIndented, canada / 4,       TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json",       4);
+BENCHMARK_CAPTURE(ParseIndented, citm_catalog / 4, TEST_DATA_DIRECTORY "/nativejson-benchmark/citm_catalog.json", 4);
+BENCHMARK_CAPTURE(ParseIndented, twitter / 4,      TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json",      4);
+
 //////////////////////////////////////////////////////////////////////////////
 // serialize JSON
 //////////////////////////////////////////////////////////////////////////////
@@ -93,7 +131,8 @@ static void Dump(benchmark::State& state, const char* filename, int indent)
 
     while (state.KeepRunning())
     {
-        j.dump(indent);
+        std::string output = j.dump(indent);
+        benchmark::DoNotOptimize(output);
     }
 
     state.SetBytesProcessed(state.iterations() * j.dump(indent).size());
@@ -214,4 +253,339 @@ static void BinaryToCbor(benchmark::State& state)
 }
 BENCHMARK(BinaryToCbor)->RangeMultiplier(2)->Range(8, 8 << 12);
 
+//////////////////////////////////////////////////////////////////////////////
+// parse binary formats
+//////////////////////////////////////////////////////////////////////////////
+
+// Only MessagePack had a read benchmark (FromMsgpack above, left untouched so
+// its numbers stay comparable across releases). The benchmarks below cover the
+// other formats, and read from a contiguous buffer as well as from a FILE*:
+// most callers pass a container, and the two adapters compile to different
+// code. The test data repository ships JSON only, so the input for each is
+// derived at setup time by serializing a parsed test file.
+
+/// binary format to benchmark; the _optimized variants add UBJSON/BJData size
+/// and type annotations, which the readers handle in a separate code path
+enum class binary_format
+{
+    cbor,
+    msgpack,
+    ubjson,
+    ubjson_optimized,
+    bjdata,
+    bjdata_optimized,
+    bson,
+    bon8
+};
+
+static std::vector<std::uint8_t> to_binary(const json& j, const binary_format format)
+{
+    switch (format)
+    {
+        case binary_format::cbor:
+            return json::to_cbor(j);
+        case binary_format::msgpack:
+            return json::to_msgpack(j);
+        case binary_format::ubjson:
+            return json::to_ubjson(j);
+        case binary_format::ubjson_optimized:
+            return json::to_ubjson(j, true, true);
+        case binary_format::bjdata:
+            return json::to_bjdata(j);
+        case binary_format::bjdata_optimized:
+            return json::to_bjdata(j, true, true);
+        case binary_format::bon8:
+            return json::to_bon8(j);
+        case binary_format::bson:
+        default:
+            return json::to_bson(j);
+    }
+}
+
+static json from_binary(const std::vector<std::uint8_t>& bytes, const binary_format format)
+{
+    switch (format)
+    {
+        case binary_format::cbor:
+            return json::from_cbor(bytes);
+        case binary_format::msgpack:
+            return json::from_msgpack(bytes);
+        case binary_format::ubjson:
+        case binary_format::ubjson_optimized:
+            return json::from_ubjson(bytes);
+        case binary_format::bjdata:
+        case binary_format::bjdata_optimized:
+            return json::from_bjdata(bytes);
+        case binary_format::bon8:
+            return json::from_bon8(bytes);
+        case binary_format::bson:
+        default:
+            return json::from_bson(bytes);
+    }
+}
+
+static json from_binary(std::FILE* file, const binary_format format)
+{
+    switch (format)
+    {
+        case binary_format::cbor:
+            return json::from_cbor(file);
+        case binary_format::msgpack:
+            return json::from_msgpack(file);
+        case binary_format::ubjson:
+        case binary_format::ubjson_optimized:
+            return json::from_ubjson(file);
+        case binary_format::bjdata:
+        case binary_format::bjdata_optimized:
+            return json::from_bjdata(file);
+        case binary_format::bon8:
+            return json::from_bon8(file);
+        case binary_format::bson:
+        default:
+            return json::from_bson(file);
+    }
+}
+
+/*!
+@brief serialize a parsed test file to @a format
+
+Returns an empty vector and marks the benchmark as skipped if the file cannot
+be represented in the format, rather than letting the exception escape: BSON
+requires an object at the top level, and several test files are arrays.
+*/
+static std::vector<std::uint8_t> binary_input(benchmark::State& state, const char* filename, const binary_format format)
+{
+    std::ifstream f(filename);
+    std::string const str((std::istreambuf_iterator<char>(f)), std::istreambuf_iterator<char>());
+    const json j = json::parse(str);
+
+    if (format == binary_format::bson && !j.is_object())
+    {
+        state.SkipWithError("BSON requires an object at the top level");
+        return {};
+    }
+
+    return to_binary(j, format);
+}
+
+static void FromBinaryBuffer(benchmark::State& state, const char* filename, const binary_format format)
+{
+    const std::vector<std::uint8_t> bytes = binary_input(state, filename, format);
+    if (bytes.empty())
+    {
+        return;
+    }
+
+    for (auto _ : state)
+    {
+        // the value is destroyed outside the timed section, because destroying
+        // a large DOM is not what this benchmark measures
+        state.PauseTiming();
+        auto* j = new json();
+        state.ResumeTiming();
+
+        *j = from_binary(bytes, format);
+
+        state.PauseTiming();
+        delete j;
+        state.ResumeTiming();
+    }
+
+    state.SetBytesProcessed(state.iterations() * bytes.size());
+}
+
+BENCHMARK_CAPTURE(FromBinaryBuffer, cbor / jeopardy, TEST_DATA_DIRECTORY "/jeopardy/jeopardy.json", binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryBuffer, cbor / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryBuffer, cbor / citm_catalog, TEST_DATA_DIRECTORY "/nativejson-benchmark/citm_catalog.json", binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryBuffer, cbor / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryBuffer, cbor / floats, TEST_DATA_DIRECTORY "/regression/floats.json", binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryBuffer, cbor / signed_ints, TEST_DATA_DIRECTORY "/regression/signed_ints.json", binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryBuffer, msgpack / jeopardy, TEST_DATA_DIRECTORY "/jeopardy/jeopardy.json", binary_format::msgpack);
+BENCHMARK_CAPTURE(FromBinaryBuffer, msgpack / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::msgpack);
+BENCHMARK_CAPTURE(FromBinaryBuffer, msgpack / citm_catalog, TEST_DATA_DIRECTORY "/nativejson-benchmark/citm_catalog.json", binary_format::msgpack);
+BENCHMARK_CAPTURE(FromBinaryBuffer, msgpack / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::msgpack);
+BENCHMARK_CAPTURE(FromBinaryBuffer, ubjson / jeopardy, TEST_DATA_DIRECTORY "/jeopardy/jeopardy.json", binary_format::ubjson);
+BENCHMARK_CAPTURE(FromBinaryBuffer, ubjson / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::ubjson);
+BENCHMARK_CAPTURE(FromBinaryBuffer, ubjson / citm_catalog, TEST_DATA_DIRECTORY "/nativejson-benchmark/citm_catalog.json", binary_format::ubjson);
+BENCHMARK_CAPTURE(FromBinaryBuffer, ubjson / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::ubjson);
+BENCHMARK_CAPTURE(FromBinaryBuffer, ubjson_optimized / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::ubjson_optimized);
+BENCHMARK_CAPTURE(FromBinaryBuffer, ubjson_optimized / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::ubjson_optimized);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bjdata / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::bjdata);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bjdata / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::bjdata);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bjdata_optimized / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::bjdata_optimized);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bjdata_optimized / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::bjdata_optimized);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bon8 / jeopardy, TEST_DATA_DIRECTORY "/jeopardy/jeopardy.json", binary_format::bon8);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bon8 / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::bon8);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bon8 / citm_catalog, TEST_DATA_DIRECTORY "/nativejson-benchmark/citm_catalog.json", binary_format::bon8);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bon8 / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::bon8);
+// BSON requires an object at the top level, so the array-rooted test files
+// (jeopardy and the regression files) cannot be captured here
+BENCHMARK_CAPTURE(FromBinaryBuffer, bson / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::bson);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bson / citm_catalog, TEST_DATA_DIRECTORY "/nativejson-benchmark/citm_catalog.json", binary_format::bson);
+BENCHMARK_CAPTURE(FromBinaryBuffer, bson / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::bson);
+
+static void FromBinaryFile(benchmark::State& state, const char* filename, const binary_format format)
+{
+    const std::vector<std::uint8_t> bytes = binary_input(state, filename, format);
+    if (bytes.empty())
+    {
+        return;
+    }
+
+    const char* tmp = "benchmark_input.bin";
+    std::ofstream o(tmp, std::ios::binary);
+    o.write(reinterpret_cast<const char*>(bytes.data()), static_cast<std::streamsize>(bytes.size()));
+    o.flush();
+    o.close();
+
+    for (auto _ : state)
+    {
+        state.PauseTiming();
+        auto* j = new json();
+        auto* file = std::fopen(tmp, "rb");
+        state.ResumeTiming();
+
+        *j = from_binary(file, format);
+
+        state.PauseTiming();
+        std::fclose(file);
+        delete j;
+        state.ResumeTiming();
+    }
+
+    state.SetBytesProcessed(state.iterations() * bytes.size());
+}
+
+BENCHMARK_CAPTURE(FromBinaryFile, cbor / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryFile, cbor / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryFile, ubjson / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::ubjson);
+BENCHMARK_CAPTURE(FromBinaryFile, ubjson / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::ubjson);
+BENCHMARK_CAPTURE(FromBinaryFile, bjdata / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::bjdata);
+BENCHMARK_CAPTURE(FromBinaryFile, bon8 / canada, TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json", binary_format::bon8);
+BENCHMARK_CAPTURE(FromBinaryFile, bon8 / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::bon8);
+BENCHMARK_CAPTURE(FromBinaryFile, bson / twitter, TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json", binary_format::bson);
+
+//////////////////////////////////////////////////////////////////////////////
+// parse binary formats: value shapes
+//////////////////////////////////////////////////////////////////////////////
+
+// The test files above are wide and shallow, but the readers' cost is per
+// container, so these cover the shapes that stress the container handling
+// itself. Every shape is wrapped in an object so that BSON, which requires an
+// object at the top level, measures the same value as the other formats.
+
+/// deeply nested arrays: one container per level, no other work
+static json make_nested()
+{
+    json nested = json::array();
+    json* p = &nested;
+    for (std::size_t i = 1; i < 1000; ++i)
+    {
+        p->push_back(json::array());
+        p = &p->operator[](0);
+    }
+
+    json j = json::object();
+    j["data"] = std::move(nested);
+    return j;
+}
+
+/// many sibling containers: maximum container churn, minimum nesting
+static json make_containers()
+{
+    json data = json::array();
+    for (std::size_t i = 0; i < 100000; ++i)
+    {
+        data.push_back(json::array({1, 2}));
+    }
+
+    json j = json::object();
+    j["data"] = std::move(data);
+    return j;
+}
+
+/// one flat array of numbers: the scalar decoding path, which must not move
+static json make_scalars()
+{
+    json data = json::array();
+    for (std::size_t i = 0; i < 1000000; ++i)
+    {
+        data.push_back(i);
+    }
+
+    json j = json::object();
+    j["data"] = std::move(data);
+    return j;
+}
+
+static void FromBinaryShape(benchmark::State& state, json (*build)(), const binary_format format)
+{
+    const std::vector<std::uint8_t> bytes = to_binary(build(), format);
+
+    for (auto _ : state)
+    {
+        state.PauseTiming();
+        auto* j = new json();
+        state.ResumeTiming();
+
+        *j = from_binary(bytes, format);
+
+        state.PauseTiming();
+        delete j;
+        state.ResumeTiming();
+    }
+
+    state.SetBytesProcessed(state.iterations() * bytes.size());
+}
+
+BENCHMARK_CAPTURE(FromBinaryShape, nested / cbor, make_nested, binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryShape, nested / msgpack, make_nested, binary_format::msgpack);
+BENCHMARK_CAPTURE(FromBinaryShape, nested / ubjson, make_nested, binary_format::ubjson);
+BENCHMARK_CAPTURE(FromBinaryShape, nested / bjdata, make_nested, binary_format::bjdata);
+BENCHMARK_CAPTURE(FromBinaryShape, nested / bson, make_nested, binary_format::bson);
+BENCHMARK_CAPTURE(FromBinaryShape, nested / bon8, make_nested, binary_format::bon8);
+BENCHMARK_CAPTURE(FromBinaryShape, containers / cbor, make_containers, binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryShape, containers / msgpack, make_containers, binary_format::msgpack);
+BENCHMARK_CAPTURE(FromBinaryShape, containers / ubjson, make_containers, binary_format::ubjson);
+BENCHMARK_CAPTURE(FromBinaryShape, containers / ubjson_optimized, make_containers, binary_format::ubjson_optimized);
+BENCHMARK_CAPTURE(FromBinaryShape, containers / bjdata, make_containers, binary_format::bjdata);
+BENCHMARK_CAPTURE(FromBinaryShape, containers / bson, make_containers, binary_format::bson);
+BENCHMARK_CAPTURE(FromBinaryShape, containers / bon8, make_containers, binary_format::bon8);
+// BSON names every array element, so a large array measures key generation
+// rather than scalar decoding and is left out here
+BENCHMARK_CAPTURE(FromBinaryShape, scalars / cbor, make_scalars, binary_format::cbor);
+BENCHMARK_CAPTURE(FromBinaryShape, scalars / msgpack, make_scalars, binary_format::msgpack);
+BENCHMARK_CAPTURE(FromBinaryShape, scalars / ubjson, make_scalars, binary_format::ubjson);
+BENCHMARK_CAPTURE(FromBinaryShape, scalars / bjdata, make_scalars, binary_format::bjdata);
+BENCHMARK_CAPTURE(FromBinaryShape, scalars / bon8, make_scalars, binary_format::bon8);
+
+/*!
+@brief parse an indefinite-length CBOR string
+
+The writer never emits this form, so the input is assembled by hand: 0x7F
+opens the string, each chunk is a one-character string, and 0xFF closes it.
+*/
+static void FromCborChunkedString(benchmark::State& state, const std::size_t chunks)
+{
+    std::vector<std::uint8_t> bytes;
+    bytes.reserve(2 * chunks + 2);
+    bytes.push_back(0x7F);
+    for (std::size_t i = 0; i < chunks; ++i)
+    {
+        bytes.push_back(0x61); // string of length 1
+        bytes.push_back(0x61); // 'a'
+    }
+    bytes.push_back(0xFF);
+
+    for (auto _ : state)
+    {
+        json j = json::from_cbor(bytes);
+        benchmark::DoNotOptimize(j);
+    }
+
+    state.SetBytesProcessed(state.iterations() * bytes.size());
+}
+
+BENCHMARK_CAPTURE(FromCborChunkedString, 10000 chunks, 10000);
+
 BENCHMARK_MAIN();

tests/fuzzing.md

@@ -1,6 +1,6 @@
 # Fuzz testing
 
-Each parser of the library (JSON, BJData, BSON, CBOR, MessagePack, and UBJSON) can be fuzz tested. Currently,
+Each parser of the library (JSON, BJData, BON8, BSON, CBOR, MessagePack, and UBJSON) can be fuzz tested. Currently,
 [libFuzzer](https://llvm.org/docs/LibFuzzer.html) and [afl++](https://github.com/AFLplusplus/AFLplusplus) are supported.
 
 ## Corpus creation
@@ -10,11 +10,11 @@ directory with some simple input files that cover several features of the parser
 for mutations.
 
 ```shell
-TEST_DATA_VERSION=3.1.0
+TEST_DATA_VERSION=3.2.0
 wget https://github.com/nlohmann/json_test_data/archive/refs/tags/v$TEST_DATA_VERSION.zip
 unzip v$TEST_DATA_VERSION.zip
 rm v$TEST_DATA_VERSION.zip
-for FORMAT in json bjdata bson cbor msgpack ubjson
+for FORMAT in json bjdata bon8 bson cbor msgpack ubjson
 do
   rm -fr corpus_$FORMAT
   mkdir corpus_$FORMAT
@@ -79,3 +79,26 @@ the same `fuzzers` target as above and also relies on the `FUZZER_ENGINE` variab
 [build script](https://github.com/google/oss-fuzz/blob/master/projects/json/build.sh) for more information.
 
 In case the build at OSS-Fuzz fails, an issue will be created automatically.
+
+### Handling OSS-Fuzz reports
+
+OSS-Fuzz files the crashes it finds in its own [issue tracker](https://issues.oss-fuzz.com), not on GitHub. So that
+each report can be traced to the change that fixed it, and each fix to the report it answers, fixes follow these
+conventions:
+
+- **Reference the OSS-Fuzz issue in the pull request**, next to any GitHub issue it closes, as `OSS-Fuzz: <id>` (for
+  example, `OSS-Fuzz: 563659413`), and in the commit message. The ID alone does not disclose the crash. If the report
+  was triaged into a GitHub issue, link the OSS-Fuzz issue there too.
+- **Turn the reproducer into a unit test.** Download the testcase from the OSS-Fuzz report, reduce it if possible, and
+  add it as a regression test to the unit test of the affected format (e.g., `tests/src/unit-bjdata.cpp`), with a
+  comment naming the OSS-Fuzz issue. This way the input is checked by every CI run rather than only by OSS-Fuzz, and
+  it stays covered even if OSS-Fuzz later closes the report as not reproducible.
+- **Keep the fuzzer drivers and the unit tests in sync.** The round-trip checks of the UBJSON and BJData drivers are
+  also run on a fixed corpus in the unit tests (see `tests/src/round_trip_corpus.hpp` and the "round-trip invariants"
+  test cases), so a regression shows up in CI first. When a driver's checks change, change the unit tests with them.
+- **Record in the report whether the bug shipped.** OSS-Fuzz asks whether a crash was a short-lived regression or
+  affects a released version; answer it when the fix is merged, as it decides whether the fix needs a release note or
+  a security advisory (see the [security policy](../.github/SECURITY.md)).
+
+After the fix is merged, OSS-Fuzz re-runs the reproducer on its next build and marks the report as verified and
+closed. If it does not, the fix is incomplete.

tests/src/fuzzer-parse_bjdata.cpp

@@ -21,16 +21,53 @@ array data, it performs the following steps:
 - j4 = from_bjdata(vec3)
 - assert(j1 == j4)
 
+Re-serializing j2/j3/j4 with the same use_size/use_type settings is checked
+for value-stability rather than byte-exact stability: from_bjdata(to_bjdata(j2))
+must equal j2 (and likewise for j3, j4). Byte-exact stability does not hold in
+general, because a BJData value can lose type fidelity across a round trip
+(e.g. a binary_t value serialized without the optimized "$U#" array header is
+parsed back as a plain array of numbers, see #5398 and the discussion on
+PR #5494) - the numeric value is preserved, but the writer's smallest-type
+selection for the now-plain numbers may legitimately pick a different, but
+equally valid, single-byte type marker than the dedicated binary-data writer
+would have. Both encodings are valid BJData and both decode to the same
+value, so this is not treated as a round-trip failure here.
+
+"Value-stable" is checked by comparing dump()s rather than with operator==
+directly: a BJData/UBJSON payload can decode to a non-finite double (NaN or
++-Infinity), and IEEE 754 NaN is never equal to itself, so operator== would
+report two structurally-identical trees as different whenever a NaN is
+involved -- not a round-trip bug, just NaN's ordinary (non-)reflexivity.
+dump() serializes any non-finite double the same deterministic way (as JSON
+`null`, since JSON itself cannot represent NaN/Infinity), so comparing
+dumps is stable under exactly the same values that break operator==.
+
+The unit tests run the same checks on a fixed corpus (see the "BJData round-trip
+invariants" test case), so keep both in sync.
+
 The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer
 drivers.
 */
 
+#include <cassert>
 #include <iostream>
 #include <sstream>
 #include <nlohmann/json.hpp>
 
+// the round-trip checks below are assertions; NDEBUG would compile them away
+#ifdef NDEBUG
+    #error "the fuzzer drivers must be built without NDEBUG"
+#endif
+
 using json = nlohmann::json;
 
+// value-stable comparison for the round-trip checks below; see the note
+// above on why this compares dump()s rather than the json values directly
+static bool is_value_stable(const json& lhs, const json& rhs)
+{
+    return lhs.dump() == rhs.dump();
+}
+
 // see http://llvm.org/docs/LibFuzzer.html
 extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size)
 {
@@ -56,10 +93,12 @@ extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size)
             json const j3 = json::from_bjdata(vec3);
             json const j4 = json::from_bjdata(vec4);
 
-            // serializations must match
-            assert(json::to_bjdata(j2, false, false) == vec2);
-            assert(json::to_bjdata(j3, true, false) == vec3);
-            assert(json::to_bjdata(j4, true, true) == vec4);
+            // re-serializing must be value-stable (see the notes above on
+            // why byte-exact stability is not guaranteed in general, and
+            // why this compares dump()s rather than the values directly)
+            assert(is_value_stable(json::from_bjdata(json::to_bjdata(j2, false, false)), j2));
+            assert(is_value_stable(json::from_bjdata(json::to_bjdata(j3, true, false)), j3));
+            assert(is_value_stable(json::from_bjdata(json::to_bjdata(j4, true, true)), j4));
         }
         catch (const json::parse_error&)
         {

tests/src/fuzzer-parse_bon8.cpp

@@ -0,0 +1,103 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+/*
+This file implements a parser test suitable for fuzz testing. Given a byte
+array data, it performs the following steps:
+
+- j1 = from_bon8(data)
+- vec = to_bon8(j1)
+- j2 = from_bon8(vec)
+- assert(j1 == j2)
+
+It also checks that reading the data from a stream, which reads strings byte by
+byte, gives the same value or error as reading it from contiguous memory, which
+copies strings in bulk.
+
+The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer
+drivers.
+*/
+
+#include <cassert>
+#include <iostream>
+#include <sstream>
+#include <nlohmann/json.hpp>
+
+// the round-trip checks below are assertions; NDEBUG would compile them away
+#ifdef NDEBUG
+    #error "the fuzzer drivers must be built without NDEBUG"
+#endif
+
+using json = nlohmann::json;
+
+namespace
+{
+// the serialization of the value read from @a input, or the error message
+template<typename InputType>
+std::string read_bon8(InputType&& input)
+{
+    try
+    {
+        const auto vec = json::to_bon8(json::from_bon8(std::forward<InputType>(input)));
+        return {vec.begin(), vec.end()};
+    }
+    catch (const json::exception& e)
+    {
+        return e.what();
+    }
+}
+} // namespace
+
+// see http://llvm.org/docs/LibFuzzer.html
+extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size)
+{
+    // contiguous and stream input must be read alike
+    {
+        std::istringstream stream(std::string(reinterpret_cast<const char*>(data), size));
+        assert(read_bon8(std::vector<uint8_t>(data, data + size)) == read_bon8(stream));
+    }
+
+    try
+    {
+        // step 1: parse input
+        std::vector<uint8_t> const vec1(data, data + size);
+        json const j1 = json::from_bon8(vec1);
+
+        try
+        {
+            // step 2: round trip
+            std::vector<uint8_t> const vec2 = json::to_bon8(j1);
+
+            // parse serialization
+            json const j2 = json::from_bon8(vec2);
+
+            // serializations must match
+            assert(json::to_bon8(j2) == vec2);
+        }
+        catch (const json::parse_error&)
+        {
+            // parsing a BON8 serialization must not fail
+            assert(false);
+        }
+    }
+    catch (const json::parse_error&)
+    {
+        // parse errors are ok, because input may be random bytes
+    }
+    catch (const json::type_error&)
+    {
+        // type errors can occur during parsing, too
+    }
+    catch (const json::out_of_range&)
+    {
+        // out of range errors may happen if provided sizes are excessive
+    }
+
+    // return 0 - non-zero return values are reserved for future use
+    return 0;
+}

tests/src/fuzzer-parse_bson.cpp

@@ -19,10 +19,16 @@ The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer
 drivers.
 */
 
+#include <cassert>
 #include <iostream>
 #include <sstream>
 #include <nlohmann/json.hpp>
 
+// the round-trip checks below are assertions; NDEBUG would compile them away
+#ifdef NDEBUG
+    #error "the fuzzer drivers must be built without NDEBUG"
+#endif
+
 using json = nlohmann::json;
 
 // see http://llvm.org/docs/LibFuzzer.html

tests/src/fuzzer-parse_cbor.cpp

@@ -19,10 +19,16 @@ The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer
 drivers.
 */
 
+#include <cassert>
 #include <iostream>
 #include <sstream>
 #include <nlohmann/json.hpp>
 
+// the round-trip checks below are assertions; NDEBUG would compile them away
+#ifdef NDEBUG
+    #error "the fuzzer drivers must be built without NDEBUG"
+#endif
+
 using json = nlohmann::json;
 
 // see http://llvm.org/docs/LibFuzzer.html

tests/src/fuzzer-parse_json.cpp

@@ -20,10 +20,16 @@ The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer
 drivers.
 */
 
+#include <cassert>
 #include <iostream>
 #include <sstream>
 #include <nlohmann/json.hpp>
 
+// the round-trip checks below are assertions; NDEBUG would compile them away
+#ifdef NDEBUG
+    #error "the fuzzer drivers must be built without NDEBUG"
+#endif
+
 using json = nlohmann::json;
 
 // see http://llvm.org/docs/LibFuzzer.html

tests/src/fuzzer-parse_msgpack.cpp

@@ -19,10 +19,16 @@ The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer
 drivers.
 */
 
+#include <cassert>
 #include <iostream>
 #include <sstream>
 #include <nlohmann/json.hpp>
 
+// the round-trip checks below are assertions; NDEBUG would compile them away
+#ifdef NDEBUG
+    #error "the fuzzer drivers must be built without NDEBUG"
+#endif
+
 using json = nlohmann::json;
 
 // see http://llvm.org/docs/LibFuzzer.html

tests/src/fuzzer-parse_ubjson.cpp

@@ -21,14 +21,23 @@ array data, it performs the following steps:
 - j4 = from_ubjson(vec3)
 - assert(j1 == j4)
 
+The unit tests run the same checks on a fixed corpus (see the "UBJSON round-trip
+invariants" test case), so keep both in sync.
+
 The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer
 drivers.
 */
 
+#include <cassert>
 #include <iostream>
 #include <sstream>
 #include <nlohmann/json.hpp>
 
+// the round-trip checks below are assertions; NDEBUG would compile them away
+#ifdef NDEBUG
+    #error "the fuzzer drivers must be built without NDEBUG"
+#endif
+
 using json = nlohmann::json;
 
 // see http://llvm.org/docs/LibFuzzer.html

tests/src/round_trip_corpus.hpp

@@ -0,0 +1,213 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+#pragma once
+
+#include <cmath> // nan
+#include <cstddef> // size_t
+#include <cstdint> // int32_t, int64_t, uint32_t, uint64_t
+#include <limits> // numeric_limits
+#include <random> // mt19937
+#include <string> // string, to_string
+#include <utility> // move
+#include <vector> // vector
+
+#include <nlohmann/json.hpp>
+
+// Values for the round-trip property tests of the UBJSON and BJData writers.
+//
+// The fuzzer drivers (tests/src/fuzzer-parse_ubjson.cpp and
+// fuzzer-parse_bjdata.cpp) check that anything the library parses can be
+// serialized, parsed back, and serialized again without loss. Those checks
+// only run at OSS-Fuzz, so a regression used to surface days later as an
+// external report. The unit tests run the same checks on this corpus in CI.
+//
+// The corpus is deterministic: std::mt19937's output sequence is fixed by
+// the standard, and it is used directly rather than through a distribution
+// (whose results are implementation-defined).
+namespace utils
+{
+
+class round_trip_corpus
+{
+  public:
+    using json = nlohmann::json;
+
+    static std::vector<json> values()
+    {
+        round_trip_corpus corpus;
+        return corpus.build();
+    }
+
+    // whether a value contains a binary value, which a BJData or UBJSON round
+    // trip may turn into an array of integers
+    static bool contains_binary(const json& j)
+    {
+        if (j.is_binary())
+        {
+            return true;
+        }
+        if (j.is_structured())
+        {
+            for (const auto& element : j)
+            {
+                if (contains_binary(element))
+                {
+                    return true;
+                }
+            }
+        }
+        return false;
+    }
+
+  private:
+    std::vector<json> atoms;
+    // a fixed seed is the point: the corpus must be the same in every run
+    std::mt19937 generator{42}; // NOLINT(cert-msc32-c,cert-msc51-cpp,bugprone-random-generator-seed)
+
+    round_trip_corpus()
+        : atoms
+    {
+        nullptr, true, false,
+        // integers at the boundaries of every UBJSON/BJData integer type
+        0, 1, -1, 127, 128, 255, 256, -128, -129,
+        32767, 32768, 65535, 65536, -32768, -32769,
+        (std::numeric_limits<std::int32_t>::min)(), (std::numeric_limits<std::int32_t>::max)(),
+        (std::numeric_limits<std::uint32_t>::max)(),
+        (std::numeric_limits<std::int64_t>::min)(), (std::numeric_limits<std::int64_t>::max)(),
+        static_cast<std::uint64_t>((std::numeric_limits<std::int64_t>::max)()) + 1u,
+        (std::numeric_limits<std::uint64_t>::max)(),
+        // floating-point numbers, including non-finite ones
+        0.0, -0.0, 1.5, -2.25, 3.4e38, (std::numeric_limits<double>::max)(),
+        std::nan(""), std::numeric_limits<double>::infinity(), -std::numeric_limits<double>::infinity(),
+        // strings, including a non-ASCII one and one longer than 255 bytes
+        "", "a", "\xC3\xA4", std::string(300, 'x'),
+        // binary values with and without subtype
+        json::binary({}), json::binary({1, 2, 255}), json::binary({0x80, 0x7F}, 42), json::binary({1}, 0)
+    }
+    {}
+
+    std::vector<json> build()
+    {
+        std::vector<json> result = atoms;
+
+        // each atom inside containers, including homogeneous ones that the
+        // writers encode as optimized (typed) containers
+        result.emplace_back(json::array());
+        result.emplace_back(json::object());
+        for (const auto& atom : atoms)
+        {
+            result.push_back(json::array({atom}));
+            result.push_back(json::array({atom, atom, atom}));
+            result.push_back(json::array({json::array({atom})}));
+            result.push_back(json::object({{"key", atom}}));
+        }
+        result.push_back(json::array({1, 1.5}));
+        result.push_back(json::array({-1, 255}));
+        result.push_back(json::array({"a", "b"}));
+
+        // deep, but well below any recursion or depth limit
+        json nested_array = 1;
+        json nested_object = 1;
+        for (int i = 0; i < 300; ++i)
+        {
+            nested_array = json::array({nested_array});
+            nested_object = json::object({{"key", nested_object}});
+        }
+        result.push_back(nested_array);
+        result.push_back(nested_object);
+
+        add_annotated_arrays(result);
+        add_random_values(result);
+        return result;
+    }
+
+    // objects in the JData annotated array format, which the BJData writer
+    // encodes as ND-arrays when the annotation describes a packed array, and
+    // as plain objects otherwise (see #5398, #5399, #5403, #5404, and #5542)
+    static void add_annotated_arrays(std::vector<json>& result)
+    {
+        const std::vector<json> types =
+        {
+            "uint8", "int8", "uint16", "int16", "uint32", "int32", "uint64", "int64",
+            "single", "double", "char", "byte", "bool", "unknown", 5, nullptr
+        };
+        const std::vector<json> sizes =
+        {
+            json::array(), {3}, {1, 3}, {3, 1}, {2, 3}, {2, 0}, {0, 2}, {2, 2, 2}, {-1, 2}, {2, 1.5},
+            "3", 3, nullptr, json::binary({})
+        };
+        const std::vector<json> data =
+        {
+            nullptr, 5, "s", json::object({{"a", 1}}), json::array(),
+            {1, 2, 3}, {1, 2, 3, 4, 5, 6}, {1, 2, 3, 4, 5, 6, 7, 8},
+            {1.5, 2.5, 3.5, 4.5, 5.5, 6.5}, {300, -300, 70000, -70000, 1, 2},
+            {"a", "b", "c", "d", "e", "f"}, {json::array({1, 2, 3}), json::array({4, 5, 6})}
+        };
+
+        for (const auto& type : types)
+        {
+            for (const auto& size : sizes)
+            {
+                for (const auto& d : data)
+                {
+                    result.push_back({{"_ArrayType_", type}, {"_ArraySize_", size}, {"_ArrayData_", d}});
+                }
+            }
+        }
+
+        // incomplete annotations and annotations with an extra key
+        result.push_back({{"_ArraySize_", {2, 3}}, {"_ArrayData_", {1, 2, 3, 4, 5, 6}}});
+        result.push_back({{"_ArrayType_", "uint8"}, {"_ArrayData_", {1, 2, 3, 4, 5, 6}}});
+        result.push_back({{"_ArrayType_", "uint8"}, {"_ArraySize_", {2, 3}}});
+        result.push_back({{"_ArrayType_", "uint8"}, {"_ArraySize_", {2, 3}}, {"_ArrayData_", {1, 2, 3, 4, 5, 6}}, {"extra", 1}});
+    }
+
+    // random containers of atoms, both homogeneous and mixed
+    void add_random_values(std::vector<json>& result)
+    {
+        for (int i = 0; i < 1000; ++i)
+        {
+            result.push_back(random_value(0));
+        }
+    }
+
+    std::size_t random_below(std::size_t bound)
+    {
+        return generator() % bound;
+    }
+
+    json random_value(int depth)
+    {
+        const auto kind = random_below(10);
+        if (depth > 3 || kind < 5)
+        {
+            return atoms[random_below(atoms.size())];
+        }
+
+        json result = kind < 8 ? json::array() : json::object();
+        const auto count = random_below(5);
+        const bool homogeneous = random_below(2) == 0;
+        const json fixed = atoms[random_below(atoms.size())];
+        for (std::size_t i = 0; i < count; ++i)
+        {
+            json element = homogeneous ? fixed : random_value(depth + 1);
+            if (result.is_array())
+            {
+                result.push_back(std::move(element));
+            }
+            else
+            {
+                result[std::to_string(i)] = std::move(element);
+            }
+        }
+        return result;
+    }
+};
+
+} // namespace utils

tests/src/skip_library_version_check.cpp

@@ -0,0 +1,61 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+// Standalone compile-and-run check for the JSON_SKIP_LIBRARY_VERSION_CHECK
+// configuration macro, which (per #5423) was never exercised anywhere in the
+// test matrix.
+//
+// include/nlohmann/detail/abi_macros.hpp normally emits a #warning if
+// NLOHMANN_JSON_VERSION_MAJOR/MINOR/PATCH are already defined (as they would
+// be by an earlier inclusion of a different version of the library) with
+// values that mismatch the version about to be defined -- unless
+// JSON_SKIP_LIBRARY_VERSION_CHECK is defined, in which case the check (and
+// that #warning) is skipped.
+//
+// This file deliberately is not named tests/src/unit-*.cpp: it is compiled
+// directly (with a modest, non-strict warning set) by the dedicated
+// ci_test_skiplibraryversioncheck target in cmake/ci.cmake, rather than being
+// folded into the library's own -Weverything/-Werror unit test matrix. That
+// is because the scenario simulated here -- mixing two different, already
+// differently-versioned inclusions of the library in one translation unit --
+// unavoidably also triggers the *compiler's own* "macro redefined" warning,
+// independent of (and unaffected by) JSON_SKIP_LIBRARY_VERSION_CHECK, which
+// only ever silences the library's own #warning. Building this file under
+// -Weverything -Werror would therefore fail for a reason unrelated to the
+// macro under test.
+#define NLOHMANN_JSON_VERSION_MAJOR 0
+#define NLOHMANN_JSON_VERSION_MINOR 0
+#define NLOHMANN_JSON_VERSION_PATCH 0
+
+#define JSON_SKIP_LIBRARY_VERSION_CHECK 1
+
+#include <nlohmann/json.hpp>
+
+int main()
+{
+    // reaching this point at all already proves that the mismatched,
+    // pre-defined version macros above did not stop compilation -- which is
+    // exactly what JSON_SKIP_LIBRARY_VERSION_CHECK is for. The library must
+    // also still be fully usable.
+    const nlohmann::json j = {{"a", 1}, {"b", {1, 2, 3}}};
+    if (j.dump() != "{\"a\":1,\"b\":[1,2,3]}")
+    {
+        return 1;
+    }
+
+    // include/nlohmann/detail/abi_macros.hpp unconditionally (re)defines the
+    // version macros to the library's real, current version right after the
+    // (here, skipped) mismatch check, regardless of the deliberately wrong
+    // stand-in values defined above.
+    if (NLOHMANN_JSON_VERSION_MAJOR == 0 && NLOHMANN_JSON_VERSION_MINOR == 0 && NLOHMANN_JSON_VERSION_PATCH == 0)
+    {
+        return 1;
+    }
+
+    return 0;
+}

tests/src/test_utils.hpp

@@ -9,12 +9,39 @@
 #pragma once
 
 #include <cstdint> // uint8_t
+#include <cstddef> // size_t
 #include <fstream> // ifstream, istreambuf_iterator, ios
 #include <vector> // vector
 
 namespace utils
 {
 
+// Some tests intentionally discard the [[nodiscard]]/JSON_HEDLEY_WARN_UNUSED_RESULT
+// return value of a call they only make to exercise its side effects (e.g. checking
+// that it does not throw). A plain (void) cast on the call expression does not
+// suppress GCC's warning for functions using the GNU __attribute__((warn_unused_result))
+// form (as opposed to the C++17 [[nodiscard]] attribute) -- passing the value into an
+// ordinary function call does.
+template<typename T>
+inline void ignore_return_value(T&& /*unused*/) noexcept {}
+
+// Advance i toward last (inclusive) by stride, always visiting last.
+// stride 7 is coprime to 256, so every low-byte residue is still hit.
+template<typename T>
+T next_integer_sample(T i, T last, T stride)
+{
+    if (i >= last)
+    {
+        return static_cast<T>(last + 1);
+    }
+    if (stride > 0 && i > static_cast<T>(last - stride))
+    {
+        return last;
+    }
+    const T n = static_cast<T>(i + stride);
+    return n < last ? n : last;
+}
+
 inline std::vector<std::uint8_t> read_binary_file(const std::string& filename)
 {
     std::ifstream file(filename, std::ios::binary);

tests/src/unit-allocator.cpp

@@ -37,6 +37,12 @@ struct bad_allocator : std::allocator<T>
 };
 } // namespace
 
+TEST_CASE("get_allocator")
+{
+    const auto alloc = nlohmann::json::get_allocator();
+    CHECK(alloc == std::allocator<nlohmann::json>());
+}
+
 TEST_CASE("bad_alloc")
 {
     SECTION("bad_alloc")
@@ -216,9 +222,136 @@ TEST_CASE("controlled bad_alloc")
             CHECK_THROWS_AS(my_json(s), std::bad_alloc&);
             next_construct_fails = false;
         }
+
+        SECTION("basic_json(const basic_json&) of a deeply nested value (#5387)")
+        {
+            // Copying a value nested deeper than the descent bound builds the
+            // copy from the top down: every value whose own copy has not been
+            // made yet stays a null value until it is. Failing an allocation
+            // part-way through is what proves such a half-built copy can still
+            // be destroyed.
+            //
+            // Which path the failure lands in depends on the build: the first
+            // allocation of a copy belongs to the outermost level, so here it
+            // is the descending one. Built with JSON_NO_THREAD_LOCAL - as the
+            // ci_test_no_thread_local target builds the whole suite - no
+            // descent is made at all and the very same failure lands in the
+            // iterative path instead, part-way through its worklist.
+            const auto check_deep_copy = [](bool objects)
+            {
+                CAPTURE(objects);
+
+                next_construct_fails = false;
+
+                // deeper than the 128 levels the copy constructor descends into
+                const std::size_t depth = 300;
+
+                my_json j = 1;
+                for (std::size_t i = 0; i < depth; ++i)
+                {
+                    if (objects)
+                    {
+                        my_json wrapper = my_json::object();
+                        wrapper["a"] = std::move(j);
+                        j = std::move(wrapper);
+                    }
+                    else
+                    {
+                        j = my_json::array({std::move(j)});
+                    }
+                }
+
+                // NOLINTNEXTLINE(performance-unnecessary-copy-initialization): the copy is what is tested
+                CHECK_NOTHROW(my_json(j));
+
+                next_construct_fails = true;
+                // NOLINTNEXTLINE(performance-unnecessary-copy-initialization): the copy is what is tested
+                CHECK_THROWS_AS(my_json(j), std::bad_alloc&);
+                next_construct_fails = false;
+            };
+
+            check_deep_copy(false);
+            check_deep_copy(true);
+        }
     }
 }
 
+namespace
+{
+// counts the allocations of pairs with a non-const first member: the object
+// types store std::pair<const Key, T>, so only the scratch space of the
+// iterative deep copy allocates std::pair<Key, T>
+std::size_t scratch_pair_allocations = 0;
+
+template<class T>
+struct is_scratch_pair : std::false_type {};
+
+template<class K, class V>
+struct is_scratch_pair<std::pair<K, V>> : std::integral_constant < bool, !std::is_const<K>::value > {};
+
+template<class T>
+struct scratch_counting_allocator : std::allocator<T>
+{
+    using std::allocator<T>::allocator;
+
+    T* allocate(std::size_t n)
+    {
+        if (is_scratch_pair<T>::value)
+        {
+            ++scratch_pair_allocations;
+        }
+        return std::allocator<T>::allocate(n);
+    }
+
+#ifdef __cpp_lib_allocate_at_least
+    // std::allocator<T>::allocate_at_least would bypass the counting, and
+    // libc++'s containers prefer it over allocate from C++23 on
+    auto allocate_at_least(std::size_t n)
+    {
+        if (is_scratch_pair<T>::value)
+        {
+            ++scratch_pair_allocations;
+        }
+        return std::allocator<T>::allocate_at_least(n);
+    }
+#endif
+
+    template <class U>
+    struct rebind
+    {
+        using other = scratch_counting_allocator<U>;
+    };
+};
+} // namespace
+
+TEST_CASE("deep copy uses the provided allocator")
+{
+    using counting_json = nlohmann::basic_json<std::map,
+          std::vector,
+          std::string,
+          bool,
+          std::int64_t,
+          std::uint64_t,
+          double,
+          scratch_counting_allocator>;
+
+    // deeper than the 128 levels the copy constructor descends into, so the
+    // innermost objects are copied by the iterative deep copy
+    counting_json j = 1;
+    for (std::size_t i = 0; i < 300; ++i)
+    {
+        counting_json wrapper = counting_json::object();
+        wrapper["a"] = std::move(j);
+        j = std::move(wrapper);
+    }
+
+    scratch_pair_allocations = 0;
+    // NOLINTNEXTLINE(performance-unnecessary-copy-initialization): the copy is what is tested
+    const counting_json copy(j);
+    CHECK(scratch_pair_allocations > 0);
+    CHECK(copy == j);
+}
+
 namespace
 {
 template<class T>

tests/src/unit-alt-string.cpp

@@ -11,8 +11,10 @@
 
 #include <nlohmann/json.hpp>
 
+#include <cstdint>
 #include <string>
 #include <utility>
+#include <vector>
 
 /* forward declarations */
 class alt_string;
@@ -22,6 +24,10 @@ void int_to_string(alt_string& target, std::size_t value); // NOLINT(misc-use-in
 /*
  * This is virtually a string class.
  * It covers std::string under the hood.
+ *
+ * It deliberately does not provide c_str(), back(), find(str, pos), replace(),
+ * or substr(): the library must not rely on them. Do not add members here
+ * without checking that the library actually needs them.
  */
 class alt_string
 {
@@ -106,11 +112,6 @@ class alt_string
         return str_impl < op.str_impl;
     }
 
-    const char* c_str() const
-    {
-        return str_impl.c_str();
-    }
-
     char& operator[](std::size_t index)
     {
         return str_impl[index];
@@ -121,16 +122,6 @@ class alt_string
         return str_impl[index];
     }
 
-    char& back()
-    {
-        return str_impl.back();
-    }
-
-    const char& back() const
-    {
-        return str_impl.back();
-    }
-
     void clear()
     {
         str_impl.clear();
@@ -146,28 +137,11 @@ class alt_string
         return str_impl.empty();
     }
 
-    std::size_t find(const alt_string& str, std::size_t pos = 0) const
-    {
-        return str_impl.find(str.str_impl, pos);
-    }
-
     std::size_t find_first_of(char c, std::size_t pos = 0) const
     {
         return str_impl.find_first_of(c, pos);
     }
 
-    alt_string substr(std::size_t pos = 0, std::size_t count = npos) const
-    {
-        const std::string s = str_impl.substr(pos, count);
-        return {s.data(), s.size()};
-    }
-
-    alt_string& replace(std::size_t pos, std::size_t count, const alt_string& str)
-    {
-        str_impl.replace(pos, count, str.str_impl);
-        return *this;
-    }
-
     void reserve( std::size_t new_cap = 0 )
     {
         str_impl.reserve(new_cap);
@@ -202,6 +176,32 @@ bool operator<(const char* op1, const alt_string& op2) noexcept
 
 TEST_CASE("alternative string type")
 {
+    SECTION("binary formats")
+    {
+        alt_json doc;
+        doc["pi"] = 3.141;
+        doc["happy"] = true;
+        doc["list"] = {1, 2, 3};
+
+        CHECK(alt_json::from_cbor(alt_json::to_cbor(doc)) == doc);
+        CHECK(alt_json::from_msgpack(alt_json::to_msgpack(doc)) == doc);
+        CHECK(alt_json::from_bon8(alt_json::to_bon8(doc)) == doc);
+        // BSON is not covered: it additionally needs string_t::find(value_type),
+        // which alt_string does not provide
+        CHECK(alt_json::from_ubjson(alt_json::to_ubjson(doc)) == doc);
+
+        // a UBJSON high-precision number is parsed into a std::string that the
+        // reader has to hand to the SAX interface as an alt_string
+        const std::vector<uint8_t> high_precision =
+        {
+            'H', 'i', 0x16, '3', '.', '1', '4', '1', '5', '9', '2', '6', '5', '3',
+            '5', '8', '9', '7', '9', '3', '2', '3', '8', '4', '6'
+        };
+        const auto number = alt_json::from_ubjson(high_precision);
+        CHECK(number.is_number_float());
+        CHECK(number.get<double>() == doctest::Approx(3.14159265358979323846));
+    }
+
     SECTION("dump")
     {
         {
@@ -332,6 +332,15 @@ TEST_CASE("alternative string type")
 
         CHECK(j.at(alt_json::json_pointer("/foo/0")) == j["foo"][0]);
         CHECK(j.at(alt_json::json_pointer("/foo/1")) == j["foo"][1]);
+
+        // RFC 6901 escaping works without string_t::find(str, pos), replace(),
+        // and substr()
+        auto j2 = alt_json::parse(R"({"a/b": 1, "m~n": 2, "~/~~//": 3})");
+        CHECK(j2.at(alt_json::json_pointer("/a~1b")) == 1);
+        CHECK(j2.at(alt_json::json_pointer("/m~0n")) == 2);
+        CHECK(j2.at(alt_json::json_pointer("/~0~1~0~0~1~1")) == 3);
+        CHECK(alt_json::json_pointer("/~0~1~0~0~1~1").to_string() == alt_string("/~0~1~0~0~1~1"));
+        CHECK(j2.flatten().unflatten() == j2);
     }
 
     SECTION("patch")

tests/src/unit-binary_formats.cpp

@@ -25,6 +25,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         const auto bjdata_1_size = json::to_bjdata(j).size();
         const auto bjdata_2_size = json::to_bjdata(j, true).size();
         const auto bjdata_3_size = json::to_bjdata(j, true, true).size();
+        const auto bon8_size = json::to_bon8(j).size();
         const auto bson_size = json::to_bson(j).size();
         const auto cbor_size = json::to_cbor(j).size();
         const auto msgpack_size = json::to_msgpack(j).size();
@@ -36,6 +37,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK(bjdata_1_size == 1112030);
         CHECK(bjdata_2_size == 1224148);
         CHECK(bjdata_3_size == 1224148);
+        CHECK(bon8_size == 1055792);
         CHECK(bson_size == 1794522);
         CHECK(cbor_size == 1055552);
         CHECK(msgpack_size == 1056145);
@@ -47,6 +49,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK((100.0 * double(bjdata_1_size) / double(json_size)) == Approx(53.199));
         CHECK((100.0 * double(bjdata_2_size) / double(json_size)) == Approx(58.563));
         CHECK((100.0 * double(bjdata_3_size) / double(json_size)) == Approx(58.563));
+        CHECK((100.0 * double(bon8_size) / double(json_size)) == Approx(50.509));
         CHECK((100.0 * double(bson_size) / double(json_size)) == Approx(85.849));
         CHECK((100.0 * double(cbor_size) / double(json_size)) == Approx(50.497));
         CHECK((100.0 * double(msgpack_size) / double(json_size)) == Approx(50.526));
@@ -64,6 +67,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         const auto bjdata_1_size = json::to_bjdata(j).size();
         const auto bjdata_2_size = json::to_bjdata(j, true).size();
         const auto bjdata_3_size = json::to_bjdata(j, true, true).size();
+        const auto bon8_size = json::to_bon8(j).size();
         const auto bson_size = json::to_bson(j).size();
         const auto cbor_size = json::to_cbor(j).size();
         const auto msgpack_size = json::to_msgpack(j).size();
@@ -75,6 +79,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK(bjdata_1_size == 425342);
         CHECK(bjdata_2_size == 429970);
         CHECK(bjdata_3_size == 429970);
+        CHECK(bon8_size == 391396);
         CHECK(bson_size == 444568);
         CHECK(cbor_size == 402814);
         CHECK(msgpack_size == 401510);
@@ -86,6 +91,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK((100.0 * double(bjdata_1_size) / double(json_size)) == Approx(91.097));
         CHECK((100.0 * double(bjdata_2_size) / double(json_size)) == Approx(92.089));
         CHECK((100.0 * double(bjdata_3_size) / double(json_size)) == Approx(92.089));
+        CHECK((100.0 * double(bon8_size) / double(json_size)) == Approx(83.828));
         CHECK((100.0 * double(bson_size) / double(json_size)) == Approx(95.215));
         CHECK((100.0 * double(cbor_size) / double(json_size)) == Approx(86.273));
         CHECK((100.0 * double(msgpack_size) / double(json_size)) == Approx(85.993));
@@ -103,6 +109,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         const auto bjdata_1_size = json::to_bjdata(j).size();
         const auto bjdata_2_size = json::to_bjdata(j, true).size();
         const auto bjdata_3_size = json::to_bjdata(j, true, true).size();
+        const auto bon8_size = json::to_bon8(j).size();
         const auto bson_size = json::to_bson(j).size();
         const auto cbor_size = json::to_cbor(j).size();
         const auto msgpack_size = json::to_msgpack(j).size();
@@ -114,6 +121,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK(bjdata_1_size == 390781);
         CHECK(bjdata_2_size == 433557);
         CHECK(bjdata_3_size == 432964);
+        CHECK(bon8_size == 317879);
         CHECK(bson_size == 479430);
         CHECK(cbor_size == 342373);
         CHECK(msgpack_size == 342473);
@@ -125,6 +133,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK((100.0 * double(bjdata_1_size) / double(json_size)) == Approx(78.109));
         CHECK((100.0 * double(bjdata_2_size) / double(json_size)) == Approx(86.659));
         CHECK((100.0 * double(bjdata_3_size) / double(json_size)) == Approx(86.541));
+        CHECK((100.0 * double(bon8_size) / double(json_size)) == Approx(63.538));
         CHECK((100.0 * double(bson_size) / double(json_size)) == Approx(95.828));
         CHECK((100.0 * double(cbor_size) / double(json_size)) == Approx(68.433));
         CHECK((100.0 * double(msgpack_size) / double(json_size)) == Approx(68.453));
@@ -142,6 +151,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         const auto bjdata_1_size = json::to_bjdata(j).size();
         const auto bjdata_2_size = json::to_bjdata(j, true).size();
         const auto bjdata_3_size = json::to_bjdata(j, true, true).size();
+        const auto bon8_size = json::to_bon8(j).size();
         const auto bson_size = json::to_bson({{"", j}}).size(); // wrap array in object for BSON
         const auto cbor_size = json::to_cbor(j).size();
         const auto msgpack_size = json::to_msgpack(j).size();
@@ -153,6 +163,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK(bjdata_1_size == 50710965);
         CHECK(bjdata_2_size == 51144830);
         CHECK(bjdata_3_size == 51144830);
+        CHECK(bon8_size == 45942080);
         CHECK(bson_size == 56008520);
         CHECK(cbor_size == 46187320);
         CHECK(msgpack_size == 46158575);
@@ -164,6 +175,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK((100.0 * double(bjdata_1_size) / double(json_size)) == Approx(96.576));
         CHECK((100.0 * double(bjdata_2_size) / double(json_size)) == Approx(97.402));
         CHECK((100.0 * double(bjdata_3_size) / double(json_size)) == Approx(97.402));
+        CHECK((100.0 * double(bon8_size) / double(json_size)) == Approx(87.494));
         CHECK((100.0 * double(bson_size) / double(json_size)) == Approx(106.665));
         CHECK((100.0 * double(cbor_size) / double(json_size)) == Approx(87.961));
         CHECK((100.0 * double(msgpack_size) / double(json_size)) == Approx(87.906));
@@ -181,6 +193,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         const auto bjdata_1_size = json::to_bjdata(j).size();
         const auto bjdata_2_size = json::to_bjdata(j, true).size();
         const auto bjdata_3_size = json::to_bjdata(j, true, true).size();
+        const auto bon8_size = json::to_bon8(j).size();
         // BSON cannot process the file as it contains code point  U+0000
         const auto cbor_size = json::to_cbor(j).size();
         const auto msgpack_size = json::to_msgpack(j).size();
@@ -192,6 +205,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK(bjdata_1_size == 148695);
         CHECK(bjdata_2_size == 150569);
         CHECK(bjdata_3_size == 150569);
+        CHECK(bon8_size == 144477);
         CHECK(cbor_size == 147095);
         CHECK(msgpack_size == 147017);
         CHECK(ubjson_1_size == 148695);
@@ -202,6 +216,7 @@ TEST_CASE("Binary Formats" * doctest::skip())
         CHECK((100.0 * double(bjdata_1_size) / double(json_size)) == Approx(88.153));
         CHECK((100.0 * double(bjdata_2_size) / double(json_size)) == Approx(89.264));
         CHECK((100.0 * double(bjdata_3_size) / double(json_size)) == Approx(89.264));
+        CHECK((100.0 * double(bon8_size) / double(json_size)) == Approx(85.653));
         CHECK((100.0 * double(cbor_size) / double(json_size)) == Approx(87.205));
         CHECK((100.0 * double(msgpack_size) / double(json_size)) == Approx(87.158));
         CHECK((100.0 * double(ubjson_1_size) / double(json_size)) == Approx(88.153));

tests/src/unit-binary_writer_sinks.cpp

@@ -0,0 +1,216 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+#include "doctest_compatibility.h"
+
+#include <nlohmann/json.hpp>
+using nlohmann::json;
+
+#include <cstdint>
+#include <limits>
+#include <string>
+#include <vector>
+
+namespace
+{
+
+// a spread of values exercising every writer path: scalars of each width, the
+// float paths, strings, binary, and containers big enough to reallocate
+std::vector<json> test_values()
+{
+    json big_array = json::array();
+    for (int i = 0; i < 5000; ++i)
+    {
+        big_array.push_back(i);
+    }
+
+    json big_object = json::object();
+    for (int i = 0; i < 1000; ++i)
+    {
+        big_object[std::to_string(i)] = i;
+    }
+
+    return
+    {
+        json(nullptr), json(true), json(false),
+        json(0), json(-1), json(255), json(-129), json(65535), json(-32769),
+        json(4294967295U), json(-2147483649LL), json(18446744073709551615ULL),
+        json(0.0), json(-0.5), json(3.1415926535897932),
+        json(""), json("hello"), json(std::string(1000, 'x')),
+        json::binary({0x00, 0x01, 0x02}, 42),
+        json::array(), json::object(),
+        json::array({1, 2, 3}), json({{"a", 1}, {"b", nullptr}}),
+        json({{"nested", {{"deep", json::array({1, "two", 3.0, nullptr})}}}}),
+        big_array, big_object
+    };
+}
+
+// BON8 has no integers above the int64 range, so to_bon8() rejects them
+bool bon8_representable(const json& j)
+{
+    return !j.is_number_unsigned() || j.get<std::uint64_t>() <= static_cast<std::uint64_t>((std::numeric_limits<std::int64_t>::max)());
+}
+
+// values to_bson() accepts: the document must be an object
+std::vector<json> bson_values()
+{
+    json big_object = json::object();
+    for (int i = 0; i < 1000; ++i)
+    {
+        big_object[std::to_string(i)] = i;
+    }
+
+    return
+    {
+        json::object(),
+        json({{"a", 1}, {"b", nullptr}, {"c", true}, {"d", 2.5}, {"e", "text"}}),
+        json({{"arr", json::array({1, 2, 3})}, {"obj", {{"k", "v"}}}}),
+        big_object
+    };
+}
+
+} // namespace
+
+// The vector-returning to_*(j) overloads write through the non-virtual
+// output_vector_sink, while to_*(j, adapter) goes through output_adapter_sink.
+// The two are separate code paths that must stay byte-for-byte identical; these
+// checks fail if either overload is ever changed without the other.
+TEST_CASE("binary writer output sinks")
+{
+    SECTION("vector sink and adapter sink agree")
+    {
+        // note: no SUBCASE inside these loops - doctest keys subcases by
+        // name/file/line, so a subcase in a loop body would only ever run for
+        // the first iteration
+        for (const auto& j : test_values())
+        {
+            CAPTURE(j.dump(-1, ' ', false, json::error_handler_t::replace));
+
+            std::vector<std::uint8_t> cbor;
+            json::to_cbor(j, cbor);
+            CHECK(json::to_cbor(j) == cbor);
+
+            std::vector<std::uint8_t> msgpack;
+            json::to_msgpack(j, msgpack);
+            CHECK(json::to_msgpack(j) == msgpack);
+
+            if (bon8_representable(j))
+            {
+                std::vector<std::uint8_t> bon8;
+                json::to_bon8(j, bon8);
+                CHECK(json::to_bon8(j) == bon8);
+            }
+
+            for (const bool use_size :
+                    {
+                        false, true
+                    })
+            {
+                for (const bool use_type :
+                        {
+                            false, true
+                        })
+                {
+                    if (use_type && !use_size)
+                    {
+                        continue; // not a supported combination
+                    }
+                    CAPTURE(use_size);
+                    CAPTURE(use_type);
+                    std::vector<std::uint8_t> ubjson;
+                    json::to_ubjson(j, ubjson, use_size, use_type);
+                    CHECK(json::to_ubjson(j, use_size, use_type) == ubjson);
+                }
+            }
+
+            for (const auto version :
+                    {
+                        json::bjdata_version_t::draft2, json::bjdata_version_t::draft3
+                    })
+            {
+                std::vector<std::uint8_t> bjdata;
+                json::to_bjdata(j, bjdata, false, false, version);
+                CHECK(json::to_bjdata(j, false, false, version) == bjdata);
+            }
+        }
+
+        for (const auto& j : bson_values())
+        {
+            CAPTURE(j.dump());
+            std::vector<std::uint8_t> bson;
+            json::to_bson(j, bson);
+            CHECK(json::to_bson(j) == bson);
+        }
+    }
+
+    SECTION("the char adapter produces the same bytes")
+    {
+        for (const auto& j : test_values())
+        {
+            CAPTURE(j.dump(-1, ' ', false, json::error_handler_t::replace));
+
+            const std::vector<std::uint8_t> expected = json::to_cbor(j);
+            std::vector<char> as_char;
+            json::to_cbor(j, as_char);
+
+            REQUIRE(as_char.size() == expected.size());
+            std::vector<std::uint8_t> as_bytes;
+            as_bytes.reserve(as_char.size());
+            for (const char c : as_char)
+            {
+                as_bytes.push_back(static_cast<std::uint8_t>(c));
+            }
+            CHECK(as_bytes == expected);
+        }
+    }
+}
+
+// binary_reserve_hint() is documented as a *lower* bound on the serialized size,
+// so that reserving it up front can never leave the returned vector holding
+// capacity beyond what the value actually needs.
+TEST_CASE("binary_reserve_hint never over-reserves")
+{
+    for (const auto& j : test_values())
+    {
+        CAPTURE(j.dump(-1, ' ', false, json::error_handler_t::replace));
+
+        const std::size_t hint = nlohmann::detail::binary_reserve_hint(j);
+
+        CHECK(hint <= json::to_cbor(j).size());
+        CHECK(hint <= json::to_msgpack(j).size());
+        CHECK(hint <= json::to_ubjson(j).size());
+        CHECK(hint <= json::to_ubjson(j, true, true).size());
+        CHECK(hint <= json::to_bjdata(j).size());
+        if (bon8_representable(j))
+        {
+            CHECK(hint <= json::to_bon8(j).size());
+        }
+    }
+
+    for (const auto& j : bson_values())
+    {
+        CAPTURE(j.dump());
+        CHECK(nlohmann::detail::binary_reserve_hint(j) <= json::to_bson(j).size());
+    }
+
+    SECTION("scalars get no hint")
+    {
+        CHECK(nlohmann::detail::binary_reserve_hint(json(nullptr)) == 0);
+        CHECK(nlohmann::detail::binary_reserve_hint(json(42)) == 0);
+        CHECK(nlohmann::detail::binary_reserve_hint(json("a string")) == 0);
+        CHECK(nlohmann::detail::binary_reserve_hint(json::binary({0x01})) == 0);
+    }
+
+    SECTION("containers are hinted from their element count")
+    {
+        CHECK(nlohmann::detail::binary_reserve_hint(json::array()) == 1);
+        CHECK(nlohmann::detail::binary_reserve_hint(json::array({1, 2, 3})) == 4);
+        CHECK(nlohmann::detail::binary_reserve_hint(json::object()) == 1);
+        CHECK(nlohmann::detail::binary_reserve_hint(json({{"a", 1}, {"b", 2}})) == 5);
+    }
+}

tests/src/unit-bjdata.cpp

@@ -19,6 +19,7 @@ using nlohmann::json;
 #include <fstream>
 #include <set>
 #include "make_test_data_available.hpp"
+#include "round_trip_corpus.hpp"
 #include "test_utils.hpp"
 
 namespace
@@ -417,7 +418,7 @@ TEST_CASE("BJData")
 
                 SECTION("-32768..-129 (int16)")
                 {
-                    for (int32_t i = -32768; i <= -129; ++i)
+                    for (int32_t i = -32768; i <= -129; i = utils::next_integer_sample(i, -129, 7))
                     {
                         CAPTURE(i)
 
@@ -577,7 +578,7 @@ TEST_CASE("BJData")
 
                 SECTION("256..32767 (int16)")
                 {
-                    for (size_t i = 256; i <= 32767; ++i)
+                    for (size_t i = 256; i <= 32767; i = utils::next_integer_sample(i, static_cast<size_t>(32767), static_cast<size_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -910,7 +911,7 @@ TEST_CASE("BJData")
 
                 SECTION("256..32767 (int16)")
                 {
-                    for (size_t i = 256; i <= 32767; ++i)
+                    for (size_t i = 256; i <= 32767; i = utils::next_integer_sample(i, static_cast<size_t>(32767), static_cast<size_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -2586,7 +2587,12 @@ TEST_CASE("BJData")
                 CHECK(json::to_bjdata(json::from_bjdata(v_d), true, true) == v_d);
                 CHECK(json::to_bjdata(json::from_bjdata(v_D), true, true) == v_D);
                 CHECK(json::to_bjdata(json::from_bjdata(v_C), true, true) == v_C);
-                CHECK(json::to_bjdata(json::from_bjdata(v_B), true, true) == v_B);
+                // v_B uses the Draft-3-only 'B' marker, so it round-trips only when
+                // Draft 3 is explicitly selected (see GitHub issue #5404); the
+                // default Draft 2 falls back to a plain object instead, covered by
+                // the "ndarray with _ArrayType_ "byte" is gated by the BJData draft
+                // version" section below
+                CHECK(json::to_bjdata(json::from_bjdata(v_B), true, true, json::bjdata_version_t::draft3) == v_B);
             }
 
             SECTION("ndarray with data not matching _ArrayType_ is written as an object")
@@ -2599,25 +2605,25 @@ TEST_CASE("BJData")
                 // that still round-trips.
 
                 // string data declared as a uint64 array
-                json const j_str = json({{"_ArrayType_", "uint64"}, {"_ArraySize_", {1}}, {"_ArrayData_", {"pointer"}}});
+                json const j_str = json({{"_ArrayType_", "uint64"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {"pointer", "value"}}});
                 const auto out_str = json::to_bjdata(j_str);
                 CHECK(out_str.at(0) == '{');
                 CHECK(json::from_bjdata(out_str) == j_str);
 
                 // integer data declared as a double array
-                json const j_float = json({{"_ArrayType_", "double"}, {"_ArraySize_", {2}}, {"_ArrayData_", {1, 2}}});
+                json const j_float = json({{"_ArrayType_", "double"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {1, 2}}});
                 const auto out_float = json::to_bjdata(j_float);
                 CHECK(out_float.at(0) == '{');
                 CHECK(json::from_bjdata(out_float) == j_float);
 
                 // a non-integer shape entry is likewise not treated as an ndarray
-                json const j_size = json({{"_ArrayType_", "uint8"}, {"_ArraySize_", {"x"}}, {"_ArrayData_", {1}}});
+                json const j_size = json({{"_ArrayType_", "uint8"}, {"_ArraySize_", {"x", 1}}, {"_ArrayData_", {1}}});
                 const auto out_size = json::to_bjdata(j_size);
                 CHECK(out_size.at(0) == '{');
                 CHECK(json::from_bjdata(out_size) == j_size);
 
                 // a negative shape entry is not a usable dimension either
-                json const j_neg = json::parse(R"({"_ArrayType_":"uint8","_ArraySize_":[-1],"_ArrayData_":[1]})");
+                json const j_neg = json::parse(R"({"_ArrayType_":"uint8","_ArraySize_":[-1,1],"_ArrayData_":[1]})");
                 const auto out_neg = json::to_bjdata(j_neg);
                 CHECK(out_neg.at(0) == '{');
                 CHECK(json::from_bjdata(out_neg) == j_neg);
@@ -2629,8 +2635,10 @@ TEST_CASE("BJData")
                 // the C++ API stores an int literal as number_integer, so _ArrayType_
                 // names the wire type rather than the storage. Both storages have to
                 // produce the same typed array for every type.
+                // "byte" is checked separately below since it additionally requires
+                // BJData Draft 3 to be selected explicitly (see GitHub issue #5404).
                 for (const char* type :
-                        {"uint8", "int8", "uint16", "int16", "uint32", "int32", "uint64", "int64", "char", "byte"
+                        {"uint8", "int8", "uint16", "int16", "uint32", "int32", "uint64", "int64", "char"
                         })
                 {
                     CAPTURE(type);
@@ -2641,15 +2649,23 @@ TEST_CASE("BJData")
                     CHECK(from_text == json::to_bjdata(json({{"_ArrayType_", type}, {"_ArraySize_", {2, 3}}, {"_ArrayData_", {1, 2, 3, 4, 5, 6}}})));
                 }
 
+                {
+                    const std::string text = R"({"_ArrayType_":"byte","_ArraySize_":[2,3],"_ArrayData_":[1,2,3,4,5,6]})";
+                    const auto from_text = json::to_bjdata(json::parse(text), true, true, json::bjdata_version_t::draft3);
+                    CHECK(from_text.at(0) == '[');
+                    CHECK(from_text == json::to_bjdata(json({{"_ArrayType_", "byte"}, {"_ArraySize_", {2, 3}}, {"_ArrayData_", {1, 2, 3, 4, 5, 6}}}),
+                    true, true, json::bjdata_version_t::draft3));
+                }
+
                 // negative values under a signed type behave the same way
-                const auto from_neg = json::to_bjdata(json::parse(R"({"_ArrayType_":"int32","_ArraySize_":[2],"_ArrayData_":[-5,7]})"));
+                const auto from_neg = json::to_bjdata(json::parse(R"({"_ArrayType_":"int32","_ArraySize_":[2,1],"_ArrayData_":[-5,7]})"));
                 CHECK(from_neg.at(0) == '[');
-                CHECK(from_neg == json::to_bjdata(json({{"_ArrayType_", "int32"}, {"_ArraySize_", {2}}, {"_ArrayData_", {-5, 7}}})));
+                CHECK(from_neg == json::to_bjdata(json({{"_ArrayType_", "int32"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {-5, 7}}})));
 
                 // and so do the floating point types
-                const auto from_float = json::to_bjdata(json::parse(R"({"_ArrayType_":"double","_ArraySize_":[2],"_ArrayData_":[1.5,2.5]})"));
+                const auto from_float = json::to_bjdata(json::parse(R"({"_ArrayType_":"double","_ArraySize_":[2,1],"_ArrayData_":[1.5,2.5]})"));
                 CHECK(from_float.at(0) == '[');
-                CHECK(from_float == json::to_bjdata(json({{"_ArrayType_", "double"}, {"_ArraySize_", {2}}, {"_ArrayData_", {1.5, 2.5}}})));
+                CHECK(from_float == json::to_bjdata(json({{"_ArrayType_", "double"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {1.5, 2.5}}})));
             }
 
             SECTION("optimized ndarray (type and vector-size as 1D array)")
@@ -2730,6 +2746,295 @@ TEST_CASE("BJData")
                 CHECK(json::from_bjdata(json::to_bjdata(j_type), true, true) == j_type);
                 CHECK(json::from_bjdata(json::to_bjdata(j_size), true, true) == j_size);
             }
+
+            SECTION("ndarray whose _ArrayType_ is not a string stays as object")
+            {
+                // the type name is looked up as a string below the annotation
+                // check; a non-string _ArrayType_ cannot name a known dtype,
+                // so calling get<string_t>() on it would throw type_error.302
+                // instead of falling back like an unrecognized type name
+                // already does (see GitHub issue #5398)
+                json const j_number = json({{"_ArrayType_", 1}, {"_ArraySize_", {2}}, {"_ArrayData_", {1, 2}}});
+                const auto out_number = json::to_bjdata(j_number);
+                CHECK(out_number.at(0) == '{');
+                CHECK(json::from_bjdata(out_number) == j_number);
+
+                json const j_null = json({{"_ArrayType_", nullptr}, {"_ArraySize_", {2}}, {"_ArrayData_", {1, 2}}});
+                const auto out_null = json::to_bjdata(j_null);
+                CHECK(out_null.at(0) == '{');
+                CHECK(json::from_bjdata(out_null) == j_null);
+
+                json const j_bool = json({{"_ArrayType_", true}, {"_ArraySize_", {2}}, {"_ArrayData_", {1, 2}}});
+                const auto out_bool = json::to_bjdata(j_bool);
+                CHECK(out_bool.at(0) == '{');
+                CHECK(json::from_bjdata(out_bool) == j_bool);
+
+                json const j_array = json({{"_ArrayType_", {"uint8"}}, {"_ArraySize_", {2}}, {"_ArrayData_", {1, 2}}});
+                const auto out_array = json::to_bjdata(j_array);
+                CHECK(out_array.at(0) == '{');
+                CHECK(json::from_bjdata(out_array) == j_array);
+
+                json const j_object = json({{"_ArrayType_", {{"a", 1}}}, {"_ArraySize_", {2}}, {"_ArrayData_", {1, 2}}});
+                const auto out_object = json::to_bjdata(j_object);
+                CHECK(out_object.at(0) == '{');
+                CHECK(json::from_bjdata(out_object) == j_object);
+            }
+
+            SECTION("re-serializing a value containing a plain-array-of-bytes is value-stable but not byte-stable")
+            {
+                // OSS-Fuzz found this input (an array whose first element is a
+                // binary_t byte, followed by an object whose _ArrayType_ is
+                // not a string) while exercising the fix for #5398 above: once
+                // the fix stops to_bjdata() from throwing type_error.302 for
+                // the third element, serialization proceeds far enough to
+                // reach a pre-existing, unrelated round-trip quirk in how a
+                // single-byte binary_t value is re-encoded.
+                std::vector<std::uint8_t> const input
+                {
+                    0x5b, 0x5b, 0x24, 0x42, 0x23, 0x5b, 0x69, 0x01, 0x5d, 0x5b, 0x5b, 0x5d, 0x7b, 0x55, 0x0b,
+                    0x5f, 0x41, 0x72, 0x72, 0x61, 0x79, 0x44, 0x61, 0x74, 0x61, 0x5f, 0x54, 0x55, 0x0b, 0x5f,
+                    0x41, 0x72, 0x72, 0x61, 0x79, 0x53, 0x69, 0x7a, 0x65, 0x5f, 0x5a, 0x55, 0x0b, 0x5f, 0x41,
+                    0x72, 0x72, 0x61, 0x79, 0x54, 0x79, 0x70, 0x65, 0x5f, 0x54, 0x7d, 0x5d
+                };
+                json const j1 = json::from_bjdata(input);
+
+                // to_bjdata() must not throw (this is what #5398 fixes)
+                std::vector<std::uint8_t> vec2;
+                CHECK_NOTHROW(vec2 = json::to_bjdata(j1, false, false));
+
+                // parsing back a plain (non-optimized) array of bytes cannot
+                // recover that it used to be a binary_t: from_bjdata() has no
+                // way to distinguish "array of uint8 numbers" from "array of
+                // bytes" unless the compact "$U#" array header is used, so
+                // the binary_t collapses into a plain JSON array
+                json const j2 = json::from_bjdata(vec2);
+                CHECK(j1 != j2);
+                CHECK(j2 == json({{91}, json::array(), {{"_ArrayData_", true}, {"_ArraySize_", nullptr}, {"_ArrayType_", true}}}));
+
+                // re-serializing j2 no longer goes through the dedicated
+                // binary_t writer (which always uses the 'U' marker for raw
+                // bytes); the now-plain number 91 goes through the generic
+                // smallest-type writer instead, which - like the rest of the
+                // UBJSON/BJData writer, and unchanged by this fix - prefers
+                // the 'i' (int8) marker over 'U' (uint8) for values that fit
+                // both. Both markers are valid BJData and both decode back to
+                // 91, so this is not byte-for-byte identical to vec2, but it
+                // is value-stable: parsing it again reproduces j2 exactly.
+                std::vector<std::uint8_t> const vec3 = json::to_bjdata(j2, false, false);
+                CHECK(json::from_bjdata(vec3) == j2);
+            }
+
+            SECTION("ndarray whose dimensions overflow stays as object")
+            {
+                // the product of the dimensions wraps around std::size_t to 0
+                // and so matches the size of the empty _ArrayData_; writing this
+                // as an ndarray would announce an element count no reader can
+                // honor, so it has to stay a plain object
+                json j_overflow = json({{"_ArrayData_", json::array()}, {"_ArraySize_", {9223372036854775808ull, 2}}, {"_ArrayType_", "uint8"}});
+                CHECK(json::from_bjdata(json::to_bjdata(j_overflow), true, true) == j_overflow);
+
+                // a single dimension that does not fit into std::size_t is
+                // rejected for the same reason (only observable where
+                // std::size_t is narrower than 64 bit)
+                json j_huge = json({{"_ArrayData_", json::array()}, {"_ArraySize_", {18446744073709551615ull, 2}}, {"_ArrayType_", "uint8"}});
+                CHECK(json::from_bjdata(json::to_bjdata(j_huge), true, true) == j_huge);
+
+                // a well-formed ndarray is still encoded as one
+                json j_ok = json({{"_ArrayData_", {1, 2, 3, 4, 5, 6}}, {"_ArraySize_", {2, 3}}, {"_ArrayType_", "uint8"}});
+                CHECK(json::to_bjdata(j_ok) == std::vector<uint8_t>({'[', '$', 'U', '#', '[', 'i', 2, 'i', 3, ']', 1, 2, 3, 4, 5, 6}));
+                CHECK(json::from_bjdata(json::to_bjdata(j_ok), true, true) == j_ok);
+            }
+
+            SECTION("ndarray whose _ArraySize_ is not an array stays as object")
+            {
+                // the shape is written verbatim as the header length, so a
+                // value that is not an array cannot produce a valid one: null
+                // would emit 'Z' and an object '{', neither of which a reader
+                // accepts after '#'. Both have to stay plain objects.
+                json const j_null = json({{"_ArrayType_", "uint8"}, {"_ArraySize_", nullptr}, {"_ArrayData_", json::array()}});
+                const auto out_null = json::to_bjdata(j_null);
+                CHECK(out_null.at(0) == '{');
+                CHECK(json::from_bjdata(out_null) == j_null);
+
+                // an object shape passes the per-entry check by iterating its
+                // values rather than dimensions, so it needs rejecting too
+                json const j_obj = json({{"_ArrayType_", "uint8"}, {"_ArraySize_", {{"a", 1}}}, {"_ArrayData_", {1}}});
+                const auto out_obj = json::to_bjdata(j_obj);
+                CHECK(out_obj.at(0) == '{');
+                CHECK(json::from_bjdata(out_obj) == j_obj);
+
+                // a scalar shape is not a dimension list either
+                json const j_num = json({{"_ArrayType_", "uint8"}, {"_ArraySize_", 1}, {"_ArrayData_", {1}}});
+                const auto out_num = json::to_bjdata(j_num);
+                CHECK(out_num.at(0) == '{');
+                CHECK(json::from_bjdata(out_num) == j_num);
+
+                // OSS-Fuzz issue 474400817: an empty object _ArraySize_ was
+                // written as the ND-array header length, which from_bjdata()
+                // could not read back
+                const std::vector<uint8_t> input =
+                {
+                    '[', '{', 'U', 11, '_', 'A', 'r', 'r', 'a', 'y', 'D', 'a', 't', 'a', '_', 'Z',
+                    'U', 11, '_', 'A', 'r', 'r', 'a', 'y', 'T', 'y', 'p', 'e', '_', 'S', 'i', 5, 'i', 'n', 't', '1', '6',
+                    'U', 11, '_', 'A', 'r', 'r', 'a', 'y', 'S', 'i', 'z', 'e', '_', '{', '}', '}', ']'
+                };
+                const json j1 = json::from_bjdata(input);
+                CHECK(j1 == json::parse(R"([{"_ArrayType_":"int16","_ArraySize_":{},"_ArrayData_":null}])"));
+                json j2;
+                CHECK_NOTHROW(j2 = json::from_bjdata(json::to_bjdata(j1, false, false)));
+                CHECK(j2 == j1);
+            }
+
+            SECTION("ndarray with out-of-range _ArrayData_ elements stays as object")
+            {
+                // each element is cast to the (possibly narrower) C++ type
+                // named by _ArrayType_ before being written; a value that
+                // does not fit that type would silently wrap instead of
+                // being reported, so such an object falls back to a plain
+                // object encoding that still round-trips (see GitHub issue #5403)
+
+                // an unsigned element that does not fit uint8
+                json const j_uint8 = json({{"_ArrayType_", "uint8"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {1, 256}}});
+                const auto out_uint8 = json::to_bjdata(j_uint8);
+                CHECK(out_uint8.at(0) == '{');
+                CHECK(json::from_bjdata(out_uint8) == j_uint8);
+
+                // a signed element that does not fit int8
+                json const j_int8 = json({{"_ArrayType_", "int8"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {1, 200}}});
+                const auto out_int8 = json::to_bjdata(j_int8);
+                CHECK(out_int8.at(0) == '{');
+                CHECK(json::from_bjdata(out_int8) == j_int8);
+
+                // a negative element is likewise out of range for an
+                // unsigned _ArrayType_
+                json const j_uint16_neg = json({{"_ArrayType_", "uint16"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {1, -1}}});
+                const auto out_uint16_neg = json::to_bjdata(j_uint16_neg);
+                CHECK(out_uint16_neg.at(0) == '{');
+                CHECK(json::from_bjdata(out_uint16_neg) == j_uint16_neg);
+
+                // a double element that overflows to infinity when narrowed
+                // to the "single" (float) precision named by _ArrayType_
+                json const j_single = json({{"_ArrayType_", "single"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {1.5, 1e40}}});
+                const auto out_single = json::to_bjdata(j_single);
+                CHECK(out_single.at(0) == '{');
+                CHECK(json::from_bjdata(out_single) == j_single);
+
+                // in-range boundary values still use the compact ndarray encoding
+                json const j_uint8_ok = json({{"_ArrayType_", "uint8"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {0, 255}}});
+                CHECK(json::to_bjdata(j_uint8_ok) == std::vector<uint8_t>({'[', '$', 'U', '#', '[', 'i', 2, 'i', 1, ']', 0, 255}));
+
+                json const j_int8_ok = json({{"_ArrayType_", "int8"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {-128, 127}}});
+                CHECK(json::to_bjdata(j_int8_ok) == std::vector<uint8_t>({'[', '$', 'i', '#', '[', 'i', 2, 'i', 1, ']', 0x80, 0x7F}));
+
+                json const j_single_ok = json({{"_ArrayType_", "single"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {1.5, -1.5}}});
+                const auto out_single_ok = json::to_bjdata(j_single_ok);
+                CHECK(out_single_ok.at(0) == '[');
+                CHECK(json::from_bjdata(out_single_ok) == json({{"_ArrayType_", "single"}, {"_ArraySize_", {2, 1}}, {"_ArrayData_", {1.5f, -1.5f}}}));
+            }
+
+            SECTION("ndarray that would not be read back as an annotated object stays as object")
+            {
+                // the reader only restores an annotated object from an ND-array
+                // with at least two non-zero dimensions that is not a 1xN row
+                // vector; any other shape is read back as a plain array. Writing
+                // such an object as an ND-array would drop its annotation, so it
+                // falls back to a plain object encoding that round-trips.
+                for (const char* text :
+                        {
+                            R"({"_ArrayType_":"int16","_ArraySize_":[],"_ArrayData_":[]})",
+                            R"({"_ArrayType_":"int16","_ArraySize_":[2],"_ArrayData_":[1,2]})",
+                            R"({"_ArrayType_":"int16","_ArraySize_":[1,2],"_ArrayData_":[1,2]})",
+                            R"({"_ArrayType_":"int16","_ArraySize_":[0],"_ArrayData_":[]})",
+                            R"({"_ArrayType_":"int16","_ArraySize_":[2,0],"_ArrayData_":[]})",
+                            R"({"_ArrayType_":"int16","_ArraySize_":[0,2],"_ArrayData_":[]})"
+                        })
+                {
+                    CAPTURE(text);
+                    const json j = json::parse(text);
+                    for (const bool use_size :
+                            {
+                                false, true
+                            })
+                    {
+                        const auto out = json::to_bjdata(j, use_size, use_size);
+                        CHECK(out.at(0) == '{');
+                        CHECK(json::from_bjdata(out) == j);
+                    }
+                }
+
+                // a genuine ND-array still uses the compact encoding and round-trips
+                const json j_2d = json::parse(R"({"_ArrayType_":"int16","_ArraySize_":[2,1],"_ArrayData_":[1,2]})");
+                const auto out_2d = json::to_bjdata(j_2d);
+                CHECK(out_2d.at(0) == '[');
+                CHECK(json::from_bjdata(out_2d) == j_2d);
+            }
+
+            SECTION("ndarray with non-array _ArrayData_ stays as object")
+            {
+                // the elements are written from _ArrayData_ as a flat list, so it
+                // has to be an array: null has size 0, any other scalar has size 1,
+                // and iterating an object visits its values, so each of these could
+                // match the dimensions and be encoded as an unrelated ND-array
+                for (const char* text :
+                        {
+                            R"({"_ArrayType_":"int16","_ArraySize_":[2,1],"_ArrayData_":null})",
+                            R"({"_ArrayType_":"int16","_ArraySize_":[2,1],"_ArrayData_":{"a":1,"b":2}})",
+                            R"({"_ArrayType_":"int16","_ArraySize_":[1],"_ArrayData_":5})",
+                            R"({"_ArrayType_":"int16","_ArraySize_":[],"_ArrayData_":null})"
+                        })
+                {
+                    CAPTURE(text);
+                    const json j = json::parse(text);
+                    const auto out = json::to_bjdata(j);
+                    CHECK(out.at(0) == '{');
+                    CHECK(json::from_bjdata(out) == j);
+                }
+
+                // OSS-Fuzz issue 563659413: an empty binary _ArraySize_ is written
+                // as a plain object and read back as an empty array, after which
+                // the object with a null _ArrayData_ was encoded as an empty
+                // ND-array and re-read as [], so a second round trip lost the value
+                const std::vector<uint8_t> input =
+                {
+                    '{', 'U', 11, '_', 'A', 'r', 'r', 'a', 'y', 'D', 'a', 't', 'a', '_', 'Z',
+                    'U', 11, '_', 'A', 'r', 'r', 'a', 'y', 'T', 'y', 'p', 'e', '_', 'S', 'i', 5, 'i', 'n', 't', '1', '6',
+                    'U', 11, '_', 'A', 'r', 'r', 'a', 'y', 'S', 'i', 'z', 'e', '_', '[', '$', 'B', '#', '[', ']', '}'
+                };
+                const json j1 = json::from_bjdata(input);
+                const json j2 = json::from_bjdata(json::to_bjdata(j1, false, false));
+                CHECK(j2 == json::parse(R"({"_ArrayType_":"int16","_ArraySize_":[],"_ArrayData_":null})"));
+                CHECK(json::from_bjdata(json::to_bjdata(j2, false, false)) == j2);
+            }
+
+            SECTION("ndarray with _ArrayType_ \"byte\" is gated by the BJData draft version")
+            {
+                // the 'B' (byte) marker used by _ArrayType_ "byte" is only defined
+                // by BJData Draft 3; Draft 2 (the default) has no such marker, so
+                // emitting it unconditionally produced a stream that a Draft 2
+                // reader could not parse as intended (see GitHub issue #5404).
+                // Two dimensions are used so that a successfully written ndarray
+                // round-trips back into the annotated object (a single dimension
+                // is, by the BJData ndarray convention, read back as a plain
+                // binary value rather than the annotated object, same as every
+                // other single-dimension ndarray of a non-"byte" type is read
+                // back as a plain array instead of the annotated object).
+                json const j_byte = json({{"_ArrayType_", "byte"}, {"_ArraySize_", {2, 3}}, {"_ArrayData_", {1, 2, 3, 4, 5, 6}}});
+
+                // default (Draft 2): falls back to a plain object and round-trips
+                const auto out_draft2 = json::to_bjdata(j_byte);
+                CHECK(out_draft2.at(0) == '{');
+                CHECK(json::from_bjdata(out_draft2) == j_byte);
+
+                // explicit Draft 2: same as the default
+                const auto out_draft2_explicit = json::to_bjdata(j_byte, true, true, json::bjdata_version_t::draft2);
+                CHECK(out_draft2_explicit.at(0) == '{');
+                CHECK(json::from_bjdata(out_draft2_explicit) == j_byte);
+
+                // Draft 3 explicitly selected: still uses the compact 'B' ndarray encoding
+                const auto out_draft3 = json::to_bjdata(j_byte, true, true, json::bjdata_version_t::draft3);
+                CHECK(out_draft3 == std::vector<uint8_t>({'[', '$', 'B', '#', '[', '$', 'i', '#', 'i', 2, 2, 3, 1, 2, 3, 4, 5, 6}));
+                CHECK(json::from_bjdata(out_draft3) == j_byte);
+            }
         }
     }
 
@@ -3242,8 +3547,10 @@ TEST_CASE("BJData")
             CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vR1), "[json.exception.parse_error.113] parse error at byte 6: syntax error while parsing BJData size: ndarray dimensional vector is not allowed", json::parse_error&);
             CHECK(json::from_bjdata(vR1, true, false).is_discarded());
 
+            // a dimension vector that opens another one is rejected where the
+            // nested '[' is read, rather than after it has been descended into
             std::vector<uint8_t> const vR2 = {'[', '$', 'i', '#', '[', '#', '[', 'i', 1, ']', ']', 1};
-            CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vR2), "[json.exception.parse_error.113] parse error at byte 11: syntax error while parsing BJData size: expected length type specification (U, i, u, I, m, l, M, L) after '#'; last byte: 0x5D", json::parse_error&);
+            CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vR2), "[json.exception.parse_error.113] parse error at byte 7: syntax error while parsing BJData size: ndarray dimensional vector is not allowed", json::parse_error&);
             CHECK(json::from_bjdata(vR2, true, false).is_discarded());
 
             std::vector<uint8_t> const vR3 = {'[', '#', '[', 'i', '2', 'i', 2, ']'};
@@ -3251,7 +3558,7 @@ TEST_CASE("BJData")
             CHECK(json::from_bjdata(vR3, true, false).is_discarded());
 
             std::vector<uint8_t> const vR4 = {'[', '$', 'i', '#', '[', '$', 'i', '#', '[', 'i', 1, ']', 1};
-            CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vR4), "[json.exception.parse_error.110] parse error at byte 14: syntax error while parsing BJData number: unexpected end of input", json::parse_error&);
+            CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vR4), "[json.exception.parse_error.113] parse error at byte 9: syntax error while parsing BJData size: ndarray dimensional vector is not allowed", json::parse_error&);
             CHECK(json::from_bjdata(vR4, true, false).is_discarded());
 
             std::vector<uint8_t> const vR5 = {'[', '$', 'i', '#', '[', '[', '[', ']', ']', ']'};
@@ -3259,12 +3566,25 @@ TEST_CASE("BJData")
             CHECK(json::from_bjdata(vR5, true, false).is_discarded());
 
             std::vector<uint8_t> const vR6 = {'[', '$', 'i', '#', '[', '$', 'i', '#', '[', 'i', '2', 'i', 2, ']'};
-            CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vR6), "[json.exception.parse_error.112] parse error at byte 14: syntax error while parsing BJData size: ndarray can not be recursive", json::parse_error&);
+            CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vR6), "[json.exception.parse_error.113] parse error at byte 9: syntax error while parsing BJData size: ndarray dimensional vector is not allowed", json::parse_error&);
             CHECK(json::from_bjdata(vR6, true, false).is_discarded());
 
             std::vector<uint8_t> const vH = {'[', 'H', '[', '#', '[', '$', 'i', '#', '[', 'i', '2', 'i', 2, ']'};
             CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vH), "[json.exception.parse_error.113] parse error at byte 3: syntax error while parsing BJData size: ndarray dimensional vector is not allowed", json::parse_error&);
             CHECK(json::from_bjdata(vH, true, false).is_discarded());
+
+            // Every "#[" of this chain used to open another dimension vector
+            // and cost several stack frames before anything was rejected, so a
+            // long enough chain crashed the process (see #5104). The nested
+            // vector is refused where it is read, so the length is irrelevant.
+            std::vector<uint8_t> vRdeep = {'['};
+            for (std::size_t i = 0; i < 100000; ++i)
+            {
+                vRdeep.push_back('#');
+                vRdeep.push_back('[');
+            }
+            CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vRdeep), "[json.exception.parse_error.113] parse error at byte 5: syntax error while parsing BJData size: ndarray dimensional vector is not allowed", json::parse_error&);
+            CHECK(json::from_bjdata(vRdeep, true, false).is_discarded());
         }
 
         SECTION("objects")
@@ -3443,6 +3763,154 @@ TEST_CASE("BJData")
     }
 }
 
+TEST_CASE("BJData input that cannot be read is discarded by every overload")
+{
+    std::vector<std::uint8_t> input = json::to_bjdata(json({{"a", {1, 2}}}));
+    input.pop_back();
+
+    json _;
+    CHECK_THROWS_AS(_ = json::from_bjdata(input.begin(), input.end()), json::parse_error&);
+    CHECK(json::from_bjdata(input, true, false).is_discarded());
+    CHECK(json::from_bjdata(input.begin(), input.end(), true, false).is_discarded());
+}
+
+TEST_CASE("BJData SAX parsing stops at every event")
+{
+    // Containers are opened and closed by the loop that reads them; a SAX
+    // handler that rejects any event - including the end of a nested
+    // container - must stop the parse right there.
+    const auto count_events = [](const std::vector<std::uint8_t>& input)
+    {
+        int events = 0;
+        while (true)
+        {
+            SaxCountdown scp(events);
+            if (json::sax_parse(input, &scp, json::input_format_t::bjdata))
+            {
+                return events;
+            }
+            ++events;
+            REQUIRE(events < 1000);
+        }
+    };
+
+    // 20 events: every container kind closes inside another one
+    const json j = json::parse(R"({"a": [1, {"b": []}], "c": {"d": [[2]]}})");
+    CHECK(count_events(json::to_bjdata(j)) == 20);
+    CHECK(count_events(json::to_bjdata(j, true)) == 20);
+    CHECK(count_events(json::to_bjdata(j, true, true)) == 20);
+
+    // an ND-array is announced as an annotated object: start_object, then
+    // _ArrayType_, _ArraySize_ and _ArrayData_ with its elements
+    const json ndarray = json::parse(R"({"_ArrayType_": "uint8", "_ArraySize_": [2, 2], "_ArrayData_": [1, 2, 3, 4]})");
+    CHECK(count_events(json::to_bjdata(ndarray, true, true)) == 16);
+}
+
+TEST_CASE("issue #5405 - array reserve for definite-length BJData arrays")
+{
+#if !defined(JSON_NOEXCEPTION)
+    // this SECTION relies on catching a thrown exception to distinguish
+    // which of two acceptable, bounded rejections a hostile header took;
+    // under JSON_NOEXCEPTION, JSON_THROW never produces a catchable C++
+    // exception (it aborts instead), so this cannot be tested that way here
+    SECTION("a huge claimed length with no element data must not over-allocate")
+    {
+        // optimized form [$type#count: type 'i' (int8), count as a four-byte
+        // little-endian 'l' (int32) of 0x7FFFFFFF (2147483647), but no
+        // element data at all. max_size() for a std::vector is far larger
+        // than this count, so it does not reject the header outright; the
+        // (capped) reservation must not attempt to allocate space for
+        // billions of elements before the missing data is detected.
+        json _;
+        const std::vector<uint8_t> input = {'[', '$', 'i', '#', 'l', 0xFF, 0xFF, 0xFF, 0x7F};
+        // On a platform where std::vector<json>::max_size() is smaller than
+        // the claimed count (e.g. 32-bit, where max_size() is bounded by a
+        // 32-bit SIZE_MAX divided by sizeof(json)), the SAX consumer's own
+        // check rejects the header outright (out_of_range.408, with the
+        // claimed count in the message) instead of accepting it and only
+        // finding it short of data once the (capped) reservation looks for
+        // element bytes that were never provided (parse_error.110). Either
+        // is an acceptable, bounded rejection of the hostile header -- the
+        // property under test is that no path attempts to allocate space
+        // for billions of elements.
+        bool threw = false;
+        try
+        {
+            _ = json::from_bjdata(input);
+        }
+        catch (const json::parse_error& e)
+        {
+            threw = true;
+            CHECK(e.id == 110);
+            CHECK(std::string(e.what()) == "[json.exception.parse_error.110] parse error at byte 10: syntax error while parsing BJData number: unexpected end of input");
+        }
+        catch (const json::out_of_range& e)
+        {
+            threw = true;
+            CHECK(e.id == 408);
+            CHECK(std::string(e.what()).find("excessive array size") != std::string::npos);
+        }
+        CHECK(threw);
+
+        // json_sax_dom_parser::start_array()'s max_size() check (unlike the
+        // scanner's own parse_error path) throws unconditionally via
+        // JSON_THROW rather than going through sax->parse_error(), so it is
+        // not gated by allow_exceptions=false on a platform where this
+        // header hits that check (e.g. 32-bit, see above) -- allow either
+        // a discarded result or the same out_of_range it throws with
+        // exceptions enabled.
+        try
+        {
+            CHECK(json::from_bjdata(input, true, false).is_discarded());
+        }
+        catch (const json::out_of_range& e)
+        {
+            CHECK(e.id == 408);
+        }
+    }
+#endif
+
+    SECTION("arrays of various sizes decode to the same value as before the reserve optimization")
+    {
+        for (const auto size :
+                {
+                    std::size_t{0}, std::size_t{1}, std::size_t{5}, // small
+                    std::size_t{16384},                             // exactly at the reserve cap
+                    std::size_t{20000}                              // above the reserve cap
+                })
+        {
+            CAPTURE(size)
+            json j = json::array();
+            for (std::size_t i = 0; i < size; ++i)
+            {
+                j.push_back(static_cast<int>(i % 1000));
+            }
+
+            // exercise both the plain and the optimized [$type#count encoding
+            const auto packed_plain = json::to_bjdata(j);
+            CHECK(json::from_bjdata(packed_plain) == j);
+
+            const auto packed_optimized = json::to_bjdata(j, true, true);
+            CHECK(json::from_bjdata(packed_optimized) == j);
+        }
+    }
+
+    SECTION("a user-defined SAX consumer is unaffected by the internal DOM reserve optimization")
+    {
+        // the reserve() call is local to json_sax_dom_parser / json_sax_dom_callback_parser;
+        // a custom SAX consumer that does not touch a DOM array sees identical events
+        json j = json::array();
+        for (int i = 0; i < 100; ++i)
+        {
+            j.push_back(i);
+        }
+        const auto packed = json::to_bjdata(j, true, true);
+
+        SaxCountdown scp(1000000); // large enough to never trigger an abort
+        CHECK(json::sax_parse(packed, &scp, json::input_format_t::bjdata));
+    }
+}
+
 TEST_CASE("Universal Binary JSON Specification Examples 1")
 {
     SECTION("Null Value")
@@ -3822,6 +4290,205 @@ TEST_CASE("all BJData first bytes")
 }
 #endif
 
+TEST_CASE("BJData and UBJSON can be written to a string")
+{
+    const std::vector<json> values =
+    {
+        {{"a", {1, 2.5, "x", nullptr}}, {"b", json::binary({1, 2})}},
+        // an annotated ND-array, and objects that only look like one
+        json::parse(R"({"_ArrayType_": "uint8", "_ArraySize_": [2, 2], "_ArrayData_": [1, 2, 3, 4]})"),
+        json::parse(R"({"_ArrayType_": 1, "_ArraySize_": [2, 2], "_ArrayData_": [1, 2, 3, 4]})"),
+        json::parse(R"({"_ArrayType_": "uint8", "_ArraySize_": 4, "_ArrayData_": [1, 2, 3, 4]})"),
+        json::parse(R"({"_ArrayType_": "uint8", "_ArraySize_": [2, -2], "_ArrayData_": [1, 2, 3, 4]})"),
+        json::parse(R"({"_ArrayType_": "uint8", "_ArraySize_": [2, 2], "_ArrayData_": [1, 2, 3]})"),
+        json::parse(R"({"_ArrayType_": "uint8", "_ArraySize_": [2, 2], "_ArrayData_": 1})"),
+    };
+
+    // compared byte by byte: building a std::string from the bytes would
+    // convert them implicitly, which -fsanitize=integer reports for bytes of
+    // 0x80 and above
+    const auto same_bytes = [](const std::vector<std::uint8_t>& bytes, const std::string & text)
+    {
+        return bytes.size() == text.size() && std::equal(bytes.begin(), bytes.end(), text.begin(), [](std::uint8_t byte, char c)
+        {
+            return byte == static_cast<std::uint8_t>(c);
+        });
+    };
+
+    for (const auto& j : values)
+    {
+        CAPTURE(j.dump());
+        for (const bool use_size :
+                {
+                    false, true
+                })
+        {
+            for (const bool use_type :
+                    {
+                        false, true
+                    })
+            {
+                if (use_type && !use_size)
+                {
+                    continue;
+                }
+                CAPTURE(use_size);
+                CAPTURE(use_type);
+
+                const auto bjdata = json::to_bjdata(j, use_size, use_type);
+                std::string bjdata_string;
+                json::to_bjdata(j, bjdata_string, use_size, use_type);
+                CHECK(same_bytes(bjdata, bjdata_string));
+
+                const auto ubjson = json::to_ubjson(j, use_size, use_type);
+                std::string ubjson_string;
+                json::to_ubjson(j, ubjson_string, use_size, use_type);
+                CHECK(same_bytes(ubjson, ubjson_string));
+            }
+        }
+    }
+}
+
+TEST_CASE("BJData use_type requires use_size")
+{
+    SECTION("non-empty object throws other_error.502")
+    {
+        const json j = {{"a", 1}, {"b", 2}};
+        CHECK_THROWS_WITH_AS(json::to_bjdata(j, false, true),
+                             "[json.exception.other_error.502] use_type requires use_size = true",
+                             json::other_error&);
+    }
+
+    SECTION("non-empty array throws other_error.502")
+    {
+        const json j = {1, 2, 3};
+        CHECK_THROWS_WITH_AS(json::to_bjdata(j, false, true),
+                             "[json.exception.other_error.502] use_type requires use_size = true",
+                             json::other_error&);
+    }
+
+    SECTION("non-empty binary value throws other_error.502")
+    {
+        const json j = json::binary({1, 2, 3});
+        CHECK_THROWS_WITH_AS(json::to_bjdata(j, false, true),
+                             "[json.exception.other_error.502] use_type requires use_size = true",
+                             json::other_error&);
+        CHECK_THROWS_WITH_AS(json::to_ubjson(j, false, true),
+                             "[json.exception.other_error.502] use_type requires use_size = true",
+                             json::other_error&);
+    }
+
+    SECTION("scalars do not throw with use_type=true, use_count=false")
+    {
+        CHECK_NOTHROW(json::to_bjdata(42, false, true));
+        CHECK_NOTHROW(json::to_bjdata(3.14, false, true));
+        CHECK_NOTHROW(json::to_bjdata("hello", false, true));
+        CHECK_NOTHROW(json::to_bjdata(true, false, true));
+        CHECK_NOTHROW(json::to_bjdata(nullptr, false, true));
+    }
+
+    SECTION("empty containers do not throw with use_type=true, use_count=false")
+    {
+        CHECK_NOTHROW(json::to_bjdata(json::array(), false, true));
+        CHECK_NOTHROW(json::to_bjdata(json::object(), false, true));
+    }
+
+    SECTION("valid combinations on non-empty containers")
+    {
+        const json j = {{"a", 1}, {"b", 2}};
+        CHECK_NOTHROW(json::to_bjdata(j, false, false));
+        CHECK_NOTHROW(json::to_bjdata(j, true, false));
+        CHECK_NOTHROW(json::to_bjdata(j, true, true));
+    }
+}
+
+TEST_CASE("BJData round-trip invariants")
+{
+    // This checks what the parse_bjdata_fuzzer driver checks (see
+    // tests/src/fuzzer-parse_bjdata.cpp), so that a regression shows up in CI
+    // rather than as an OSS-Fuzz report: every value from_bjdata() returns
+    // (j1) can be serialized with any combination of options, the result can
+    // be parsed back (j2), and serializing j2 again with the same options
+    // yields a value-equal result.
+    //
+    // Beyond the driver, this also checks that j2 equals j1 and that
+    // serializing j2 reproduces the exact bytes, both except for values that
+    // contain a binary value: a binary value is only written as a binary
+    // value with Draft 3's optimized binary array, and otherwise read back as
+    // an array of integers, for which the writer may choose different (but
+    // equally valid) type markers when it is serialized again (see #5494).
+    //
+    // Values are compared with dump() rather than operator==, because a NaN
+    // never compares equal to itself.
+    struct options
+    {
+        bool use_size;
+        bool use_type;
+        json::bjdata_version_t version;
+    };
+    const std::vector<options> all_options =
+    {
+        {false, false, json::bjdata_version_t::draft2},
+        {true, false, json::bjdata_version_t::draft2},
+        {true, true, json::bjdata_version_t::draft2},
+        {false, false, json::bjdata_version_t::draft3},
+        {true, false, json::bjdata_version_t::draft3},
+        {true, true, json::bjdata_version_t::draft3},
+    };
+
+    for (const auto& j0 : utils::round_trip_corpus::values())
+    {
+        // turn the corpus value into a value as from_bjdata() returns it
+        for (const auto& initial : all_options)
+        {
+            const json j1 = json::from_bjdata(json::to_bjdata(j0, initial.use_size, initial.use_type, initial.version));
+            const bool has_binary = utils::round_trip_corpus::contains_binary(j1);
+
+            for (const auto& o : all_options)
+            {
+                INFO("j1 = " << j1.dump() << ", use_size = " << o.use_size << ", use_type = " << o.use_type
+                     << ", draft3 = " << (o.version == json::bjdata_version_t::draft3));
+
+                const std::vector<std::uint8_t> vec = json::to_bjdata(j1, o.use_size, o.use_type, o.version);
+                json j2;
+                // anything the library writes must be parsable by the library
+                REQUIRE_NOTHROW(j2 = json::from_bjdata(vec));
+                const std::vector<std::uint8_t> vec2 = json::to_bjdata(j2, o.use_size, o.use_type, o.version);
+                CHECK(json::from_bjdata(vec2).dump() == j2.dump());
+
+                if (!has_binary)
+                {
+                    CHECK(j2.dump() == j1.dump());
+                    CHECK(vec2 == vec);
+                }
+            }
+        }
+    }
+}
+
+TEST_CASE("BJData round trip of a binary value is value-stable, not byte-stable")
+{
+    // OSS-Fuzz issue 474480402: a Draft 3 optimized binary array is read as a
+    // binary value, which to_bjdata() writes in the default Draft 2 mode as a
+    // plain array of uint8 numbers. That is read back as an array of numbers,
+    // for which the writer then picks the smallest type marker, int8 ('i'),
+    // so re-serializing changes the bytes, but not the value. This is the
+    // exception described in the "Round trips" note of the BJData
+    // documentation, and why the fuzzer checks value stability (see #5494).
+    const std::vector<uint8_t> input = {'[', '$', 'B', '#', 'U', 1, 0x20};
+    const json j1 = json::from_bjdata(input);
+    CHECK(j1 == json::binary({0x20}));
+
+    const std::vector<uint8_t> vec = json::to_bjdata(j1, false, false);
+    CHECK(vec == std::vector<uint8_t>({'[', 'U', 0x20, ']'}));
+    const json j2 = json::from_bjdata(vec);
+    CHECK(j2 == json::array({0x20}));
+
+    const std::vector<uint8_t> vec2 = json::to_bjdata(j2, false, false);
+    CHECK(vec2 == std::vector<uint8_t>({'[', 'i', 0x20, ']'}));
+    CHECK(json::from_bjdata(vec2) == j2);
+}
+
 TEST_CASE("BJData roundtrips" * doctest::skip())
 {
     SECTION("input from self-generated BJData files")
@@ -3874,45 +4541,27 @@ TEST_CASE("BJData roundtrips" * doctest::skip())
         {
             CAPTURE(filename)
 
+            std::ifstream f_json(filename);
+            const json j1 = json::parse(f_json);
+            auto packed = utils::read_binary_file(filename + ".bjdata");
+
             {
                 INFO_WITH_TEMP(filename + ": std::vector<uint8_t>");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                const json j1 = json::parse(f_json);
-
-                // parse BJData file
-                auto packed = utils::read_binary_file(filename + ".bjdata");
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_bjdata(packed));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": std::ifstream");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                const json j1 = json::parse(f_json);
-
-                // parse BJData file
                 std::ifstream f_bjdata(filename + ".bjdata", std::ios::binary);
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_bjdata(f_bjdata));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": output to output adapters");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                json const j1 = json::parse(f_json);
-
-                // parse BJData file
-                auto packed = utils::read_binary_file(filename + ".bjdata");
-
                 {
                     INFO_WITH_TEMP(filename + ": output adapters: std::vector<uint8_t>");
                     std::vector<uint8_t> vec;

tests/src/unit-brace-init-copy-semantics.cpp

@@ -0,0 +1,167 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+#include "doctest_compatibility.h"
+
+// This file tests the opt-in JSON_BRACE_INIT_COPY_SEMANTICS, so it defines the
+// macro itself rather than relying on a -D flag, and runs in every build.
+#ifdef JSON_BRACE_INIT_COPY_SEMANTICS
+    #undef JSON_BRACE_INIT_COPY_SEMANTICS
+#endif
+
+#define JSON_BRACE_INIT_COPY_SEMANTICS 1
+
+#include <nlohmann/json.hpp>
+using nlohmann::json;
+
+#include <array>
+#include <list>
+#include <map>
+#include <string>
+#include <tuple>
+#include <utility>
+#include <vector>
+
+#define STRINGIZE_EX(x) #x
+#define STRINGIZE(x) STRINGIZE_EX(x)
+
+TEST_CASE("JSON_BRACE_INIT_COPY_SEMANTICS")
+{
+    SECTION("the macro is part of the ABI tag")
+    {
+        const std::string ns = STRINGIZE(NLOHMANN_JSON_NAMESPACE);
+        // other tags may come before it, e.g. json_abi_ldvcmp_bics
+        CHECK(ns.find("_bics") != std::string::npos);
+    }
+
+    SECTION("single-element brace initialization copies the element (#5074)")
+    {
+        json const j_obj = {{"key", "value"}, {"num", 42}};
+        json const j_arr = {1, 2, 3};
+
+        // object: brace init copies instead of wrapping
+        json const j1{j_obj};
+        CHECK(j1.is_object());
+        CHECK(j1 == j_obj);
+
+        // array: brace init copies instead of wrapping
+        json const j2{j_arr};
+        CHECK(j2.is_array());
+        CHECK(j2.size() == 3);
+        CHECK(j2 == j_arr);
+
+        // this applies to any single element, not only to JSON values
+        json const j3{true};
+        CHECK(j3.is_boolean());
+
+        json const j4{42};
+        CHECK(j4.is_number_integer());
+
+        json const j5 = {1};
+        CHECK(j5 == 1);
+
+        json const j6 = {"text"};
+        CHECK(j6 == "text");
+
+        json const j7 = {{1, 2}};
+        CHECK(j7 == json::array({1, 2}));
+    }
+
+    SECTION("what the macro does not change")
+    {
+        // lists with more than one element are unaffected
+        json const j1 = {1, 2};
+        CHECK(j1.is_array());
+        CHECK(j1.size() == 2);
+
+        // a single [string, value] pair still describes an object
+        json const j2 = {{"key", "value"}};
+        CHECK(j2.is_object());
+        CHECK(j2["key"] == "value");
+
+        // json::array() always creates an array
+        json const j3 = json::array({1});
+        CHECK(j3.is_array());
+        CHECK(j3.size() == 1);
+        CHECK(j3[0] == 1);
+
+        json const j_obj = {{"key", "value"}};
+        json const j4 = json::array({j_obj});
+        CHECK(j4.is_array());
+        CHECK(j4.size() == 1);
+        CHECK(j4[0] == j_obj);
+    }
+
+    SECTION("conversions build the same values as without the macro")
+    {
+        SECTION("one-element std::tuple")
+        {
+            json const j1 = std::tuple<int> {5};
+            CHECK(j1.dump() == "[5]");
+            CHECK(std::get<0>(j1.get<std::tuple<int>>()) == 5);
+
+            json const j2 = std::tuple<std::string> {"text"};
+            CHECK(j2.dump() == "[\"text\"]");
+            CHECK(std::get<0>(j2.get<std::tuple<std::string>>()) == "text");
+
+            json const j3 = std::tuple<json> {json::array({1, 2})};
+            CHECK(j3.dump() == "[[1,2]]");
+
+            // as without the macro, a [string, value] pair becomes an object
+            // member (see the known limitation documented for std::pair)
+            json const j4 = std::tuple<std::pair<std::string, int>> {{"a", 1}};
+            CHECK(j4.dump() == "{\"a\":1}");
+        }
+
+        SECTION("tuples with more elements")
+        {
+            json const j1 = std::tuple<int, std::string> {1, "a"};
+            CHECK(j1.dump() == "[1,\"a\"]");
+
+            json const j2 = std::tuple<> {};
+            CHECK(j2.dump() == "[]");
+        }
+
+        SECTION("one-element containers")
+        {
+            json const j1 = std::vector<int> {1};
+            CHECK(j1.dump() == "[1]");
+            CHECK(j1.get<std::vector<int>>() == std::vector<int> {1});
+
+            std::array<int, 1> const arr = {{1}};
+            json const j2 = arr;
+            CHECK(j2.dump() == "[1]");
+
+            json const j3 = std::list<std::string> {"a"};
+            CHECK(j3.dump() == "[\"a\"]");
+
+            json const j4 = std::map<std::string, int> {{"a", 1}};
+            CHECK(j4.dump() == "{\"a\":1}");
+
+            json const j5 = std::map<int, int> {{1, 2}};
+            CHECK(j5.dump() == "[[1,2]]");
+        }
+
+        SECTION("std::pair")
+        {
+            json const j = std::pair<int, int> {1, 2};
+            CHECK(j.dump() == "[1,2]");
+            CHECK((j.get<std::pair<int, int>>() == std::pair<int, int> {1, 2}));
+        }
+
+        SECTION("items()")
+        {
+            json j_obj = {{"key", 1}};
+            for (const auto& el : j_obj.items())
+            {
+                json const j = el;
+                CHECK(j.dump() == "{\"key\":1}");
+            }
+        }
+    }
+}

tests/src/unit-bson.cpp

@@ -38,6 +38,54 @@ class huge_binary_t : public std::vector<std::uint8_t>
 using huge_binary_json = nlohmann::basic_json <
                          std::map, std::vector, std::string, bool, std::int64_t, std::uint64_t,
                          double, std::allocator, nlohmann::adl_serializer, huge_binary_t, void >;
+
+// a string type that can be made to report a size beyond INT32_MAX without
+// allocating that much memory, so BSON length overflow can be tested for
+// strings and (embedded) documents as well, following the same idea as
+// huge_binary_t.
+//
+// Unlike huge_binary_t (which is only ever used as the BSON *value* type),
+// this type doubles as basic_json's StringType and is therefore also used
+// for *object keys* (e.g. "s" or "nested" below). Only the designated test
+// value is meant to lie about its size - if every huge_string_t (including
+// keys) reported a huge size, the running totals computed while walking the
+// BSON document (see calc_bson_sizes in binary_writer.hpp)
+// would need more than 32 bits, and on platforms where std::size_t is only
+// 32 bits wide that arithmetic would silently wrap around, producing wrong
+// (or even unguarded) lengths. The fake size is therefore opt-in via
+// as_huge(), and plain strings - in particular object keys - keep reporting
+// their real, small size.
+class huge_string_t : public std::string
+{
+  public:
+    using std::string::string;
+    huge_string_t(const std::string& s) : std::string(s) {} // NOLINT(google-explicit-constructor,hicpp-explicit-conversions)
+
+    // returns a copy of @a s whose size() pretends to be huge
+    static huge_string_t as_huge(const std::string& s)
+    {
+        huge_string_t result(s);
+        result.pretend_huge = true;
+        return result;
+    }
+
+    size_type size() const noexcept
+    {
+        if (pretend_huge)
+        {
+            // one byte more than the BSON length field can represent
+            return static_cast<size_type>((std::numeric_limits<std::int32_t>::max)()) + 1;
+        }
+        return std::string::size();
+    }
+
+  private:
+    bool pretend_huge = false;
+};
+
+using huge_string_json = nlohmann::basic_json <
+                         std::map, std::vector, huge_string_t, bool, std::int64_t, std::uint64_t,
+                         double, std::allocator, nlohmann::adl_serializer, std::vector<std::uint8_t>, void >;
 } // namespace
 
 TEST_CASE("BSON")
@@ -105,10 +153,36 @@ TEST_CASE("BSON")
 
     SECTION("lengths exceeding INT32_MAX cannot be serialized to BSON")
     {
-        huge_binary_json j;
-        j["b"] = huge_binary_json::binary(huge_binary_t{});
+        // out_of_range.412 is thrown from a single shared helper
+        // (to_bson_length) that guards the BSON length fields of binary
+        // values, strings, and (embedded) documents alike
+        SECTION("binary")
+        {
+            huge_binary_json j;
+            j["b"] = huge_binary_json::binary(huge_binary_t{});
 
-        CHECK_THROWS_WITH_AS(huge_binary_json::to_bson(j), "[json.exception.out_of_range.412] BSON length 2147483661 exceeds maximum of 2147483647", huge_binary_json::out_of_range&);
+            CHECK_THROWS_WITH_AS(huge_binary_json::to_bson(j), "[json.exception.out_of_range.412] BSON length 2147483661 exceeds maximum of 2147483647", huge_binary_json::out_of_range&);
+        }
+
+        SECTION("string")
+        {
+            huge_string_json j;
+            j["s"] = huge_string_t::as_huge("value");
+
+            CHECK_THROWS_WITH_AS(huge_string_json::to_bson(j), "[json.exception.out_of_range.412] BSON length 2147483661 exceeds maximum of 2147483647", huge_string_json::out_of_range&);
+        }
+
+        SECTION("document")
+        {
+            // an oversized string nested one level deep makes the
+            // *embedded* document's own length exceed INT32_MAX as well
+            huge_string_json nested;
+            nested["s"] = huge_string_t::as_huge("value");
+            huge_string_json j;
+            j["nested"] = nested;
+
+            CHECK_THROWS_WITH_AS(huge_string_json::to_bson(j), "[json.exception.out_of_range.412] BSON length 2147483674 exceeds maximum of 2147483647", huge_string_json::out_of_range&);
+        }
     }
 
     SECTION("string length must be at least 1")
@@ -193,6 +267,23 @@ TEST_CASE("BSON")
             CHECK(json::from_bson(result, true, false) == j);
         }
 
+        SECTION("non-empty object with bool from a non-0/1 byte (lenient parsing)")
+        {
+            // documented lenient behavior (see gh-5333): any non-zero byte
+            // is accepted as `true`, not just 0x01
+            std::vector<std::uint8_t> const input =
+            {
+                0x0D, 0x00, 0x00, 0x00, // size (little endian)
+                0x08,               // entry: boolean
+                'e', 'n', 't', 'r', 'y', '\x00',
+                0x02,           // value = 0x02 (neither 0x00 nor 0x01)
+                0x00                    // end marker
+            };
+
+            const json expected = { { "entry", true } };
+            CHECK(json::from_bson(input) == expected);
+        }
+
         SECTION("non-empty object with double")
         {
             json const j =
@@ -499,6 +590,29 @@ TEST_CASE("BSON")
             CHECK(json::from_bson(result, true, false) == j);
         }
 
+        SECTION("array elements with non-conforming keys (lenient parsing)")
+        {
+            // documented lenient behavior (see gh-5333): BSON array element
+            // keys are not checked against the required decimal sequence
+            // "0", "1", "2", ... - elements are taken in encoded order
+            std::vector<std::uint8_t> const input =
+            {
+                0x26, 0x00, 0x00, 0x00, // size (little endian)
+                0x04, 'e', 'n', 't', 'r', 'y', '\x00', // entry: embedded array
+
+                0x1A, 0x00, 0x00, 0x00, // size (little endian)
+                0x10, '5', 0x00, 0x0A, 0x00, 0x00, 0x00, // key "5" (bogus)      -> 10
+                0x10, 'x', 0x00, 0x14, 0x00, 0x00, 0x00, // key "x" (non-numeric) -> 20
+                0x10, '1', 0x00, 0x1E, 0x00, 0x00, 0x00, // key "1" (out of order) -> 30
+                0x00, // end marker (embedded array)
+
+                0x00 // end marker
+            };
+
+            const json expected = { { "entry", json::array({10, 20, 30}) } };
+            CHECK(json::from_bson(input) == expected);
+        }
+
         SECTION("non-empty object with binary member")
         {
             const size_t N = 10;
@@ -594,6 +708,31 @@ TEST_CASE("BSON")
             CHECK(json::from_bson(result, true, false) == j);
         }
 
+        SECTION("binary member with subtype 0x02 (old binary) keeps its inner length prefix (lenient parsing)")
+        {
+            // documented lenient behavior (see gh-5333): the payload for
+            // binary subtype 0x02 ("old binary") is returned as-is,
+            // including its own inner 4-byte length prefix; it is not
+            // stripped or reinterpreted
+            std::vector<std::uint8_t> const input =
+            {
+                0x17, 0x00, 0x00, 0x00, // size (little endian)
+                0x05, 'e', 'n', 't', 'r', 'y', '\x00', // entry: binary
+
+                0x06, 0x00, 0x00, 0x00, // size of binary (little endian)
+                0x02, // "old binary" subtype
+                0x02, 0x00, 0x00, 0x00, // inner length prefix (part of the old-binary payload)
+                0x68, 0x69, // payload ('h', 'i')
+
+                0x00 // end marker
+            };
+
+            // the inner length prefix is part of the (unmodified) payload
+            const std::vector<std::uint8_t> expected_payload = {0x02, 0x00, 0x00, 0x00, 0x68, 0x69};
+            const json expected = { { "entry", json::binary(expected_payload, 0x02) } };
+            CHECK(json::from_bson(input) == expected);
+        }
+
         SECTION("Some more complex document")
         {
             json const j =
@@ -652,6 +791,15 @@ TEST_CASE("BSON")
     }
 }
 
+TEST_CASE("regression test - BSON binary subtype rejects a value that doesn't fit a single byte")
+{
+    json const doc255 = {{"b", json::binary({1, 2}, 255)}};
+    CHECK(json::from_bson(json::to_bson(doc255))["b"].get_binary().subtype() == 255);
+
+    CHECK_THROWS_AS(json::to_bson(json{{"b", json::binary({1, 2}, 256)}}), json::out_of_range);
+    CHECK_THROWS_WITH_AS(json::to_bson(json{{"b", json::binary({1, 2}, 300)}}), "[json.exception.out_of_range.415] subtype 300 is too large for the BSON binary subtype (max 255)", json::out_of_range);
+}
+
 TEST_CASE("BSON input/output_adapters")
 {
     const json json_representation =
@@ -918,6 +1066,45 @@ TEST_CASE("Incomplete BSON Input")
     }
 }
 
+// the test catches the exceptions of invalid input
+#if !defined(JSON_NOEXCEPTION)
+TEST_CASE("BSON keys from contiguous and stream input")
+{
+    // contiguous input reads a key up to its \x00-byte in one step, a stream
+    // reads it byte by byte; both must give the same value or error for the
+    // complete document and for every truncation of it
+    const json j = {{"", true}, {"k", {1, 2, 3}}, {std::string(40, 'x'), {{"nested key", "value"}}}};
+    const std::vector<std::uint8_t> bson = json::to_bson(j);
+    CHECK(json::from_bson(bson) == j);
+
+    for (std::size_t length = 0; length <= bson.size(); ++length)
+    {
+        CAPTURE(length)
+        const std::vector<std::uint8_t> input(bson.begin(), bson.begin() + static_cast<std::ptrdiff_t>(length));
+        std::string from_vector;
+        std::string from_stream;
+        try
+        {
+            from_vector = json::from_bson(input).dump();
+        }
+        catch (const json::parse_error& e)
+        {
+            from_vector = e.what();
+        }
+        try
+        {
+            std::istringstream stream(std::string(input.begin(), input.end()));
+            from_stream = json::from_bson(stream).dump();
+        }
+        catch (const json::parse_error& e)
+        {
+            from_stream = e.what();
+        }
+        CHECK(from_vector == from_stream);
+    }
+}
+#endif
+
 TEST_CASE("Negative size of binary value")
 {
     // invalid BSON: the size of the binary value is -1
@@ -1011,6 +1198,129 @@ TEST_CASE("BSON document size mismatch")
     }
 }
 
+TEST_CASE("BSON nesting does not consume the call stack")
+{
+    // An embedded document or array used to be read by calling back into the
+    // document reader, so the native call stack grew with the nesting depth of
+    // the input (#5104). The open documents are kept on a heap stack now.
+    //
+    // Deeply nested values must not be compared, copied or dumped here: those
+    // operations are still recursive and would reintroduce the crash.
+
+    // A document nested deeply enough to have crashed. The bytes are built
+    // here rather than with to_bson(), because the writer still recurses once
+    // per level and would overflow the stack before the reader is ever
+    // reached. Every level is
+    //     <int32 size> 0x03 'a' 0x00 <inner document> 0x00
+    // so a level is eight bytes larger than the one it holds, and the sizes
+    // can be filled in from the outside in.
+    const std::size_t depth = 30000;
+    std::vector<uint8_t> input;
+    input.reserve(5 + (8 * depth));
+    for (std::size_t i = 0; i < depth; ++i)
+    {
+        const auto size = static_cast<std::uint32_t>(5 + (8 * (depth - i)));
+        input.push_back(static_cast<uint8_t>(size & 0xFF));
+        input.push_back(static_cast<uint8_t>((size >> 8) & 0xFF));
+        input.push_back(static_cast<uint8_t>((size >> 16) & 0xFF));
+        input.push_back(static_cast<uint8_t>((size >> 24) & 0xFF));
+        input.push_back(0x03); // embedded document
+        input.push_back('a');
+        input.push_back(0x00);
+    }
+    // the innermost document is empty, then one terminator closes each level
+    input.insert(input.end(), {0x05, 0x00, 0x00, 0x00, 0x00});
+    input.insert(input.end(), depth, 0x00);
+
+    SECTION("a well-formed deep document is read through the SAX interface")
+    {
+        SaxCountdown accept_all(1000000);
+        CHECK(json::sax_parse(input, &accept_all, json::input_format_t::bson));
+    }
+
+    SECTION("a well-formed deep document is read into a value")
+    {
+        json j = json::from_bson(input);
+
+        // walked rather than compared: comparing, copying or dumping a value
+        // this deep is still recursive
+        std::size_t measured = 0;
+        const json* q = &j;
+        while (q->is_object() && !q->empty())
+        {
+            q = &q->begin().value();
+            ++measured;
+        }
+        CHECK(measured == depth);
+    }
+
+    SECTION("embedded documents and arrays are still read the same way")
+    {
+        const json values = {{"a", {{"b", {{"c", 1}}}}}};
+        CHECK(json::from_bson(json::to_bson(values)) == values);
+
+        const json array = {{"a", {1, 2, 3}}};
+        CHECK(json::from_bson(json::to_bson(array)) == array);
+
+        const json mixed = {{"a", {json{{"x", 1}}, json{{"y", 2}}}}};
+        CHECK(json::from_bson(json::to_bson(mixed)) == mixed);
+
+        CHECK(json::from_bson(json::to_bson(json::object())) == json::object());
+    }
+
+    SECTION("a size that does not match is still reported per document")
+    {
+        // the embedded document claims one byte too many
+        std::vector<uint8_t> const bad =
+        {
+            0x15, 0x00, 0x00, 0x00, 0x03, 'a', 0x00,
+            0x0D, 0x00, 0x00, 0x00, 0x08, 'b', 0x00, 0x01, 0x00,
+            0x00
+        };
+        json _;
+        CHECK_THROWS_AS(_ = json::from_bson(bad), json::parse_error&);
+        CHECK(json::from_bson(bad, true, false).is_discarded());
+    }
+}
+
+TEST_CASE("BSON input that cannot be read is discarded by every overload")
+{
+    std::vector<std::uint8_t> input = json::to_bson(json({{"a", {1, 2}}}));
+    input.pop_back();
+
+    json _;
+    CHECK_THROWS_AS(_ = json::from_bson(input.begin(), input.end()), json::parse_error&);
+    CHECK(json::from_bson(input, true, false).is_discarded());
+    CHECK(json::from_bson(input.begin(), input.end(), true, false).is_discarded());
+    CHECK(json::from_bson(input.data(), input.size(), true, false).is_discarded());
+    CHECK(json::from_bson({input.data(), input.size()}, true, false).is_discarded());
+}
+
+TEST_CASE("BSON SAX parsing stops at every event")
+{
+    // Containers are opened and closed by the loop that reads them; a SAX
+    // handler that rejects any event - including the end of a nested
+    // container - must stop the parse right there.
+    const auto count_events = [](const std::vector<std::uint8_t>& input)
+    {
+        int events = 0;
+        while (true)
+        {
+            SaxCountdown scp(events);
+            if (json::sax_parse(input, &scp, json::input_format_t::bson))
+            {
+                return events;
+            }
+            ++events;
+            REQUIRE(events < 1000);
+        }
+    };
+
+    // 20 events: every container kind closes inside another one
+    const json j = json::parse(R"({"a": [1, {"b": []}], "c": {"d": [[2]]}})");
+    CHECK(count_events(json::to_bson(j)) == 20);
+}
+
 TEST_CASE("BSON numerical data")
 {
     SECTION("number")
@@ -1464,3 +1774,140 @@ TEST_CASE("BSON roundtrips" * doctest::skip())
         }
     }
 }
+
+TEST_CASE("BSON: deeply nested values")
+{
+    SECTION("documents and arrays round-trip at every depth")
+    {
+        // nested documents and arrays, with siblings on every level, so
+        // every length prefix covers entries of both kinds
+        json value = "leaf";
+        for (std::size_t depth = 0; depth <= 300; ++depth)
+        {
+            CAPTURE(depth);
+            const json document = {{"value", value}, {"n", depth}};
+            CHECK(json::from_bson(json::to_bson(document)) == document);
+
+value = depth % 2 == 0 ? json{{"a", std::move(value)}, {"b", {1, "x"}}} :
+            json::array({std::move(value), depth, json::object()});
+        }
+    }
+
+    SECTION("a key containing U+0000 is rejected before anything is written")
+    {
+        json value = json::object({{std::string("bad\0key", 7), 1}});
+        for (std::size_t depth = 0; depth < 200; ++depth)
+        {
+            value = json{{"a", {{"b", 1}}}, {"z", std::move(value)}};
+        }
+        std::vector<std::uint8_t> output;
+        CHECK_THROWS_AS(json::to_bson(value, output), json::out_of_range&);
+        CHECK(output.empty());
+    }
+
+    SECTION("values nested too deeply for the call stack (#5392)")
+    {
+        // serializing recursed once per nesting level, and computed every
+        // nested document's length by walking everything below it again.
+        // The values are only parsed, serialized and walked, never copied or
+        // compared, since those recurse too.
+        const std::size_t depth = 100000;
+        for (const bool objects :
+                {
+                    false, true
+                })
+        {
+            CAPTURE(objects);
+            std::string text = "{\"a\":";
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                text += objects ? "{\"a\":" : "[";
+            }
+            text += "1";
+            text.append(depth, objects ? '}' : ']');
+            text += "}";
+
+            const auto bson = json::to_bson(json::parse(text));
+            const auto result = json::from_bson(bson);
+            const json* p = &result.at("a");
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                p = objects ? &p->at("a") : &p->at(0);
+            }
+            CHECK(*p == 1);
+        }
+    }
+}
+
+TEST_CASE("Invalid document size handling")
+{
+    SECTION("document size must be at least 5")
+    {
+        std::vector<std::uint8_t> const v = {0x04, 0x00, 0x00, 0x00, 0x00};
+        json _;
+        CHECK_THROWS_WITH_AS(_ = json::from_bson(v), "[json.exception.parse_error.112] parse error at byte 5: syntax error while parsing BSON document: document size 4 does not match the number of bytes read (5)", json::parse_error&);
+        CHECK(json::from_bson(v, true, false).is_discarded());
+    }
+
+    SECTION("declared document size must match consumed bytes (extra trailing element)")
+    {
+        // Declares 5-byte empty document but appends an int32 element after the declared end.
+        std::vector<std::uint8_t> const v =
+        {
+            0x05, 0x00, 0x00, 0x00,
+            0x10, 'a', 'd', 'm', 'i', 'n', 0x00,
+            0x01, 0x00, 0x00, 0x00,
+            0x00
+        };
+        json _;
+        CHECK_THROWS_WITH_AS(_ = json::from_bson(v), "[json.exception.parse_error.112] parse error at byte 16: syntax error while parsing BSON document: document size 5 does not match the number of bytes read (16)", json::parse_error&);
+        CHECK(json::from_bson(v, true, false).is_discarded());
+    }
+
+    SECTION("declared document size must match consumed bytes (premature terminator)")
+    {
+        // Declares 32-byte document but only contains the size field followed by an immediate terminator.
+        std::vector<std::uint8_t> const v =
+        {
+            0x20, 0x00, 0x00, 0x00,
+            0x00
+        };
+        json _;
+        CHECK_THROWS_WITH_AS(_ = json::from_bson(v), "[json.exception.parse_error.112] parse error at byte 5: syntax error while parsing BSON document: document size 32 does not match the number of bytes read (5)", json::parse_error&);
+        CHECK(json::from_bson(v, true, false).is_discarded());
+    }
+
+    SECTION("array declared size must match consumed bytes")
+    {
+        // Outer object contains an array "a" that declares 5 bytes (empty) but
+        // actually contains an int32 element before its terminator.
+        std::vector<std::uint8_t> const v =
+        {
+            0x14, 0x00, 0x00, 0x00,                         // object size = 20
+            0x04, 'a', 0x00,                                // key "a", array type
+            0x05, 0x00, 0x00, 0x00,                         // array declared size = 5 (empty)
+            0x10, '0', 0x00, 0x01, 0x00, 0x00, 0x00,        // extra int32 element "0" = 1
+            0x00,                                           // array terminator
+            0x00                                            // object terminator
+        };
+        json _;
+        CHECK_THROWS_WITH_AS(_ = json::from_bson(v), "[json.exception.parse_error.112] parse error at byte 19: syntax error while parsing BSON document: document size 5 does not match the number of bytes read (12)", json::parse_error&);
+        CHECK(json::from_bson(v, true, false).is_discarded());
+    }
+
+    SECTION("BSON string must end with 0x00")
+    {
+        // Length-prefixed string whose terminator byte is 'X' (0x58), not 0x00.
+        std::vector<std::uint8_t> const v =
+        {
+            0x0F, 0x00, 0x00, 0x00,
+            0x02, 's', 0x00,
+            0x02, 0x00, 0x00, 0x00,
+            'A', 'X',
+            0x00
+        };
+        json _;
+        CHECK_THROWS_WITH_AS(_ = json::from_bson(v), "[json.exception.parse_error.112] parse error at byte 13: syntax error while parsing BSON string: BSON string is not null-terminated", json::parse_error&);
+        CHECK(json::from_bson(v, true, false).is_discarded());
+    }
+}

tests/src/unit-byte_container_with_subtype.cpp

@@ -42,6 +42,39 @@ TEST_CASE("byte_container_with_subtype")
         CHECK(container.subtype() == static_cast<subtype_type>(-1));
     }
 
+    SECTION("move semantics")
+    {
+        // the rvalue-reference constructor (without a subtype) must actually move
+        // the passed-in container rather than copy it; comparing the buffer address
+        // before and after is a stronger check than just observing the source is
+        // empty afterward, since a copy-then-clear could also leave it empty
+        {
+            std::vector<std::uint8_t> bytes = {{0xCA, 0xFE, 0xBA, 0xBE}};
+            const auto* const data_ptr = bytes.data();
+
+            nlohmann::byte_container_with_subtype<std::vector<std::uint8_t>> container(std::move(bytes));
+
+            CHECK(container.size() == 4);
+            CHECK(container.data() == data_ptr);
+            CHECK(!container.has_subtype());
+            CHECK(bytes.empty()); // NOLINT(bugprone-use-after-move,clang-analyzer-cplusplus.Move,hicpp-invalid-access-moved)
+        }
+
+        // same check for the rvalue-reference constructor that also takes a subtype
+        {
+            std::vector<std::uint8_t> bytes = {{0xCA, 0xFE, 0xBA, 0xBE}};
+            const auto* const data_ptr = bytes.data();
+
+            nlohmann::byte_container_with_subtype<std::vector<std::uint8_t>> container(std::move(bytes), 42);
+
+            CHECK(container.size() == 4);
+            CHECK(container.data() == data_ptr);
+            CHECK(container.has_subtype());
+            CHECK(container.subtype() == 42);
+            CHECK(bytes.empty()); // NOLINT(bugprone-use-after-move,clang-analyzer-cplusplus.Move,hicpp-invalid-access-moved)
+        }
+    }
+
     SECTION("comparisons")
     {
         std::vector<std::uint8_t> const bytes = {{0xCA, 0xFE, 0xBA, 0xBE}};

tests/src/unit-cbor.cpp

@@ -15,6 +15,7 @@ using nlohmann::json;
 #include <sstream>
 #include <iomanip>
 #include <limits>
+#include <list>
 #include <set>
 #include "make_test_data_available.hpp"
 #include "test_utils.hpp"
@@ -290,7 +291,7 @@ TEST_CASE("CBOR")
 
                 SECTION("-65536..-257")
                 {
-                    for (int32_t i = -65536; i <= -257; ++i)
+                    for (int32_t i = -65536; i <= -257; i = utils::next_integer_sample(i, -257, 7))
                     {
                         CAPTURE(i)
 
@@ -478,7 +479,7 @@ TEST_CASE("CBOR")
 
                 SECTION("256..65535")
                 {
-                    for (size_t i = 256; i <= 65535; ++i)
+                    for (size_t i = 256; i <= 65535; i = utils::next_integer_sample(i, static_cast<size_t>(65535), static_cast<size_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -613,7 +614,7 @@ TEST_CASE("CBOR")
 
                 SECTION("-32768..-129 (int 16)")
                 {
-                    for (int16_t i = -32768; i <= static_cast<std::int16_t>(-129); ++i)
+                    for (int16_t i = -32768; i <= static_cast<std::int16_t>(-129); i = utils::next_integer_sample(i, static_cast<int16_t>(-129), static_cast<int16_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -718,7 +719,7 @@ TEST_CASE("CBOR")
 
                 SECTION("256..65535 (two-byte uint16_t)")
                 {
-                    for (size_t i = 256; i <= 65535; ++i)
+                    for (size_t i = 256; i <= 65535; i = utils::next_integer_sample(i, static_cast<size_t>(65535), static_cast<size_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -1833,6 +1834,59 @@ TEST_CASE("CBOR")
             CHECK(json::from_cbor(std::vector<uint8_t>({0xa1, 0xff, 0x01}), true, false).is_discarded());
         }
 
+        SECTION("invalid UTF-8 in string (see #5529)")
+        {
+            // a two-character text string (major type 3) whose bytes are not
+            // valid UTF-8 (0xC0 0xAE is an overlong encoding of '.') must be
+            // rejected at decode time, matching every other kind of
+            // malformed binary input, rather than only failing later when
+            // the resulting value is dumped
+            json _;
+            CHECK_THROWS_WITH_AS(_ = json::from_cbor(std::vector<uint8_t>({0x62, 0xc0, 0xae})), "[json.exception.parse_error.113] parse error at byte 3: syntax error while parsing CBOR string: invalid string: ill-formed UTF-8 byte", json::parse_error&);
+            CHECK(json::from_cbor(std::vector<uint8_t>({0x62, 0xc0, 0xae}), true, false).is_discarded());
+
+            // a CBOR byte string (major type 2) with the very same bytes is
+            // NOT text and must still be accepted as-is
+            CHECK_NOTHROW(_ = json::from_cbor(std::vector<uint8_t>({0x42, 0xc0, 0xae})));
+            CHECK(_ == json::binary(std::vector<std::uint8_t>({0xc0, 0xae})));
+
+            // valid UTF-8 must still round-trip
+            const json j = "h\xc3\xa9llo, w\xc3\xb6rld! \xe6\x97\xa5\xe6\x9c\xac\xe8\xaa\x9e"; // héllo, wörld! 日本語
+            CHECK(json::from_cbor(json::to_cbor(j)) == j);
+        }
+
+        SECTION("invalid UTF-8 in indefinite-length string")
+        {
+            json _;
+
+            // every chunk must be valid UTF-8 on its own (RFC 8949, Section
+            // 3.2.3), so a code point split across two chunks is rejected
+            CHECK_THROWS_WITH_AS(_ = json::from_cbor(std::vector<uint8_t>({0x7f, 0x61, 0xc3, 0x61, 0xa9, 0xff})), "[json.exception.parse_error.113] parse error at byte 3: syntax error while parsing CBOR string: invalid string: ill-formed UTF-8 byte", json::parse_error&);
+            CHECK(json::from_cbor(std::vector<uint8_t>({0x7f, 0x61, 0xc3, 0x61, 0xa9, 0xff}), true, false).is_discarded());
+
+            // an ill-formed later chunk is rejected after valid ones
+            CHECK_THROWS_WITH_AS(_ = json::from_cbor(std::vector<uint8_t>({0x7f, 0x62, 0xc3, 0xa9, 0x62, 0xc0, 0xae, 0xff})), "[json.exception.parse_error.113] parse error at byte 7: syntax error while parsing CBOR string: invalid string: ill-formed UTF-8 byte", json::parse_error&);
+
+            // valid multi-byte chunks are accepted
+            CHECK(json::from_cbor(std::vector<uint8_t>({0x7f, 0x62, 0xc3, 0xa9, 0x62, 0xc3, 0xb6, 0xff})) == "\xc3\xa9\xc3\xb6");
+        }
+
+        SECTION("many chunks in indefinite-length string")
+        {
+            // only the newly read chunk is validated, not the whole string
+            // collected so far; validating the latter made this input take
+            // quadratic time (about ten seconds for 100000 chunks)
+            constexpr std::size_t chunks = 100000;
+            std::vector<uint8_t> v{0x7f};
+            for (std::size_t i = 0; i < chunks; ++i)
+            {
+                v.push_back(0x61);
+                v.push_back('a');
+            }
+            v.push_back(0xff);
+            CHECK(json::from_cbor(v) == std::string(chunks, 'a'));
+        }
+
         SECTION("strict mode")
         {
             std::vector<uint8_t> const vec = {0xf6, 0xf6};
@@ -2035,6 +2089,291 @@ TEST_CASE("CBOR definite length equal to the indefinite-length sentinel")
     }
 }
 
+TEST_CASE("CBOR nesting does not consume the call stack")
+{
+    // Containers used to be read by calling back into the value reader once
+    // per element, and a tag by calling it for the tagged value, so the native
+    // call stack grew with the nesting depth of the input. Each of the three
+    // costs a single byte to encode -- 0x9F, 0x81 and 0xC2 -- so a payload of
+    // repeated bytes crashed the process (#5104). The containers are kept on a
+    // heap stack now, and a tag is read in a loop.
+    //
+    // Deeply nested values must not be compared, copied or dumped here: those
+    // operations are still recursive and would reintroduce the crash.
+    json _;
+
+    SECTION("indefinite-length containers")
+    {
+        const std::vector<uint8_t> input(500000, 0x9F);
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input), "[json.exception.parse_error.110] parse error at byte 500001: syntax error while parsing CBOR value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_cbor(input, true, false).is_discarded());
+    }
+
+    SECTION("definite-length containers")
+    {
+        const std::vector<uint8_t> input(500000, 0x81);
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input), "[json.exception.parse_error.110] parse error at byte 500001: syntax error while parsing CBOR value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_cbor(input, true, false).is_discarded());
+    }
+
+    SECTION("tags")
+    {
+        // a tag is not a value of its own, so a chain of them used to recurse
+        const std::vector<uint8_t> input(500000, 0xC2);
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input, true, true, json::cbor_tag_handler_t::ignore), "[json.exception.parse_error.110] parse error at byte 500001: syntax error while parsing CBOR value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_cbor(input, true, false, json::cbor_tag_handler_t::ignore).is_discarded());
+    }
+
+    SECTION("stored tags")
+    {
+        // a tag over something other than a byte string is read like for
+        // ignore, so a chain of them must not recurse either (#5316)
+        std::vector<uint8_t> input;
+        for (std::size_t i = 0; i < 500000; ++i)
+        {
+            input.push_back(0xD8);
+            input.push_back(0x18);
+        }
+        input.push_back(0x01);
+        CHECK(json::from_cbor(input, true, true, json::cbor_tag_handler_t::store) == 1);
+    }
+
+    SECTION("a well-formed deep value is read through the SAX interface")
+    {
+        std::vector<uint8_t> input(200000, 0x9F);
+        input.insert(input.end(), 200000, 0xFF);
+
+        SaxCountdown accept_all(1000000);
+        CHECK(json::sax_parse(input, &accept_all, json::input_format_t::cbor));
+    }
+
+    SECTION("a well-formed deep value is read into a value")
+    {
+        const std::size_t depth = 10000;
+        std::vector<uint8_t> input(depth, 0x81);
+        input.push_back(0x00);
+
+        json j = json::from_cbor(input);
+
+        std::size_t measured = 0;
+        const json* p = &j;
+        while (p->is_array() && !p->empty())
+        {
+            p = &p->front();
+            ++measured;
+        }
+        CHECK(measured == depth);
+        CHECK(p->is_number());
+    }
+
+    SECTION("containers are still read the same way")
+    {
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x80})) == json::array());
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xA0})) == json::object());
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x9F, 0xFF})) == json::array());
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xBF, 0xFF})) == json::object());
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x9F, 0x01, 0x02, 0xFF})) == json({1, 2}));
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xBF, 0x61, 'a', 0x01, 0xFF})) == json({{"a", 1}}));
+        // definite and indefinite forms nested inside each other
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x9F, 0x82, 0x01, 0x02, 0xA1, 0x61, 'k', 0xBF, 0xFF, 0xFF})) == json({{1, 2}, {{"k", json::object()}}}));
+    }
+
+    SECTION("tagged values are still read the same way")
+    {
+        const auto ignore = json::cbor_tag_handler_t::ignore;
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xC2, 0x01}), true, true, ignore) == json(1));
+        // a chain of tags resolves to the value that follows it
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xC2, 0xC2, 0xC2, 0x01}), true, true, ignore) == json(1));
+        // a tag inside a container, and one in front of a container
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x82, 0xC2, 0x01, 0x02}), true, true, ignore) == json({1, 2}));
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xC2, 0x82, 0x01, 0x02}), true, true, ignore) == json({1, 2}));
+    }
+}
+
+TEST_CASE("CBOR input that cannot be read is discarded by every overload")
+{
+    std::vector<std::uint8_t> input = json::to_cbor(json({{"a", {1, 2}}}));
+    input.pop_back();
+
+    json _;
+    CHECK_THROWS_AS(_ = json::from_cbor(input.begin(), input.end()), json::parse_error&);
+    CHECK(json::from_cbor(input, true, false).is_discarded());
+    CHECK(json::from_cbor(input.begin(), input.end(), true, false).is_discarded());
+    CHECK(json::from_cbor(input.data(), input.size(), true, false).is_discarded());
+    CHECK(json::from_cbor({input.data(), input.size()}, true, false).is_discarded());
+
+    // a string that ends early, read through iterators that are not
+    // contiguous and have to be copied from one element at a time
+    const std::list<std::uint8_t> truncated_string = {0x63, 'a', 'b'};
+    CHECK(json::from_cbor(truncated_string.begin(), truncated_string.end(), true, false).is_discarded());
+    const std::list<std::uint8_t> complete_string = {0x63, 'a', 'b', 'c'};
+    CHECK(json::from_cbor(complete_string.begin(), complete_string.end()) == "abc");
+}
+
+TEST_CASE("CBOR SAX parsing stops at every event")
+{
+    // Containers are opened and closed by the loop that reads them; a SAX
+    // handler that rejects any event - including the end of a nested
+    // container - must stop the parse right there.
+    const auto count_events = [](const std::vector<std::uint8_t>& input)
+    {
+        int events = 0;
+        while (true)
+        {
+            SaxCountdown scp(events);
+            if (json::sax_parse(input, &scp, json::input_format_t::cbor))
+            {
+                return events;
+            }
+            ++events;
+            REQUIRE(events < 1000);
+        }
+    };
+
+    // 20 events: every container kind closes inside another one
+    const json j = json::parse(R"({"a": [1, {"b": []}], "c": {"d": [[2]]}})");
+    CHECK(count_events(json::to_cbor(j)) == 20);
+    CHECK(count_events(std::vector<std::uint8_t>({0xBF, 0x61, 'a', 0x9F, 0x01, 0xFF, 0xFF})) == 6);
+}
+
+TEST_CASE("CBOR indefinite-length strings do not recurse per chunk")
+{
+    // Reading an indefinite-length string or byte array used to call itself
+    // once per chunk, so a payload of repeated 0x7F (or 0x5F) bytes exhausted
+    // the call stack before any of the input was rejected. The open levels are
+    // counted now, and the levels below prove the reader still reads the same
+    // values and reports the same errors at the same byte offsets.
+    json _;
+
+    SECTION("many open levels are reported, not crashed on")
+    {
+        const std::vector<uint8_t> input(200000, 0x7F);
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input), "[json.exception.parse_error.110] parse error at byte 200001: syntax error while parsing CBOR string: unexpected end of input", json::parse_error&);
+        CHECK(json::from_cbor(input, true, false).is_discarded());
+    }
+
+    SECTION("many open levels are reported, not crashed on (binary)")
+    {
+        const std::vector<uint8_t> input(200000, 0x5F);
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input), "[json.exception.parse_error.110] parse error at byte 200001: syntax error while parsing CBOR binary: unexpected end of input", json::parse_error&);
+        CHECK(json::from_cbor(input, true, false).is_discarded());
+    }
+
+    SECTION("chunks are still concatenated")
+    {
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x7F, 0xFF})) == json(""));
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x7F, 0x61, 0x61, 0xFF})) == json("a"));
+        // nested indefinite-length strings are concatenated across levels
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x7F, 0x7F, 0x61, 0x61, 0xFF, 0x61, 0x62, 0xFF})) == json("ab"));
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x7F, 0x7F, 0x7F, 0x61, 0x7A, 0xFF, 0xFF, 0xFF})) == json("z"));
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xA1, 0x7F, 0x61, 0x61, 0xFF, 0x01})) == json({{"a", 1}}));
+    }
+
+    SECTION("chunks are still concatenated (binary)")
+    {
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x5F, 0x41, 0x61, 0xFF})) == json::binary({0x61}));
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x5F, 0x5F, 0x41, 0x61, 0xFF, 0x41, 0x62, 0xFF})) == json::binary({0x61, 0x62}));
+    }
+
+    SECTION("a chunk that is not a string is still rejected")
+    {
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(std::vector<uint8_t>({0x7F, 0x7F, 0x00})), "[json.exception.parse_error.113] parse error at byte 3: syntax error while parsing CBOR string: expected length specification (0x60-0x7B) or indefinite string type (0x7F); last byte: 0x00", json::parse_error&);
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(std::vector<uint8_t>({0x5F, 0x5F, 0x00})), "[json.exception.parse_error.113] parse error at byte 3: syntax error while parsing CBOR binary: expected length specification (0x40-0x5B) or indefinite binary array type (0x5F); last byte: 0x00", json::parse_error&);
+    }
+
+    SECTION("a break marker outside an indefinite-length string is not a string")
+    {
+        // 0xFF only closes a string that was opened; on its own it is not one
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(std::vector<uint8_t>({0xA1, 0xFF, 0x01})), "[json.exception.parse_error.113] parse error at byte 2: syntax error while parsing CBOR string: expected length specification (0x60-0x7B) or indefinite string type (0x7F); last byte: 0xFF", json::parse_error&);
+    }
+}
+
+TEST_CASE("issue #5405 - array reserve for definite-length CBOR arrays")
+{
+#if !defined(JSON_NOEXCEPTION)
+    // this SECTION relies on catching a thrown exception to distinguish
+    // which of two acceptable, bounded rejections a hostile header took;
+    // under JSON_NOEXCEPTION, JSON_THROW never produces a catchable C++
+    // exception (it aborts instead), so this cannot be tested that way here
+    SECTION("a huge claimed length with no element data must not over-allocate")
+    {
+        // 0x9A: array with a four-byte length; claims 0xFFFFFFFF (4294967295)
+        // elements but provides none. max_size() for a std::vector is far
+        // larger than this count, so it does not reject the header outright;
+        // the (capped) reservation must not attempt to allocate space for
+        // billions of elements before the missing data is detected.
+        json _;
+        const std::vector<uint8_t> input = {0x9A, 0xFF, 0xFF, 0xFF, 0xFF};
+        // On a platform where std::size_t is narrower than 64 bits (e.g.
+        // 32-bit), the claimed count 0xFFFFFFFF coincides with that
+        // platform's detail::unknown_size() sentinel (SIZE_MAX), so the
+        // format-level size check rejects it outright (out_of_range.408,
+        // "excessive ... size") before the SAX consumer's own max_size()
+        // check would even run; on a 64-bit platform it passes both of
+        // those checks and is only found short of data once the (capped)
+        // reservation looks for element bytes that were never provided
+        // (parse_error.110). Either is an acceptable, bounded rejection of
+        // the hostile header -- the property under test is that no path
+        // attempts to allocate space for billions of elements.
+        bool threw = false;
+        try
+        {
+            _ = json::from_cbor(input);
+        }
+        catch (const json::parse_error& e)
+        {
+            threw = true;
+            CHECK(e.id == 110);
+            CHECK(std::string(e.what()) == "[json.exception.parse_error.110] parse error at byte 6: syntax error while parsing CBOR value: unexpected end of input");
+        }
+        catch (const json::out_of_range& e)
+        {
+            threw = true;
+            CHECK(e.id == 408);
+            CHECK(std::string(e.what()).find("excessive") != std::string::npos);
+        }
+        CHECK(threw);
+        CHECK(json::from_cbor(input, true, false).is_discarded());
+    }
+#endif
+
+    SECTION("arrays of various sizes decode to the same value as before the reserve optimization")
+    {
+        for (const auto size :
+                {
+                    std::size_t{0}, std::size_t{1}, std::size_t{5}, // small
+                    std::size_t{16384},                             // exactly at the reserve cap
+                    std::size_t{20000}                              // above the reserve cap
+                })
+        {
+            CAPTURE(size)
+            json j = json::array();
+            for (std::size_t i = 0; i < size; ++i)
+            {
+                j.push_back(static_cast<int>(i % 1000));
+            }
+
+            const auto packed = json::to_cbor(j);
+            CHECK(json::from_cbor(packed) == j);
+        }
+    }
+
+    SECTION("a user-defined SAX consumer is unaffected by the internal DOM reserve optimization")
+    {
+        // the reserve() call is local to json_sax_dom_parser / json_sax_dom_callback_parser;
+        // a custom SAX consumer that does not touch a DOM array sees identical events
+        json j = json::array();
+        for (int i = 0; i < 100; ++i)
+        {
+            j.push_back(i);
+        }
+        const auto packed = json::to_cbor(j);
+
+        SaxCountdown scp(1000000); // large enough to never trigger an abort
+        CHECK(json::sax_parse(packed, &scp, json::input_format_t::cbor));
+    }
+}
+
 TEST_CASE("CBOR roundtrips" * doctest::skip())
 {
     SECTION("input from flynn")
@@ -2204,60 +2543,34 @@ TEST_CASE("CBOR roundtrips" * doctest::skip())
         {
             CAPTURE(filename)
 
+            std::ifstream f_json(filename);
+            const json j1 = json::parse(f_json);
+            const auto packed = utils::read_binary_file(filename + ".cbor");
+
             {
                 INFO_WITH_TEMP(filename + ": std::vector<uint8_t>");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                const json j1 = json::parse(f_json);
-
-                // parse CBOR file
-                const auto packed = utils::read_binary_file(filename + ".cbor");
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_cbor(packed));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": std::ifstream");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                const json j1 = json::parse(f_json);
-
-                // parse CBOR file
                 std::ifstream f_cbor(filename + ".cbor", std::ios::binary);
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_cbor(f_cbor));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": uint8_t* and size");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                const json j1 = json::parse(f_json);
-
-                // parse CBOR file
-                const auto packed = utils::read_binary_file(filename + ".cbor");
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_cbor({packed.data(), packed.size()}));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": output to output adapters");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                json const j1 = json::parse(f_json);
-
-                // parse CBOR file
-                const auto packed = utils::read_binary_file(filename + ".cbor");
-
                 if (exclude_packed.count(filename) == 0u)
                 {
                     {
@@ -2565,11 +2878,16 @@ TEST_CASE("Tagged values")
     const json j = "s";
     auto v = json::to_cbor(j);
 
-    SECTION("0xC6..0xD4")
+    const json j_bin_payload = json::binary(std::vector<std::uint8_t> {0x01, 0x02, 0x03});
+    auto v_bin_payload = json::to_cbor(j_bin_payload);
+
+    SECTION("0xC0..0xD7")
     {
         for (const auto b : std::vector<std::uint8_t>
     {
-        0xC6, 0xC7, 0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF, 0xD0, 0xD1, 0xD2, 0xD3, 0xD4
+        0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5,
+        0xC6, 0xC7, 0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF, 0xD0, 0xD1, 0xD2, 0xD3, 0xD4,
+        0xD5, 0xD6, 0xD7
     })
         {
             CAPTURE(b);
@@ -2589,6 +2907,12 @@ TEST_CASE("Tagged values")
 
             auto j_tagged_stored = json::from_cbor(v_tagged, true, true, json::cbor_tag_handler_t::store);
             CHECK(j_tagged_stored == j);
+
+            auto v_binary_tagged = v_bin_payload;
+            v_binary_tagged.insert(v_binary_tagged.begin(), b);
+            auto j_binary_tagged_stored = json::from_cbor(v_binary_tagged, true, true, json::cbor_tag_handler_t::store);
+            CHECK(j_binary_tagged_stored == j_bin_payload);
+            CHECK(!j_binary_tagged_stored.get_binary().has_subtype());
         }
     }
 
@@ -2744,6 +3068,77 @@ TEST_CASE("Tagged values")
             CHECK_THROWS_AS(_ = json::from_cbor(v_tagged, true, true, json::cbor_tag_handler_t::error), json::parse_error);
             CHECK_THROWS_AS(_ = json::from_cbor(v_tagged, true, true, json::cbor_tag_handler_t::ignore), json::parse_error);
         }
+
+        SECTION("issue #5316 - cbor_tag_handler_t::store on non-binary tagged items")
+        {
+            // 55799({"a": 1}) -- CBOR self-describe magic followed by a map
+            const std::vector<std::uint8_t> v_map{0xD9, 0xD9, 0xF7, 0xA1, 0x61, 0x61, 0x01};
+            CHECK(json::from_cbor(v_map, true, true, json::cbor_tag_handler_t::ignore) == json({{"a", 1}}));
+            CHECK(json::from_cbor(v_map, true, true, json::cbor_tag_handler_t::store) == json({{"a", 1}}));
+
+            // Tag 24 over unsigned integer 5
+            const std::vector<std::uint8_t> v_int{0xD8, 0x18, 0x05};
+            CHECK(json::from_cbor(v_int, true, true, json::cbor_tag_handler_t::ignore) == 5);
+            CHECK(json::from_cbor(v_int, true, true, json::cbor_tag_handler_t::store) == 5);
+
+            // Tag 24 over text string "foo"
+            const std::vector<std::uint8_t> v_str{0xD8, 0x18, 0x63, 'f', 'o', 'o'};
+            CHECK(json::from_cbor(v_str, true, true, json::cbor_tag_handler_t::ignore) == "foo");
+            CHECK(json::from_cbor(v_str, true, true, json::cbor_tag_handler_t::store) == "foo");
+
+            // Tag 24 over array [1, 2]
+            const std::vector<std::uint8_t> v_arr{0xD8, 0x18, 0x82, 0x01, 0x02};
+            CHECK(json::from_cbor(v_arr, true, true, json::cbor_tag_handler_t::ignore) == json({1, 2}));
+            CHECK(json::from_cbor(v_arr, true, true, json::cbor_tag_handler_t::store) == json({1, 2}));
+
+            // Tag 24 over boolean true
+            const std::vector<std::uint8_t> v_bool{0xD8, 0x18, 0xF5};
+            CHECK(json::from_cbor(v_bool, true, true, json::cbor_tag_handler_t::ignore) == true);
+            CHECK(json::from_cbor(v_bool, true, true, json::cbor_tag_handler_t::store) == true);
+
+            // Tag 24 over null
+            const std::vector<std::uint8_t> v_null{0xD8, 0x18, 0xF6};
+            CHECK(json::from_cbor(v_null, true, true, json::cbor_tag_handler_t::ignore) == nullptr);
+            CHECK(json::from_cbor(v_null, true, true, json::cbor_tag_handler_t::store) == nullptr);
+
+            // Nested tags: tag 55799 over tag 24 over integer 42
+            const std::vector<std::uint8_t> v_nested{0xD9, 0xD9, 0xF7, 0xD8, 0x18, 0x18, 0x2A};
+            CHECK(json::from_cbor(v_nested, true, true, json::cbor_tag_handler_t::ignore) == 42);
+            CHECK(json::from_cbor(v_nested, true, true, json::cbor_tag_handler_t::store) == 42);
+
+            // Tag 24 over byte string continues to store subtype as before
+            const std::vector<std::uint8_t> v_bin{0xD8, 0x18, 0x42, 0xCA, 0xFE};
+            auto j_bin_store = json::from_cbor(v_bin, true, true, json::cbor_tag_handler_t::store);
+            CHECK(j_bin_store.is_binary());
+            CHECK(j_bin_store.get_binary().has_subtype());
+            CHECK(j_bin_store.get_binary().subtype() == 24);
+            CHECK(j_bin_store.get_binary() == json::binary({0xCA, 0xFE}, 24).get_binary());
+
+            // Tagged values inside a container under store: [24(1), 25(h'0001')]
+            const std::vector<std::uint8_t> v_container{0x82, 0xD8, 0x18, 0x01, 0xD8, 0x19, 0x42, 0x00, 0x01};
+            auto j_container_store = json::from_cbor(v_container, true, true, json::cbor_tag_handler_t::store);
+            CHECK(j_container_store.is_array());
+            CHECK(j_container_store.size() == 2);
+            CHECK(j_container_store[0] == 1);
+            CHECK(j_container_store[1].is_binary());
+            CHECK(j_container_store[1].get_binary().has_subtype());
+            CHECK(j_container_store[1].get_binary().subtype() == 25);
+            CHECK(j_container_store[1].get_binary() == json::binary({0x00, 0x01}, 25).get_binary());
+
+            // Tagged values as object values under store: {"a": 55799(1), "b": 24(h'01')}
+            const std::vector<std::uint8_t> v_object{0xA2, 0x61, 'a', 0xD9, 0xD9, 0xF7, 0x01, 0x61, 'b', 0xD8, 0x18, 0x41, 0x01};
+            CHECK(json::from_cbor(v_object, true, true, json::cbor_tag_handler_t::store) == json({{"a", 1}, {"b", json::binary({0x01}, 24)}}));
+
+            // two tags in a row before a byte string: the inner tag is stored
+            // (this uses item_read and then the byte-string path)
+            const std::vector<std::uint8_t> v_nested_byte_string{0xD8, 0x18, 0xD8, 0x19, 0x42, 0x00, 0x01};
+            CHECK(json::from_cbor(v_nested_byte_string, true, true, json::cbor_tag_handler_t::store) == json::binary({0x00, 0x01}, 25));
+
+            // errors after a stored tag are now the same as with ignore
+            json _;
+            CHECK_THROWS_WITH_AS(_ = json::from_cbor(std::vector<std::uint8_t> {0xD8, 0x18}, true, true, json::cbor_tag_handler_t::store), "[json.exception.parse_error.110] parse error at byte 3: syntax error while parsing CBOR value: unexpected end of input", json::parse_error&);
+            CHECK_THROWS_WITH_AS(_ = json::from_cbor(std::vector<std::uint8_t> {0xD8, 0x18, 0x1C}, true, true, json::cbor_tag_handler_t::store), "[json.exception.parse_error.112] parse error at byte 3: syntax error while parsing CBOR value: invalid byte: 0x1C", json::parse_error&);
+        }
     }
 
     SECTION("negative integer overflow")

tests/src/unit-class_const_iterator.cpp

@@ -43,6 +43,13 @@ TEST_CASE("const_iterator class")
             json::const_iterator const it(&j);
             json::const_iterator it2(&j);
             it2 = it;
+
+            // assigning an iterator to itself leaves it unchanged
+            json const a = {1, 2, 3};
+            json::const_iterator it3 = a.cbegin() + 1;
+            const json::const_iterator& same = it3;
+            it3 = same;
+            CHECK(*it3 == 2);
         }
 
         SECTION("copy constructor from non-const iterator")

tests/src/unit-class_lexer.cpp

@@ -12,6 +12,12 @@
 #include <nlohmann/json.hpp>
 using nlohmann::json;
 
+#include <cfloat> // FLT_EVAL_METHOD
+#include <cstdlib> // strtod
+#include <sstream> // stringstream
+#include <string> // string
+#include <vector> // vector
+
 namespace
 {
 // shortcut to scan a string literal
@@ -224,3 +230,473 @@ TEST_CASE("lexer class")
         CHECK((scan_string("/**//**//**/", true) == json::lexer::token_type::end_of_input));
     }
 }
+
+TEST_CASE("lexer number fast path")
+{
+    // The contiguous fast path (used for pointer/string input) must agree with
+    // the streaming byte path (used for std::istream) on token type, numeric
+    // value, and round-trip text for every well-formed number, and reject the
+    // same malformed numbers with the same message.
+    SECTION("contiguous vs streaming parity")
+    {
+        const std::vector<std::string> numbers =
+        {
+            "0", "-0", "1", "-1", "42", "-42", "10", "100", "1234567890",
+            "0.0", "-0.0", "3.14", "-3.14", "0.5", "-0.001", "123.456789",
+            "1e0", "1E0", "1e10", "1e-10", "1e+10", "1.5e3", "-2.5E-4",
+            "9223372036854775807",             // INT64_MAX -> unsigned
+            "9223372036854775808",             // INT64_MAX + 1 -> unsigned
+            "18446744073709551615",            // UINT64_MAX -> unsigned
+            "18446744073709551616",            // UINT64_MAX + 1 -> float
+            "-9223372036854775808",            // INT64_MIN -> integer
+            "-9223372036854775809",            // INT64_MIN - 1 -> float
+            "123456789012345678901234567890",  // huge -> float
+            "0.30000000000000004", "2.2250738585072014e-308", "1e308",
+            // high-precision / wide-exponent values that exercise the
+            // std::from_chars (Eisel-Lemire) path beyond the Clinger subset
+            "1.7976931348623157e308", "1.2345678901234567e-250",
+            "9007199254740993", "5e-324", "1e-320"
+        };
+
+        for (const auto& n : numbers)
+        {
+            const std::string doc = "[" + n + "]";
+
+            // contiguous fast path
+            const json a = json::parse(doc);
+            // streaming byte path
+            std::stringstream ss(doc);
+            const json b = json::parse(ss);
+
+            CAPTURE(n);
+            CHECK(a == b);
+            CHECK(a.dump() == b.dump());
+            CHECK(a[0].type() == b[0].type());
+        }
+    }
+
+    SECTION("significant-digit gate for the Clinger fast path")
+    {
+        // Clinger's fast path needs a significand below 2^53, so it cannot
+        // succeed once the mantissa has 17 or more significant digits (the
+        // significand would be at least 10^16). The lexer skips the attempt
+        // there. That is only allowed to save work: every value must still come
+        // out bit-exactly, and both scanners must agree. In particular the gate
+        // must not fire for tokens whose leading zeros merely look like extra
+        // digits - "0.1234567890123456" has 16 significant digits, not 17.
+        const std::vector<std::string> numbers =
+        {
+            "1234567890123456",                  // 16 significant digits
+            "12345678901234567",                 // 17 -> attempt skipped
+            "123456789012345678",                // 18 -> attempt skipped
+            "0.1234567890123456",                // 16: the leading "0" is not significant
+            "0.12345678901234567",               // 17
+            "0.00000000000000001",               // 1, in a long token
+            "0.000000000000000012345678901234",  // 14, in a long token
+            "-0.0000000000000000000001",         // 1, negative
+            "1.0000000000000000",                // 17: trailing zeros are significant here
+            "10000000000000000",                 // 17
+            "9007199254740992",                  // 2^53
+            "9007199254740993",                  // 2^53 + 1
+            "-65.613616999999977",               // canada.json shape
+            "1.2345678901234567e-250",           // 17 with an exponent
+            "1.234567890123456e-250",            // 16 with an exponent
+            "1e10", "0.0", "-0.0", "0e0", "0.000123"
+        };
+
+        for (const auto& n : numbers)
+        {
+            CAPTURE(n);
+            const std::string doc = "[" + n + "]";
+
+            const json a = json::parse(doc);   // contiguous fast path
+            std::stringstream ss(doc);
+            const json b = json::parse(ss);    // streaming byte path
+
+            CHECK(a[0].type() == b[0].type());
+            CHECK(a == b);
+
+            if (a[0].is_number_float())
+            {
+                const double expected = std::strtod(n.c_str(), nullptr);
+                CHECK(a[0].get<double>() == expected);
+                CHECK(b[0].get<double>() == expected);
+            }
+        }
+    }
+
+    SECTION("token type classification")
+    {
+        CHECK((scan_string("0") == json::lexer::token_type::value_unsigned));
+        CHECK((scan_string("-1") == json::lexer::token_type::value_integer));
+        CHECK((scan_string("1.5") == json::lexer::token_type::value_float));
+        CHECK((scan_string("1e5") == json::lexer::token_type::value_float));
+        CHECK((scan_string("18446744073709551615") == json::lexer::token_type::value_unsigned));
+        CHECK((scan_string("18446744073709551616") == json::lexer::token_type::value_float));
+        CHECK((scan_string("-9223372036854775808") == json::lexer::token_type::value_integer));
+        CHECK((scan_string("-9223372036854775809") == json::lexer::token_type::value_float));
+    }
+
+    SECTION("malformed numbers are rejected identically")
+    {
+        for (const char* bad :
+                {"-", "1.", "1e", "1e+", "1.2e", "01", "-01", "1..2", "1.2.3"
+                })
+        {
+            CAPTURE(bad);
+            // the contiguous fast path must decline and let the byte path report
+            const std::string doc = std::string("[") + bad + "]";
+            CHECK_FALSE(json::accept(doc));
+            std::stringstream ss(doc);
+            CHECK_FALSE(json::accept(ss));
+        }
+    }
+
+#if !defined(JSON_NOEXCEPTION)
+    // these sections parse invalid input, which aborts when exceptions are off
+    SECTION("exhaustive grammar parity with the streaming path")
+    {
+        // The JSON number grammar is encoded twice: once as the scan_number()
+        // state machine and once as the contiguous fast path. Enumerate every
+        // short string over the number alphabet and require the two encodings to
+        // agree exactly - on acceptance, on the reported error, and on the parsed
+        // value - so they cannot drift apart.
+        const std::string alphabet = "01.eE+-";
+
+        // full outcome of parsing @a doc, so a mismatch in type, value, or error
+        // message is caught, not just a mismatch in acceptance
+        const auto outcome = [](const std::string & doc, bool streaming) -> std::string
+        {
+            try
+            {
+                if (streaming)
+                {
+                    std::stringstream ss(doc);
+                    const json j = json::parse(ss);
+                    return std::string(j[0].type_name()) + '|' + j.dump();
+                }
+                const json j = json::parse(doc);
+                return std::string(j[0].type_name()) + '|' + j.dump();
+            }
+            catch (const json::parse_error& e)
+            {
+                return {e.what()};
+            }
+        };
+
+        std::vector<std::string> mismatches;
+        std::vector<std::string> tokens{""};
+        for (std::size_t length = 1; length <= 4; ++length)
+        {
+            std::vector<std::string> next;
+            next.reserve(tokens.size() * alphabet.size());
+            for (const auto& prefix : tokens)
+            {
+                for (const char c : alphabet)
+                {
+                    next.push_back(prefix + c);
+                }
+            }
+            tokens = next;
+
+            for (const auto& token : tokens)
+            {
+                const std::string doc = "[" + token + "]";
+                if (outcome(doc, false) != outcome(doc, true))
+                {
+                    mismatches.push_back(doc);
+                }
+            }
+        }
+
+        // 7 + 49 + 343 + 2401 tokens
+        CHECK(tokens.size() == 2401);
+        CAPTURE(mismatches);
+        CHECK(mismatches.empty());
+    }
+
+    SECTION("error positions match the streaming path")
+    {
+        // Rejecting identically is not enough: the fast path must also report the
+        // error at the same position as the byte path. A number directly followed
+        // by a newline is the interesting case, because the byte path reaches the
+        // newline (which resets the column) and then ungets it.
+        // returns the parse_error message, or "" if the document parsed
+        const auto contiguous_error = [](const std::string & doc) -> std::string
+        {
+            try
+            {
+                const json j = json::parse(doc);
+                static_cast<void>(j);
+            }
+            catch (const json::parse_error& e)
+            {
+                return {e.what()};
+            }
+            return {};
+        };
+        const auto streaming_error = [](const std::string & doc) -> std::string
+        {
+            try
+            {
+                std::stringstream ss(doc);
+                const json j = json::parse(ss);
+                static_cast<void>(j);
+            }
+            catch (const json::parse_error& e)
+            {
+                return {e.what()};
+            }
+            return {};
+        };
+
+        for (const char* bad :
+                {"[01\n]", "[00\n]", "[-01\n]", "{1\n}", "[1\n2]", "[1.2.3\n]",
+                 "[1 \n2]", "[\n1\n2]", "1\n2", "[01\r\n]", "[1e\n]", "[-\n]"
+                })
+        {
+            CAPTURE(bad);
+            const std::string doc = bad;
+            const std::string contiguous_what = contiguous_error(doc);
+
+            CHECK_FALSE(contiguous_what.empty());
+            CHECK(contiguous_what == streaming_error(doc));
+        }
+
+        // A number terminated by a newline must report the same position as the
+        // same number terminated by anything else: scan_number() reads the
+        // terminator and ungets it, so the reported column is the one reached
+        // after the number's last character - not the 0 that an unget() across
+        // the newline used to leave behind.
+        CHECK(contiguous_error("[01\n]") == contiguous_error("[01 ]"));
+        CHECK(contiguous_error("[01\n]") ==
+              "[json.exception.parse_error.101] parse error at line 1, column 3: "
+              "syntax error while parsing array - unexpected number literal; expected ']'");
+
+        // the same for a multi-character token, where the column of the last
+        // character (the '3' of "-2.5e3") differs from the column it starts at
+        CHECK(contiguous_error("null -2.5e3\nfalse") == contiguous_error("null -2.5e3 false"));
+        CHECK(contiguous_error("null -2.5e3\nfalse") ==
+              "[json.exception.parse_error.101] parse error at line 1, column 11: "
+              "syntax error while parsing value - unexpected number literal; expected end of input");
+    }
+#endif
+}
+
+TEST_CASE("lexer string fast path")
+{
+    // Build a byte string from explicit values: a hex escape in a string
+    // literal swallows every following hex digit, which makes sequences like
+    // "\xC3\xA9b" mean something other than they look like.
+    const auto bytes = [](std::initializer_list<int> values)
+    {
+        std::string result;
+        for (const int value : values)
+        {
+            result.push_back(static_cast<char>(value));
+        }
+        return result;
+    };
+
+#if !defined(JSON_NOEXCEPTION)
+    // the full outcome of parsing @a doc: the parsed value, or the exact error
+    // message, so a mismatch in either is caught. Only usable with exceptions
+    // on: parsing invalid input aborts when they are off.
+    const auto outcome = [](const std::string & doc, bool streaming) -> std::string
+    {
+        try
+        {
+            if (streaming)
+            {
+                std::stringstream ss(doc);
+                const json j = json::parse(ss);
+                return j.dump();
+            }
+            const json j = json::parse(doc);
+            return j.dump();
+        }
+        // not just parse_error: if a bulk scanner ever let ill-formed UTF-8
+        // through, dump() would throw type_error.316, and that has to surface
+        // as a reported mismatch rather than as an uncaught exception
+        catch (const json::exception& e)
+        {
+            return {e.what()};
+        }
+    };
+#endif
+
+    // once at the start of the string, once past the first 8-byte SWAR word, so
+    // the bulk scanner sees each case with and without a run behind it
+    const std::vector<std::size_t> offsets{0, 9};
+
+#if !defined(JSON_NOEXCEPTION)
+    SECTION("exhaustive contiguous vs streaming parity")
+    {
+        // ordinary ASCII, both specials, a control byte, characters that make
+        // the preceding backslash a valid escape, a UTF-8 lead byte of each
+        // length, a continuation byte, and a byte that is never valid
+        const std::vector<std::string> alphabet =
+        {
+            "a", "\"", "\\", "n", "u", "0", bytes({0x01}),
+            bytes({0xC3}), bytes({0xA9}), bytes({0xE4}), bytes({0xF0}),
+            bytes({0x80}), bytes({0xFF})
+        };
+
+        std::vector<std::string> mismatches;
+        std::vector<std::string> tokens{""};
+        for (std::size_t length = 1; length <= 3; ++length)
+        {
+            std::vector<std::string> next;
+            next.reserve(tokens.size() * alphabet.size());
+            for (const auto& prefix : tokens)
+            {
+                for (const auto& symbol : alphabet)
+                {
+                    next.push_back(prefix + symbol);
+                }
+            }
+            tokens = next;
+
+            for (const auto& token : tokens)
+            {
+                for (const std::size_t offset : offsets)
+                {
+                    const std::string doc = "[\"" + std::string(offset, 'a') + token + "\"]";
+                    if (outcome(doc, false) != outcome(doc, true))
+                    {
+                        mismatches.push_back(doc);
+                    }
+                }
+            }
+        }
+
+        // 13 + 169 + 2197 tokens, each at two offsets
+        CHECK(tokens.size() == 2197);
+        CAPTURE(mismatches);
+        CHECK(mismatches.empty());
+    }
+
+    SECTION("special bytes at every offset of the SWAR stride")
+    {
+        // The bulk scanner consumes 8 bytes at a time and then a tail; place
+        // every kind of byte that ends a run at each offset across two words,
+        // so multibyte sequences also straddle the word boundary.
+        const std::vector<std::string> specials =
+        {
+            "\"", "\\", bytes({0x01}), bytes({0x1F}), bytes({0x7F}),
+            bytes({0xC3, 0xA9}), bytes({0xE4, 0xB8, 0xAD}), bytes({0xF0, 0x9F, 0x98, 0x80}),
+            bytes({0xFF}), bytes({0xC3}), bytes({0xE4, 0xB8})
+        };
+
+        std::vector<std::string> mismatches;
+        for (std::size_t offset = 0; offset <= 17; ++offset)
+        {
+            for (const auto& special : specials)
+            {
+                const std::string doc = "[\"" + std::string(offset, 'a') + special + "\"]";
+                if (outcome(doc, false) != outcome(doc, true))
+                {
+                    mismatches.push_back(doc);
+                }
+            }
+        }
+        CAPTURE(mismatches);
+        CHECK(mismatches.empty());
+    }
+#endif
+
+    // json::accept() never throws, so the ranges stay covered without exceptions
+    SECTION("UTF-8 ranges are accepted and rejected as documented")
+    {
+        // The bulk validator must accept exactly what the byte-at-a-time
+        // scanner accepts, so pin the boundaries of every range it recognizes.
+        // aggregate, only ever brace-initialized below; default member
+        // initializers would stop it being an aggregate in C++11
+        struct utf8_case // NOLINT(cppcoreguidelines-pro-type-member-init,hicpp-member-init)
+        {
+            std::string sequence;
+            bool valid;
+            const char* description;
+        };
+        const std::vector<utf8_case> cases =
+        {
+            {bytes({0xC2, 0x80}), true, "U+0080, shortest two-byte"},
+            {bytes({0xDF, 0xBF}), true, "U+07FF, longest two-byte"},
+            {bytes({0xC1, 0xBF}), false, "overlong two-byte"},
+            {bytes({0xC2, 0x7F}), false, "two-byte with bad continuation"},
+            {bytes({0xE0, 0xA0, 0x80}), true, "U+0800, shortest three-byte"},
+            {bytes({0xE0, 0x9F, 0xBF}), false, "overlong three-byte"},
+            {bytes({0xED, 0x9F, 0xBF}), true, "U+D7FF, just below the surrogates"},
+            {bytes({0xED, 0xA0, 0x80}), false, "surrogate U+D800"},
+            {bytes({0xED, 0xBF, 0xBF}), false, "surrogate U+DFFF"},
+            {bytes({0xEE, 0x80, 0x80}), true, "U+E000, just above the surrogates"},
+            {bytes({0xEF, 0xBF, 0xBF}), true, "U+FFFF"},
+            {bytes({0xF0, 0x90, 0x80, 0x80}), true, "U+10000, shortest four-byte"},
+            {bytes({0xF0, 0x8F, 0xBF, 0xBF}), false, "overlong four-byte"},
+            {bytes({0xF4, 0x8F, 0xBF, 0xBF}), true, "U+10FFFF, highest code point"},
+            {bytes({0xF4, 0x90, 0x80, 0x80}), false, "above U+10FFFF"},
+            {bytes({0xF5, 0x80, 0x80, 0x80}), false, "lead byte out of range"},
+            {bytes({0x80}), false, "bare continuation byte"},
+            {bytes({0xFF}), false, "byte that never appears in UTF-8"},
+            {bytes({0xC3}), false, "truncated two-byte"},
+            {bytes({0xE4, 0xB8}), false, "truncated three-byte"},
+            {bytes({0xF0, 0x9F, 0x98}), false, "truncated four-byte"}
+        };
+
+        for (const auto& test_case : cases)
+        {
+            CAPTURE(test_case.description);
+            for (const std::size_t offset : offsets)
+            {
+                CAPTURE(offset);
+                const std::string doc = "[\"" + std::string(offset, 'a') + test_case.sequence + "\"]";
+                CHECK(json::accept(doc) == test_case.valid);
+#if !defined(JSON_NOEXCEPTION)
+                CHECK(outcome(doc, false) == outcome(doc, true));
+#endif
+            }
+        }
+    }
+}
+
+TEST_CASE("parse_float_fast declines what it cannot convert exactly")
+{
+    // The lexer only hands well-formed numbers to parse_float_fast, so the
+    // malformed ones below can only be passed to it directly. Declining is
+    // always safe: the caller then falls back to a slower, exact conversion.
+    const auto fast = [](const std::string & s, double & out)
+    {
+        return nlohmann::detail::parse_float_fast(s.data(), s.data() + s.size(), '.', out);
+    };
+    double out = 0;
+
+#if defined(FLT_EVAL_METHOD) && FLT_EVAL_METHOD != 0
+    // without true double precision, the fast path declines everything
+    CHECK_FALSE(fast("1.5", out));
+#else
+    CHECK(fast("1.5", out));
+    CHECK(out == 1.5);
+    CHECK(fast("+2.5e1", out));
+    CHECK(out == 25.0);
+    CHECK(fast("-25E-1", out));
+    CHECK(out == -2.5);
+    CHECK(fast("1e", out));
+    CHECK(out == 1.0);
+#endif
+
+    // not a number
+    CHECK_FALSE(fast("", out));
+    CHECK_FALSE(fast("-", out));
+    CHECK_FALSE(fast(".", out));
+    CHECK_FALSE(fast("1.2.3", out));
+    CHECK_FALSE(fast("1x", out));
+    CHECK_FALSE(fast("1e+", out));
+    CHECK_FALSE(fast("1e1x", out));
+
+    // numbers that are not represented exactly on the fast path
+    CHECK_FALSE(fast("12345678901234567890", out));
+    CHECK_FALSE(fast("1e10000", out));
+    CHECK_FALSE(fast("9007199254740993", out));
+    CHECK_FALSE(fast("1e23", out));
+    CHECK_FALSE(fast("1e-23", out));
+}

tests/src/unit-comparison.cpp

@@ -15,6 +15,14 @@
 
 #include "doctest_compatibility.h"
 
+#include <algorithm>
+
+#include <cstdint>
+#include <map>
+#include <string>
+#include <utility>
+#include <vector>
+
 #define JSON_TESTS_PRIVATE
 #include <nlohmann/json.hpp>
 using nlohmann::json;
@@ -255,6 +263,135 @@ TEST_CASE("lexicographical comparison operators")
             {f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_, f_}, // 21
         };
 
+        SECTION("signed/unsigned mixed comparison above INT64_MAX")
+        {
+            const json above_int64_max = static_cast<std::uint64_t>((std::numeric_limits<std::int64_t>::max)()) + 1ULL;
+            const json max_uint64 = (std::numeric_limits<std::uint64_t>::max)();
+            const json negative_one = -1;
+            const json one = 1;
+            const json max_int64 = (std::numeric_limits<std::int64_t>::max)();
+
+            CHECK_FALSE(above_int64_max == negative_one);
+            CHECK(above_int64_max != negative_one);
+            CHECK(negative_one < above_int64_max);
+            CHECK(negative_one <= above_int64_max);
+            CHECK_FALSE(negative_one > above_int64_max);
+            CHECK_FALSE(negative_one >= above_int64_max);
+            CHECK_FALSE(above_int64_max < negative_one);
+            CHECK_FALSE(above_int64_max <= negative_one);
+            CHECK(above_int64_max > negative_one);
+            CHECK(above_int64_max >= negative_one);
+            CHECK(negative_one != above_int64_max);
+            CHECK_FALSE(negative_one == above_int64_max);
+
+            CHECK_FALSE(max_uint64 == negative_one);
+            CHECK(max_uint64 != negative_one);
+            CHECK(negative_one < max_uint64);
+            CHECK(negative_one <= max_uint64);
+            CHECK_FALSE(negative_one > max_uint64);
+            CHECK_FALSE(negative_one >= max_uint64);
+            CHECK_FALSE(max_uint64 < negative_one);
+            CHECK_FALSE(max_uint64 <= negative_one);
+            CHECK(max_uint64 > negative_one);
+            CHECK(max_uint64 >= negative_one);
+            CHECK(negative_one != max_uint64);
+            CHECK_FALSE(negative_one == max_uint64);
+
+            CHECK_FALSE(one == above_int64_max);
+            CHECK(one != above_int64_max);
+            CHECK(one < above_int64_max);
+            CHECK(one <= above_int64_max);
+            CHECK_FALSE(one > above_int64_max);
+            CHECK_FALSE(one >= above_int64_max);
+            CHECK_FALSE(above_int64_max < one);
+            CHECK_FALSE(above_int64_max <= one);
+            CHECK(above_int64_max > one);
+            CHECK(above_int64_max >= one);
+
+            CHECK_FALSE(max_int64 == above_int64_max);
+            CHECK(max_int64 != above_int64_max);
+            CHECK(max_int64 < above_int64_max);
+            CHECK(max_int64 <= above_int64_max);
+            CHECK_FALSE(max_int64 > above_int64_max);
+            CHECK_FALSE(max_int64 >= above_int64_max);
+            CHECK_FALSE(above_int64_max < max_int64);
+            CHECK_FALSE(above_int64_max <= max_int64);
+            CHECK(above_int64_max > max_int64);
+            CHECK(above_int64_max >= max_int64);
+
+#if JSON_HAS_THREE_WAY_COMPARISON
+            // JSON_HAS_CPP_20 (do not remove; see note at top of file)
+            CHECK((negative_one <=> above_int64_max) == std::partial_ordering::less); // *NOPAD*
+            CHECK((above_int64_max <=> negative_one) == std::partial_ordering::greater); // *NOPAD*
+            CHECK((negative_one <=> max_uint64) == std::partial_ordering::less); // *NOPAD*
+            CHECK((max_uint64 <=> negative_one) == std::partial_ordering::greater); // *NOPAD*
+            CHECK((one <=> above_int64_max) == std::partial_ordering::less); // *NOPAD*
+            CHECK((above_int64_max <=> one) == std::partial_ordering::greater); // *NOPAD*
+            CHECK((max_int64 <=> above_int64_max) == std::partial_ordering::less); // *NOPAD*
+            CHECK((above_int64_max <=> max_int64) == std::partial_ordering::greater); // *NOPAD*
+#endif
+        }
+
+        SECTION("integer/float mixed comparison is exact")
+        {
+            // Widening the integer to a double loses precision past the
+            // mantissa, so 2^63-2 and 2^63-1 both used to compare equal to the
+            // double 2^63 while differing from each other. That makes equality
+            // intransitive and the ordering not a strict weak ordering.
+            const json below_two_63 = static_cast<std::int64_t>(9223372036854775806LL);
+            const json max_int64 = (std::numeric_limits<std::int64_t>::max)();
+            const json two_63 = 9223372036854775808.0;
+
+            CHECK_FALSE(below_two_63 == two_63);
+            CHECK_FALSE(max_int64 == two_63);
+            CHECK(below_two_63 != max_int64);
+            CHECK(below_two_63 < max_int64);
+            CHECK(below_two_63 < two_63);
+            CHECK(max_int64 < two_63);
+            CHECK(two_63 > max_int64);
+            CHECK_FALSE(two_63 < max_int64);
+
+            // the same past the unsigned range
+            const json max_uint64 = (std::numeric_limits<std::uint64_t>::max)();
+            const json two_64 = 18446744073709551616.0;
+            CHECK_FALSE(max_uint64 == two_64);
+            CHECK(max_uint64 < two_64);
+            CHECK(two_64 > max_uint64);
+
+            // values a double represents exactly still compare equal
+            CHECK(json(1) == json(1.0));
+            CHECK(json(1u) == json(1.0));
+            CHECK(json(-3) == json(-3.0));
+            CHECK(json(1) < json(1.5));
+            CHECK(json(1.5) < json(2));
+            CHECK(json(2) > json(1.5));
+            CHECK(json(-1) > json(-1.5));
+            CHECK(json(-1.5) < json(-1));
+            CHECK(json(-2) < json(-1.5));
+
+            // a float below the range of the integer type
+            CHECK(json(0) > json(-1e30));
+            CHECK(json(-1e30) < json(0));
+            CHECK(json(0u) > json(-0.5));
+            CHECK(json(-0.5) < json(0u));
+
+            // a NaN operand stays unordered against either integer kind
+            CHECK_FALSE(json(1) == json(nan));
+            CHECK_FALSE(json(1) < json(nan));
+            CHECK_FALSE(json(nan) < json(1));
+            CHECK_FALSE(json(1u) == json(nan));
+
+#if JSON_HAS_THREE_WAY_COMPARISON
+            // JSON_HAS_CPP_20 (do not remove; see note at top of file)
+            CHECK((max_int64 <=> two_63) == std::partial_ordering::less); // *NOPAD*
+            CHECK((two_63 <=> max_int64) == std::partial_ordering::greater); // *NOPAD*
+            CHECK((below_two_63 <=> max_int64) == std::partial_ordering::less); // *NOPAD*
+            CHECK((max_uint64 <=> two_64) == std::partial_ordering::less); // *NOPAD*
+            CHECK((json(1) <=> json(1.0)) == std::partial_ordering::equivalent); // *NOPAD*
+            CHECK((json(1) <=> json(nan)) == std::partial_ordering::unordered); // *NOPAD*
+#endif
+        }
+
         SECTION("compares unordered")
         {
             std::vector<std::vector<bool>> expected =
@@ -613,3 +750,205 @@ TEST_CASE("regression #3868 - heterogeneous comparisons compile under C++20 (P24
     }
 }
 #endif
+
+namespace
+{
+// orders keys ascending or descending, as chosen when a map is created
+template<class Key>
+class directed_less
+{
+  public:
+    directed_less() = default;
+
+    explicit directed_less(const bool descending) noexcept
+        : m_descending(descending)
+    {}
+
+    bool operator()(const Key& lhs, const Key& rhs) const
+    {
+        return m_descending ? rhs < lhs : lhs < rhs;
+    }
+
+  private:
+    bool m_descending = false;
+};
+
+// An object type that, like std::unordered_map, enumerates its entries in no
+// fixed order - ascending or descending by key, depending on how the map was
+// created - and whose operator== does not depend on that order.
+// std::unordered_map itself cannot be used here: the standard does not
+// require it to accept an incomplete mapped type such as basic_json, and
+// libstdc++ 6 to 9 as well as the EDG front ends of icpc and nvc++ reject
+// basic_json<std::unordered_map>. std::map, the default object type, works
+// with all supported compilers.
+template<class Key, class Value, class /*Compare*/, class Allocator>
+struct unordered_object_t : std::map<Key, Value, directed_less<Key>, Allocator>
+{
+    using base_type = std::map<Key, Value, directed_less<Key>, Allocator>;
+    using base_type::base_type;
+
+    friend bool operator==(const unordered_object_t& lhs, const unordered_object_t& rhs)
+    {
+        return lhs.size() == rhs.size() && std::all_of(lhs.begin(), lhs.end(), [&rhs](const std::pair<const Key, Value>& entry)
+        {
+            const auto it = rhs.find(entry.first);
+            return it != rhs.end() && it->second == entry.second;
+        });
+    }
+
+    friend bool operator!=(const unordered_object_t& lhs, const unordered_object_t& rhs)
+    {
+        return !(lhs == rhs);
+    }
+};
+using unordered_json = nlohmann::basic_json<unordered_object_t>;
+
+// the entries "0" to "9", enumerated in ascending or in descending order
+unordered_json make_unordered_object(const bool descending)
+{
+    unordered_json j = unordered_json::object_t(directed_less<std::string>(descending));
+    for (int i = 0; i < 10; ++i)
+    {
+        j[std::to_string(i)] = i;
+    }
+    return j;
+}
+
+template<typename Json>
+Json nest(Json j, const std::size_t depth)
+{
+    for (std::size_t i = 0; i < depth; ++i)
+    {
+        Json outer = Json::object();
+        outer["x"] = std::move(j);
+        j = std::move(outer);
+    }
+    return j;
+}
+} // namespace
+
+TEST_CASE("equality of objects whose entries have no fixed order")
+{
+    // Values nested deeper than a bound are compared without the call stack,
+    // entry by entry. That must agree with the object type's own operator==,
+    // which for unordered_object_t (as for std::unordered_map) does not
+    // depend on the order of the entries, and for ordered_map does.
+    REQUIRE(make_unordered_object(true).begin().key() == "9");
+    REQUIRE(make_unordered_object(false).begin().key() == "0");
+
+    for (const std::size_t depth : std::vector<std::size_t> {0, 200})
+    {
+        CAPTURE(depth);
+
+        const unordered_json descending = nest(make_unordered_object(true), depth);
+        const unordered_json ascending = nest(make_unordered_object(false), depth);
+        CHECK(descending == ascending);
+        CHECK_FALSE(descending != ascending);
+
+        // a copy is equal to its original
+        const unordered_json copy = descending; // NOLINT(performance-unnecessary-copy-initialization)
+        CHECK(copy == descending);
+
+        // a different value, a different key, or another entry still count
+        unordered_json other_value = make_unordered_object(true);
+        other_value["5"] = 42;
+        CHECK_FALSE(nest(other_value, depth) == ascending);
+
+        unordered_json other_key = make_unordered_object(true);
+        other_key.erase("5");
+        other_key["50"] = 5;
+        CHECK_FALSE(nest(other_key, depth) == ascending);
+
+        unordered_json more_entries = make_unordered_object(true);
+        more_entries["10"] = 10;
+        CHECK_FALSE(nest(more_entries, depth) == ascending);
+        CHECK_FALSE(ascending == nest(more_entries, depth));
+
+        // ordered_json compares its entries in sequence
+        const nlohmann::ordered_json ab = nest(nlohmann::ordered_json({{"a", 1}, {"b", 2}}), depth);
+        const nlohmann::ordered_json ba = nest(nlohmann::ordered_json({{"b", 2}, {"a", 1}}), depth);
+        CHECK_FALSE(ab == ba);
+        CHECK(ab != ba);
+    }
+}
+
+TEST_CASE("containers are compared element by element")
+{
+    // Containers nested deeper than a bound are compared without the call
+    // stack, by code of their own; every relation is checked both at the top
+    // level and below that bound.
+    const auto deep = [](const json & j, const std::size_t depth)
+    {
+        json result = j;
+        for (std::size_t i = 0; i < depth; ++i)
+        {
+            result = json::array({std::move(result)});
+        }
+        return result;
+    };
+
+    for (const std::size_t depth : std::vector<std::size_t> {0, 200})
+    {
+        CAPTURE(depth);
+
+        // objects with different keys
+        {
+            const json a = deep({{"a", 1}}, depth);
+            const json b = deep({{"b", 1}}, depth);
+            CHECK_FALSE(a == b);
+            CHECK(a != b);
+            CHECK(a < b);
+            CHECK(b > a);
+            CHECK_FALSE(b < a);
+#if JSON_HAS_THREE_WAY_COMPARISON
+            // JSON_HAS_CPP_20 (do not remove; see note at top of file)
+            CHECK((a <=> b) == std::partial_ordering::less); // *NOPAD*
+            CHECK((b <=> a) == std::partial_ordering::greater); // *NOPAD*
+            CHECK((a <=> a) == std::partial_ordering::equivalent); // *NOPAD*
+#endif
+        }
+
+        // a container that is a prefix of the other one
+        {
+            // the one that runs out of elements first is the smaller one
+            const json shorter = deep({1}, depth);
+            const json longer = deep({1, 2}, depth);
+            CHECK(shorter < longer);
+            CHECK(longer > shorter);
+            CHECK_FALSE(longer < shorter);
+            CHECK_FALSE(shorter == longer);
+
+            const json smaller_object = deep({{"a", 1}}, depth);
+            const json larger_object = deep({{"a", 1}, {"b", 2}}, depth);
+            CHECK(smaller_object < larger_object);
+            CHECK(larger_object > smaller_object);
+            CHECK_FALSE(smaller_object == larger_object);
+#if JSON_HAS_THREE_WAY_COMPARISON
+            // JSON_HAS_CPP_20 (do not remove; see note at top of file)
+            CHECK((shorter <=> longer) == std::partial_ordering::less); // *NOPAD*
+            CHECK((longer <=> shorter) == std::partial_ordering::greater); // *NOPAD*
+#endif
+        }
+
+        // elements that cannot be ordered
+        {
+            const double nan = std::numeric_limits<double>::quiet_NaN();
+            const json lhs = deep({nan, 1}, depth);
+            const json rhs = deep({nan, 2}, depth);
+
+            CHECK_FALSE(lhs == lhs);
+            CHECK_FALSE(rhs < lhs);
+#if JSON_HAS_THREE_WAY_COMPARISON
+            // JSON_HAS_CPP_20 (do not remove; see note at top of file)
+            // operator<=> stops there, as std::lexicographical_compare_three_way
+            // does, and operator< is derived from it
+            CHECK((lhs <=> rhs) == std::partial_ordering::unordered); // *NOPAD*
+            CHECK_FALSE(lhs < rhs);
+#else
+            // operator< skips a pair of elements that cannot be ordered, as
+            // std::lexicographical_compare does, and the next pair decides
+            CHECK(lhs < rhs);
+#endif
+        }
+    }
+}

tests/src/unit-convenience.cpp

@@ -98,8 +98,10 @@ void check_escaped(const char* original, const char* escaped = "", bool ensure_a
 void check_escaped(const char* original, const char* escaped, const bool ensure_ascii)
 {
     std::stringstream ss;
-    json::serializer s(nlohmann::detail::output_adapter<char>(ss), ' ');
-    s.dump_escaped(original, ensure_ascii);
+    nlohmann::detail::output_stream_adapter<char> adapter(ss);
+    json::serializer s(adapter, ' ', false, ensure_ascii);
+    s.dump_escaped(original);
+    s.flush(); // dump_escaped writes into the serializer's internal buffer
     CHECK(ss.str() == escaped);
 }
 } // namespace

tests/src/unit-conversions.cpp

@@ -1389,6 +1389,37 @@ TEST_CASE("value conversion")
                 // CHECK(m5["one"] == "eins");
             }
 
+            SECTION("reserve is called on containers that support it (#5406)")
+            {
+                // build a larger object so that a missing/incorrect reserve()
+                // call would be more likely to corrupt or drop elements
+                json j_large;
+                for (int i = 0; i < 100; ++i)
+                {
+                    j_large[std::to_string(i)] = i;
+                }
+
+                SECTION("std::unordered_map (supports reserve)")
+                {
+                    const auto m = j_large.get<std::unordered_map<std::string, int>>();
+                    CHECK(m.size() == 100);
+                    for (int i = 0; i < 100; ++i)
+                    {
+                        CHECK(m.at(std::to_string(i)) == i);
+                    }
+                }
+
+                SECTION("std::map (no reserve, fallback path)")
+                {
+                    const auto m = j_large.get<std::map<std::string, int>>();
+                    CHECK(m.size() == 100);
+                    for (int i = 0; i < 100; ++i)
+                    {
+                        CHECK(m.at(std::to_string(i)) == i);
+                    }
+                }
+            }
+
             SECTION("std::multimap")
             {
                 j1.get<std::multimap<std::string, int>>();
@@ -1761,6 +1792,40 @@ TEST_CASE("std::filesystem::path")
 }
 #endif
 
+// the ADL to_json overload for std::u8string only exists under the same guard
+// as std::filesystem::path support (it is otherwise only reached indirectly,
+// via std::filesystem::path::u8string()) -- mirror both #if conditions from
+// include/nlohmann/detail/conversions/to_json.hpp exactly
+#if JSON_HAS_FILESYSTEM || JSON_HAS_EXPERIMENTAL_FILESYSTEM
+#if defined(__cpp_lib_char8_t)
+TEST_CASE("std::u8string")
+{
+    SECTION("ascii")
+    {
+        const std::u8string s = u8"Path";
+        json const j = s;
+
+        CHECK(j.template get<std::string>() == "Path");
+    }
+
+    SECTION("utf-8")
+    {
+        // use \u universal-character-names (rather than raw \x byte escapes
+        // or literal non-ASCII source bytes) to compose the multi-byte UTF-8
+        // encoding -- MSVC treats \x escapes used that way inside a u8
+        // literal as a nonstandard extension (warning C5321), which some of
+        // our CI configs promote to an error; \u is portable and produces
+        // the exact same encoded bytes without depending on the source
+        // file's encoding
+        const std::u8string s = u8"P\u011B\u0161ina";
+        json const j = s;
+
+        CHECK(j.template get<std::string>() == "P\xc4\x9b\xc5\xa1ina");
+    }
+}
+#endif
+#endif
+
 TEST_CASE("std::optional")
 {
     SECTION("null")

tests/src/unit-custom-array-type.cpp

@@ -0,0 +1,150 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+#include "doctest_compatibility.h"
+
+#include <nlohmann/json.hpp>
+
+#include <deque>
+#include <map>
+#include <memory>
+#include <string>
+#include <type_traits>
+#include <vector>
+
+namespace
+{
+
+// std::deque has no capacity() member function, which the library only needs
+// to detect a reallocation for JSON_DIAGNOSTICS
+using deque_json = nlohmann::basic_json<std::map, std::deque>;
+
+// a std::vector whose at() is hidden: the library performs its own bounds
+// check and must not fall back to the container's checked accessor
+template<class T, class Allocator = std::allocator<T>>
+class vector_without_at : public std::vector<T, Allocator>
+{
+  public:
+    vector_without_at() = default;
+
+    // the array of an initializer list is built from a range
+    template<class InputIt>
+    vector_without_at(InputIt first, InputIt last) : std::vector<T, Allocator>(first, last) {}
+
+    void at() = delete;
+};
+
+using no_at_json = nlohmann::basic_json<std::map, vector_without_at>;
+
+} // namespace
+
+TEST_CASE("array type without capacity()")
+{
+    SECTION("the iterators take their exception specification from the container")
+    {
+        // basic_json's iterators move exactly as the container iterators do:
+        // their move operations are defaulted without a declared noexcept,
+        // because an array or object type whose iterator is not nothrow move
+        // constructible would otherwise have them deleted (std::deque's is not
+        // with libstdc++ before 11, and neither are MSVC's debug iterators)
+        CHECK(std::is_nothrow_move_constructible<nlohmann::json::iterator>::value ==
+              (std::is_nothrow_move_constructible<nlohmann::json::object_t::iterator>::value
+               && std::is_nothrow_move_constructible<nlohmann::json::array_t::iterator>::value));
+        CHECK(std::is_nothrow_move_assignable<nlohmann::json::iterator>::value ==
+              (std::is_nothrow_move_assignable<nlohmann::json::object_t::iterator>::value
+               && std::is_nothrow_move_assignable<nlohmann::json::array_t::iterator>::value));
+        CHECK(std::is_nothrow_move_constructible<nlohmann::json::const_iterator>::value ==
+              (std::is_nothrow_move_constructible<nlohmann::json::object_t::const_iterator>::value
+               && std::is_nothrow_move_constructible<nlohmann::json::array_t::const_iterator>::value));
+
+        // and they are movable at all, which is what dropping the declared
+        // noexcept buys for a std::deque array
+        CHECK(std::is_move_constructible<deque_json::iterator>::value);
+        CHECK(std::is_move_assignable<deque_json::iterator>::value);
+    }
+
+    SECTION("adding elements")
+    {
+        deque_json j = deque_json::array();
+        j.push_back(1);
+        j.push_back("two");
+        j.emplace_back(3);
+        j += 4;
+
+        CHECK(j.size() == 4);
+        CHECK(j == deque_json({1, "two", 3, 4}));
+        CHECK(j.back() == 4);
+        CHECK(j.front() == 1);
+    }
+
+    SECTION("accessing and modifying elements")
+    {
+        auto j = deque_json::parse(R"([1,2,3])");
+
+        CHECK(j[1] == 2);
+        CHECK(j.at(2) == 3);
+
+        // growing through operator[] fills up with null values
+        j[5] = 6;
+        CHECK(j.size() == 6);
+        CHECK(j[4].is_null());
+        CHECK(j[5] == 6);
+
+        j.erase(0);
+        CHECK(j == deque_json({2, 3, nullptr, nullptr, 6}));
+
+        auto it = j.erase(j.begin());
+        CHECK(*it == 3);
+
+        j.insert(j.begin(), 1);
+        CHECK(j.front() == 1);
+    }
+
+    SECTION("serialization and deserialization")
+    {
+        const auto j = deque_json::parse(R"({"a":[1,[2,3]],"b":[]})");
+        CHECK(j.dump() == R"({"a":[1,[2,3]],"b":[]})");
+        CHECK(deque_json::parse(j.dump()) == j);
+        CHECK(deque_json::from_cbor(deque_json::to_cbor(j)) == j);
+
+        // empty containers are flattened to null and cannot be restored
+        const auto nested = deque_json::parse(R"({"a":[1,[2,3]]})");
+        CHECK(nested.flatten().unflatten() == nested);
+    }
+
+    SECTION("references stay valid while the array grows")
+    {
+        deque_json j = deque_json::array();
+        j.push_back(1);
+        auto& first = j[0];
+        for (int i = 0; i < 100; ++i)
+        {
+            j.push_back(i);
+        }
+        CHECK(&first == &j[0]);
+        CHECK(first == 1);
+    }
+}
+
+TEST_CASE("array type without at()")
+{
+    // built in memory rather than parsed, so that the exception message does
+    // not gain a byte range with JSON_DIAGNOSTIC_POSITIONS
+    no_at_json j = {1, 2, 3};
+    const auto& jc = j;
+
+    CHECK(j.at(0) == 1);
+    CHECK(j.at(2) == 3);
+    CHECK(jc.at(2) == 3);
+
+    CHECK_THROWS_WITH_AS(j.at(3), "[json.exception.out_of_range.401] array index 3 is out of range", no_at_json::out_of_range);
+    CHECK_THROWS_WITH_AS(jc.at(3), "[json.exception.out_of_range.401] array index 3 is out of range", no_at_json::out_of_range);
+
+    CHECK(j.at(no_at_json::json_pointer("/1")) == 2);
+    CHECK_THROWS_AS(j.at(no_at_json::json_pointer("/3")), no_at_json::out_of_range);
+}

tests/src/unit-custom-binary-type.cpp

@@ -0,0 +1,91 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+#include "doctest_compatibility.h"
+
+#include <nlohmann/json.hpp>
+
+#include <cstdint>
+#include <functional>
+#include <map>
+#include <memory>
+#include <string>
+#include <vector>
+
+#ifdef JSON_HAS_CPP_17
+    #include <cstddef>
+#endif
+
+namespace
+{
+
+// a BinaryType whose value type is signed: the elements must still be
+// processed as the numbers 0..255
+using char_binary_json = nlohmann::basic_json <
+                         std::map, std::vector, std::string, bool, std::int64_t, std::uint64_t,
+                         double, std::allocator, nlohmann::adl_serializer, std::vector<char>, void >;
+
+#ifdef JSON_HAS_CPP_17
+    // a BinaryType whose value type is not an integer type at all
+    using byte_binary_json = nlohmann::basic_json <
+    std::map, std::vector, std::string, bool, std::int64_t, std::uint64_t,
+    double, std::allocator, nlohmann::adl_serializer, std::vector<std::byte>, void >;
+#endif
+
+} // namespace
+
+TEST_CASE("binary type whose value type is not std::uint8_t")
+{
+    SECTION("a signed value type does not dump negative numbers")
+    {
+        const std::vector<char> chars{'\0', '\x01', '\xFF'};
+        CHECK(char_binary_json::binary(chars).dump() == R"({"bytes":[0,1,255],"subtype":null})");
+        CHECK(char_binary_json::binary(chars, 42).dump() == R"({"bytes":[0,1,255],"subtype":42})");
+        CHECK(char_binary_json::binary({}).dump() == R"({"bytes":[],"subtype":null})");
+    }
+
+    SECTION("a value is converted to the binary type if it is binary or an array")
+    {
+        const std::vector<char> chars{'\0', '\x01', '\x7F'};
+        CHECK(char_binary_json::binary(chars).get<std::vector<char>>() == chars);
+        CHECK(char_binary_json({0, 1, 127}).get<std::vector<char>>() == chars);
+        CHECK_THROWS_WITH_AS(char_binary_json(1).get<std::vector<char>>(),
+                             "[json.exception.type_error.302] type must be binary or array, but is number",
+                             char_binary_json::type_error&);
+    }
+
+    SECTION("the default binary type is unchanged")
+    {
+        CHECK(nlohmann::json::binary({0, 1, 255}, 42).dump() == R"({"bytes":[0,1,255],"subtype":42})");
+    }
+
+#ifdef JSON_HAS_CPP_17
+    SECTION("dumping a value type that is not an integer")
+    {
+        const std::vector<std::byte> bytes{std::byte{0}, std::byte{1}, std::byte{0xFF}};
+        CHECK(byte_binary_json::binary(bytes).dump() == R"({"bytes":[0,1,255],"subtype":null})");
+        CHECK(byte_binary_json::binary(bytes, 42).dump() == R"({"bytes":[0,1,255],"subtype":42})");
+        CHECK(byte_binary_json::binary({}).dump() == R"({"bytes":[],"subtype":null})");
+    }
+
+    SECTION("hashing and the binary formats")
+    {
+        const std::vector<std::byte> bytes{std::byte{0}, std::byte{1}, std::byte{0xFF}};
+        const auto j = byte_binary_json::binary(bytes);
+
+        CHECK(std::hash<byte_binary_json> {}(j) == std::hash<byte_binary_json> {}(j));
+        CHECK(byte_binary_json::from_cbor(byte_binary_json::to_cbor(j)) == j);
+        CHECK(byte_binary_json::from_msgpack(byte_binary_json::to_msgpack(j)) == j);
+
+        // UBJSON has no binary type, so binary values are written as an array
+        CHECK(byte_binary_json::from_ubjson(byte_binary_json::to_ubjson(j)) == byte_binary_json({0, 1, 255}));
+        // the same holds for BON8
+        CHECK(byte_binary_json::from_bon8(byte_binary_json::to_bon8(j)) == byte_binary_json({0, 1, 255}));
+    }
+#endif
+}

tests/src/unit-custom-object-type.cpp

@@ -0,0 +1,324 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+#include "doctest_compatibility.h"
+
+#include <nlohmann/json.hpp>
+
+#include <cstdint>
+#include <map>
+#include <string>
+#include <type_traits>
+#include <utility>
+#include <vector>
+
+
+namespace
+{
+
+// An ObjectType that does *not* define a key_compare member type, which is
+// what every hash map looks like to the library.
+//
+// A hash map is deliberately not used here: object_t is probed for
+// key_compare inside the definition of basic_json, that is, while basic_json
+// is still an incomplete type, and whether a hash map can be instantiated
+// with an incomplete mapped type depends on the standard library (libstdc++ 9
+// needs the size of the mapped type for its node type and rejects it). So the
+// object type wraps a std::map instead of inheriting from it: an earlier
+// version derived from std::map and shadowed the inherited key_compare type
+// with a same-named member function, relying on ordinary member hiding to
+// make key_compare unreachable as a type. MSVC 2017 (AppVeyor, /std:c++17)
+// does not honor that hiding for a typename-qualified lookup performed from
+// outside the class and still resolves key_compare to the base's comparator
+// type, so the library's probe incorrectly found one. Composition sidesteps
+// the question entirely: with no base class, there is no key_compare to find
+// under any lookup rule.
+template<class Key, class T, class Compare, class Allocator>
+class no_key_compare_map
+{
+    using map_t = std::map<Key, T, Compare, Allocator>;
+    map_t data;
+
+  public:
+    using key_type = typename map_t::key_type;
+    using mapped_type = typename map_t::mapped_type;
+    using value_type = typename map_t::value_type;
+    using size_type = typename map_t::size_type;
+    using allocator_type = typename map_t::allocator_type;
+    using iterator = typename map_t::iterator;
+    using const_iterator = typename map_t::const_iterator;
+
+    // -Weffc++ asks for the member to be initialized in the member
+    // initialization list, which a defaulted constructor does not do; the
+    // exception specification a defaulted one would have carried has to be
+    // written out as well, or -Wnoexcept objects where the standard library
+    // takes noexcept(construct(...))
+    no_key_compare_map() noexcept(std::is_nothrow_default_constructible<map_t>::value) : data() {}
+
+    // converting between two basic_json types builds the object from a range
+    template<class InputIt>
+    no_key_compare_map(InputIt first, InputIt last) : data(first, last) {}
+
+    iterator begin() noexcept
+    {
+        return data.begin();
+    }
+    iterator end() noexcept
+    {
+        return data.end();
+    }
+    const_iterator begin() const noexcept
+    {
+        return data.begin();
+    }
+    const_iterator end() const noexcept
+    {
+        return data.end();
+    }
+    const_iterator cbegin() const noexcept
+    {
+        return data.cbegin();
+    }
+    const_iterator cend() const noexcept
+    {
+        return data.cend();
+    }
+
+    bool empty() const noexcept
+    {
+        return data.empty();
+    }
+    size_type size() const noexcept
+    {
+        return data.size();
+    }
+    size_type max_size() const noexcept
+    {
+        return data.max_size();
+    }
+    void clear() noexcept
+    {
+        data.clear();
+    }
+
+    iterator find(const key_type& key)
+    {
+        return data.find(key);
+    }
+    const_iterator find(const key_type& key) const
+    {
+        return data.find(key);
+    }
+    size_type count(const key_type& key) const
+    {
+        return data.count(key);
+    }
+
+    std::pair<iterator, bool> emplace(const key_type& key, const mapped_type& value)
+    {
+        return data.emplace(key, value);
+    }
+
+    std::pair<iterator, bool> insert(const value_type& value)
+    {
+        return data.insert(value);
+    }
+
+    template<class InputIt>
+    void insert(InputIt first, InputIt last)
+    {
+        data.insert(first, last);
+    }
+
+    mapped_type& operator[](const key_type& key)
+    {
+        return data[key];
+    }
+
+    mapped_type& at(const key_type& key)
+    {
+        return data.at(key);
+    }
+    const mapped_type& at(const key_type& key) const
+    {
+        return data.at(key);
+    }
+
+    iterator erase(iterator pos)
+    {
+        return data.erase(pos);
+    }
+    iterator erase(iterator first, iterator last)
+    {
+        return data.erase(first, last);
+    }
+    size_type erase(const key_type& key)
+    {
+        return data.erase(key);
+    }
+
+    void swap(no_key_compare_map& other) noexcept(noexcept(data.swap(other.data)))
+    {
+        data.swap(other.data);
+    }
+
+    friend bool operator==(const no_key_compare_map& lhs, const no_key_compare_map& rhs)
+    {
+        return lhs.data == rhs.data;
+    }
+    friend bool operator<(const no_key_compare_map& lhs, const no_key_compare_map& rhs)
+    {
+        return lhs.data < rhs.data;
+    }
+};
+
+using no_key_compare_json = nlohmann::basic_json<no_key_compare_map>;
+
+// An ObjectType whose erase(iterator) returns void rather than the following
+// iterator, as for instance Abseil's hash maps do
+template<class Key, class T, class Compare, class Allocator>
+struct void_erase_map : std::map<Key, T, Compare, Allocator>
+{
+    using base_t = std::map<Key, T, Compare, Allocator>;
+    using iterator = typename base_t::iterator;
+    using base_t::erase;
+
+    void erase(iterator pos)
+    {
+        base_t::erase(pos);
+    }
+};
+
+using void_erase_json = nlohmann::basic_json<void_erase_map>;
+
+} // namespace
+
+TEST_CASE("object type whose erase() returns void")
+{
+    SECTION("erasing every element through the returned iterator")
+    {
+        void_erase_json j;
+        for (int i = 0; i < 8; ++i)
+        {
+            j["k" + std::to_string(i)] = i;
+        }
+
+        std::size_t erased = 0;
+        for (auto it = j.begin(); it != j.end(); ++erased)
+        {
+            it = j.erase(it);
+        }
+        CHECK(erased == 8);
+        CHECK(j.empty());
+    }
+
+    SECTION("erasing in the middle returns the following element")
+    {
+        void_erase_json j;
+        for (int i = 0; i < 4; ++i)
+        {
+            j["k" + std::to_string(i)] = i;
+        }
+
+        auto it = j.begin();
+        ++it;
+        const auto after = j.erase(it);
+        CHECK(j.size() == 3);
+        CHECK(after.key() == "k2");
+        CHECK(after.value() == 2);
+        CHECK(!j.contains("k1"));
+    }
+
+    SECTION("the other erase overloads are unaffected")
+    {
+        void_erase_json j;
+        j["a"] = 1;
+        j["b"] = 2;
+        j["c"] = 3;
+
+        CHECK(j.erase("a") == 1);
+        CHECK(j.erase("nope") == 0);
+        j.erase(j.begin(), j.end());
+        CHECK(j.empty());
+    }
+}
+
+TEST_CASE("object type without key_compare")
+{
+    SECTION("object_comparator_t falls back to default_object_comparator_t")
+    {
+        CHECK(std::is_same < no_key_compare_json::object_comparator_t,
+              no_key_compare_json::default_object_comparator_t >::value);
+    }
+
+    SECTION("object types defining key_compare are unaffected")
+    {
+        CHECK(std::is_same<nlohmann::json::object_comparator_t,
+              nlohmann::json::object_t::key_compare>::value);
+        CHECK(std::is_same<nlohmann::ordered_json::object_comparator_t,
+              nlohmann::ordered_json::object_t::key_compare>::value);
+    }
+
+    SECTION("creating and accessing values")
+    {
+        no_key_compare_json j;
+        j["one"] = 1;
+        j["two"] = "zwei";
+        j["three"]["nested"] = true;
+
+        CHECK(j.size() == 3);
+        CHECK(j.at("one") == 1);
+        CHECK(j["two"] == "zwei");
+        CHECK(j["three"]["nested"] == true);
+        CHECK(j.contains("one"));
+        CHECK(!j.contains("four"));
+        CHECK(j.find("one") != j.end());
+        CHECK(j.count("one") == 1);
+        CHECK(j.erase("one") == 1);
+        CHECK(j.size() == 2);
+    }
+
+    SECTION("serialization and deserialization")
+    {
+        const auto j = no_key_compare_json::parse(R"({"a":[1,2,3],"b":{"c":null}})");
+        CHECK(j["a"].size() == 3);
+        CHECK(j["a"][2] == 3);
+        CHECK(j["b"]["c"].is_null());
+        CHECK(no_key_compare_json::parse(j.dump()) == j);
+    }
+
+    SECTION("binary formats")
+    {
+        const auto j = no_key_compare_json::parse(R"({"a":[1,2,3],"b":"x"})");
+        CHECK(no_key_compare_json::from_cbor(no_key_compare_json::to_cbor(j)) == j);
+        CHECK(no_key_compare_json::from_msgpack(no_key_compare_json::to_msgpack(j)) == j);
+        CHECK(no_key_compare_json::from_bon8(no_key_compare_json::to_bon8(j)) == j);
+    }
+
+    SECTION("flatten and unflatten")
+    {
+        // "o" has a key that looks like an array index, so unflatten() must
+        // not turn it into an array
+        const auto j = no_key_compare_json::parse(
+                           R"({"c":[1,2,3],"d":{"e":"s"},"n":[[0,1],[2]],"o":{"2":"x"}})");
+        CHECK(j.flatten().unflatten() == j);
+    }
+
+    SECTION("conversion to and from nlohmann::json")
+    {
+        const auto j = no_key_compare_json::parse(R"({"a":1,"b":[true,null]})");
+        const nlohmann::json converted(j);
+
+        CHECK(converted.is_object());
+        CHECK(converted["a"] == 1);
+        CHECK(converted["b"][0] == true);
+        CHECK(converted["b"][1].is_null());
+        CHECK(no_key_compare_json(converted) == j);
+    }
+}
+

tests/src/unit-deserialization.cpp

@@ -8,6 +8,14 @@
 
 #include "doctest_compatibility.h"
 
+// capture whether JSON_STRICT_NUL_HANDLING was enabled on the command line
+// (e.g. -DJSON_STRICT_NUL_HANDLING=1) *before* including json.hpp, since the
+// library #undefs JSON_STRICT_NUL_HANDLING itself once the header has been
+// fully processed (see include/nlohmann/detail/macro_unscope.hpp)
+#if defined(JSON_STRICT_NUL_HANDLING) && (JSON_STRICT_NUL_HANDLING == 1)
+    #define JSON_TEST_STRICT_NUL_HANDLING_ENABLED 1
+#endif
+
 #include <nlohmann/json.hpp>
 using nlohmann::json;
 #ifdef JSON_TEST_NO_GLOBAL_UDLS
@@ -17,6 +25,7 @@ using nlohmann::json;
 #include <iostream>
 #include <iterator>
 #include <sstream>
+#include <string>
 #include <valarray>
 
 #if defined(_WIN32)
@@ -323,6 +332,23 @@ TEST_CASE("deserialization")
             CHECK(j == json({"foo", 1, 2, 3, false, {{"one", 1}}}));
         }
 
+        SECTION("operator>> with a NUL byte after the value (issue #5530)")
+        {
+            // operator>> parses non-strictly (it does not require the whole
+            // stream to be consumed), so a NUL byte following a complete
+            // value is simply left unread on the stream and never reaches
+            // the "expected end of input" check that JSON_STRICT_NUL_HANDLING
+            // affects; this holds regardless of the macro (verified below for
+            // the opt-in state as well)
+            std::string data = "123";
+            data.push_back('\0');
+            std::istringstream ss(data);
+            json j;
+            ss >> j;
+            CHECK(j == json(123));
+            CHECK(ss.good());
+        }
+
         SECTION("user-defined string literal")
         {
             CHECK("[\"foo\",1,2,3,false,{\"one\":1}]"_json == json({"foo", 1, 2, 3, false, {{"one", 1}}}));
@@ -405,6 +431,27 @@ TEST_CASE("deserialization")
             CHECK_THROWS_WITH_AS(ss >> j, "[json.exception.parse_error.101] parse error at line 1, column 29: syntax error while parsing array - unexpected end of input; expected ']'", json::parse_error&);
         }
 
+#if defined(JSON_TEST_STRICT_NUL_HANDLING_ENABLED)
+        SECTION("operator>> with a NUL byte where a value is expected (JSON_STRICT_NUL_HANDLING == 1, issue #5530)")
+        {
+            // a trailing NUL byte *after* a complete value is unaffected by the
+            // macro (see the successful-deserialization "operator>> with a NUL
+            // byte after the value" section above): operator>> parses
+            // non-strictly and never reaches the "expected end of input" check
+            // that the macro changes. A NUL byte where a *value* is expected,
+            // however, goes through the same token dispatch as any other input
+            // and is affected: with the macro enabled it now raises
+            // parse_error.101 (like any other unrecognized byte) instead of
+            // being silently treated the same as an empty stream.
+            std::string const data(1, '\0');
+            std::istringstream ss(data);
+            json j;
+            CHECK_THROWS_WITH_AS(ss >> j,
+                                 "[json.exception.parse_error.101] parse error at line 1, column 1: syntax error while parsing value - invalid literal; last read: '<U+0000>'",
+                                 json::parse_error&);
+        }
+#endif
+
         SECTION("user-defined string literal")
         {
             CHECK_THROWS_WITH_AS("[\"foo\",1,2,3,false,{\"one\":1}"_json, "[json.exception.parse_error.101] parse error at line 1, column 29: syntax error while parsing array - unexpected end of input; expected ']'", json::parse_error&);
@@ -453,7 +500,11 @@ TEST_CASE("deserialization")
 
             SECTION("from std::array")
             {
-                std::array<uint8_t, 5> const v { {'t', 'r', 'u', 'e'} };
+                // sized to exactly the length of "true": a size of 5 would leave
+                // a value-initialized trailing 0x00 element that is only
+                // silently accepted as end-of-input by default and would fail
+                // under JSON_STRICT_NUL_HANDLING
+                std::array<uint8_t, 4> const v { {'t', 'r', 'u', 'e'} };
                 CHECK(json::parse(v) == json(true));
                 CHECK(json::accept(v));
 
@@ -549,7 +600,9 @@ TEST_CASE("deserialization")
 
             SECTION("from std::array")
             {
-                std::array<uint8_t, 5> v { {'t', 'r', 'u', 'e'} };
+                // sized to exactly the length of "true", see the analogous
+                // "from std::array" section above for why
+                std::array<uint8_t, 4> v { {'t', 'r', 'u', 'e'} };
                 CHECK(json::parse(std::begin(v), std::end(v)) == json(true));
                 CHECK(json::accept(std::begin(v), std::end(v)));
 
@@ -1181,6 +1234,59 @@ TEST_CASE("deserialization")
         }
     }
 
+    SECTION("stream position after extraction without JSON_PRECISE_STREAM_POSITION (#5340)")
+    {
+        // By default, the character that terminates a number is consumed, so
+        // the stream is left one byte too far after a number (and only after a
+        // number). JSON_PRECISE_STREAM_POSITION changes this; see
+        // unit-precise-stream-position.cpp. These checks pin the default.
+        const auto remaining = [](std::istream & is) -> std::string
+        {
+            return {std::istreambuf_iterator<char>(is), std::istreambuf_iterator<char>()};
+        };
+
+        SECTION("the character after a number is consumed")
+        {
+            std::istringstream ss("1true");
+            json j;
+            ss >> j;
+            CHECK(j == 1);
+            CHECK(remaining(ss) == "rue");
+        }
+
+        SECTION("the character after other values is not consumed")
+        {
+            std::istringstream ss("[1]true");
+            json j;
+            ss >> j;
+            CHECK(j == json::parse("[1]"));
+            CHECK(remaining(ss) == "true");
+        }
+
+        SECTION("comma-separated numbers can be read one by one")
+        {
+            std::istringstream ss("1,2,3");
+            json j1;
+            json j2;
+            json j3;
+            ss >> j1 >> j2 >> j3;
+            CHECK(j1 == 1);
+            CHECK(j2 == 2);
+            CHECK(j3 == 3);
+        }
+
+        SECTION("std::getline after a number skips the line break")
+        {
+            std::istringstream ss("42\nfoo");
+            json j;
+            std::string line;
+            ss >> j;
+            std::getline(ss, line);
+            CHECK(j == 42);
+            CHECK(line == "foo");
+        }
+    }
+
     // build with C++20
     // JSON_HAS_CPP_20
 #if defined(__cpp_char8_t)

tests/src/unit-diagnostic-positions-only.cpp

@@ -1,44 +0,0 @@
-//     __ _____ _____ _____
-//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
-// |  |  |__   |  |  | | | |  version 3.12.0
-// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
-//
-// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
-// SPDX-License-Identifier: MIT
-
-#include "doctest_compatibility.h"
-
-#ifdef JSON_DIAGNOSTICS
-    #undef JSON_DIAGNOSTICS
-#endif
-
-#define JSON_DIAGNOSTICS 0
-#define JSON_DIAGNOSTIC_POSITIONS 1
-#include <nlohmann/json.hpp>
-
-using json = nlohmann::json;
-
-TEST_CASE("Better diagnostics with positions only")
-{
-    SECTION("invalid type")
-    {
-        const std::string json_invalid_string = R"(
-        {
-            "address": {
-                "street": "Fake Street",
-                "housenumber": "1"
-            }
-        }
-        )";
-        json j = json::parse(json_invalid_string);
-        CHECK_THROWS_WITH_AS(j.at("address").at("housenumber").get<int>(),
-                             "[json.exception.type_error.302] (bytes 108-111) type must be number, but is string", json::type_error);
-    }
-
-    SECTION("invalid type without positions")
-    {
-        const json j = "foo";
-        CHECK_THROWS_WITH_AS(j.get<int>(),
-                             "[json.exception.type_error.302] type must be number, but is string", json::type_error);
-    }
-}

tests/src/unit-diagnostic-positions.cpp

@@ -8,7 +8,9 @@
 
 #include "doctest_compatibility.h"
 
-#define JSON_DIAGNOSTICS 1
+#ifndef JSON_DIAGNOSTICS
+    #define JSON_DIAGNOSTICS 1
+#endif
 #define JSON_DIAGNOSTIC_POSITIONS 1
 #include <nlohmann/json.hpp>
 
@@ -27,8 +29,13 @@ TEST_CASE("Better diagnostics with positions")
         }
         )";
         json j = json::parse(json_invalid_string);
+#if JSON_DIAGNOSTICS
         CHECK_THROWS_WITH_AS(j.at("address").at("housenumber").get<int>(),
                              "[json.exception.type_error.302] (/address/housenumber) (bytes 108-111) type must be number, but is string", json::type_error);
+#else
+        CHECK_THROWS_WITH_AS(j.at("address").at("housenumber").get<int>(),
+                             "[json.exception.type_error.302] (bytes 108-111) type must be number, but is string", json::type_error);
+#endif
     }
 
     SECTION("invalid type without positions")
@@ -38,13 +45,149 @@ TEST_CASE("Better diagnostics with positions")
                              "[json.exception.type_error.302] type must be number, but is string", json::type_error);
     }
 
+    SECTION("positions of strings containing escape sequences")
+    {
+        // escape sequences make the token longer than the string it parses to,
+        // so the positions must not be derived from the parsed value's length
+        const auto check = [](const std::string & text, const std::string & token)
+        {
+            CAPTURE(text)
+            CAPTURE(token)
+            const json j = json::parse(text);
+            const json& v = j.at("a");
+            CHECK(text.substr(v.start_pos(), v.end_pos() - v.start_pos()) == token);
+        };
+
+        check(R"({"a":"plain"})", R"("plain")");
+        check(R"({"a":"tab\there"})", R"("tab\there")");
+        check(R"({"a":"\n\n\n\n\n\n"})", R"("\n\n\n\n\n\n")");
+        check(R"({"a":"\""})", R"("\"")");
+        check(R"({"a":"\\"})", R"("\\")");
+        check(R"({"a":"é"})", R"("é")");
+        check(R"({"a":"🌞"})", R"("🌞")");
+        check("{\"a\":\"\xc3\xa9\"}", "\"\xc3\xa9\"");  // multi-byte UTF-8, no escapes
+
+        // a string at the root, where an escape would otherwise push the
+        // reported start position past the opening quote
+        const std::string root = R"("a\tb")";
+        const json j = json::parse(root);
+        CHECK(j.start_pos() == 0);
+        CHECK(j.end_pos() == root.size());
+    }
+
+    SECTION("copying keeps the positions of nested values (#5387)")
+    {
+        // Values nested deeper than the copy constructor's descent bound are
+        // copied without the call stack, on a path that has to carry the
+        // positions over itself; shallower ones copy their containers, which
+        // bring the positions along. Both sides of the bound are checked here.
+        const auto check_copy = [](std::size_t depth, bool objects)
+        {
+            CAPTURE(depth)
+            CAPTURE(objects)
+
+            const std::string opening = objects ? R"({"a":)" : "[";
+            const std::string closing = objects ? "}" : "]";
+
+            std::string text;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                text += opening;
+            }
+            text += "12";
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                text += closing;
+            }
+
+            const json original = json::parse(text);
+            const json copy(original); // NOLINT(performance-unnecessary-copy-initialization)
+
+            const json* o = &original;
+            const json* c = &copy;
+            for (std::size_t level = 0; level <= depth; ++level)
+            {
+                CAPTURE(level)
+                REQUIRE(c->start_pos() == o->start_pos());
+                REQUIRE(c->end_pos() == o->end_pos());
+
+                if (level < depth)
+                {
+                    o = objects ? &o->at("a") : &o->at(0);
+                    c = objects ? &c->at("a") : &c->at(0);
+                }
+            }
+        };
+
+        const auto check_arrays = [&check_copy](std::size_t depth)
+        {
+            check_copy(depth, false);
+        };
+        const auto check_objects = [&check_copy](std::size_t depth)
+        {
+            check_copy(depth, true);
+        };
+
+        check_arrays(1);
+        check_arrays(127);
+        check_arrays(128);
+        check_arrays(129);
+        check_arrays(300);
+
+        check_objects(1);
+        check_objects(127);
+        check_objects(128);
+        check_objects(129);
+        check_objects(300);
+    }
+
     SECTION("JSON patch add to primitive parent (#4292)")
     {
         // the JSON Patch "add" target /foo/bar/baz has a string parent
         // (/foo/bar); the position of that parent is reported in the message
         const json doc = json::parse(R"({"foo":{"bar":"a string"}})");
         const json patch = json::parse(R"([{"op":"add","path":"/foo/bar/baz","value":1}])");
+#if JSON_DIAGNOSTICS
         CHECK_THROWS_WITH_AS(doc.patch(patch),
                              "[json.exception.out_of_range.411] (/foo/bar) (bytes 14-24) cannot add value: the JSON Patch 'add' target's parent is of type string, but must be an object or array", json::out_of_range);
+#else
+        CHECK_THROWS_WITH_AS(doc.patch(patch),
+                             "[json.exception.out_of_range.411] (bytes 14-24) cannot add value: the JSON Patch 'add' target's parent is of type string, but must be an object or array", json::out_of_range);
+#endif
+    }
+}
+
+TEST_CASE("values read from a binary format have no positions")
+{
+    // only the JSON lexer knows where a value started and ended
+    const json source = {{"a", {1, "x", json::binary({1})}}, {"b", {{"c", true}}}, {"d", nullptr}, {"e", 1.5}};
+    const std::vector<std::uint8_t> cbor = json::to_cbor(source);
+
+    const auto check_no_positions = [](const json & j)
+    {
+        CHECK(j.start_pos() == std::string::npos);
+        CHECK(j.end_pos() == std::string::npos);
+        CHECK(j.at("a").start_pos() == std::string::npos);
+        CHECK(j.at("a").at(1).end_pos() == std::string::npos);
+        CHECK(j.at("b").at("c").start_pos() == std::string::npos);
+    };
+
+    SECTION("DOM parser")
+    {
+        const json j = json::from_cbor(cbor);
+        CHECK(j == source);
+        check_no_positions(j);
+    }
+
+    SECTION("DOM parser with a callback")
+    {
+        json j;
+        nlohmann::detail::json_sax_dom_callback_parser<json, decltype(nlohmann::detail::input_adapter(cbor))> sdp(j, [](int /*unused*/, json::parse_event_t /*unused*/, const json& /*unused*/) noexcept
+        {
+            return true;
+        });
+        CHECK(json::sax_parse(cbor, &sdp, json::input_format_t::cbor));
+        CHECK(j == source);
+        check_no_positions(j);
     }
 }

tests/src/unit-diagnostics.cpp

@@ -273,5 +273,242 @@ TEST_CASE("Regression tests for extended diagnostics")
         CHECK(j1["numbers"]["two"] == 2);
         CHECK(j1["string"] == "t");
     }
+
+    SECTION("Regression test for issue #5387 - copying keeps the parents of nested values")
+    {
+        // A value nested deeper than the copy constructor's descent bound is
+        // copied without the call stack. Every container that path creates has
+        // to have the parents of its children set, or the JSON Pointer in the
+        // diagnostic is cut short.
+        const std::size_t depth = 300;
+
+        SECTION("objects")
+        {
+            json j = "not a number";
+            std::string pointer;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                j = json{{"a", j}};
+                pointer += "/a";
+            }
+
+            json const copy(j); // NOLINT(performance-unnecessary-copy-initialization)
+
+            const json* inner = &copy;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                inner = &inner->at("a");
+            }
+
+            std::string const expected = "[json.exception.type_error.302] (" + pointer + ") type must be number, but is string";
+            int i = 0;
+            CHECK_THROWS_WITH_AS(i = inner->get<int>(), expected.c_str(), json::type_error);
+            CHECK(i == 0);
+        }
+
+        SECTION("arrays")
+        {
+            json j = "not a number";
+            std::string pointer;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                j = json::array({j});
+                pointer += "/0";
+            }
+
+            json const copy(j); // NOLINT(performance-unnecessary-copy-initialization)
+
+            const json* inner = &copy;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                inner = &inner->at(0);
+            }
+
+            std::string const expected = "[json.exception.type_error.302] (" + pointer + ") type must be number, but is string";
+            int i = 0;
+            CHECK_THROWS_WITH_AS(i = inner->get<int>(), expected.c_str(), json::type_error);
+            CHECK(i == 0);
+        }
+    }
+
+    SECTION("Regression test - swap(array_t&)/swap(object_t&) must update JSON_DIAGNOSTICS parent pointers")
+    {
+        // swap(array_t&)
+        {
+            json j = json::array();
+            json::array_t arr = {json::array({1})};
+            j.swap(arr);
+
+            // parent pointers of the moved-in elements must point into j, not
+            // into the now-defunct free-standing array_t
+            CHECK_THROWS_WITH_AS(j[0][0].get<std::string>(), "[json.exception.type_error.302] (/0/0) type must be string, but is number", json::type_error);
+
+            // must not trigger assert_invariant() in a debug/assert-enabled build
+            json const k = j;
+            CHECK(k == j);
+        }
+
+        // swap(object_t&)
+        {
+            json o = json::object();
+            json::object_t obj = {{"a", json::array({1})}};
+            o.swap(obj);
+
+            CHECK_THROWS_WITH_AS(o["a"][0].get<std::string>(), "[json.exception.type_error.302] (/a/0) type must be string, but is number", json::type_error);
+
+            // must not trigger assert_invariant() in a debug/assert-enabled build
+            json const p = o;
+            CHECK(p == o);
+        }
+    }
+
+    SECTION("Regression test - erase() and update() must keep JSON_DIAGNOSTICS parent pointers of ordered_json members")
+    {
+        // ordered_json keeps its members in a vector: erasing a member
+        // re-constructs all members after it in place, and adding a key may
+        // reallocate the vector; both reset the parent pointers of the members
+        // that were moved
+        using nlohmann::ordered_json;
+
+        const auto check_parents = [](const ordered_json & j)
+        {
+            // const access, so operator[] cannot repair the parent pointers
+            CHECK_THROWS_WITH_AS(j["z"]["x"].at(0), "[json.exception.type_error.304] (/z/x) cannot use at() with number", ordered_json::type_error);
+
+            // must not trigger assert_invariant() in a debug/assert-enabled build
+            ordered_json const copy = j; // NOLINT(performance-unnecessary-copy-initialization)
+            CHECK(copy == j);
+        };
+
+        // erase(key)
+        {
+            ordered_json j = {{"a", 1}, {"z", {{"x", 1}}}};
+            CHECK(j.erase("a") == 1);
+            check_parents(j);
+        }
+
+        // erase(iterator)
+        {
+            ordered_json j = {{"a", 1}, {"z", {{"x", 1}}}};
+            j.erase(j.begin());
+            check_parents(j);
+        }
+
+        // erase(iterator, iterator)
+        {
+            ordered_json j = {{"a", 1}, {"b", 2}, {"z", {{"x", 1}}}};
+            j.erase(j.begin(), j.find("z"));
+            check_parents(j);
+        }
+
+        // patch() removes via erase(iterator)
+        {
+            ordered_json j = {{"a", 1}, {"z", {{"x", 1}}}};
+            j.patch_inplace(ordered_json::parse(R"([{"op": "remove", "path": "/a"}])"));
+            check_parents(j);
+        }
+
+        // update(j)
+        {
+            ordered_json j = {{"z", {{"x", 1}}}};
+            j.update({{"a", 1}, {"b", 2}});
+            check_parents(j);
+        }
+
+        // update(j, true), the outer and the nested vector both grow
+        {
+            ordered_json j = {{"z", {{"x", 1}}}};
+            j.update({{"z", {{"y", 2}}}, {"a", 1}}, true);
+            check_parents(j);
+        }
+
+        // update(j, true) around its descent bound, where the nested vectors
+        // grow while the objects are merged without recursing
+        for (const std::size_t depth :
+                {
+                    nlohmann::detail::recursion_depth_limit() - 1, nlohmann::detail::recursion_depth_limit(), nlohmann::detail::recursion_depth_limit() + 2
+                })
+        {
+            ordered_json j = {{"z", {{"x", 1}}}};
+            ordered_json patch = {{"a", 1}, {"b", 2}, {"c", {{"d", 3}}}};
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                j = ordered_json{{"k", 0}, {"n", std::move(j)}};
+                patch = ordered_json{{"n", std::move(patch)}, {"l", 1}, {"m", 2}};
+            }
+            j.update(patch, true);
+
+            // must not trigger assert_invariant() on any level in a
+            // debug/assert-enabled build
+            ordered_json const copy = j; // NOLINT(performance-unnecessary-copy-initialization)
+            CHECK(copy == j);
+        }
+
+        // merge_patch() inserts "c" and removes "d" at /a/c, then inserts "e"
+        // at /a, which copies /a/c
+        {
+            auto j = ordered_json::parse(R"({"a": {"c": {"d": {}}}})");
+            j.merge_patch(ordered_json::parse(R"({"a": {"c": {"c": "s", "d": null}, "e": "s"}})"));
+            CHECK(j.dump() == R"({"a":{"c":{"c":"s"},"e":"s"}})");
+
+            auto const& constJ = j;
+#if JSON_DIAGNOSTIC_POSITIONS
+            CHECK_THROWS_WITH_AS(constJ["a"]["c"]["c"].at(0), "[json.exception.type_error.304] (/a/c/c) (bytes 18-21) cannot use at() with string", ordered_json::type_error);
+#else
+            CHECK_THROWS_WITH_AS(constJ["a"]["c"]["c"].at(0), "[json.exception.type_error.304] (/a/c/c) cannot use at() with string", ordered_json::type_error);
+#endif
+            ordered_json const copy = j;
+            CHECK(copy == j);
+        }
+    }
 }
 
+TEST_CASE("Better diagnostics past the descent bound of update() and merge_patch()")
+{
+    // Both merge objects nested more than detail::recursion_depth_limit()
+    // (128) levels deep without recursing; the values they add or replace
+    // there must still know their parents.
+    // The values are built rather than parsed, so that the expected messages
+    // carry no byte positions under JSON_DIAGNOSTIC_POSITIONS.
+    const std::size_t depth = 200;
+    json target = {{"x", 1}};
+    json patch = {{"y", 2}};
+    std::string path;
+    for (std::size_t i = 0; i < depth; ++i)
+    {
+        target = json{{"a", std::move(target)}};
+        patch = json{{"a", std::move(patch)}};
+        path += "/a";
+    }
+    const std::string expected_x = "[json.exception.type_error.304] (" + path + "/x) cannot use at() with number";
+    const std::string expected_y = "[json.exception.type_error.304] (" + path + "/y) cannot use at() with number";
+
+    SECTION("update()")
+    {
+        json j = target;
+        j.update(patch, true);
+
+        // walk down through const references, which leave m_parent alone
+        const json* p = &j;
+        for (std::size_t i = 0; i < depth; ++i)
+        {
+            p = &p->at("a");
+        }
+        CHECK_THROWS_WITH_AS(p->at("x").at(0), expected_x.c_str(), json::type_error);
+        CHECK_THROWS_WITH_AS(p->at("y").at(0), expected_y.c_str(), json::type_error);
+    }
+
+    SECTION("merge_patch()")
+    {
+        json j = target;
+        j.merge_patch(patch);
+
+        const json* p = &j;
+        for (std::size_t i = 0; i < depth; ++i)
+        {
+            p = &p->at("a");
+        }
+        CHECK_THROWS_WITH_AS(p->at("x").at(0), expected_x.c_str(), json::type_error);
+        CHECK_THROWS_WITH_AS(p->at("y").at(0), expected_y.c_str(), json::type_error);
+    }
+}

tests/src/unit-element_access2.cpp

@@ -1517,6 +1517,16 @@ TEST_CASE_TEMPLATE("element access 2 (throwing tests)", Json, nlohmann::json, nl
                     CHECK(j.value("/not/existing"_json_pointer, Json({{"foo", "bar"}})) == Json({{"foo", "bar"}}));
                     CHECK(j.value("/not/existing"_json_pointer, Json({10, 100})) == Json({10, 100}));
 
+                    // an array index that is out of range, too large to be
+                    // represented, or "-", and a token below a scalar
+                    CHECK(j.value("/array/3"_json_pointer, 2) == 2);
+                    CHECK(j.value("/array/-"_json_pointer, 2) == 2);
+                    CHECK(j.value("/array/99999999999999999999999999"_json_pointer, 2) == 2);
+                    CHECK(j.value("/integer/0"_json_pointer, 2) == 2);
+                    CHECK(j.value("/string/x"_json_pointer, 2) == 2);
+                    CHECK(j.value("/null/x"_json_pointer, 2) == 2);
+                    CHECK(j.value("/array/0"_json_pointer, 2) == 1);
+
                     CHECK(j_const.value("/not/existing"_json_pointer, 2) == 2);
                     CHECK(j_const.value("/not/existing"_json_pointer, 2u) == 2u);
                     CHECK(j_const.value("/not/existing"_json_pointer, false) == false);

tests/src/unit-explicit_instantiation.cpp

@@ -0,0 +1,37 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+// cmake/test.cmake selects the C++ standard versions with which to build a
+// unit test based on the presence of JSON_HAS_CPP_<VERSION> macros.
+// The regression below only showed on C++17, so build this file for every
+// standard like the other regression tests:
+// JSON_HAS_CPP_17 JSON_HAS_CPP_20 (do not remove; see note at top of file)
+
+#include "doctest_compatibility.h"
+
+#include <nlohmann/json.hpp>
+using json = nlohmann::json;
+
+/////////////////////////////////////////////////////////////////////
+// for #4825 - explicitly instantiating basic_json must compile; this
+// forces instantiation of binary_writer::write_bjdata_ndarray, whose
+// static_cast<string_t> was ambiguous under explicit instantiation on
+// C++17. Merely compiling this translation unit is the regression test.
+//
+// The instantiation compiles every member function, so it has a file of its
+// own: in unit-regression3.cpp it made the object too large for the MinGW
+// linker to relocate (see #5511).
+/////////////////////////////////////////////////////////////////////
+template class nlohmann::basic_json<>;
+
+TEST_CASE("explicit instantiation of basic_json (#4825)")
+{
+    const json j = {1, "two", 3.0};
+    CHECK(j.size() == 3);
+    CHECK(json::from_bjdata(json::to_bjdata(j)) == j);
+}

tests/src/unit-hash.cpp

@@ -13,6 +13,78 @@ using json = nlohmann::json;
 using ordered_json = nlohmann::ordered_json;
 
 #include <set>
+#include <string>
+
+namespace
+{
+// how detail::hash defines the hash of an array or object: the seeds of the
+// elements, combined in order. Recursive, so only usable on values nested a
+// few hundred levels deep - which is exactly what is needed to check that the
+// iterative path taken below detail::recursion_depth_limit() computes the same.
+template<typename BasicJsonType>
+std::size_t reference_hash(const BasicJsonType& j)
+{
+    using nlohmann::detail::combine;
+    using string_t = typename BasicJsonType::string_t;
+
+    if (!j.is_structured())
+    {
+        return std::hash<BasicJsonType> {}(j);
+    }
+
+    auto seed = combine(static_cast<std::size_t>(j.type()), j.size());
+    for (const auto& element : j.items())
+    {
+        if (j.is_object())
+        {
+            seed = combine(seed, std::hash<string_t> {}(element.key()));
+        }
+        seed = combine(seed, reference_hash(element.value()));
+    }
+    return seed;
+}
+
+// a value nested `depth` levels deep, with siblings on every level
+template<typename BasicJsonType>
+BasicJsonType nested(const std::size_t depth, const bool objects)
+{
+    BasicJsonType value = "leaf";
+    for (std::size_t i = 0; i < depth; ++i)
+    {
+        if (objects)
+        {
+            value = BasicJsonType{{"before", i}, {"nested", std::move(value)}, {"after", {i, "x"}}};
+        }
+        else
+        {
+            value = BasicJsonType::array({i, std::move(value), BasicJsonType::object({{"k", i}})});
+        }
+    }
+    return value;
+}
+
+std::string nested_text(const std::size_t depth, const bool objects)
+{
+    std::string text;
+    if (objects)
+    {
+        text.reserve((6 * depth) + 1);
+        for (std::size_t i = 0; i < depth; ++i)
+        {
+            text += "{\"a\":";
+        }
+        text += "1";
+        text.append(depth, '}');
+    }
+    else
+    {
+        text.assign(depth, '[');
+        text += "1";
+        text.append(depth, ']');
+    }
+    return text;
+}
+} // namespace
 
 TEST_CASE("hash<nlohmann::json>")
 {
@@ -111,3 +183,44 @@ TEST_CASE("hash<nlohmann::ordered_json>")
 
     CHECK(hashes.size() == 21);
 }
+
+TEST_CASE("hash of deeply nested values")
+{
+    SECTION("hashing past the descent bound computes the same values")
+    {
+        // every depth on either side of where the iterative path takes over
+        for (std::size_t depth = 0; depth <= (2 * nlohmann::detail::recursion_depth_limit()) + 10; ++depth)
+        {
+            CAPTURE(depth);
+            const auto arrays = nested<json>(depth, false);
+            const auto objects = nested<json>(depth, true);
+            const auto ordered = nested<ordered_json>(depth, true);
+            CHECK(std::hash<json> {}(arrays) == reference_hash(arrays));
+            CHECK(std::hash<json> {}(objects) == reference_hash(objects));
+            CHECK(std::hash<ordered_json> {}(ordered) == reference_hash(ordered));
+        }
+    }
+
+    SECTION("values nested too deeply for the call stack (#5545)")
+    {
+        // recursing once per level used to exhaust the call stack here; the
+        // values are only parsed and hashed, never copied or compared, since
+        // those recurse as well
+        const std::size_t depth = 100000;
+        for (const bool objects :
+                {
+                    false, true
+                })
+        {
+            CAPTURE(objects);
+            const auto text = nested_text(depth, objects);
+            const auto a = json::parse(text);
+            const auto b = json::parse(text);
+            CHECK(std::hash<json> {}(a) == std::hash<json> {}(b));
+
+            const auto c = ordered_json::parse(text);
+            const auto d = ordered_json::parse(text);
+            CHECK(std::hash<ordered_json> {}(c) == std::hash<ordered_json> {}(d));
+        }
+    }
+}

tests/src/unit-items-cpp17.cpp

@@ -0,0 +1,42 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+// This file contains the C++17-only part of unit-items.cpp (structured
+// bindings support for json::items()). It is kept in a separate
+// translation unit so the (much larger) unit-items.cpp does not need to
+// be compiled a second time just for this one SECTION.
+
+#include "doctest_compatibility.h"
+
+#include <nlohmann/json.hpp>
+using nlohmann::json;
+
+#ifdef JSON_HAS_CPP_17
+#include <map>
+#include <string>
+
+TEST_CASE("items()")
+{
+    SECTION("object")
+    {
+        SECTION("structured bindings")
+        {
+            json j = { {"A", 1}, {"B", 2} };
+
+            std::map<std::string, int> m;
+
+            for (auto const&[key, value] : j.items())
+            {
+                m.emplace(key, value);
+            }
+
+            CHECK(j.get<decltype(m)>() == m);
+        }
+    }
+}
+#endif

tests/src/unit-items.cpp

@@ -862,22 +862,6 @@ TEST_CASE("items()")
 
             CHECK(counter == 3);
         }
-
-#ifdef JSON_HAS_CPP_17
-        SECTION("structured bindings")
-        {
-            json j = { {"A", 1}, {"B", 2} };
-
-            std::map<std::string, int> m;
-
-            for (auto const&[key, value] : j.items())
-            {
-                m.emplace(key, value);
-            }
-
-            CHECK(j.get<decltype(m)>() == m);
-        }
-#endif
     }
 
     SECTION("const object")

tests/src/unit-json_patch.cpp

@@ -672,6 +672,102 @@ TEST_CASE("JSON patch")
         }
     }
 
+    SECTION("patch_inplace")
+    {
+        SECTION("happy path: patch_inplace mirrors patch() on success")
+        {
+            // mirrors "A.5. Replacing a Value" above, but applies the patch with
+            // patch_inplace() to a mutable copy instead of using patch()'s
+            // returned copy
+            json doc = R"(
+                    {
+                        "baz": "qux",
+                        "foo": "bar"
+                    }
+                )"_json;
+
+            json const patch = R"(
+                    [
+                        { "op": "replace", "path": "/baz", "value": "boo" }
+                    ]
+                )"_json;
+
+            json const expected = R"(
+                    {
+                        "baz": "boo",
+                        "foo": "bar"
+                    }
+                )"_json;
+
+            doc.patch_inplace(patch);
+            CHECK(doc == expected);
+        }
+
+        // this test relies on the "test" operation actually throwing so the
+        // partial-application state can be observed right after the throw
+        // point; under JSON_NOEXCEPTION, JSON_THROW() calls std::abort()
+        // instead (there is no C++ exception to throw), and doctest's
+        // CHECK_THROWS_AS() is compiled out to a no-op that never even
+        // invokes the given expression (see doctest's "--no-throw" test
+        // filter, which ci_test_noexceptions passes) -- so patch()/
+        // patch_inplace() would never be called at all and the follow-up
+        // state assertions below would fail against the untouched original
+#if !defined(JSON_NOEXCEPTION)
+        SECTION("distinguishing contract vs patch(): partial application on failure")
+        {
+            // Unlike patch(), which is all-or-nothing because it applies the
+            // patch to an internal copy that is simply discarded when an
+            // exception is thrown (leaving the original untouched no matter
+            // what), patch_inplace() mutates the document it is called on
+            // directly and immediately, operation by operation. So if a JSON
+            // Patch fails partway through, whatever operations already
+            // succeeded remain applied -- the document is left in a partially
+            // patched state. This is empirically verified current behavior,
+            // not just documented intent, and is pinned here as such.
+            json const original = R"(
+                    {
+                        "baz": "qux",
+                        "foo": "bar"
+                    }
+                )"_json;
+
+            // the first operation ("replace") succeeds; the second ("test")
+            // fails because the value at "/baz" no longer (and never did)
+            // equal "not boo"
+            json const patch = R"(
+                    [
+                        { "op": "replace", "path": "/baz", "value": "boo" },
+                        { "op": "test", "path": "/baz", "value": "not boo" }
+                    ]
+                )"_json;
+
+            // patch() never modifies the object it is called on -- it always
+            // operates on (and returns) a separate copy, so the original is
+            // left completely untouched, regardless of success or failure.
+            // copy_for_patch is intentionally a real copy, not a reference
+            // to `original`: the whole point of this check is to catch a
+            // hypothetical future regression where patch() *does* mutate its
+            // receiver. Using a reference here would make the assertion
+            // below compare `original` to itself -- trivially true even if
+            // such a bug existed -- which is exactly what a static analyzer
+            // can't see when it suggests "this copy is never modified, use
+            // a reference instead".
+            json copy_for_patch = original; // NOLINT(performance-unnecessary-copy-initialization)
+            CHECK_THROWS_AS(copy_for_patch.patch(patch), json::other_error&);
+            CHECK(copy_for_patch == original);
+
+            // patch_inplace(), in contrast, already applied the successful
+            // "replace" operation to the document before the "test" operation
+            // threw -- that change is not rolled back
+            json doc = original;
+            CHECK_THROWS_AS(doc.patch_inplace(patch), json::other_error&);
+            CHECK(doc != original);
+            CHECK(doc.at("baz") == "boo");
+            CHECK(doc.at("foo") == "bar");
+        }
+#endif // !defined(JSON_NOEXCEPTION)
+    }
+
     SECTION("errors")
     {
         SECTION("unknown operation")
@@ -1388,3 +1484,365 @@ TEST_CASE("JSON patch - add to a primitive parent (regression #4292)")
         CHECK_THROWS_AS(doc.patch(patch), json::out_of_range&);
     }
 }
+
+TEST_CASE("JSON patch - remove with primitive or null parent (regression #5396)")
+{
+    // Regression test for https://github.com/nlohmann/json/issues/5396
+    //
+    // RFC 6902 (§4.2) requires the target location of a "remove" operation
+    // to exist. When the target's parent resolves to a primitive value or
+    // null, the operation must fail. Previously operation_remove silently
+    // did nothing in this case (neither the "is_object" nor the "is_array"
+    // branch matched, and there was no final "else"), so the patch appeared
+    // to succeed without changing the document. It now throws
+    // out_of_range.413.
+
+    SECTION("parent is a primitive (number)")
+    {
+        json const doc = {{"a", 1}};
+        json const patch = {{{"op", "remove"}, {"path", "/a/b"}}};
+#if JSON_DIAGNOSTICS
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.413] (/a) cannot remove value: the JSON Patch 'remove' target's parent is of type number, but must be an object or array", json::out_of_range&);
+#else
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.413] cannot remove value: the JSON Patch 'remove' target's parent is of type number, but must be an object or array", json::out_of_range&);
+#endif
+    }
+
+    SECTION("parent is a primitive (string)")
+    {
+        json const doc = {{"foo", {{"bar", "a string"}}}};
+        json const patch = {{{"op", "remove"}, {"path", "/foo/bar/baz"}}};
+#if JSON_DIAGNOSTICS
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.413] (/foo/bar) cannot remove value: the JSON Patch 'remove' target's parent is of type string, but must be an object or array", json::out_of_range&);
+#else
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.413] cannot remove value: the JSON Patch 'remove' target's parent is of type string, but must be an object or array", json::out_of_range&);
+#endif
+    }
+
+    SECTION("top-level document is null")
+    {
+        json const doc = nullptr;
+        json const patch = {{{"op", "remove"}, {"path", "/a"}}};
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.413] cannot remove value: the JSON Patch 'remove' target's parent is of type null, but must be an object or array", json::out_of_range&);
+    }
+
+    SECTION("legitimate removes still work")
+    {
+        // object member
+        json const doc1 = {{"a", 1}, {"b", 2}};
+        json const patch1 = {{{"op", "remove"}, {"path", "/a"}}};
+        CHECK(doc1.patch(patch1) == json({{"b", 2}}));
+
+        // array element
+        json const doc2 = R"([1, 2, 3])"_json;
+        json const patch2 = {{{"op", "remove"}, {"path", "/1"}}};
+        CHECK(doc2.patch(patch2) == R"([1, 3])"_json);
+    }
+}
+
+TEST_CASE("JSON patch - move where 'from' is a proper prefix of 'path' (regression #5397)")
+{
+    // Regression test for https://github.com/nlohmann/json/issues/5397
+    //
+    // RFC 6902 (§4.4) forbids "from" from being a proper prefix of "path"
+    // for a "move" operation: "a location cannot be moved into one of its
+    // children." "move" is implemented as remove-then-add; for an object
+    // target this happened to throw anyway as a side effect of the "add"
+    // step re-resolving through the now-removed parent, but for an array
+    // target the removal shifted subsequent indices, so "path" silently
+    // re-resolved to a different element and the operation "succeeded"
+    // with a corrupted result. It now throws out_of_range.414 for both
+    // object and array targets.
+
+    SECTION("array target (from the issue)")
+    {
+        json const doc = R"([[1,2],[3]])"_json;
+        json const patch = {{{"op", "move"}, {"from", "/0"}, {"path", "/0/0"}}};
+#if JSON_DIAGNOSTIC_POSITIONS
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] (bytes 0-11) cannot move value: 'from' path '/0' is a proper prefix of 'path' '/0/0'", json::out_of_range&);
+#else
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] cannot move value: 'from' path '/0' is a proper prefix of 'path' '/0/0'", json::out_of_range&);
+#endif
+    }
+
+    SECTION("object target")
+    {
+        json const doc = R"({"a": {"b": 1}})"_json;
+        json const patch = {{{"op", "move"}, {"from", "/a"}, {"path", "/a/b"}}};
+#if JSON_DIAGNOSTIC_POSITIONS
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] (bytes 0-15) cannot move value: 'from' path '/a' is a proper prefix of 'path' '/a/b'", json::out_of_range&);
+#else
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] cannot move value: 'from' path '/a' is a proper prefix of 'path' '/a/b'", json::out_of_range&);
+#endif
+    }
+
+    SECTION("from == path is not a proper prefix and must not be rejected")
+    {
+        // "from" equal to "path" is a no-op move; it is not a *proper*
+        // prefix relationship, so this new check must not reject it.
+        json const doc = R"({"a": 1, "b": 2})"_json;
+        json const patch = {{{"op", "move"}, {"from", "/a"}, {"path", "/a"}}};
+        CHECK(doc.patch(patch) == doc);
+    }
+
+    SECTION("raw string prefix that is not a pointer-token prefix must be allowed")
+    {
+        // "/ab" is a string-prefix of "/abc/x" as raw text, but "ab" and
+        // "abc" are different reference tokens, so this is NOT a
+        // pointer-token prefix relationship and the move must succeed.
+        // This is the key case proving the check compares tokens, not
+        // raw pointer text (a naive std::string prefix/rfind check on
+        // the undecoded pointer would wrongly reject this).
+        json const doc = R"({"ab": 1, "abc": {"x": 2}})"_json;
+        json const patch = {{{"op", "move"}, {"from", "/ab"}, {"path", "/abc/x"}}};
+        json const result = R"({"abc": {"x": 1}})"_json;
+        CHECK(doc.patch(patch) == result);
+    }
+
+    SECTION("escaped reference tokens are compared unescaped")
+    {
+        // "from" is the single token "a/b" (escaped as "a~1b"); "path"
+        // addresses member "x" of that same value, so "from" is a
+        // proper (token-level) prefix of "path" and must be rejected.
+        json const doc = R"({"a/b": {"x": 1}})"_json;
+        json const patch = {{{"op", "move"}, {"from", "/a~1b"}, {"path", "/a~1b/x"}}};
+#if JSON_DIAGNOSTIC_POSITIONS
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] (bytes 0-17) cannot move value: 'from' path '/a~1b' is a proper prefix of 'path' '/a~1b/x'", json::out_of_range&);
+#else
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] cannot move value: 'from' path '/a~1b' is a proper prefix of 'path' '/a~1b/x'", json::out_of_range&);
+#endif
+    }
+
+    SECTION("ordinary valid moves still work")
+    {
+        // unrelated top-level members
+        json const doc1 = R"({"a": 1, "b": 2})"_json;
+        json const patch1 = {{{"op", "move"}, {"from", "/a"}, {"path", "/c"}}};
+        CHECK(doc1.patch(patch1) == R"({"b": 2, "c": 1})"_json);
+
+        // sibling paths that share a textual prefix but are unrelated
+        json const doc2 = R"({"a": {"x": 1}, "b": {"y": 2}})"_json;
+        json const patch2 = {{{"op", "move"}, {"from", "/a/x"}, {"path", "/b/z"}}};
+        CHECK(doc2.patch(patch2) == R"({"a": {}, "b": {"y": 2, "z": 1}})"_json);
+
+        // "path" is a proper prefix of "from" (the reverse relationship,
+        // which RFC 6902 does not forbid)
+        json const doc3 = R"({"a": {"b": 1}})"_json;
+        json const patch3 = {{{"op", "move"}, {"from", "/a/b"}, {"path", "/a"}}};
+        CHECK(doc3.patch(patch3) == R"({"a": 1})"_json);
+    }
+
+    SECTION("root 'from' is a proper prefix of every non-root 'path'")
+    {
+        // the whole document is a proper prefix of any location inside it
+        json const doc = R"({"a": 1})"_json;
+        json const patch = {{{"op", "move"}, {"from", ""}, {"path", "/a"}}};
+#if JSON_DIAGNOSTIC_POSITIONS
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] (bytes 0-8) cannot move value: 'from' path '' is a proper prefix of 'path' '/a'", json::out_of_range&);
+#else
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] cannot move value: 'from' path '' is a proper prefix of 'path' '/a'", json::out_of_range&);
+#endif
+    }
+
+    SECTION("root 'path' is never a proper prefix violation for a non-root 'from'")
+    {
+        // the reverse of the above: moving a non-root location to the root
+        // is the "path is a prefix of from" relationship, which RFC 6902
+        // permits (already covered generally above; this pins the root
+        // case specifically, since root is the one path with no reference
+        // tokens at all)
+        json const doc = R"({"a": {"b": 1}})"_json;
+        json const patch = {{{"op", "move"}, {"from", "/a"}, {"path", ""}}};
+        CHECK(doc.patch(patch) == R"({"b": 1})"_json);
+    }
+
+    SECTION("the array-append token '-' is an ordinary child token")
+    {
+        // "-" (append-to-array) addresses a location *inside* the array,
+        // so "from" pointing at the array is still a proper prefix of
+        // "path" ending in "-" and must be rejected like any other child.
+        json const doc = R"({"a": [1, 2]})"_json;
+        json const patch = {{{"op", "move"}, {"from", "/a"}, {"path", "/a/-"}}};
+#if JSON_DIAGNOSTIC_POSITIONS
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] (bytes 0-13) cannot move value: 'from' path '/a' is a proper prefix of 'path' '/a/-'", json::out_of_range&);
+#else
+        CHECK_THROWS_WITH_AS(doc.patch(patch), "[json.exception.out_of_range.414] cannot move value: 'from' path '/a' is a proper prefix of 'path' '/a/-'", json::out_of_range&);
+#endif
+    }
+}
+
+TEST_CASE("JSON patch - diff emits array removals in descending index order")
+{
+    SECTION("array shrunk to empty")
+    {
+        json const source = {0, 1, 2, 3, 4};
+        json const target = json::array();
+
+        json const patch = json::diff(source, target);
+
+        json const expected = R"(
+            [
+                {"op": "remove", "path": "/4"},
+                {"op": "remove", "path": "/3"},
+                {"op": "remove", "path": "/2"},
+                {"op": "remove", "path": "/1"},
+                {"op": "remove", "path": "/0"}
+            ]
+        )"_json;
+
+        CHECK(patch == expected);
+        CHECK(source.patch(patch) == target);
+    }
+
+    SECTION("array partially shrunk, after a replacement at a common index")
+    {
+        json const source = {0, 1, 2, 3, 4};
+        json const target = {0, 9};
+
+        json const patch = json::diff(source, target);
+
+        // the replacement comes first, then the removals, highest index first
+        json const expected = R"(
+            [
+                {"op": "replace", "path": "/1", "value": 9},
+                {"op": "remove", "path": "/4"},
+                {"op": "remove", "path": "/3"},
+                {"op": "remove", "path": "/2"}
+            ]
+        )"_json;
+
+        CHECK(patch == expected);
+        CHECK(source.patch(patch) == target);
+    }
+
+    SECTION("nested array shrunk")
+    {
+        json const source = {{"a", {0, 1, 2}}};
+        json const target = {{"a", json::array()}};
+
+        json const patch = json::diff(source, target);
+
+        json const expected = R"(
+            [
+                {"op": "remove", "path": "/a/2"},
+                {"op": "remove", "path": "/a/1"},
+                {"op": "remove", "path": "/a/0"}
+            ]
+        )"_json;
+
+        CHECK(patch == expected);
+        CHECK(source.patch(patch) == target);
+    }
+
+    SECTION("many removals still round-trip")
+    {
+        json source = json::array();
+        for (int i = 0; i < 1000; ++i)
+        {
+            source.push_back(i);
+        }
+        json const target = json::array();
+
+        json const patch = json::diff(source, target);
+
+        CHECK(patch.size() == 1000);
+        CHECK(patch.front().at("path") == "/999");
+        CHECK(patch.back().at("path") == "/0");
+        CHECK(source.patch(patch) == target);
+    }
+}
+
+TEST_CASE("JSON patch - every operation on ordered_json")
+{
+    using nlohmann::ordered_json;
+
+    const ordered_json doc = {{"foo", "bar"}, {"arr", {1, 2, 3}}, {"obj", {{"a", 1}}}};
+
+    SECTION("successful operations")
+    {
+        const ordered_json patch = ordered_json::parse(R"([
+            {"op": "add", "path": "/obj/b", "value": 2},
+            {"op": "add", "path": "/arr/1", "value": 9},
+            {"op": "add", "path": "/arr/-", "value": 4},
+            {"op": "remove", "path": "/arr/0"},
+            {"op": "remove", "path": "/obj/a"},
+            {"op": "replace", "path": "/foo", "value": "baz"},
+            {"op": "move", "from": "/foo", "path": "/moved"},
+            {"op": "copy", "from": "/obj", "path": "/copied"},
+            {"op": "test", "path": "/copied/b", "value": 2}
+        ])");
+
+        const ordered_json expected = ordered_json::parse(R"({
+            "arr": [9, 2, 3, 4], "obj": {"b": 2}, "moved": "baz", "copied": {"b": 2}
+        })");
+
+        CHECK(doc.patch(patch) == expected);
+
+        // adding to the root replaces the document
+        CHECK(doc.patch(ordered_json::parse(R"([{"op": "add", "path": "", "value": [1]}])")) == ordered_json({1}));
+    }
+
+    SECTION("failing operations")
+    {
+        ordered_json _;
+#if JSON_DIAGNOSTICS
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "add", "path": "/arr/4", "value": 1}])")),
+                             "[json.exception.out_of_range.401] (/arr) array index 4 is out of range", ordered_json::out_of_range&);
+#else
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "add", "path": "/arr/4", "value": 1}])")),
+                             "[json.exception.out_of_range.401] array index 4 is out of range", ordered_json::out_of_range&);
+#endif
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "add", "path": "/nope/x", "value": 1}])")),
+                             "[json.exception.out_of_range.403] key 'nope' not found", ordered_json::out_of_range&);
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "remove", "path": "/obj/nope"}])")),
+                             "[json.exception.out_of_range.403] key 'nope' not found", ordered_json::out_of_range&);
+#if JSON_DIAGNOSTICS
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "remove", "path": "/arr/3"}])")),
+                             "[json.exception.out_of_range.401] (/arr) array index 3 is out of range", ordered_json::out_of_range&);
+#else
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "remove", "path": "/arr/3"}])")),
+                             "[json.exception.out_of_range.401] array index 3 is out of range", ordered_json::out_of_range&);
+#endif
+#if JSON_DIAGNOSTICS
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "test", "path": "/foo", "value": "qux"}])")),
+                             "[json.exception.other_error.501] (/0) unsuccessful: {\"op\":\"test\",\"path\":\"/foo\",\"value\":\"qux\"}", ordered_json::other_error&);
+#elif JSON_DIAGNOSTIC_POSITIONS
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "test", "path": "/foo", "value": "qux"}])")),
+                             "[json.exception.other_error.501] (bytes 1-47) unsuccessful: {\"op\":\"test\",\"path\":\"/foo\",\"value\":\"qux\"}", ordered_json::other_error&);
+#else
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "test", "path": "/foo", "value": "qux"}])")),
+                             "[json.exception.other_error.501] unsuccessful: {\"op\":\"test\",\"path\":\"/foo\",\"value\":\"qux\"}", ordered_json::other_error&);
+#endif
+#if JSON_DIAGNOSTICS
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "add", "path": "/foo"}])")),
+                             "[json.exception.parse_error.105] parse error: (/0) operation 'add' must have member 'value'", ordered_json::parse_error&);
+#elif JSON_DIAGNOSTIC_POSITIONS
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "add", "path": "/foo"}])")),
+                             "[json.exception.parse_error.105] parse error: (bytes 1-30) operation 'add' must have member 'value'", ordered_json::parse_error&);
+#else
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "add", "path": "/foo"}])")),
+                             "[json.exception.parse_error.105] parse error: operation 'add' must have member 'value'", ordered_json::parse_error&);
+#endif
+        CHECK_THROWS_WITH_AS(_ = doc.patch(ordered_json::parse(R"([{"op": "move", "from": "/obj", "path": "/obj/a/b"}])")),
+                             "[json.exception.out_of_range.414] cannot move value: 'from' path '/obj' is a proper prefix of 'path' '/obj/a/b'", ordered_json::out_of_range&);
+    }
+
+    SECTION("diff reproduces the target")
+    {
+        const ordered_json source = {{"a", 1}, {"b", 2}, {"c", {{"x", 1}}}, {"l", {1, 2, 3}}};
+        const std::vector<ordered_json> targets =
+        {
+            // a key removed, a key added, a nested change, a shorter array
+            {{"a", 1}, {"c", {{"x", 2}}}, {"l", {1}}, {"d", 4}},
+            // the same keys in another order
+            {{"c", {{"x", 1}}}, {"a", 1}, {"b", 2}, {"l", {1, 2, 3}}},
+            // new keys ahead of the common ones
+            {{"new", true}, {"a", 1}, {"b", 3}, {"c", {{"x", 1}}}, {"l", {1, 2, 3}}},
+        };
+        for (const auto& target : targets)
+        {
+            CAPTURE(target.dump());
+            CHECK(source.patch(ordered_json::diff(source, target)) == target);
+        }
+    }
+}

tests/src/unit-json_pointer.cpp

@@ -319,6 +319,44 @@ TEST_CASE("JSON pointers")
 
                 CHECK_THROWS_WITH_AS(j[jp] = 1, throw_msg.c_str(), json::out_of_range&);
                 CHECK_THROWS_WITH_AS(j_const[jp] == 1, throw_msg.c_str(), json::out_of_range&);
+
+                // #5395: contains() must not throw for a reference token that is a
+                // syntactically valid array index but numerically exceeds ULLONG_MAX
+                // (causing strtoull() to set errno to ERANGE) -- it should just report
+                // that the pointer does not resolve to an element
+                CHECK(!j.contains(jp));
+                CHECK(!j_const.contains(jp));
+            }
+
+            {
+                // #5395: same as above, but using the exact reproduction from the issue
+                json::json_pointer const jp("/99999999999999999999");
+                std::string const throw_msg = "[json.exception.out_of_range.404] unresolved reference token '99999999999999999999'";
+
+                CHECK_THROWS_WITH_AS(j[jp] = 1, throw_msg.c_str(), json::out_of_range&);
+                CHECK_THROWS_WITH_AS(j_const[jp] == 1, throw_msg.c_str(), json::out_of_range&);
+                CHECK_THROWS_WITH_AS(j.at(jp) = 1, throw_msg.c_str(), json::out_of_range&);
+                CHECK_THROWS_WITH_AS(j_const.at(jp) == 1, throw_msg.c_str(), json::out_of_range&);
+
+                CHECK(!j.contains(jp));
+                CHECK(!j_const.contains(jp));
+            }
+
+            {
+                // #5395: a reference token that is numerically representable in
+                // unsigned long long but exceeds size_type's max (e.g. ULLONG_MAX
+                // itself on typical 64-bit platforms, where size_type's max equals
+                // ULLONG_MAX) must not make contains() throw either
+                json::json_pointer const jp("/18446744073709551615");
+                std::string const throw_msg = "[json.exception.out_of_range.410] array index 18446744073709551615 exceeds size_type";
+
+                CHECK_THROWS_WITH_AS(j[jp] = 1, throw_msg.c_str(), json::out_of_range&);
+                CHECK_THROWS_WITH_AS(j_const[jp] == 1, throw_msg.c_str(), json::out_of_range&);
+                CHECK_THROWS_WITH_AS(j.at(jp) = 1, throw_msg.c_str(), json::out_of_range&);
+                CHECK_THROWS_WITH_AS(j_const.at(jp) == 1, throw_msg.c_str(), json::out_of_range&);
+
+                CHECK(!j.contains(jp));
+                CHECK(!j_const.contains(jp));
             }
 
             // on some machines, the check below is not constant
@@ -334,6 +372,10 @@ TEST_CASE("JSON pointers")
 
                 CHECK_THROWS_WITH_AS(j[jp] = 1, throw_msg.c_str(), json::out_of_range&);
                 CHECK_THROWS_WITH_AS(j_const[jp] == 1, throw_msg.c_str(), json::out_of_range&);
+
+                // #5395: contains() must not throw for a reference token exceeding size_type's max
+                CHECK(!j.contains(jp));
+                CHECK(!j_const.contains(jp));
             }
 
             DOCTEST_MSVC_SUPPRESS_WARNING_POP
@@ -531,6 +573,16 @@ TEST_CASE("JSON pointers")
         // explicit roundtrip check
         CHECK(j.flatten().unflatten() == j);
 
+        // an object is only unflattened to an array if one of its keys is the
+        // reference token 0; this must not depend on which key is seen first
+        CHECK(json({{"/2", "x"}}).unflatten() == json({{"2", "x"}}));
+        CHECK(json({{"/10", "y"}, {"/2", "z"}}).unflatten() == json({{"10", "y"}, {"2", "z"}}));
+        CHECK(json({{"/0", 1}, {"/1", 2}}).unflatten() == json({1, 2}));
+        CHECK(json({{"/1", 2}, {"/0", 1}}).unflatten() == json({1, 2}));
+        CHECK(json({{"/0", 1}, {"/2", 3}}).unflatten() == json({1, nullptr, 3}));
+        CHECK(json({{"/a/1", 2}, {"/a/0", 1}}).unflatten() == json({{"a", {1, 2}}}));
+        CHECK(json({{"/a/1", 2}, {"/a/x", 1}}).unflatten() == json({{"a", {{"1", 2}, {"x", 1}}}}));
+
         // roundtrip for primitive values
         json j_null;
         CHECK(j_null.flatten().unflatten() == j_null);
@@ -886,3 +938,16 @@ TEST_CASE("JSON pointers")
     }
 #endif
 }
+
+TEST_CASE("unescaping keeps a '~' that does not start an escape sequence")
+{
+    // the parser of a JSON pointer rejects such reference tokens before it
+    // unescapes them, so this is only reachable by calling unescape directly
+    std::string s = "a~2b~";
+    nlohmann::detail::unescape(s);
+    CHECK(s == "a~2b~");
+
+    s = "~0~1~";
+    nlohmann::detail::unescape(s);
+    CHECK(s == "~/~");
+}

tests/src/unit-large_json.cpp

@@ -12,12 +12,13 @@
 using nlohmann::json;
 
 #include <algorithm>
+#include <string>
 
 TEST_CASE("tests on very large JSONs")
 {
     SECTION("issue #1419 - Segmentation fault (stack overflow) due to unbounded recursion")
     {
-        const auto depth = 5000000;
+        const auto depth = 500000;
 
         std::string s(static_cast<std::size_t>(2 * depth), '[');
         std::fill(s.begin() + depth, s.end(), ']');
@@ -27,3 +28,201 @@ TEST_CASE("tests on very large JSONs")
     }
 }
 
+namespace
+{
+
+// Descend a chain of single-element containers and return the value at its end,
+// reporting the number of levels traversed in @a depth.
+//
+// The values in the test case below are nested far deeper than the call stack
+// can follow, so they must not be inspected with operator== or dump(): both are
+// still recursive and would overflow the stack themselves.
+const json* innermost_value(const json& j, std::size_t& depth)
+{
+    const json* current = &j;
+    depth = 0;
+
+    while ((current->is_array() || current->is_object()) && !current->empty())
+    {
+        current = current->is_array()
+                  ? &current->front()
+                  : &current->begin().value();
+        ++depth;
+    }
+
+    return current;
+}
+
+} // namespace
+
+TEST_CASE("tests on deeply nested JSONs")
+{
+    // deep enough to exhaust the call stack, but small enough to stay cheap:
+    // parsing is iterative, so building the values below costs little
+    const std::size_t depth = 100000;
+
+    SECTION("issue #5387 - stack overflow in the copy constructor")
+    {
+        SECTION("array")
+        {
+            const json j = json::parse(std::string(depth, '[') + '0' + std::string(depth, ']'));
+
+            const json copy(j); // NOLINT(performance-unnecessary-copy-initialization): the copy is what is tested
+
+            std::size_t copy_depth = 0;
+            CHECK(*innermost_value(copy, copy_depth) == 0);
+            CHECK(copy_depth == depth);
+        }
+
+        SECTION("object")
+        {
+            std::string s;
+            s.reserve((6 * depth) + 1);
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                s += "{\"a\":";
+            }
+            s += '1';
+            s.append(depth, '}');
+
+            const json j = json::parse(s);
+
+            const json copy(j); // NOLINT(performance-unnecessary-copy-initialization): the copy is what is tested
+
+            std::size_t copy_depth = 0;
+            CHECK(*innermost_value(copy, copy_depth) == 1);
+            CHECK(copy_depth == depth);
+        }
+
+        SECTION("copy assignment")
+        {
+            // operator=(basic_json) takes its argument by value, so the deep
+            // copy happens in the copy constructor
+            const json j = json::parse(std::string(depth, '[') + '0' + std::string(depth, ']'));
+
+            json target;
+            target = j;
+
+            std::size_t target_depth = 0;
+            CHECK(*innermost_value(target, target_depth) == 0);
+            CHECK(target_depth == depth);
+        }
+
+        SECTION("depths around the bound of the recursive descent")
+        {
+            // The copy constructor descends into a bounded number of levels and
+            // completes whatever is below that without the call stack. Cover
+            // every depth around that bound, so that the two ways of copying
+            // are known to meet cleanly - wherever the bound is set.
+            for (std::size_t d = 1; d <= 300; ++d)
+            {
+                CAPTURE(d);
+
+                const json array = json::parse(std::string(d, '[') + '0' + std::string(d, ']'));
+                const json array_copy(array); // NOLINT(performance-unnecessary-copy-initialization): the copy is what is tested
+                std::size_t array_depth = 0;
+                CHECK(*innermost_value(array_copy, array_depth) == 0);
+                CHECK(array_depth == d);
+
+                std::string object_text;
+                for (std::size_t i = 0; i < d; ++i)
+                {
+                    object_text += "{\"a\":";
+                }
+                object_text += '1';
+                object_text.append(d, '}');
+
+                const json object = json::parse(object_text);
+                const json object_copy(object); // NOLINT(performance-unnecessary-copy-initialization): the copy is what is tested
+                std::size_t object_depth = 0;
+                CHECK(*innermost_value(object_copy, object_depth) == 1);
+                CHECK(object_depth == d);
+            }
+        }
+
+        SECTION("a value that is deep in one place only")
+        {
+            json j = json::object();
+            j["shallow"] = 1;
+            j["deep"] = json::parse(std::string(depth, '[') + '0' + std::string(depth, ']'));
+            j["also_shallow"] = json::array({1, 2, 3});
+
+            const json copy(j);
+
+            CHECK(copy["shallow"] == 1);
+            CHECK(copy["also_shallow"] == json::array({1, 2, 3}));
+
+            std::size_t deep_depth = 0;
+            CHECK(*innermost_value(copy["deep"], deep_depth) == 0);
+            CHECK(deep_depth == depth);
+        }
+
+        SECTION("comparing")
+        {
+            // Comparing used to descend once per level, and an ordered
+            // comparison used to compare every pair of elements twice, once in
+            // each direction, which took exponentially long in the nesting
+            // depth. Both are gone: these finish in milliseconds, where the
+            // second used to take longer than anyone would wait even for a
+            // value nested only a few dozen levels deep.
+            const std::string text = std::string(depth, '[') + '0' + std::string(depth, ']');
+            const json j = json::parse(text);
+            const json same = json::parse(text);
+            const json larger = json::parse(std::string(depth, '[') + '1' + std::string(depth, ']'));
+
+            CHECK(j == same);
+            CHECK_FALSE(j == larger);
+            CHECK(j != larger);
+
+            CHECK(j < larger);
+            CHECK_FALSE(larger < j);
+            CHECK(larger > j);
+            CHECK(j <= same);
+            CHECK(j >= same);
+
+            // a value that ends earlier is the smaller one
+            const json shorter = json::parse(std::string(depth - 1, '[') + '0' + std::string(depth - 1, ']'));
+            CHECK_FALSE(j == shorter);
+        }
+
+        SECTION("comparing objects")
+        {
+            std::string text;
+            text.reserve((6 * depth) + 1);
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                text += "{\"a\":";
+            }
+            text += '1';
+            text.append(depth, '}');
+
+            const json j = json::parse(text);
+            const json same = json::parse(text);
+
+            CHECK(j == same);
+            CHECK_FALSE(j != same);
+            CHECK(j <= same);
+            CHECK(j >= same);
+        }
+
+        SECTION("the copy is independent of the original")
+        {
+            const json j = json::parse(std::string(depth, '[') + '0' + std::string(depth, ']'));
+
+            json copy(j);
+
+            // reach the innermost value without recursing and replace it
+            json* current = &copy;
+            while (current->is_array() && !current->empty())
+            {
+                current = &current->front();
+            }
+            *current = 42;
+
+            std::size_t unused = 0;
+            CHECK(*innermost_value(copy, unused) == 42);
+            CHECK(*innermost_value(j, unused) == 0);
+        }
+    }
+}
+

tests/src/unit-locale-cpp.cpp

@@ -158,6 +158,17 @@ TEST_CASE("locale-dependent test (LC_NUMERIC=de_DE)")
             json::sax_parse("12.34", &sax);
             CHECK(sax.float_string_copy == "12.34");
         }
+
+        SECTION("serializing a long double")
+        {
+            // a floating-point type that is not a float or a double is written
+            // with snprintf, whose locale-specific decimal point and thousands
+            // separator are undone afterwards
+            using long_double_json = nlohmann::basic_json<std::map, std::vector, std::string, bool, std::int64_t, std::uint64_t, long double>;
+            CHECK(long_double_json(12345.5L).dump() == "12345.5");
+            CHECK(long_double_json(1.0L).dump() == "1.0");
+            CHECK(long_double_json(-0.25L).dump() == "-0.25");
+        }
     }
     else
     {

tests/src/unit-merge_patch.cpp

@@ -14,6 +14,60 @@ using nlohmann::json;
     using namespace nlohmann::literals; // NOLINT(google-build-using-namespace)
 #endif
 
+#include <string>
+
+namespace
+{
+// RFC 7396's MergePatch, written recursively as in the RFC; only usable on
+// values nested a few hundred levels deep
+void reference_merge_patch(json& target, const json& patch)
+{
+    if (!patch.is_object())
+    {
+        target = patch;
+        return;
+    }
+    if (!target.is_object())
+    {
+        target = json::object();
+    }
+    for (auto it = patch.begin(); it != patch.end(); ++it)
+    {
+        if (it.value().is_null())
+        {
+            target.erase(it.key());
+        }
+        else
+        {
+            reference_merge_patch(target[it.key()], it.value());
+        }
+    }
+}
+
+// objects nested `depth` levels deep under the key "a", with members that
+// differ by `variant` on the way down
+std::string nested_objects(const std::size_t depth, const int variant)
+{
+    std::string text;
+    for (std::size_t i = 0; i < depth; ++i)
+    {
+        text += "{";
+        if ((i + static_cast<std::size_t>(variant)) % 3 == 0)
+        {
+            text += "\"s" + std::to_string(variant) + "\":" + std::to_string(i) + ",";
+        }
+        if (variant == 2 && i % 5 == 0)
+        {
+            text += "\"s0\":null,";
+        }
+        text += "\"a\":";
+    }
+    text += variant == 1 ? R"({"x":1,"y":null})" : "{\"y\":2}";
+    text.append(depth, '}');
+    return text;
+}
+} // namespace
+
 TEST_CASE("JSON Merge Patch")
 {
     SECTION("examples from RFC 7396")
@@ -242,3 +296,81 @@ TEST_CASE("JSON Merge Patch")
         }
     }
 }
+
+TEST_CASE("JSON Merge Patch on deeply nested values")
+{
+    SECTION("patching past the descent bound gives the same result")
+    {
+        // every depth on either side of where the iterative version takes
+        // over (detail::recursion_depth_limit(), 128)
+        for (std::size_t depth = 0; depth <= 300; ++depth)
+        {
+            CAPTURE(depth);
+            for (int variant = 0; variant < 3; ++variant)
+            {
+                CAPTURE(variant);
+                const json patch = json::parse(nested_objects(depth, variant));
+
+                json result = json::parse(nested_objects(depth, (variant + 1) % 3));
+                json expected = result;
+                result.merge_patch(patch);
+                reference_merge_patch(expected, patch);
+                CHECK(result == expected);
+
+                // a target that is not an object, and an empty one
+                json from_null;
+                from_null.merge_patch(patch);
+                json expected_from_null;
+                reference_merge_patch(expected_from_null, patch);
+                CHECK(from_null == expected_from_null);
+            }
+        }
+    }
+
+    SECTION("patches nested too deeply for the call stack (#5393)")
+    {
+        // applying a patch used to recurse once per nesting level. The result
+        // is only walked, never copied or compared, since those recurse too.
+        const std::size_t depth = 100000;
+        json target = json::parse(nested_objects(depth, 0));
+        target.merge_patch(json::parse(nested_objects(depth, 1)));
+
+        const json* p = &target;
+        for (std::size_t i = 0; i < depth; ++i)
+        {
+            p = &p->at("a");
+        }
+        // {"y":2} patched with {"x":1,"y":null}
+        CHECK(p->size() == 1);
+        CHECK(p->at("x") == 1);
+    }
+}
+
+TEST_CASE("JSON Merge Patch and update on ordered_json")
+{
+    using nlohmann::ordered_json;
+
+    SECTION("merge_patch")
+    {
+        ordered_json target = ordered_json::parse(R"({"a": {"b": 1, "c": 2}, "d": 3, "e": [1]})");
+        target.merge_patch(ordered_json::parse(R"({"a": {"b": null, "f": 4}, "d": {"x": {"y": null}}, "e": null, "g": {"h": 5}})"));
+        CHECK(target == ordered_json::parse(R"({"a": {"c": 2, "f": 4}, "d": {"x": {}}, "g": {"h": 5}})"));
+
+        // a patch that is not an object replaces the target
+        target.merge_patch(ordered_json({1, 2}));
+        CHECK(target == ordered_json({1, 2}));
+        // an object patch turns a target that is not an object into one
+        target.merge_patch(ordered_json::parse(R"({"k": {"l": null}})"));
+        CHECK(target == ordered_json::parse(R"({"k": {}})"));
+    }
+
+    SECTION("update with merge_objects")
+    {
+        ordered_json target = ordered_json::parse(R"({"a": {"b": 1, "c": {"d": 2}}, "e": 3})");
+        target.update(ordered_json::parse(R"({"a": {"c": {"x": 1}, "f": 4}, "e": {"y": 5}, "g": 6})"), true);
+        CHECK(target == ordered_json::parse(R"({"a": {"b": 1, "c": {"d": 2, "x": 1}, "f": 4}, "e": {"y": 5}, "g": 6})"));
+
+        target.update(ordered_json::parse(R"({"a": 1})"), false);
+        CHECK(target == ordered_json::parse(R"({"a": 1, "e": {"y": 5}, "g": 6})"));
+    }
+}

tests/src/unit-modifiers.cpp

@@ -11,6 +11,53 @@
 #include <nlohmann/json.hpp>
 using nlohmann::json;
 
+#include <string>
+
+namespace
+{
+// update(source, true) as documented, written recursively; only usable on
+// values nested a few hundred levels deep
+void reference_update(json& target, const json& source)
+{
+    for (auto it = source.begin(); it != source.end(); ++it)
+    {
+        const auto existing = target.find(it.key());
+        if (it.value().is_object() && existing != target.end() && existing->is_object())
+        {
+            reference_update(*existing, it.value());
+        }
+        else
+        {
+            target[it.key()] = it.value();
+        }
+    }
+}
+
+// objects nested `depth` levels deep under the key "a", with members that
+// differ by `variant` on the way down
+std::string nested_objects(const std::size_t depth, const int variant)
+{
+    std::string text;
+    for (std::size_t i = 0; i < depth; ++i)
+    {
+        text += "{";
+        if ((i + static_cast<std::size_t>(variant)) % 3 == 0)
+        {
+            text += "\"s" + std::to_string(variant) + "\":" + std::to_string(i) + ",";
+        }
+        if (variant == 2 && i % 5 == 0)
+        {
+            // an object replacing a primitive, which is not merged
+            text += R"("s0":{"o":1},)";
+        }
+        text += "\"a\":";
+    }
+    text += variant == 1 ? "{\"x\":1}" : "{\"y\":2}";
+    text.append(depth, '}');
+    return text;
+}
+} // namespace
+
 TEST_CASE("modifiers")
 {
     SECTION("clear()")
@@ -641,6 +688,20 @@ TEST_CASE("modifiers")
                 CHECK_THROWS_WITH_AS(j_array.insert(j_array.end(), j_other_array.begin(), j_other_array2.end()), "[json.exception.invalid_iterator.210] iterators do not fit",
                                      json::invalid_iterator&);
             }
+
+            SECTION("iterators not pointing into an array")
+            {
+                json j_object2 = {{"k", 1}, {"l", 2}};
+                json j_primitive = 5;
+                json j_null;
+
+                CHECK_THROWS_WITH_AS(j_array.insert(j_array.begin(), j_object2.begin(), j_object2.end()), "[json.exception.invalid_iterator.202] iterators first and last must point to arrays",
+                                     json::invalid_iterator&);
+                CHECK_THROWS_WITH_AS(j_array.insert(j_array.begin(), j_primitive.begin(), j_primitive.end()), "[json.exception.invalid_iterator.202] iterators first and last must point to arrays",
+                                     json::invalid_iterator&);
+                CHECK_THROWS_WITH_AS(j_array.insert(j_array.begin(), j_null.begin(), j_null.end()), "[json.exception.invalid_iterator.202] iterators first and last must point to arrays",
+                                     json::invalid_iterator&);
+            }
         }
 
         SECTION("range for object")
@@ -801,6 +862,30 @@ TEST_CASE("modifiers")
                     j1.update(j2, true);
                     CHECK(j1 == json({{"string", "t"}, {"numbers", 1}}));
                 }
+
+                SECTION("overwrite primitive with object")
+                {
+                    json j1 = {{"k", 1}};
+                    json const j2 = {{"k", {{"x", 2}}}};
+                    j1.update(j2, true);
+                    CHECK(j1 == json({{"k", {{"x", 2}}}}));
+                }
+
+                SECTION("overwrite array with object")
+                {
+                    json j1 = {{"k", {1, 2}}};
+                    json const j2 = {{"k", {{"x", 2}}}};
+                    j1.update(j2, true);
+                    CHECK(j1 == json({{"k", {{"x", 2}}}}));
+                }
+
+                SECTION("overwrite nested primitive with object")
+                {
+                    json j1 = {{"k", {{"inner", 1}}}};
+                    json const j2 = {{"k", {{"inner", {{"x", 2}}}}}};
+                    j1.update(j2, true);
+                    CHECK(j1 == json({{"k", {{"inner", {{"x", 2}}}}}}));
+                }
             }
         }
     }
@@ -950,3 +1035,44 @@ TEST_CASE("modifiers")
         }
     }
 }
+
+TEST_CASE("update() on deeply nested values")
+{
+    SECTION("merging past the descent bound gives the same result")
+    {
+        // every depth on either side of where the iterative version takes
+        // over (detail::recursion_depth_limit(), 128)
+        for (std::size_t depth = 0; depth <= 300; ++depth)
+        {
+            CAPTURE(depth);
+            for (int variant = 0; variant < 3; ++variant)
+            {
+                CAPTURE(variant);
+                const json source = json::parse(nested_objects(depth, variant));
+                json result = json::parse(nested_objects(depth, (variant + 1) % 3));
+                json expected = result;
+                result.update(source, true);
+                reference_update(expected, source);
+                CHECK(result == expected);
+            }
+        }
+    }
+
+    SECTION("objects nested too deeply for the call stack (#5545)")
+    {
+        // merging used to recurse once per nesting level. The result is only
+        // walked, never copied or compared, since those recurse too.
+        const std::size_t depth = 100000;
+        json target = json::parse(nested_objects(depth, 0));
+        target.update(json::parse(nested_objects(depth, 1)), true);
+
+        const json* p = &target;
+        for (std::size_t i = 0; i < depth; ++i)
+        {
+            p = &p->at("a");
+        }
+        CHECK(p->size() == 2);
+        CHECK(p->at("x") == 1);
+        CHECK(p->at("y") == 2);
+    }
+}

tests/src/unit-msgpack.cpp

@@ -14,6 +14,7 @@ using nlohmann::json;
     using namespace nlohmann::literals; // NOLINT(google-build-using-namespace)
 #endif
 
+#include <cstdint> // SIZE_MAX, UINT32_MAX
 #include <fstream>
 #include <sstream>
 #include <iomanip>
@@ -255,7 +256,7 @@ TEST_CASE("MessagePack")
 
                 SECTION("256..65535 (int 16)")
                 {
-                    for (size_t i = 256; i <= 65535; ++i)
+                    for (size_t i = 256; i <= 65535; i = utils::next_integer_sample(i, static_cast<size_t>(65535), static_cast<size_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -440,7 +441,7 @@ TEST_CASE("MessagePack")
 
                 SECTION("-32768..-129 (int 16)")
                 {
-                    for (int16_t i = -32768; i <= static_cast<std::int16_t>(-129); ++i)
+                    for (int16_t i = -32768; i <= static_cast<std::int16_t>(-129); i = utils::next_integer_sample(i, static_cast<int16_t>(-129), static_cast<int16_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -646,7 +647,7 @@ TEST_CASE("MessagePack")
 
                 SECTION("256..65535 (uint 16)")
                 {
-                    for (size_t i = 256; i <= 65535; ++i)
+                    for (size_t i = 256; i <= 65535; i = utils::next_integer_sample(i, static_cast<size_t>(65535), static_cast<size_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -1554,6 +1555,27 @@ TEST_CASE("MessagePack")
             CHECK(json::from_msgpack(std::vector<uint8_t>({0x81, 0xff, 0x01}), true, false).is_discarded());
         }
 
+        SECTION("invalid UTF-8 in string (see #5529)")
+        {
+            // a fixstr of length 2 (0xA0 | 2) whose bytes are not valid UTF-8
+            // (0xC0 0xAE is an overlong encoding of '.') must be rejected at
+            // decode time, matching every other kind of malformed binary
+            // input, rather than only failing later when the resulting
+            // value is dumped
+            json _;
+            CHECK_THROWS_WITH_AS(_ = json::from_msgpack(std::vector<uint8_t>({0xa2, 0xc0, 0xae})), "[json.exception.parse_error.113] parse error at byte 3: syntax error while parsing MessagePack string: invalid string: ill-formed UTF-8 byte", json::parse_error&);
+            CHECK(json::from_msgpack(std::vector<uint8_t>({0xa2, 0xc0, 0xae}), true, false).is_discarded());
+
+            // a MessagePack bin8 blob with the very same bytes is NOT text
+            // and must still be accepted as-is
+            CHECK_NOTHROW(_ = json::from_msgpack(std::vector<uint8_t>({0xc4, 0x02, 0xc0, 0xae})));
+            CHECK(_ == json::binary(std::vector<std::uint8_t>({0xc0, 0xae})));
+
+            // valid UTF-8 must still round-trip
+            const json j = "h\xc3\xa9llo, w\xc3\xb6rld! \xe6\x97\xa5\xe6\x9c\xac\xe8\xaa\x9e"; // héllo, wörld! 日本語
+            CHECK(json::from_msgpack(json::to_msgpack(j)) == j);
+        }
+
         SECTION("strict mode")
         {
             std::vector<uint8_t> const vec = {0xc0, 0xc0};
@@ -1597,7 +1619,206 @@ TEST_CASE("MessagePack")
     }
 }
 
+TEST_CASE("issue #5405 - array reserve for definite-length MessagePack arrays")
+{
+#if !defined(JSON_NOEXCEPTION)
+    // this SECTION relies on catching a thrown exception to distinguish
+    // which of two acceptable, bounded rejections a hostile header took;
+    // under JSON_NOEXCEPTION, JSON_THROW never produces a catchable C++
+    // exception (it aborts instead), so this cannot be tested that way here
+    SECTION("a huge claimed length with no element data must not over-allocate")
+    {
+        // 0xdd: array 32 (four-byte length); claims 0xFFFFFFFF (4294967295)
+        // elements but provides none. max_size() for a std::vector is far
+        // larger than this count, so it does not reject the header outright;
+        // the (capped) reservation must not attempt to allocate space for
+        // billions of elements before the missing data is detected.
+        json _;
+        const std::vector<uint8_t> input = {0xdd, 0xFF, 0xFF, 0xFF, 0xFF};
+        // On a platform where std::size_t is narrower than 64 bits (e.g.
+        // 32-bit), the claimed count 0xFFFFFFFF coincides with that
+        // platform's SIZE_MAX, which some size-narrowing checks treat the
+        // same as detail::unknown_size(); it may then be rejected before
+        // the SAX consumer's own max_size() check (out_of_range.408) rather
+        // than being accepted and only found short of data once the
+        // (capped) reservation looks for element bytes that were never
+        // provided (parse_error.110). Either is an acceptable, bounded
+        // rejection of the hostile header -- the property under test is
+        // that no path attempts to allocate space for billions of elements.
+        bool threw = false;
+        try
+        {
+            _ = json::from_msgpack(input);
+        }
+        catch (const json::parse_error& e)
+        {
+            threw = true;
+            CHECK(e.id == 110);
+            CHECK(std::string(e.what()) == "[json.exception.parse_error.110] parse error at byte 6: syntax error while parsing MessagePack value: unexpected end of input");
+        }
+        catch (const json::out_of_range& e)
+        {
+            threw = true;
+            CHECK(e.id == 408);
+            CHECK(std::string(e.what()).find("excessive") != std::string::npos);
+        }
+        CHECK(threw);
+        CHECK(json::from_msgpack(input, true, false).is_discarded());
+    }
+#endif
+
+    SECTION("arrays of various sizes decode to the same value as before the reserve optimization")
+    {
+        for (const auto size :
+                {
+                    std::size_t{0}, std::size_t{1}, std::size_t{5}, // small
+                    std::size_t{16384},                             // exactly at the reserve cap
+                    std::size_t{20000}                              // above the reserve cap
+                })
+        {
+            CAPTURE(size)
+            json j = json::array();
+            for (std::size_t i = 0; i < size; ++i)
+            {
+                j.push_back(static_cast<int>(i % 1000));
+            }
+
+            const auto packed = json::to_msgpack(j);
+            CHECK(json::from_msgpack(packed) == j);
+        }
+    }
+
+    SECTION("a user-defined SAX consumer is unaffected by the internal DOM reserve optimization")
+    {
+        // the reserve() call is local to json_sax_dom_parser / json_sax_dom_callback_parser;
+        // a custom SAX consumer that does not touch a DOM array sees identical events
+        json j = json::array();
+        for (int i = 0; i < 100; ++i)
+        {
+            j.push_back(i);
+        }
+        const auto packed = json::to_msgpack(j);
+
+        SaxCountdown scp(1000000); // large enough to never trigger an abort
+        CHECK(json::sax_parse(packed, &scp, json::input_format_t::msgpack));
+    }
+}
+
+TEST_CASE("regression test - MessagePack ext type rejects a subtype that doesn't fit a single byte")
+{
+    // subtype 0-255 must still round-trip correctly (regression guard, pre-existing behavior)
+    CHECK(json::from_msgpack(json::to_msgpack(json::binary({1, 2}, 0))).get_binary().subtype() == 0);
+    CHECK(json::from_msgpack(json::to_msgpack(json::binary({1, 2}, 200))).get_binary().subtype() == 200);
+    CHECK(json::from_msgpack(json::to_msgpack(json::binary({1, 2}, 255))).get_binary().subtype() == 255);
+
+    // a subtype > 255 must throw instead of silently truncating
+    CHECK_THROWS_AS(json::to_msgpack(json::binary({1, 2}, 256)), json::out_of_range);
+    CHECK_THROWS_WITH_AS(json::to_msgpack(json::binary({1, 2}, 70000)), "[json.exception.out_of_range.415] subtype 70000 is too large for the MessagePack ext type (max 255)", json::out_of_range);
+
+    // a binary value with no subtype at all must be unaffected
+    CHECK(json::from_msgpack(json::to_msgpack(json::binary({1, 2}))).get_binary().has_subtype() == false);
+}
+
 // use this testcase outside [hide] to run it with Valgrind
+TEST_CASE("MessagePack nesting does not consume the call stack")
+{
+    // Reading a container used to call back into the value reader once per
+    // element, so the native call stack grew with the nesting depth of the
+    // input: one frame per byte for repeated 0x91 (a one-element array), which
+    // crashes the process long before the input is exhausted (#5104). The
+    // containers are kept on a heap stack now.
+    //
+    // Note that deeply nested values must not be compared, copied or dumped
+    // here: those operations are still recursive, and would reintroduce the
+    // very crash this checks for. Depth is measured by descending instead.
+
+    SECTION("an unterminated chain is reported, not crashed on")
+    {
+        json _;
+        const std::vector<uint8_t> input(300000, 0x91);
+        CHECK_THROWS_WITH_AS(_ = json::from_msgpack(input), "[json.exception.parse_error.110] parse error at byte 300001: syntax error while parsing MessagePack value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_msgpack(input, true, false).is_discarded());
+    }
+
+    SECTION("a well-formed deep value is read through the SAX interface")
+    {
+        std::vector<uint8_t> input(300000, 0x91);
+        input.push_back(0x01); // innermost value
+
+        SaxCountdown accept_all(600001);
+        CHECK(json::sax_parse(input, &accept_all, json::input_format_t::msgpack));
+    }
+
+    SECTION("a well-formed deep value is read into a value")
+    {
+        const std::size_t depth = 10000;
+        std::vector<uint8_t> input(depth, 0x91);
+        input.push_back(0x01);
+
+        json j = json::from_msgpack(input);
+
+        std::size_t measured = 0;
+        const json* p = &j;
+        while (p->is_array() && !p->empty())
+        {
+            p = &p->front();
+            ++measured;
+        }
+        CHECK(measured == depth);
+        CHECK(p->is_number());
+    }
+
+    SECTION("containers are still read the same way")
+    {
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x90})) == json::array());
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x80})) == json::object());
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x92, 0x90, 0x80})) == json({json::array(), json::object()}));
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x91, 0x91, 0x91, 0x90})) == json({{{json::array()}}}));
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0x81, 0xA1, 'a', 0x81, 0xA1, 'b', 0x92, 0x01, 0x02})) == json({{"a", {{"b", {1, 2}}}}}));
+        // array 16 and map 32, i.e. the counted forms
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0xDC, 0x00, 0x02, 0x01, 0x02})) == json({1, 2}));
+        CHECK(json::from_msgpack(std::vector<uint8_t>({0xDF, 0x00, 0x00, 0x00, 0x01, 0xA1, 'k', 0xC3})) == json({{"k", true}}));
+    }
+}
+
+TEST_CASE("MessagePack input that cannot be read is discarded by every overload")
+{
+    std::vector<std::uint8_t> input = json::to_msgpack(json({{"a", {1, 2}}}));
+    input.pop_back();
+
+    json _;
+    CHECK_THROWS_AS(_ = json::from_msgpack(input.begin(), input.end()), json::parse_error&);
+    CHECK(json::from_msgpack(input, true, false).is_discarded());
+    CHECK(json::from_msgpack(input.begin(), input.end(), true, false).is_discarded());
+    CHECK(json::from_msgpack(input.data(), input.size(), true, false).is_discarded());
+    CHECK(json::from_msgpack({input.data(), input.size()}, true, false).is_discarded());
+}
+
+TEST_CASE("MessagePack SAX parsing stops at every event")
+{
+    // Containers are opened and closed by the loop that reads them; a SAX
+    // handler that rejects any event - including the end of a nested
+    // container - must stop the parse right there.
+    const auto count_events = [](const std::vector<std::uint8_t>& input)
+    {
+        int events = 0;
+        while (true)
+        {
+            SaxCountdown scp(events);
+            if (json::sax_parse(input, &scp, json::input_format_t::msgpack))
+            {
+                return events;
+            }
+            ++events;
+            REQUIRE(events < 1000);
+        }
+    };
+
+    // 20 events: every container kind closes inside another one
+    const json j = json::parse(R"({"a": [1, {"b": []}], "c": {"d": [[2]]}})");
+    CHECK(count_events(json::to_msgpack(j)) == 20);
+}
+
 TEST_CASE("single MessagePack roundtrip")
 {
     SECTION("sample.json")
@@ -1822,60 +2043,34 @@ TEST_CASE("MessagePack roundtrips" * doctest::skip())
         {
             CAPTURE(filename)
 
+            std::ifstream f_json(filename);
+            const json j1 = json::parse(f_json);
+            auto packed = utils::read_binary_file(filename + ".msgpack");
+
             {
                 INFO_WITH_TEMP(filename + ": std::vector<uint8_t>");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                const json j1 = json::parse(f_json);
-
-                // parse MessagePack file
-                auto packed = utils::read_binary_file(filename + ".msgpack");
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_msgpack(packed));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": std::ifstream");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                const json j1 = json::parse(f_json);
-
-                // parse MessagePack file
                 std::ifstream f_msgpack(filename + ".msgpack", std::ios::binary);
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_msgpack(f_msgpack));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": uint8_t* and size");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                const json j1 = json::parse(f_json);
-
-                // parse MessagePack file
-                auto packed = utils::read_binary_file(filename + ".msgpack");
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_msgpack({packed.data(), packed.size()}));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": output to output adapters");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                json const j1 = json::parse(f_json);
-
-                // parse MessagePack file
-                auto packed = utils::read_binary_file(filename + ".msgpack");
-
                 if (exclude_packed.count(filename) == 0u)
                 {
                     {
@@ -1968,3 +2163,256 @@ TEST_CASE("MessagePack with std::byte")
     }
 }
 #endif
+
+// the fake sizes below do not fit into a 32-bit std::size_t
+// with clang and libstdc++ 10, the std::filesystem::path conversion that
+// C++17 builds consider for every string type is ambiguous for a class
+// derived from std::string, so the string case is not tested there
+#if !(defined(__clang__) && defined(_GLIBCXX_RELEASE) && _GLIBCXX_RELEASE < 11)
+    #define JSON_TEST_BEYOND_UINT32_STRING 1
+#endif
+
+#if SIZE_MAX > UINT32_MAX
+template<typename T, typename A = std::allocator<T>>
+struct huge_array : std::vector<T, A>
+{
+    using base = std::vector<T, A>;
+    using base::base;
+
+    bool fake_size = false;
+
+    std::size_t size() const noexcept
+    {
+        if (fake_size)
+        {
+            return (std::numeric_limits<std::uint32_t>::max)() + 1ULL;
+        }
+
+        return base::size();
+    }
+};
+
+using huge_array_json = nlohmann::basic_json <
+                        std::map, huge_array, std::string, bool, std::int64_t, std::uint64_t,
+                        double, std::allocator, nlohmann::adl_serializer,
+                        std::vector<std::uint8_t>, void >;
+
+TEST_CASE("MessagePack Size above uint32 for array")
+{
+    huge_array_json j = huge_array_json::array();
+
+    j.push_back(1);
+    j.push_back(2);
+    j.push_back(3);
+
+    auto& array = j.get_ref<huge_array_json::array_t&>();
+    array.fake_size = true;
+
+    // write into a caller-owned vector: to_msgpack(j) reserves space based on
+    // the (faked) element count, which fails with bad_alloc on Windows
+    std::vector<std::uint8_t> result;
+    CHECK_THROWS_WITH_AS(
+        huge_array_json::to_msgpack(j, result),
+        "[json.exception.out_of_range.412] MessagePack length 4294967296 exceeds maximum of 4294967295",
+        json::out_of_range&);
+
+    array.fake_size = false;
+}
+
+template<typename K, typename V,
+         typename C = std::less<K>,
+         typename A = std::allocator<std::pair<const K, V>>>
+                                     struct huge_map : std::map<K, V, C, A>
+{
+    using base = std::map<K, V, C, A>;
+    using base::base;
+
+    bool fake_size = false;
+
+    std::size_t size() const noexcept
+    {
+        if (fake_size)
+        {
+            return static_cast<std::size_t>(UINT32_MAX) + 1ULL;
+        }
+
+        return base::size();
+    }
+};
+
+using huge_object_json = nlohmann::basic_json <
+                         huge_map,
+                         std::vector,
+                         std::string,
+                         bool,
+                         std::int64_t,
+                         std::uint64_t,
+                         double,
+                         std::allocator,
+                         nlohmann::adl_serializer,
+                         std::vector<std::uint8_t>,
+                         void >;
+
+TEST_CASE("MessagePack Size above uint32 for object")
+{
+
+    huge_object_json j = huge_object_json::object();
+
+    j["one"] = 1;
+    j["two"] = 2;
+
+    auto& object = j.get_ref<huge_object_json::object_t&>();
+    object.fake_size = true;
+
+    // write into a caller-owned vector: to_msgpack(j) reserves space based on
+    // the (faked) element count, which fails with bad_alloc on Windows
+    std::vector<std::uint8_t> result;
+    CHECK_THROWS_WITH_AS(
+        huge_object_json::to_msgpack(j, result),
+        "[json.exception.out_of_range.412] MessagePack length 4294967296 exceeds maximum of 4294967295",
+        json::out_of_range&);
+
+    object.fake_size = false;
+}
+
+#ifdef JSON_TEST_BEYOND_UINT32_STRING
+struct huge_string : std::string
+{
+    using std::string::string;
+
+    std::size_t size() const noexcept // NOLINT(readability-convert-member-functions-to-static)
+    {
+        return static_cast<std::size_t>(UINT32_MAX) + 1ULL;
+    }
+};
+
+using huge_string_json = nlohmann::basic_json <
+                         std::map,
+                         std::vector,
+                         huge_string,
+                         bool,
+                         std::int64_t,
+                         std::uint64_t,
+                         double,
+                         std::allocator,
+                         nlohmann::adl_serializer,
+                         std::vector<std::uint8_t>,
+                         void >;
+
+TEST_CASE("MessagePack Size above uint32 for string")
+{
+    const huge_string_json j = "hello";
+
+    CHECK_THROWS_WITH_AS(
+        huge_string_json::to_msgpack(j),
+        "[json.exception.out_of_range.412] MessagePack length 4294967296 exceeds maximum of 4294967295",
+        json::out_of_range&);
+}
+#endif
+
+struct huge_binary : std::vector<std::uint8_t>
+{
+    using std::vector<std::uint8_t>::vector;
+
+    std::size_t size() const noexcept // NOLINT(readability-convert-member-functions-to-static)
+    {
+        return static_cast<std::size_t>(UINT32_MAX) + 1ULL;
+    }
+};
+
+using huge_binary_json = nlohmann::basic_json <
+                         std::map,
+                         std::vector,
+                         std::string,
+                         bool,
+                         std::int64_t,
+                         std::uint64_t,
+                         double,
+                         std::allocator,
+                         nlohmann::adl_serializer,
+                         huge_binary,
+                         void >;
+
+TEST_CASE("MessagePack Size above uint32 for binary")
+{
+
+    huge_binary_json j = huge_binary_json::binary(huge_binary{});
+
+    j.get_binary().push_back(0x01);
+    j.get_binary().push_back(0x02);
+
+    CHECK_THROWS_WITH_AS(
+        huge_binary_json::to_msgpack(j),
+        "[json.exception.out_of_range.412] MessagePack length 4294967296 exceeds maximum of 4294967295",
+        json::out_of_range&);
+}
+#endif
+
+namespace
+{
+// types that report a size beyond UINT32_MAX without allocating that much
+// memory, so the MessagePack length limit can be tested cheaply; see the
+// similar types in unit-bson.cpp
+std::size_t beyond_uint32_size()
+{
+    return static_cast<std::size_t>((std::numeric_limits<std::uint32_t>::max)()) + 1;
+}
+
+class beyond_uint32_binary_t : public std::vector<std::uint8_t>
+{
+  public:
+    using std::vector<std::uint8_t>::vector;
+
+    size_type size() const noexcept // NOLINT(readability-convert-member-functions-to-static)
+    {
+        return beyond_uint32_size();
+    }
+};
+
+#ifdef JSON_TEST_BEYOND_UINT32_STRING
+class beyond_uint32_string_t : public std::string
+{
+  public:
+    using std::string::string;
+
+    size_type size() const noexcept // NOLINT(readability-convert-member-functions-to-static)
+    {
+        return beyond_uint32_size();
+    }
+};
+
+using beyond_uint32_string_json = nlohmann::basic_json <
+                                  std::map, std::vector, beyond_uint32_string_t, bool, std::int64_t, std::uint64_t,
+                                  double, std::allocator, nlohmann::adl_serializer, std::vector<std::uint8_t>, void >;
+#endif
+
+using beyond_uint32_binary_json = nlohmann::basic_json <
+                                  std::map, std::vector, std::string, bool, std::int64_t, std::uint64_t,
+                                  double, std::allocator, nlohmann::adl_serializer, beyond_uint32_binary_t, void >;
+} // namespace
+
+TEST_CASE("MessagePack lengths beyond UINT32_MAX cannot be serialized")
+{
+    // MessagePack stores the length of a string, binary value, array, or
+    // object in at most 32 bits; a larger one used to be written without any
+    // length at all
+#if SIZE_MAX > UINT32_MAX
+    {
+        const char* const expected = "[json.exception.out_of_range.412] MessagePack length 4294967296 exceeds maximum of 4294967295";
+
+        const beyond_uint32_binary_json binary = beyond_uint32_binary_json::binary(beyond_uint32_binary_t{});
+        CHECK_THROWS_WITH_AS(beyond_uint32_binary_json::to_msgpack(binary), expected, beyond_uint32_binary_json::out_of_range&);
+
+        const beyond_uint32_binary_json ext = beyond_uint32_binary_json::binary(beyond_uint32_binary_t{}, 42);
+        CHECK_THROWS_WITH_AS(beyond_uint32_binary_json::to_msgpack(ext), expected, beyond_uint32_binary_json::out_of_range&);
+
+#ifdef JSON_TEST_BEYOND_UINT32_STRING
+        // created from its type rather than from a beyond_uint32_string_t:
+        // that would consider the std::filesystem::path conversion, which
+        // libstdc++ 10 cannot decide for a class derived from std::string
+        const beyond_uint32_string_json string(beyond_uint32_string_json::value_t::string);
+        CHECK_THROWS_WITH_AS(beyond_uint32_string_json::to_msgpack(string), expected, beyond_uint32_string_json::out_of_range&);
+#endif
+    }
+#endif
+}

tests/src/unit-no-macro-leak.cpp

@@ -0,0 +1,34 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+// This file makes sure that none of the internal JSON_HEDLEY_* macros (vendored
+// from https://nemequ.github.io/hedley/, see
+// include/nlohmann/thirdparty/hedley/hedley.hpp) leak into the including
+// translation unit. include/nlohmann/detail/macro_unscope.hpp is supposed to
+// #undef every JSON_HEDLEY_* macro (via hedley_undef.hpp) once json.hpp has
+// been fully processed. See https://github.com/nlohmann/json/issues/5408,
+// where JSON_HEDLEY_PRAGMA, JSON_HEDLEY_PREDICT_TRUE, JSON_HEDLEY_PREDICT_FALSE,
+// and JSON_HEDLEY_CLANG_HAS_DECLSPEC_ATTRIBUTE escaped this cleanup because
+// hedley_undef.hpp had no matching #undef for them.
+//
+// hedley_undef_checks.inc (included below) is generated at CMake configure/
+// build time by cmake/scripts/gen_hedley_undef_check.cmake, which derives the
+// full list of JSON_HEDLEY_* macro names directly from hedley.hpp. That way
+// this test covers every macro Hedley actually defines -- not a hardcoded
+// snapshot that would silently go stale the next time `make update_hedley`
+// runs -- and can never drift from the vendored header.
+
+#include "doctest_compatibility.h"
+
+#include <nlohmann/json.hpp>
+
+TEST_CASE("JSON_HEDLEY macros do not leak after including json.hpp")
+{
+#include "hedley_undef_checks.inc"
+    CHECK(true); // keep an assertion when nothing leaked
+}

tests/src/unit-no-mem-leak-on-adl-serialize.cpp

@@ -70,7 +70,7 @@ TEST_CASE("check_for_mem_leak_on_adl_to_json-2")
     }
 }
 
-TEST_CASE("check_for_mem_leak_on_adl_to_json-2")
+TEST_CASE("check_for_mem_leak_on_adl_to_json-3")
 {
     try
     {

tests/src/unit-no_io_and_user_exceptions.cpp

@@ -0,0 +1,91 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+// This translation unit is a dedicated, small compile-and-run check for two
+// configuration macros that (per #5423) were never exercised anywhere in the
+// test matrix:
+// - JSON_NO_IO, which removes the library's <istream>/<ostream> support
+//   (operator<<, operator>>, and the stream-based overloads of dump()/parse())
+// - the JSON_THROW_USER / JSON_TRY_USER / JSON_CATCH_USER trio, which lets a
+//   user replace the library's internal exception handling
+//
+// Both macros are about excluding/replacing a facility the library would
+// otherwise pull in on its own, and defining one has no bearing on the other,
+// so -- to keep the test matrix small -- they are exercised together in a
+// single dedicated file instead of two.
+//
+// JSON_NO_IO requires this file itself to never rely on <iostream>/<sstream>;
+// only string-based parsing/dumping is used below.
+#define JSON_NO_IO 1
+
+// The user-supplied exception macros below are a *conforming* replacement:
+// they simply forward to the real throw/try/catch keywords (via a counter so
+// the test can assert each macro was actually invoked, not just defined), so
+// every exception-related behavior the library relies on internally --
+// including rethrowing std::out_of_range as json::out_of_range in at() --
+// keeps working exactly as it would with the library's own default macros.
+static int json_throw_user_call_count = 0; // NOLINT(cppcoreguidelines-avoid-non-const-global-variables)
+
+#define JSON_THROW_USER(exception) do { ++json_throw_user_call_count; throw (exception); } while (false) // NOLINT(cppcoreguidelines-macro-usage)
+#define JSON_TRY_USER try // NOLINT(cppcoreguidelines-macro-usage)
+#define JSON_CATCH_USER(exception) catch (exception) // NOLINT(cppcoreguidelines-macro-usage)
+
+#include "doctest_compatibility.h"
+
+#include <nlohmann/json.hpp>
+using json = nlohmann::json;
+
+TEST_CASE("JSON_NO_IO")
+{
+    // everything that does not touch <istream>/<ostream> must keep working:
+    // parsing from and dumping to std::string
+    const json j = json::parse(R"({"a":[1,2,3],"b":true})");
+    CHECK(j.dump() == R"({"a":[1,2,3],"b":true})");
+    CHECK(j.at("a").size() == 3);
+    CHECK(j.at("b").get<bool>() == true);
+}
+
+// this test relies on CHECK_THROWS_AS() actually invoking the guarded
+// expression so json_throw_user_call_count gets bumped and can be observed
+// afterwards; doctest's "--no-throw" test filter (which ci_test_noexceptions
+// passes, together with a global -DJSON_NOEXCEPTION added to CMAKE_CXX_FLAGS
+// for every translation unit in that build, this file included) compiles
+// CHECK_THROWS_AS() out to a no-op that never even invokes the given
+// expression -- so json::parse()/at() below would never be called at all and
+// the call-count assertions would fail even though our JSON_THROW_USER
+// override (which always really throws, regardless of JSON_NOEXCEPTION) would
+// have worked fine on its own
+#if !defined(JSON_NOEXCEPTION)
+TEST_CASE("JSON_THROW_USER, JSON_TRY_USER, JSON_CATCH_USER")
+{
+    json_throw_user_call_count = 0;
+
+    // json::parse() is [[nodiscard]] (JSON_HEDLEY_WARN_UNUSED_RESULT); under
+    // GCC in C++11 mode that expands to __attribute__((warn_unused_result)),
+    // which -- unlike a [[nodiscard]] attribute proper -- GCC does not
+    // consider satisfied by doctest's CHECK_THROWS_AS() wrapping the
+    // expression in a (void) cast, so the discarded return value would still
+    // be flagged under -Werror=unused-result; assign it to discard it instead,
+    // matching the established `json _ = json::parse(...)` pattern used
+    // elsewhere in the test suite (see unit-class_parser.cpp)
+    json _; // NOLINT(readability-identifier-naming)
+
+    // a parse error goes through JSON_THROW directly, i.e., through our
+    // JSON_THROW_USER override
+    CHECK_THROWS_AS(_ = json::parse("this is not JSON"), json::parse_error&);
+    CHECK(json_throw_user_call_count > 0);
+
+    // at() on an out-of-range array index internally catches std::out_of_range
+    // (JSON_TRY_USER/JSON_CATCH_USER) and rethrows it as json::out_of_range
+    // (JSON_THROW_USER again), so this exercises all three macros together
+    const int count_before = json_throw_user_call_count;
+    const json arr = json::array({1, 2, 3});
+    CHECK_THROWS_AS(arr.at(10), json::out_of_range&);
+    CHECK(json_throw_user_call_count > count_before);
+}
+#endif

tests/src/unit-ordered_json.cpp

@@ -81,3 +81,118 @@ TEST_CASE("regression test for issue #3732 - iteration_proxy_value<iter_impl<ord
     };
     static_cast<void>(fn);
 }
+
+TEST_CASE("copying an ordered_json with nested values")
+{
+    // ordered_map is backed by a vector, so copying an object that has
+    // structured values takes a different route than copying a std::map-backed
+    // one; see https://github.com/nlohmann/json/issues/5387
+    ordered_json oj;
+    oj["z"] = 1;
+    oj["a"]["y"] = 2;
+    oj["a"]["b"]["x"] = 3;
+    oj["m"] = {1, 2, {{"w", 4}}};
+
+    const ordered_json copy(oj);
+
+    SECTION("the copy is equal to the original")
+    {
+        CHECK(copy == oj);
+        CHECK(copy.dump() == oj.dump());
+    }
+
+    SECTION("the key order is preserved at every level")
+    {
+        CHECK(copy.dump() == R"({"z":1,"a":{"y":2,"b":{"x":3}},"m":[1,2,{"w":4}]})");
+    }
+
+    SECTION("the copy is independent of the original")
+    {
+        ordered_json mutated(oj);
+        mutated["a"]["b"]["x"] = 99;
+
+        CHECK(oj["a"]["b"]["x"] == 3);
+        CHECK(mutated["a"]["b"]["x"] == 99);
+    }
+}
+
+TEST_CASE("regression test - diff() must account for ordered_json member order")
+{
+    SECTION("pure reorder, no value changes")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}};
+        ordered_json b = {{"b", 2}, {"a", 1}};
+        CHECK(a != b); // order-sensitive equality
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("new key must land at the front")
+    {
+        ordered_json c = {{"b", 2}};
+        ordered_json e = {{"a", 1}, {"b", 2}};
+        CHECK(c.patch(ordered_json::diff(c, e)) == e);
+    }
+
+    SECTION("reorder plus a value change on one of the reordered keys")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}};
+        ordered_json b = {{"b", 20}, {"a", 1}};
+        CHECK(a != b);
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("reorder plus a deleted key")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}, {"c", 3}};
+        ordered_json b = {{"b", 2}, {"a", 1}};
+        CHECK(a != b);
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("reorder plus a nested value that itself needs a recursive diff")
+    {
+        ordered_json a = {{"a", {{"x", 1}, {"y", 2}}}, {"b", 2}};
+        ordered_json b = {{"b", 2}, {"a", {{"x", 1}, {"y", 99}}}};
+        CHECK(a != b);
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("three or more keys shuffled into a different order")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}, {"c", 3}, {"d", 4}};
+        ordered_json b = {{"d", 4}, {"b", 2}, {"a", 1}, {"c", 3}};
+        CHECK(a != b);
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("matching order still produces a minimal patch (fast path unaffected)")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}, {"c", 3}};
+        ordered_json b = {{"a", 1}, {"b", 20}, {"c", 3}};
+        auto p = ordered_json::diff(a, b);
+        // only the changed value should be touched, not a wholesale remove+add
+        CHECK(p.size() == 1);
+        CHECK(p[0]["op"] == "replace");
+        CHECK(p[0]["path"] == "/b");
+        CHECK(a.patch(p) == b);
+    }
+
+    SECTION("plain json (std::map-backed) is unaffected by same-key-different-insertion-order")
+    {
+        json a;
+        a["b"] = 2;
+        a["a"] = 1;
+
+        json b;
+        b["a"] = 1;
+        b["b"] = 2;
+
+        // std::map iteration is always sorted by key, so a == b regardless of
+        // insertion order, and diff() must still produce the same minimal
+        // (empty) result as before this fix
+        CHECK(a == b);
+        auto p = json::diff(a, b);
+        CHECK(p.empty());
+        CHECK(a.patch(p) == b);
+    }
+}

tests/src/unit-ordered_json2.cpp

@@ -0,0 +1,505 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-FileCopyrightText: 2018 Vitaliy Manushkin <agri@akamo.info>
+// SPDX-License-Identifier: MIT
+
+// This file closes a test-coverage gap described in GitHub issue #5421:
+// nlohmann::ordered_json (and other non-default basic_json specializations,
+// such as the alt_string-based one from unit-alt-string.cpp) were never
+// exercised through the binary formats (CBOR/MessagePack/UBJSON/BSON/BJData)
+// or through flatten()/unflatten()/diff()/patch()/merge_patch().
+
+#include "doctest_compatibility.h"
+
+#include <nlohmann/json.hpp>
+
+#include <cstdint>
+#include <string>
+#include <utility>
+#include <vector>
+
+using nlohmann::json;
+using nlohmann::ordered_json;
+
+/////////////////////////////////////////////////////////////////////////////
+// alt_json: a second, independent copy of the custom-string_t basic_json
+// specialization defined in unit-alt-string.cpp.
+//
+// It is duplicated here (rather than shared via a header) because every
+// unit-*.cpp file in this test suite is compiled into its own standalone
+// executable (see tests/CMakeLists.txt), so there is no ODR concern in
+// having the same class name defined in multiple translation units.
+//
+// Two members had to be added relative to the original alt_string
+// (a constructor from std::string, and a find(char, pos) overload) because
+// the original type was never used with the binary writers/readers before
+// this file: BSON's array/document writer converts std::to_string() results
+// and checks for embedded NUL characters via find(char), and the UBJSON/BSON
+// high-precision-number path constructs the SAX string_t argument from a
+// std::string. Neither path is exercised anywhere else in the test suite for
+// this type, which is presumably why the gap was never noticed.
+/////////////////////////////////////////////////////////////////////////////
+
+class alt_string;
+bool operator<(const char* op1, const alt_string& op2) noexcept; // NOLINT(misc-use-internal-linkage)
+void int_to_string(alt_string& target, std::size_t value); // NOLINT(misc-use-internal-linkage)
+
+class alt_string
+{
+  public:
+    using value_type = std::string::value_type;
+
+    static constexpr auto npos = (std::numeric_limits<std::size_t>::max)();
+
+    alt_string(const char* str): str_impl(str) {}
+    alt_string(const char* str, std::size_t count): str_impl(str, count) {}
+    alt_string(std::string str): str_impl(std::move(str)) {}
+    alt_string(size_t count, char chr): str_impl(count, chr) {}
+    alt_string() = default;
+
+    alt_string& append(char ch)
+    {
+        str_impl.push_back(ch);
+        return *this;
+    }
+
+    alt_string& append(const alt_string& str)
+    {
+        str_impl.append(str.str_impl);
+        return *this;
+    }
+
+    alt_string& append(const char* s, std::size_t length)
+    {
+        str_impl.append(s, length);
+        return *this;
+    }
+
+    void push_back(char c)
+    {
+        str_impl.push_back(c);
+    }
+
+    template <typename op_type>
+    bool operator==(const op_type& op) const
+    {
+        return str_impl == op;
+    }
+
+    bool operator==(const alt_string& op) const
+    {
+        return str_impl == op.str_impl;
+    }
+
+    template <typename op_type>
+    bool operator!=(const op_type& op) const
+    {
+        return str_impl != op;
+    }
+
+    bool operator!=(const alt_string& op) const
+    {
+        return str_impl != op.str_impl;
+    }
+
+    std::size_t size() const noexcept
+    {
+        return str_impl.size();
+    }
+
+    void resize(std::size_t n)
+    {
+        str_impl.resize(n);
+    }
+
+    void resize(std::size_t n, char c)
+    {
+        str_impl.resize(n, c);
+    }
+
+    template <typename op_type>
+    bool operator<(const op_type& op) const noexcept
+    {
+        return str_impl < op;
+    }
+
+    bool operator<(const alt_string& op) const noexcept
+    {
+        return str_impl < op.str_impl;
+    }
+
+    const char* c_str() const
+    {
+        return str_impl.c_str();
+    }
+
+    char& operator[](std::size_t index)
+    {
+        return str_impl[index];
+    }
+
+    const char& operator[](std::size_t index) const
+    {
+        return str_impl[index];
+    }
+
+    char& back()
+    {
+        return str_impl.back();
+    }
+
+    const char& back() const
+    {
+        return str_impl.back();
+    }
+
+    void clear()
+    {
+        str_impl.clear();
+    }
+
+    const value_type* data() const
+    {
+        return str_impl.data();
+    }
+
+    bool empty() const
+    {
+        return str_impl.empty();
+    }
+
+    std::size_t find(const alt_string& str, std::size_t pos = 0) const
+    {
+        return str_impl.find(str.str_impl, pos);
+    }
+
+    // needed by binary_writer's BSON support, which probes string keys for
+    // embedded NUL characters via find(char)
+    std::size_t find(char c, std::size_t pos = 0) const
+    {
+        return str_impl.find(c, pos);
+    }
+
+    std::size_t find_first_of(char c, std::size_t pos = 0) const
+    {
+        return str_impl.find_first_of(c, pos);
+    }
+
+    alt_string substr(std::size_t pos = 0, std::size_t count = npos) const
+    {
+        const std::string s = str_impl.substr(pos, count);
+        return {s.data(), s.size()};
+    }
+
+    alt_string& replace(std::size_t pos, std::size_t count, const alt_string& str)
+    {
+        str_impl.replace(pos, count, str.str_impl);
+        return *this;
+    }
+
+    void reserve(std::size_t new_cap = 0)
+    {
+        str_impl.reserve(new_cap);
+    }
+
+  private:
+    std::string str_impl {}; // NOLINT(readability-redundant-member-init)
+
+    friend bool operator<(const char* /*op1*/, const alt_string& /*op2*/) noexcept;
+};
+
+void int_to_string(alt_string& target, std::size_t value)
+{
+    target = std::to_string(value).c_str();
+}
+
+using alt_json = nlohmann::basic_json <
+                 std::map,
+                 std::vector,
+                 alt_string,
+                 bool,
+                 std::int64_t,
+                 std::uint64_t,
+                 double,
+                 std::allocator,
+                 nlohmann::adl_serializer >;
+
+bool operator<(const char* op1, const alt_string& op2) noexcept
+{
+    return op1 < op2.str_impl;
+}
+
+namespace
+{
+
+// collects the object keys of j, in iteration order
+std::vector<std::string> collect_keys(const ordered_json& j)
+{
+    std::vector<std::string> result;
+    for (auto it = j.cbegin(); it != j.cend(); ++it)
+    {
+        result.push_back(it.key());
+    }
+    return result;
+}
+
+// a nested object/array value with keys inserted in non-alphabetical order,
+// used to check both round-trip equality and (for ordered_json) that
+// insertion order survives a trip through a binary format
+ordered_json make_rich_ordered_json()
+{
+    ordered_json j;
+    j["zebra"] = 1;
+    j["apple"] = ordered_json::array({1, 2, 3});
+    j["mango"]["z_nested"] = true;
+    j["mango"]["a_nested"] = nullptr;
+    j["banana"] = "some text";
+    j["cherry"] = 3.14;
+    return j;
+}
+
+alt_json make_rich_alt_json()
+{
+    alt_json j;
+    j["zebra"] = 1;
+    j["apple"] = alt_json::array({1, 2, 3});
+    j["mango"]["z_nested"] = true;
+    j["mango"]["a_nested"] = nullptr;
+    j["banana"] = "some text";
+    j["cherry"] = 3.14;
+    return j;
+}
+
+} // namespace
+
+TEST_CASE("ordered_json across binary formats")
+{
+    const ordered_json original = make_rich_ordered_json();
+    const std::vector<std::string> original_keys = collect_keys(original);
+    const std::vector<std::string> original_mango_keys = collect_keys(original["mango"]);
+
+    SECTION("CBOR")
+    {
+        const auto bytes = ordered_json::to_cbor(original);
+        const auto restored = ordered_json::from_cbor(bytes);
+        CHECK(restored == original);
+        CHECK(collect_keys(restored) == original_keys);
+        CHECK(collect_keys(restored["mango"]) == original_mango_keys);
+    }
+
+    SECTION("MessagePack")
+    {
+        const auto bytes = ordered_json::to_msgpack(original);
+        const auto restored = ordered_json::from_msgpack(bytes);
+        CHECK(restored == original);
+        CHECK(collect_keys(restored) == original_keys);
+        CHECK(collect_keys(restored["mango"]) == original_mango_keys);
+    }
+
+    SECTION("UBJSON")
+    {
+        const auto bytes = ordered_json::to_ubjson(original);
+        const auto restored = ordered_json::from_ubjson(bytes);
+        CHECK(restored == original);
+        CHECK(collect_keys(restored) == original_keys);
+        CHECK(collect_keys(restored["mango"]) == original_mango_keys);
+    }
+
+    SECTION("BSON")
+    {
+        const auto bytes = ordered_json::to_bson(original);
+        const auto restored = ordered_json::from_bson(bytes);
+        CHECK(restored == original);
+        CHECK(collect_keys(restored) == original_keys);
+        CHECK(collect_keys(restored["mango"]) == original_mango_keys);
+    }
+
+    SECTION("BJData")
+    {
+        const auto bytes = ordered_json::to_bjdata(original);
+        const auto restored = ordered_json::from_bjdata(bytes);
+        CHECK(restored == original);
+        CHECK(collect_keys(restored) == original_keys);
+        CHECK(collect_keys(restored["mango"]) == original_mango_keys);
+    }
+
+    SECTION("BON8")
+    {
+        const auto bytes = ordered_json::to_bon8(original);
+        const auto restored = ordered_json::from_bon8(bytes);
+        CHECK(restored == original);
+        CHECK(collect_keys(restored) == original_keys);
+        CHECK(collect_keys(restored["mango"]) == original_mango_keys);
+    }
+}
+
+TEST_CASE("alt_json (custom string_t) across binary formats")
+{
+    const alt_json original = make_rich_alt_json();
+
+    SECTION("CBOR")
+    {
+        const auto bytes = alt_json::to_cbor(original);
+        const auto restored = alt_json::from_cbor(bytes);
+        CHECK(restored == original);
+    }
+
+    SECTION("MessagePack")
+    {
+        const auto bytes = alt_json::to_msgpack(original);
+        const auto restored = alt_json::from_msgpack(bytes);
+        CHECK(restored == original);
+    }
+
+    SECTION("UBJSON")
+    {
+        const auto bytes = alt_json::to_ubjson(original);
+        const auto restored = alt_json::from_ubjson(bytes);
+        CHECK(restored == original);
+    }
+
+    SECTION("BON8")
+    {
+        const auto bytes = alt_json::to_bon8(original);
+        const auto restored = alt_json::from_bon8(bytes);
+        CHECK(restored == original);
+    }
+
+    SECTION("BSON")
+    {
+        const auto bytes = alt_json::to_bson(original);
+        const auto restored = alt_json::from_bson(bytes);
+        CHECK(restored == original);
+    }
+
+    SECTION("BJData")
+    {
+        const auto bytes = alt_json::to_bjdata(original);
+        const auto restored = alt_json::from_bjdata(bytes);
+        CHECK(restored == original);
+    }
+}
+
+TEST_CASE("ordered_json operator== is sensitive to key order")
+{
+    // Unlike nlohmann::json (whose object_t is a std::map, so equality never
+    // depends on insertion order), ordered_json's object_t (ordered_map) is a
+    // std::vector<std::pair<Key, T>> under the hood, and does not define its
+    // own operator==: it inherits std::vector's element-wise comparison. As a
+    // result, two ordered_json objects holding the very same key/value pairs
+    // in different insertion order compare *unequal*. This is the property
+    // that makes the round-trip `CHECK(restored == original)` checks above a
+    // meaningful order-preservation check by themselves (the explicit
+    // collect_keys() comparisons make that check explicit/readable, and
+    // guard against this operator== behavior ever changing).
+    ordered_json a;
+    a["x"] = 1;
+    a["y"] = 2;
+
+    ordered_json b;
+    b["y"] = 2;
+    b["x"] = 1;
+
+    CHECK(a.size() == b.size());
+    CHECK(a["x"] == b["x"]);
+    CHECK(a["y"] == b["y"]);
+    CHECK_FALSE(a == b);
+}
+
+TEST_CASE("duplicate keys in a binary-encoded object")
+{
+    // CBOR encoding of a map with two entries under the same key "a": {"a": 1, "a": 2}
+    const std::vector<std::uint8_t> cbor_bytes
+    {
+        0xA2, 0x61, 'a', 0x01, 0x61, 'a', 0x02
+    };
+
+    // Both json (std::map, via operator[]) and ordered_json (ordered_map, via
+    // operator[]) build binary-decoded objects by looking up/creating the
+    // entry for each incoming key and then assigning the value into it. This
+    // means a repeated key does *not* produce two entries in either case;
+    // instead, the *first* occurrence's position is kept (relevant only for
+    // ordered_json) while the *last* occurrence's value wins (for both) --
+    // this matches operator[]'s "assign the referenced slot" semantics, and
+    // is worth noting because it differs from the initializer-list
+    // construction path (`ordered_json{{"a",1},{"a",2}}`), which builds
+    // through insert()/emplace() and therefore keeps the *first* value, not
+    // the last (see the "There are no dup keys..." case in
+    // unit-ordered_json.cpp).
+    const auto j = json::from_cbor(cbor_bytes);
+    const auto oj = ordered_json::from_cbor(cbor_bytes);
+
+    CHECK(j.size() == 1);
+    CHECK(oj.size() == 1);
+    CHECK(j["a"] == 2);
+    CHECK(oj["a"] == 2);
+    CHECK(j == json(oj));
+}
+
+TEST_CASE("ordered_json through flatten/unflatten")
+{
+    const ordered_json original = make_rich_ordered_json();
+    const std::vector<std::string> original_keys = collect_keys(original);
+    const std::vector<std::string> original_mango_keys = collect_keys(original["mango"]);
+
+    const ordered_json flat = original.flatten();
+    const ordered_json unflattened = flat.unflatten();
+
+    CHECK(unflattened == original);
+    // flatten() walks the value depth-first in iteration order and
+    // unflatten() re-inserts each flattened key via operator[] in the flat
+    // object's iteration order, so for ordered_json the original key order
+    // (both top-level and nested) is preserved end-to-end.
+    CHECK(collect_keys(unflattened) == original_keys);
+    CHECK(collect_keys(unflattened["mango"]) == original_mango_keys);
+}
+
+TEST_CASE("ordered_json through diff/patch/patch_inplace")
+{
+    ordered_json original;
+    original["one"] = 1;
+    original["two"] = 2;
+    original["three"] = 3;
+
+    ordered_json target = original;
+    target["one"] = 100;      // replace
+    target.erase("two");      // remove
+    target["four"] = 4;       // add
+
+    const ordered_json patch = ordered_json::diff(original, target);
+
+    SECTION("patch")
+    {
+        const ordered_json patched = original.patch(patch);
+        CHECK(patched == target);
+    }
+
+    SECTION("patch_inplace")
+    {
+        ordered_json copy = original;
+        copy.patch_inplace(patch);
+        CHECK(copy == target);
+    }
+}
+
+TEST_CASE("ordered_json through merge_patch")
+{
+    ordered_json original;
+    original["a"] = 1;
+    original["b"] = 2;
+
+    const ordered_json patch = {{"b", nullptr}, {"c", 3}};
+
+    original.merge_patch(patch);
+
+    ordered_json expected;
+    expected["a"] = 1;
+    expected["c"] = 3;
+
+    CHECK(original == expected);
+    CHECK(collect_keys(original) == collect_keys(expected));
+}

tests/src/unit-precise-stream-position.cpp

@@ -0,0 +1,237 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+#include "doctest_compatibility.h"
+
+// This file tests the opt-in JSON_PRECISE_STREAM_POSITION, so it defines the
+// macro itself rather than relying on a -D flag, and runs in every build. The
+// default behavior is pinned in unit-deserialization.cpp.
+#ifdef JSON_PRECISE_STREAM_POSITION
+    #undef JSON_PRECISE_STREAM_POSITION
+#endif
+
+#define JSON_PRECISE_STREAM_POSITION 1
+
+#include <nlohmann/json.hpp>
+using nlohmann::json;
+
+#include <cstddef>
+#include <sstream>
+#include <streambuf>
+#include <string>
+#include <utility>
+#include <vector>
+
+#define STRINGIZE_EX(x) #x
+#define STRINGIZE(x) STRINGIZE_EX(x)
+
+namespace
+{
+// A streambuf that keeps no get area at all and refuses every putback: with an
+// empty get area, sungetc() always ends up in pbackfail(). Used to check that
+// the character terminating a number is left in the input without relying on
+// the streambuf being able to put a consumed character back.
+class no_putback_streambuf : public std::streambuf
+{
+  public:
+    explicit no_putback_streambuf(std::string s) : m_data(std::move(s)) {}
+
+  protected:
+    // peek at the next character without consuming it
+    int_type underflow() override
+    {
+        if (m_pos >= m_data.size())
+        {
+            return traits_type::eof();
+        }
+        return traits_type::to_int_type(m_data[m_pos]);
+    }
+
+    // consume the next character
+    int_type uflow() override
+    {
+        if (m_pos >= m_data.size())
+        {
+            return traits_type::eof();
+        }
+        return traits_type::to_int_type(m_data[m_pos++]);
+    }
+
+    int_type pbackfail(int_type /*c*/) override
+    {
+        return traits_type::eof();
+    }
+
+  private:
+    std::string m_data;
+    std::size_t m_pos = 0;
+};
+
+// read the characters that are left in a stream
+std::string remaining(std::istream& is)
+{
+    std::string result;
+    char c = 0;
+    while (is.get(c))
+    {
+        result += c;
+    }
+    return result;
+}
+}  // namespace
+
+TEST_CASE("JSON_PRECISE_STREAM_POSITION")
+{
+    SECTION("the macro is part of the ABI tag")
+    {
+        const std::string ns = STRINGIZE(NLOHMANN_JSON_NAMESPACE);
+        // other tags may come before it, e.g. json_abi_diag_psp
+        CHECK(ns.find("_psp") != std::string::npos);
+    }
+
+    SECTION("a number does not consume the character that terminates it")
+    {
+        // a number is only terminated by the character following it; that
+        // character must be given back so the stream is positioned right
+        // after the value
+        const std::vector<std::pair<std::string, std::string>> tests =
+        {
+            {"1true", "true"},
+            {"1[2]", "[2]"},
+            {"1{}", "{}"},
+            {R"(1"a")", R"("a")"},
+            {"1 true", " true"},
+            {"12,", ","},
+            {"-0.5e3x", "x"},
+            {"1null", "null"}
+        };
+
+        for (const auto& test : tests)
+        {
+            CAPTURE(test.first);
+            std::istringstream ss(test.first);
+            json j;
+            ss >> j;
+            CHECK(j == json::parse(test.first.substr(0, test.first.size() - test.second.size())));
+            CHECK(remaining(ss) == test.second);
+        }
+    }
+
+    SECTION("values that are self-delimiting are unaffected")
+    {
+        const std::vector<std::pair<std::string, std::string>> tests =
+        {
+            {"truefalse", "false"},
+            {"[1][2]", "[2]"},
+            {R"({"a":1}{"b":2})", R"({"b":2})"},
+            {R"("a""b")", R"("b")"},
+            {"null null", " null"}
+        };
+
+        for (const auto& test : tests)
+        {
+            CAPTURE(test.first);
+            std::istringstream ss(test.first);
+            json j;
+            ss >> j;
+            CHECK(remaining(ss) == test.second);
+        }
+    }
+
+    SECTION("a number at the end of the input leaves nothing behind")
+    {
+        for (const std::string s :
+                {"1", "12", "-3.5e2", " 7 "
+                })
+        {
+            CAPTURE(s);
+            std::istringstream ss(s);
+            json j;
+            ss >> j;
+            CHECK(remaining(ss).find_first_not_of(" \t\n\r") == std::string::npos);
+        }
+    }
+
+    SECTION("repeated extraction of concatenated values")
+    {
+        std::istringstream ss(R"(1true[2]3"x"{"a":4}5)");
+        const std::vector<json> expected =
+        {
+            json(1), json(true), json::parse("[2]"), json(3),
+            json("x"), json::parse(R"({"a":4})"), json(5)
+        };
+
+        for (const auto& e : expected)
+        {
+            json j;
+            ss >> j;
+            CHECK(j == e);
+        }
+    }
+
+    SECTION("differences to the default behavior")
+    {
+        // both of these work by accident without the macro, because the
+        // character after a number is swallowed; see unit-deserialization.cpp
+
+        SECTION("a separator after a number is not skipped")
+        {
+            std::istringstream ss("1,2");
+            json j;
+            ss >> j;
+            CHECK(j == 1);
+            CHECK_THROWS_AS(ss >> j, json::parse_error&);
+        }
+
+        SECTION("std::getline after a number sees the line break")
+        {
+            std::istringstream ss("42\nfoo");
+            json j;
+            std::string line;
+            ss >> j;
+            std::getline(ss, line);
+            CHECK(j == 42);
+            CHECK(line.empty());
+            std::getline(ss, line);
+            CHECK(line == "foo");
+        }
+    }
+
+    SECTION("sax_parse with strict == false")
+    {
+        std::istringstream ss("1true");
+        json j;
+        nlohmann::detail::json_sax_dom_parser<json, nlohmann::detail::input_stream_adapter> sdp(j, true);
+        CHECK(json::sax_parse(ss, &sdp, nlohmann::detail::input_format_t::json, false));
+        CHECK(j == 1);
+        CHECK(remaining(ss) == "true");
+    }
+
+    SECTION("strict parsing still rejects trailing data")
+    {
+        std::istringstream ss("1true");
+        json _;
+        CHECK_THROWS_WITH_AS(_ = json::parse(ss),
+                             "[json.exception.parse_error.101] parse error at line 1, column 5: syntax error while parsing value - unexpected true literal; expected end of input", json::parse_error&);
+
+        std::istringstream ss2("1true");
+        CHECK_FALSE(json::accept(ss2));
+    }
+
+    SECTION("a streambuf that cannot put back is not needed")
+    {
+        // the terminating character is never consumed, so no putback
+        // position is required
+        no_putback_streambuf buf("1true");
+        std::istream is(&buf);
+        json j;
+        is >> j;
+        CHECK(j == json(1));
+        CHECK(remaining(is) == "true");
+    }
+}

tests/src/unit-regression1.cpp

@@ -29,10 +29,7 @@ using nlohmann::json;
 #include <limits>
 #include <cstdio>
 #include "make_test_data_available.hpp"
-
-#ifdef JSON_HAS_CPP_17
-    #include <variant>
-#endif
+#include "test_utils.hpp"
 
 #include "fifo_map.hpp"
 
@@ -1373,7 +1370,8 @@ TEST_CASE("regression tests 1")
         std::array<uint8_t, 28> key1 = {{ 103, 92, 117, 48, 48, 48, 55, 92, 114, 215, 126, 214, 95, 92, 34, 174, 40, 71, 38, 174, 40, 71, 38, 223, 134, 247, 127, 0 }};
         std::string const key1_str(reinterpret_cast<char*>(key1.data()));
         json const j = key1_str;
-        CHECK_THROWS_WITH_AS(j.dump(), "[json.exception.type_error.316] invalid UTF-8 byte at index 10: 0x7E", json::type_error&);
+        // dump() is nodiscard; the exception is thrown by dump() itself before it would return
+        CHECK_THROWS_WITH_AS(utils::ignore_return_value(j.dump()), "[json.exception.type_error.316] invalid UTF-8 byte at index 10: 0x7E", json::type_error&);
     }
 
 #if JSON_USE_IMPLICIT_CONVERSIONS

tests/src/unit-regression2.cpp

@@ -31,6 +31,8 @@ using ordered_json = nlohmann::ordered_json;
 #include <type_traits>
 #include <utility>
 
+#include "test_utils.hpp"
+
 #ifdef JSON_HAS_CPP_17
     #include <any>
     #include <variant>
@@ -239,209 +241,6 @@ class my_allocator : public std::allocator<T>
     };
 };
 
-/////////////////////////////////////////////////////////////////////
-// for #3077
-/////////////////////////////////////////////////////////////////////
-
-class FooAlloc
-{};
-
-class Foo
-{
-  public:
-    explicit Foo(const FooAlloc& /* unused */ = FooAlloc()) {}
-
-    bool value = false;
-};
-
-class FooBar
-{
-  public:
-    Foo foo{}; // NOLINT(readability-redundant-member-init)
-};
-
-inline void from_json(const nlohmann::json& j, FooBar& fb) // NOLINT(misc-use-internal-linkage)
-{
-    j.at("value").get_to(fb.foo.value);
-}
-
-/////////////////////////////////////////////////////////////////////
-// for #3171
-/////////////////////////////////////////////////////////////////////
-
-struct for_3171_base // NOLINT(cppcoreguidelines-special-member-functions)
-{
-    for_3171_base(const std::string& /*unused*/ = {}) {}
-    virtual ~for_3171_base();
-
-    for_3171_base(const for_3171_base& other) // NOLINT(hicpp-use-equals-default,modernize-use-equals-default)
-        : str(other.str)
-    {}
-
-    for_3171_base& operator=(const for_3171_base& other)
-    {
-        if (this != &other)
-        {
-            str = other.str;
-        }
-        return *this;
-    }
-
-    for_3171_base(for_3171_base&& other) noexcept
-        : str(std::move(other.str))
-    {}
-
-    for_3171_base& operator=(for_3171_base&& other) noexcept
-    {
-        if (this != &other)
-        {
-            str = std::move(other.str);
-        }
-        return *this;
-    }
-
-    virtual void _from_json(const json& j)
-    {
-        j.at("str").get_to(str);
-    }
-
-    std::string str{}; // NOLINT(readability-redundant-member-init)
-};
-
-for_3171_base::~for_3171_base() = default;
-
-struct for_3171_derived : public for_3171_base
-{
-    for_3171_derived() = default;
-    ~for_3171_derived() override;
-    explicit for_3171_derived(const std::string& /*unused*/) { }
-
-    for_3171_derived(const for_3171_derived& other) // NOLINT(hicpp-use-equals-default,modernize-use-equals-default)
-        : for_3171_base(other)
-    {}
-
-    for_3171_derived& operator=(const for_3171_derived& other)
-    {
-        if (this != &other)
-        {
-            for_3171_base::operator=(other); // Call base class assignment operator
-        }
-        return *this;
-    }
-
-    for_3171_derived(for_3171_derived&& other) noexcept
-        : for_3171_base(std::move(other))
-    {}
-
-    for_3171_derived& operator=(for_3171_derived&& other) noexcept
-    {
-        if (this != &other)
-        {
-            for_3171_base::operator=(std::move(other)); // Call base class move assignment operator
-        }
-        return *this;
-    }
-};
-
-for_3171_derived::~for_3171_derived() = default;
-
-inline void from_json(const json& j, for_3171_base& tb) // NOLINT(misc-use-internal-linkage)
-{
-    tb._from_json(j);
-}
-
-/////////////////////////////////////////////////////////////////////
-// for #3312
-/////////////////////////////////////////////////////////////////////
-
-#ifdef JSON_HAS_CPP_20
-struct for_3312
-{
-    std::string name;
-};
-
-inline void from_json(const json& j, for_3312& obj) // NOLINT(misc-use-internal-linkage)
-{
-    j.at("name").get_to(obj.name);
-}
-#endif
-
-/////////////////////////////////////////////////////////////////////
-// for #3204
-/////////////////////////////////////////////////////////////////////
-
-struct for_3204_foo
-{
-    for_3204_foo() = default;
-    explicit for_3204_foo(std::string /*unused*/) {} // NOLINT(performance-unnecessary-value-param)
-};
-
-struct for_3204_bar
-{
-    enum constructed_from_t // NOLINT(cppcoreguidelines-use-enum-class)
-    {
-        constructed_from_none = 0,
-        constructed_from_foo = 1,
-        constructed_from_json = 2
-    };
-
-    explicit for_3204_bar(std::function<void(for_3204_foo)> /*unused*/) noexcept // NOLINT(performance-unnecessary-value-param)
-        : constructed_from(constructed_from_foo) {}
-    explicit for_3204_bar(std::function<void(json)> /*unused*/) noexcept // NOLINT(performance-unnecessary-value-param)
-        : constructed_from(constructed_from_json) {}
-
-    constructed_from_t constructed_from = constructed_from_none;
-};
-
-/////////////////////////////////////////////////////////////////////
-// for #3333
-/////////////////////////////////////////////////////////////////////
-
-struct for_3333 final
-{
-    for_3333(int x_ = 0, int y_ = 0) : x(x_), y(y_) {}
-
-    template <class T>
-    for_3333(const T& /*unused*/)
-    {
-        CHECK(false);
-    }
-
-    int x = 0;
-    int y = 0;
-};
-
-template <>
-inline for_3333::for_3333(const json& j)
-    : for_3333(j.value("x", 0), j.value("y", 0))
-{}
-
-/////////////////////////////////////////////////////////////////////
-// for #3810
-/////////////////////////////////////////////////////////////////////
-
-struct Example_3810
-{
-    int bla{};
-
-    Example_3810() = default;
-};
-
-NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE(Example_3810, bla) // NOLINT(misc-use-internal-linkage)
-
-/////////////////////////////////////////////////////////////////////
-// for #4740
-/////////////////////////////////////////////////////////////////////
-
-#ifdef JSON_HAS_CPP_17
-struct Example_4740
-{
-    std::optional<std::string> host = std::nullopt;
-    std::optional<int> port = std::nullopt;
-    NLOHMANN_DEFINE_TYPE_INTRUSIVE_WITH_DEFAULT(Example_4740, host, port)
-};
-#endif
-
 TEST_CASE("regression tests 2")
 {
     SECTION("issue #1001 - Fix memory leak during parser callback")
@@ -533,6 +332,7 @@ TEST_CASE("regression tests 2")
         CHECK(float_json::from_cbor(float_json::to_cbor(j)) == j);
         CHECK(float_json::from_msgpack(float_json::to_msgpack(j)) == j);
         CHECK(float_json::from_ubjson(float_json::to_ubjson(j)) == j);
+        CHECK(float_json::from_bon8(float_json::to_bon8(j)) == j);
 
         float_json j2 = {1000.0, 2000.0, 3000.0};
         CHECK(float_json::from_ubjson(float_json::to_ubjson(j2, true, true)) == j2);
@@ -639,7 +439,8 @@ TEST_CASE("regression tests 2")
                 s += static_cast<char>(i);
             }
             dump_test["1"] = s;
-            dump_test.dump(-1, ' ', true, nlohmann::json::error_handler_t::replace);
+            // dump() is nodiscard; this only checks that dumping does not throw/crash
+            utils::ignore_return_value(dump_test.dump(-1, ' ', true, nlohmann::json::error_handler_t::replace));
         }
     }
 
@@ -731,12 +532,14 @@ TEST_CASE("regression tests 2")
     {
         const std::array<unsigned char, 23> data = {{0x81, 0xA4, 0x64, 0x61, 0x74, 0x61, 0xC4, 0x0F, 0x33, 0x30, 0x30, 0x32, 0x33, 0x34, 0x30, 0x31, 0x30, 0x37, 0x30, 0x35, 0x30, 0x31, 0x30}};
         const json j = json::from_msgpack(data.data(), data.size());
+        // dump() is nodiscard; this only checks that dumping does not throw
         CHECK_NOTHROW(
-            j.dump(4,                             // Indent
-                   ' ',                           // Indent char
-                   false,                         // Ensure ascii
-                   json::error_handler_t::strict  // Error
-                  ));
+            utils::ignore_return_value(
+                j.dump(4,                             // Indent
+                       ' ',                           // Indent char
+                       false,                         // Ensure ascii
+                       json::error_handler_t::strict  // Error
+                      )));
     }
 
     SECTION("PR #2181 - regression bug with lvalue")
@@ -804,7 +607,11 @@ TEST_CASE("regression tests 2")
     SECTION("issue #2546 - parsing containers of std::byte")
     {
         const char DATA[] = R"("Hello, world!")"; // NOLINT(misc-const-correctness,cppcoreguidelines-avoid-c-arrays,hicpp-avoid-c-arrays,modernize-avoid-c-arrays)
-        const auto s = std::as_bytes(std::span(DATA));
+        // exclude the trailing '\0' that string-literal initialization adds to
+        // DATA: std::span(DATA) would span the full array extent (including
+        // that NUL), which is only silently accepted as end-of-input by default
+        // and would fail under JSON_STRICT_NUL_HANDLING
+        const auto s = std::as_bytes(std::span(DATA, sizeof(DATA) - 1));
         const json j = json::parse(s);
         CHECK(j.dump() == "\"Hello, world!\"");
     }
@@ -959,600 +766,108 @@ TEST_CASE("regression tests 2")
         CHECK(j == k);
     }
 
-#if JSON_HAS_FILESYSTEM || JSON_HAS_EXPERIMENTAL_FILESYSTEM
-    // JSON_HAS_CPP_17 (do not remove; see note at top of file)
-    SECTION("issue #3070 - Version 3.10.3 breaks backward-compatibility with 3.10.2 ")
+}
+
+TEST_CASE("regression test - parser callback must not lose a duplicate key's prior value")
+{
+    // a callback that rejects only the scalar value 2
+    const json::parser_callback_t drop_value_2 = [](int /*depth*/, json::parse_event_t ev, json & v) noexcept
     {
-        nlohmann::detail::std_fs::path text_path("/tmp/text.txt");
-        const json j(text_path);
+        return !(ev == json::parse_event_t::value && v == 2);
+    };
 
-        const auto j_path = j.get<nlohmann::detail::std_fs::path>();
-        CHECK(j_path == text_path);
-
-#if DOCTEST_CLANG || DOCTEST_GCC >= DOCTEST_COMPILER(8, 4, 0)
-        // only known to work on Clang and GCC >=8.4
-        CHECK_THROWS_WITH_AS(nlohmann::detail::std_fs::path(json(1)), "[json.exception.type_error.302] type must be string, but is number", json::type_error);
-#endif
-    }
-#endif
-
-    SECTION("issue #3077 - explicit constructor with default does not compile")
+    SECTION("duplicate key, second (scalar) value rejected - prior value is restored")
     {
-        json j;
-        j[0]["value"] = true;
-        std::vector<FooBar> foo;
-        j.get_to(foo);
+        const json j = json::parse(R"({"a":1,"a":2})", drop_value_2);
+        CHECK(j.dump() == "{\"a\":1}");
     }
 
-    SECTION("issue #3108 - ordered_json doesn't support range based erase")
+    SECTION("duplicate key, second value is an object rejected at object_end - prior value is restored")
     {
-        ordered_json j = {1, 2, 2, 4};
-
-        auto last = std::unique(j.begin(), j.end());
-        j.erase(last, j.end());
-
-        CHECK(j.dump() == "[1,2,4]");
-
-        j.erase(std::remove_if(j.begin(), j.end(), [](const ordered_json & val)
+        const json j = json::parse(R"({"a":1,"a":{"x":2}})",
+                                   [](int depth, json::parse_event_t ev, json& /*parsed*/) noexcept
         {
-            return val == 2;
-        }), j.end());
-
-        CHECK(j.dump() == "[1,4]");
+            return !(ev == json::parse_event_t::object_end && depth == 1);
+        });
+        CHECK(j.dump() == "{\"a\":1}");
     }
 
-    SECTION("issue #3343 - json and ordered_json are not interchangeable")
+    SECTION("duplicate key, second value is an array rejected at array_end - prior value is restored")
     {
-        json::object_t jobj({ { "product", "one" } });
-        ordered_json::object_t ojobj({{"product", "one"}});
-
-        auto jit = jobj.begin();
-        auto ojit = ojobj.begin();
-
-        CHECK(jit->first == ojit->first);
-        CHECK(jit->second.get<std::string>() == ojit->second.get<std::string>());
-    }
-
-    SECTION("issue #3171 - if class is_constructible from std::string wrong from_json overload is being selected, compilation failed")
-    {
-        const json j{{ "str", "value"}};
-
-        // failed with: error: no match for ‘operator=’ (operand types are ‘for_3171_derived’ and ‘const nlohmann::basic_json<>::string_t’
-        //                                               {aka ‘const std::__cxx11::basic_string<char>’})
-        //                  s = *j.template get_ptr<const typename BasicJsonType::string_t*>();
-        auto td = j.get<for_3171_derived>();
-
-        CHECK(td.str == "value");
-    }
-
-#ifdef JSON_HAS_CPP_20
-    SECTION("issue #3312 - Parse to custom class from unordered_json breaks on G++11.2.0 with C++20")
-    {
-        // see test for #3171
-        const ordered_json j = {{"name", "class"}};
-        for_3312 obj{};
-
-        j.get_to(obj);
-
-        CHECK(obj.name == "class");
-    }
-#endif
-
-#if defined(JSON_HAS_CPP_17) && JSON_USE_IMPLICIT_CONVERSIONS
-    SECTION("issue #3428 - Error occurred when converting nlohmann::json to std::any")
-    {
-        const json j;
-        const std::any a1 = j;
-        std::any&& a2 = j;
-
-        CHECK(a1.type() == typeid(j));
-        CHECK(a2.type() == typeid(j));
-    }
-#endif
-
-    SECTION("issue #3204 - ambiguous regression")
-    {
-        const for_3204_bar bar_from_foo([](for_3204_foo) noexcept {}); // NOLINT(performance-unnecessary-value-param)
-        const for_3204_bar bar_from_json([](json) noexcept {}); // NOLINT(performance-unnecessary-value-param)
-
-        CHECK(bar_from_foo.constructed_from == for_3204_bar::constructed_from_foo);
-        CHECK(bar_from_json.constructed_from == for_3204_bar::constructed_from_json);
-    }
-
-    SECTION("issue #3333 - Ambiguous conversion from nlohmann::basic_json<> to custom class")
-    {
-        const json j
+        const json j = json::parse(R"({"a":1,"a":[9,9]})",
+                                   [](int depth, json::parse_event_t ev, json& /*parsed*/) noexcept
         {
-            {"x", 1},
-            {"y", 2}
-        };
-        const for_3333 p = j;
-
-        CHECK(p.x == 1);
-        CHECK(p.y == 2);
+            return !(ev == json::parse_event_t::array_end && depth == 1);
+        });
+        CHECK(j.dump() == "{\"a\":1}");
     }
 
-    SECTION("issue #3810 - ordered_json doesn't support construction from C array of custom type")
+    SECTION("duplicate key, second value accepted (scalar) - last value wins")
     {
-        Example_3810 states[45]; // NOLINT(cppcoreguidelines-avoid-c-arrays,hicpp-avoid-c-arrays,modernize-avoid-c-arrays)
-
-        // fix "not used" warning
-        states[0].bla = 1;
-
-        const auto* const expected = R"([{"bla":1},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0}])";
-
-        // This works:
-        nlohmann::json j;
-        j["test"] = states;
-        CHECK(j["test"].dump() == expected);
-
-        // This doesn't compile:
-        nlohmann::ordered_json oj;
-        oj["test"] = states;
-        CHECK(oj["test"].dump() == expected);
-    }
-
-#ifdef JSON_HAS_CPP_17
-    SECTION("issue #4740 - build issue with std::optional")
-    {
-        const auto t1 = Example_4740();
-        const auto j1 = nlohmann::json(t1);
-        CHECK(j1.dump() == "{\"host\":null,\"port\":null}");
-        const auto t2 = j1.get<Example_4740>();
-        CHECK(!t2.host.has_value());
-        CHECK(!t2.port.has_value());
-
-        // improve coverage
-        auto t3 = Example_4740();
-        t3.port = 80;
-        t3.host = "example.com";
-        const auto j2 = nlohmann::json(t3);
-        CHECK(j2.dump() == "{\"host\":\"example.com\",\"port\":80}");
-        const auto t4 = j2.get<Example_4740>();
-        CHECK(t4.host.has_value());
-        CHECK(t4.port.has_value());
-    }
-#endif
-
-#if !defined(_MSVC_LANG)
-    // MSVC returns garbage on invalid enum values, so this test is excluded
-    // there.
-    SECTION("issue #4762 - json exception 302 with unhelpful explanation : type must be number, but is number")
-    {
-        // In #4762, the main issue was that a json object with an invalid type
-        // returned "number" as type_name(), because this was the default case.
-        // This test makes sure we now return "invalid" instead.
-        json j;
-        j.m_data.m_type = static_cast<json::value_t>(100); // NOLINT(clang-analyzer-optin.core.EnumCastOutOfRange)
-        CHECK(j.type_name() == "invalid");
-    }
-#endif
-
-#ifdef JSON_HAS_CPP_17
-    SECTION("issue #4804: from_cbor incompatible with std::vector<std::byte> as binary_t")
-    {
-        const std::vector<std::uint8_t> data = {0x80};
-        const auto decoded = json_4804::from_cbor(data);
-        CHECK((decoded == json_4804::array()));
-    }
-
-    SECTION("discussion #4209 - custom BinaryType direct assignment and round-tripping")
-    {
-        // Test that assigning a custom BinaryType directly creates a binary value, not an array
-        const std::vector<std::byte> original{std::byte{1}, std::byte{2}, std::byte{3}};
-        const json_4804 j = original;
-        CHECK(j.is_binary());
-        CHECK(!j.is_array());
-
-        // Test round-tripping: extracting the binary value back as the custom container type
-        const auto extracted = j.get<std::vector<std::byte>>();
-        CHECK(extracted == original);
-
-        // Test that the default json alias behavior is unchanged: std::vector<uint8_t> -> array
-        const json default_json = std::vector<std::uint8_t> {1, 2, 3};
-        CHECK(default_json.is_array());
-        CHECK(!default_json.is_binary());
-    }
-
-    SECTION("discussion #4209 - custom BinaryType extraction from parsed array")
-    {
-        // Test that extracting a custom BinaryType from a parsed JSON array still works
-        // (not just from a binary-typed node)
-        const auto j = json_4804::parse("[1,2,3]");
-        CHECK(j.is_array());
-        CHECK(!j.is_binary());
-
-        // Extracting as custom BinaryType should work from arrays
-        const auto extracted = j.get<std::vector<std::byte>>();
-        CHECK(extracted.size() == 3);
-        CHECK(extracted[0] == std::byte{1});
-        CHECK(extracted[1] == std::byte{2});
-        CHECK(extracted[2] == std::byte{3});
-    }
-
-    SECTION("issue #5046 - implicit conversion of return json to std::optional no longer implicit")
-    {
-        const json jval{};
-        auto GetValue = [](const json & valRoot) -> std::optional<json>
-        {
-            if (valRoot.contains("default"))
-            {
-                return valRoot.at("default");
-            }
-            return std::nullopt;
-        };
-        auto result = GetValue(jval);
-        CHECK(!result.has_value());
-    }
-#endif
-
-#if JSON_HAS_RANGES == 1
-    SECTION("issue #4440 - assert when using std::views::filter and GCC 10")
-    {
-        auto noOpFilter = std::views::filter([](auto&&) noexcept
+        const json j = json::parse(R"({"a":1,"a":2})", [](int, json::parse_event_t, json&) noexcept
         {
             return true;
         });
-        json j = {1, 2, 3};
-        auto filtered = j | noOpFilter;
-        CHECK(*filtered.begin() == 1);
+        CHECK(j.dump() == "{\"a\":2}");
     }
-#endif
 
-#if JSON_HAS_RANGES && !defined(__MINGW32__)
-    SECTION("issue #4916 - constructing array from C++20 ranges view does not work")
+    SECTION("duplicate key, second value accepted (object) - last value wins")
     {
-        std::vector<int> nums{1, 2, 37, 42, 21};
-        auto filteredNums = nums | std::views::filter([](int i)
+        const json j = json::parse(R"({"a":1,"a":{"x":2}})", [](int, json::parse_event_t, json&) noexcept
         {
-            return i > 10;
+            return true;
         });
-        json const j(filteredNums);
-        CHECK(j.type() == json::value_t::array);
-        CHECK(j == json({37, 42, 21}));
+        CHECK(j.dump() == "{\"a\":{\"x\":2}}");
     }
-#endif
 
-    // owning_view is not available in libstdc++ < 12
-#if JSON_HAS_RANGES && !defined(__MINGW32__) && !(defined(__GLIBCXX__) && _GLIBCXX_RELEASE < 12)
-    SECTION("issue #4916 - constructing array from prvalue C++20 ranges view (owning_view)")
+    SECTION("brand new (non-duplicate) key, value rejected - member is fully absent")
     {
-        json const j(std::vector<int> {1, 2, 37, 42, 21} | std::views::filter([](int i)
-        {
-            return i > 10;
-        }));
-        CHECK(j.type() == json::value_t::array);
-        CHECK(j == json({37, 42, 21}));
+        const json j = json::parse(R"({"a":1,"b":2})", drop_value_2);
+        CHECK(j.dump() == "{\"a\":1}");
     }
-#endif
 
-#if JSON_HAS_RANGES && !defined(__MINGW32__)
-    SECTION("issue #4916 - constructing array from C++20 transform view (prvalue elements)")
+    SECTION("duplicate key nested two levels deep")
     {
-        std::vector<int> nums{1, 2, 3};
-        auto t = nums | std::views::transform([](int i) noexcept
-        {
-            return i * 2;
-        });
-        json const j(t);
-        CHECK(j.type() == json::value_t::array);
-        CHECK(j == json({2, 4, 6}));
+        const json j = json::parse(R"({"outer":{"a":1,"a":2}})", drop_value_2);
+        CHECK(j.dump() == "{\"outer\":{\"a\":1}}");
     }
-#endif
-}
 
-TEST_CASE_TEMPLATE("issue #4798 - nlohmann::json::to_msgpack() encode float NaN as double", T, double, float) // NOLINT(readability-math-missing-parentheses, bugprone-throwing-static-initialization)
-{
-    // With issue #4798, we encode NaN, infinity, and -infinity as float instead
-    // of double to allow for smaller encodings.
-    const json jx = std::numeric_limits<T>::quiet_NaN();
-    const json jy = std::numeric_limits<T>::infinity();
-    const json jz = -std::numeric_limits<T>::infinity();
-
-    /////////////////////////////////////////////////////////////////////////
-    // MessagePack
-    /////////////////////////////////////////////////////////////////////////
-
-    // expected MessagePack values
-    const std::vector<std::uint8_t> msgpack_x = {{0xCA, 0x7F, 0xC0, 0x00, 0x00}};
-    const std::vector<std::uint8_t> msgpack_y = {{0xCA, 0x7F, 0x80, 0x00, 0x00}};
-    const std::vector<std::uint8_t> msgpack_z = {{0xCA, 0xFF, 0x80, 0x00, 0x00}};
-
-    CHECK(json::to_msgpack(jx) == msgpack_x);
-    CHECK(json::to_msgpack(jy) == msgpack_y);
-    CHECK(json::to_msgpack(jz) == msgpack_z);
-
-    CHECK(std::isnan(json::from_msgpack(msgpack_x).get<T>()));
-    CHECK(json::from_msgpack(msgpack_y).get<T>() == std::numeric_limits<T>::infinity());
-    CHECK(json::from_msgpack(msgpack_z).get<T>() == -std::numeric_limits<T>::infinity());
-
-    // Make sure the other MessagePakc encodings for NaN, infinity, and
-    // -infinity are still supported.
-    const std::vector<std::uint8_t> msgpack_x_2 = {{0xCB, 0x7F, 0xF8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
-    const std::vector<std::uint8_t> msgpack_y_2 = {{0xCB, 0x7F, 0xF0, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
-    const std::vector<std::uint8_t> msgpack_z_2 = {{0xCB, 0xFF, 0xF0, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
-    CHECK(std::isnan(json::from_msgpack(msgpack_x_2).get<T>()));
-    CHECK(json::from_msgpack(msgpack_y_2).get<T>() == std::numeric_limits<T>::infinity());
-    CHECK(json::from_msgpack(msgpack_z_2).get<T>() == -std::numeric_limits<T>::infinity());
-
-    /////////////////////////////////////////////////////////////////////////
-    // CBOR
-    /////////////////////////////////////////////////////////////////////////
-
-    // expected CBOR values
-    const std::vector<std::uint8_t> cbor_x = {{0xF9, 0x7E, 0x00}};
-    const std::vector<std::uint8_t> cbor_y = {{0xF9, 0x7C, 0x00}};
-    const std::vector<std::uint8_t> cbor_z = {{0xF9, 0xfC, 0x00}};
-
-    CHECK(json::to_cbor(jx) == cbor_x);
-    CHECK(json::to_cbor(jy) == cbor_y);
-    CHECK(json::to_cbor(jz) == cbor_z);
-
-    CHECK(std::isnan(json::from_cbor(cbor_x).get<T>()));
-    CHECK(json::from_cbor(cbor_y).get<T>() == std::numeric_limits<T>::infinity());
-    CHECK(json::from_cbor(cbor_z).get<T>() == -std::numeric_limits<T>::infinity());
-
-    // Make sure the other CBOR encodings for NaN, infinity, and -infinity are
-    // still supported.
-    const std::vector<std::uint8_t> cbor_x_2 = {{0xFA, 0x7F, 0xC0, 0x00, 0x00}};
-    const std::vector<std::uint8_t> cbor_y_2 = {{0xFA, 0x7F, 0x80, 0x00, 0x00}};
-    const std::vector<std::uint8_t> cbor_z_2 = {{0xFA, 0xFF, 0x80, 0x00, 0x00}};
-    const std::vector<std::uint8_t> cbor_x_3 = {{0xFB, 0x7F, 0xF8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
-    const std::vector<std::uint8_t> cbor_y_3 = {{0xFB, 0x7F, 0xF0, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
-    const std::vector<std::uint8_t> cbor_z_3 = {{0xFB, 0xFF, 0xF0, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
-    CHECK(std::isnan(json::from_cbor(cbor_x_2).get<T>()));
-    CHECK(json::from_cbor(cbor_y_2).get<T>() == std::numeric_limits<T>::infinity());
-    CHECK(json::from_cbor(cbor_z_2).get<T>() == -std::numeric_limits<T>::infinity());
-    CHECK(std::isnan(json::from_cbor(cbor_x_3).get<T>()));
-    CHECK(json::from_cbor(cbor_y_3).get<T>() == std::numeric_limits<T>::infinity());
-    CHECK(json::from_cbor(cbor_z_3).get<T>() == -std::numeric_limits<T>::infinity());
-}
-
-TEST_CASE("regression test #5074 - portable workaround for single-element brace init")
-{
-    json const j_obj = {{"key", "value"}};
-
-    json const j = json::array({j_obj});
-    CHECK(j.is_array());
-    CHECK(j.size() == 1);
-    CHECK(j[0] == j_obj);
-}
-
-#if defined(JSON_BRACE_INIT_COPY_SEMANTICS) && (JSON_BRACE_INIT_COPY_SEMANTICS == 1)
-TEST_CASE("regression test #5074 - single-element brace init with JSON_BRACE_INIT_COPY_SEMANTICS")
-{
-    // with JSON_BRACE_INIT_COPY_SEMANTICS: single-element brace init copies/moves
-    json const j_obj = {{"key", "value"}, {"num", 42}};
-    json const j_arr = {1, 2, 3};
-
-    // object: brace init copies instead of wrapping
-    json const j1{j_obj};
-    CHECK(j1.is_object());
-    CHECK(j1 == j_obj);
-
-    // array: brace init copies instead of wrapping
-    json const j2{j_arr};
-    CHECK(j2.is_array());
-    CHECK(j2.size() == 3);
-    CHECK(j2 == j_arr);
-
-    // primitives still work as initializer lists
-    json const j3{true};
-    CHECK(j3.is_boolean());
-
-    json const j4{42};
-    CHECK(j4.is_number_integer());
-}
-#endif
-
-struct Example_5122
-{
-    float b = 2;
-    nlohmann::ordered_map<std::string, std::string> c{}; // NOLINT(readability-redundant-member-init): needed for GCC -Weffc++
-    int a = 1;
-    NLOHMANN_DEFINE_TYPE_INTRUSIVE_WITH_DEFAULT(Example_5122, b, c, a)
-};
-
-TEST_CASE("regression test #5122 - from_json into types holding nlohmann::ordered_map")
-{
-    Example_5122 src;
-    src.c.emplace("first", "1");
-    src.c.emplace("second", "2");
-
-    ordered_json const j = src;
-    Example_5122 const dst = j.get<Example_5122>();
-
-    CHECK(dst.b == src.b);
-    CHECK(dst.a == src.a);
-    REQUIRE(dst.c.size() == src.c.size());
-    auto src_it = src.c.begin();
-    auto dst_it = dst.c.begin();
-    for (; src_it != src.c.end(); ++src_it, ++dst_it)
+    SECTION("three occurrences of the same key - middle rejected, last accepted")
     {
-        CHECK(dst_it->first == src_it->first);
-        CHECK(dst_it->second == src_it->second);
+        const json j = json::parse(R"({"k":1,"k":2,"k":3})", drop_value_2);
+        CHECK(j.dump() == "{\"k\":3}");
     }
 }
 
-// -Wself-assign-overloaded was introduced in Clang 7. Gate the pragma on
-// __has_warning so older Clang versions do not error with "unknown warning
-// group". The __has_warning check has to stay inside the __clang__ branch
-// because GCC does not provide it and would tokenize-error on the argument.
-#if defined(__clang__) && defined(__has_warning)
-    #if __has_warning("-Wself-assign-overloaded")
-        DOCTEST_CLANG_SUPPRESS_WARNING_PUSH
-        DOCTEST_CLANG_SUPPRESS_WARNING("-Wself-assign-overloaded")
-    #endif
-#endif
-
-TEST_CASE("regression test #5122 - nlohmann::ordered_map copy-assignment is self-assignment safe")
+TEST_CASE("regression test - excessive binary container size honors allow_exceptions=false")
 {
-    nlohmann::ordered_map<std::string, std::string> m;
-    m.emplace("first", "1");
-    m.emplace("second", "2");
+    // CBOR array with declared length 2^63
+    const std::vector<std::uint8_t> cbor = {0x9b, 0x80, 0, 0, 0, 0, 0, 0, 0};
+    // CBOR map with declared length 2^63
+    const std::vector<std::uint8_t> cbor_m = {0xbb, 0x80, 0, 0, 0, 0, 0, 0, 0};
+    // UBJSON array with declared length 2^63-1
+    const std::vector<std::uint8_t> ubj = {'[', '#', 'L', 0x7f, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
+    // BJData array with declared length 2^63-1 (little endian)
+    const std::vector<std::uint8_t> bjd = {'[', '#', 'L', 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x7f};
 
-    // Insertion order is preserved by ordered_map, so we can check it directly.
-    m = m;
+    // allow_exceptions=false must report failure instead of throwing/aborting
+    CHECK(json::from_cbor(cbor, true, false).is_discarded());
+    CHECK(json::from_cbor(cbor_m, true, false).is_discarded());
+    CHECK(json::from_ubjson(ubj, true, false).is_discarded());
+    CHECK(json::from_bjdata(bjd, true, false).is_discarded());
 
-    REQUIRE(m.size() == 2);
-    auto it = m.begin();
-    CHECK(it->first == "first");
-    CHECK(it->second == "1");
-    ++it;
-    CHECK(it->first == "second");
-    CHECK(it->second == "2");
-}
+    // allow_exceptions=true (the default) must still throw exactly as before.
+    // The exact message text is not checked here: on platforms where
+    // std::size_t is 32-bit, the CBOR reader's own length-narrowing check
+    // (get_cbor_container_size(), unrelated to this fix) intercepts a
+    // declared length of 2^63 before it ever reaches the check this test
+    // targets, with different (but equally valid, and already correct)
+    // wording -- see unit-cbor.cpp for coverage of that message.
+    json _;
+    CHECK_THROWS_AS(_ = json::from_cbor(cbor), json::out_of_range);
 
-#if defined(__clang__) && defined(__has_warning)
-    #if __has_warning("-Wself-assign-overloaded")
-        DOCTEST_CLANG_SUPPRESS_WARNING_POP
-    #endif
-#endif
-
-TEST_CASE("regression test #5122 - nlohmann::ordered_map move-assignment transfers contents")
-{
-    nlohmann::ordered_map<std::string, std::string> src;
-    src.emplace("first", "1");
-    src.emplace("second", "2");
-
-    nlohmann::ordered_map<std::string, std::string> dst;
-    dst.emplace("stale", "x");
-    dst = std::move(src);
-
-    REQUIRE(dst.size() == 2);
-    auto it = dst.begin();
-    CHECK(it->first == "first");
-    CHECK(it->second == "1");
-    ++it;
-    CHECK(it->first == "second");
-    CHECK(it->second == "2");
-
-    // Re-assigning into the moved-from object must leave it in a usable state.
-    src = nlohmann::ordered_map<std::string, std::string> {};
-    src.emplace("after-move", "3");
-    REQUIRE(src.size() == 1);
-    CHECK(src.begin()->first == "after-move");
-}
-
-// Stand-in for a third-party library (e.g., Eigen as of 3.4, which added
-// STL-compatible begin()/end() to its vector types), living in its own
-// namespace with its own to_json overload for its vector type.
-namespace issue_4320_eigen
-{
-// "array-compatible" from the library's point of view (it has begin()/end()),
-// but for which this (fake) third-party namespace provides its own to_json.
-struct vector3
-{
-    double v[3]; // NOLINT(cppcoreguidelines-avoid-c-arrays,hicpp-avoid-c-arrays,modernize-avoid-c-arrays,cppcoreguidelines-use-default-member-init,modernize-use-default-member-init)
-    vector3(double x, double y, double z) : v{x, y, z} {} // NOLINT(hicpp-member-init,cppcoreguidelines-pro-type-member-init)
-    double x() const
-    {
-        return v[0];
-    }
-    double y() const
-    {
-        return v[1];
-    }
-    double z() const
-    {
-        return v[2];
-    }
-    double* begin()
-    {
-        return v;
-    }
-    double* end()
-    {
-        return v + 3;
-    }
-    const double* begin() const
-    {
-        return v;
-    }
-    const double* end() const
-    {
-        return v + 3;
-    }
-};
-
-inline void to_json(json& j, const vector3& v) // NOLINT(misc-use-internal-linkage)
-{
-    j = {{"x", v.x()}, {"y", v.y()}, {"z", v.z()}};
-}
-} // namespace issue_4320_eigen
-
-// The user's own namespace, using the (fake) Eigen type as an implementation
-// detail behind a payload type that has nothing to do with vectors/arrays.
-namespace issue_4320
-{
-// Publicly derives from issue_4320_eigen::vector3 but does *not* define its
-// own to_json - it is only ever used as a temporary to reach the base
-// class's to_json via ADL.
-struct vector3_wrapper : issue_4320_eigen::vector3
-{
-    using issue_4320_eigen::vector3::vector3;
-};
-
-struct payload
-{
-    double x, y, z;
-};
-
-inline vector3_wrapper to_eigen(const payload& p) // NOLINT(misc-use-internal-linkage)
-{
-    return {p.x, p.y, p.z};
-}
-
-inline void to_json(json& j, const payload& p) // NOLINT(misc-use-internal-linkage)
-{
-    // Unqualified call, passing a *derived* vector3_wrapper: relies on ADL
-    // finding issue_4320_eigen::to_json(json&, const vector3&) through the
-    // vector3 base class, via a derived-to-base conversion. Must NOT resolve
-    // to the library's own generic array-compatible to_json (an exact-match
-    // template for vector3_wrapper, since it also has begin()/end()), which
-    // would serialize this as [x, y, z] instead of {"x":x, "y":y, "z":z}.
-    to_json(j, to_eigen(p));
-}
-} // namespace issue_4320
-
-TEST_CASE("issue #4320 - custom base class must not leak nlohmann::detail into ADL")
-{
-    // Before the fix, basic_json unconditionally derived from a type living in
-    // nlohmann::detail (json_default_base), which made nlohmann::detail an
-    // associated namespace of every basic_json for ADL purposes. That leaked
-    // the library's internal generic-array to_json overload into unqualified
-    // to_json() calls made from user code, silently bypassing user-defined
-    // to_json overloads reached via a derived-to-base conversion.
-    const issue_4320::payload p{1.0, 2.0, 3.0};
-
-    json j;
-    to_json(j, p);
-    CHECK(j == json({{"x", 1.0}, {"y", 2.0}, {"z", 3.0}}));
-}
-
-TEST_CASE("issue #5338 - truncated CBOR tagged binary subtype is rejected")
-{
-    const std::vector<std::vector<std::uint8_t>> truncated_tags =
-    {
-        {0xD8},
-        {0xD9, 0x00},
-        {0xDA, 0x00, 0x00, 0x00},
-        {0xDB, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}
-    };
-
-    for (const auto& data : truncated_tags)
-    {
-        CAPTURE(data);
-        for (const auto tag_handler :
-                {
-                    json::cbor_tag_handler_t::ignore, json::cbor_tag_handler_t::store
-                })
-        {
-            CAPTURE(tag_handler);
-            const auto result = json::from_cbor(data, true, false, tag_handler);
-            CHECK(result.is_discarded());
-        }
-    }
+    // regression guard: a genuinely truncated CBOR input must remain discarded
+    CHECK(json::from_cbor(std::vector<std::uint8_t> {0x9b, 0, 0, 0, 0, 0, 0, 0, 0x02}, true, false).is_discarded());
 }
 
 DOCTEST_CLANG_SUPPRESS_WARNING_POP

tests/src/unit-regression3.cpp

@@ -0,0 +1,923 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
+// |  |  |__   |  |  | | | |  version 3.12.0
+// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
+//
+// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
+// SPDX-License-Identifier: MIT
+
+// cmake/test.cmake selects the C++ standard versions with which to build a
+// unit test based on the presence of JSON_HAS_CPP_<VERSION> macros.
+// When using macros that are only defined for particular versions of the standard
+// (e.g., JSON_HAS_FILESYSTEM for C++17 and up), please mention the corresponding
+// version macro in a comment close by, like this:
+// JSON_HAS_CPP_<VERSION> (do not remove; see note at top of file)
+
+#include "doctest_compatibility.h"
+
+// for some reason including this after the json header leads to linker errors with VS 2017...
+#include <locale>
+
+// skip tests if JSON_DisableEnumSerialization=ON (#4384): std::byte is a
+// scoped enum, so get<std::byte>() (needed below to get<std::vector<std::byte>>()
+// from a plain JSON array, not just from an already-binary value) relies on
+// enum serialization being enabled
+#if defined(JSON_DISABLE_ENUM_SERIALIZATION) && (JSON_DISABLE_ENUM_SERIALIZATION == 1)
+    #define SKIP_TESTS_FOR_ENUM_SERIALIZATION
+#endif
+
+#define JSON_TESTS_PRIVATE
+#include <nlohmann/json.hpp>
+using json = nlohmann::json;
+using ordered_json = nlohmann::ordered_json;
+#ifdef JSON_TEST_NO_GLOBAL_UDLS
+    using namespace nlohmann::literals; // NOLINT(google-build-using-namespace)
+#endif
+
+#include <cstdio>
+#include <deque>
+#include <list>
+#include <type_traits>
+#include <utility>
+
+#ifdef JSON_HAS_CPP_17
+    #include <any>
+    #include <variant>
+#endif
+
+#ifdef JSON_HAS_CPP_17
+    #if __has_include(<optional>)
+        #include <optional>
+    #elif __has_include(<experimental/optional>)
+    #endif
+
+    /////////////////////////////////////////////////////////////////////
+    // for #4804
+    /////////////////////////////////////////////////////////////////////
+    using json_4804 = nlohmann::basic_json<std::map,        // ObjectType
+    std::vector,     // ArrayType
+    std::string,     // StringType
+    bool,            // BooleanType
+    std::int64_t,    // NumberIntegerType
+    std::uint64_t,   // NumberUnsignedType
+    double,          // NumberFloatType
+    std::allocator,  // AllocatorType
+    nlohmann::adl_serializer,  // JSONSerializer
+    std::vector<std::byte>,    // BinaryType
+    void                       // CustomBaseClass
+    >;
+#endif
+
+#ifdef JSON_HAS_CPP_20
+    #if __has_include(<span>)
+        #include <span>
+    #endif
+#endif
+
+// the explicit instantiation for #4825 is in unit-explicit_instantiation.cpp
+
+/////////////////////////////////////////////////////////////////////
+// for #4440
+/////////////////////////////////////////////////////////////////////
+#if JSON_HAS_RANGES == 1
+    #include <ranges>
+#endif
+
+// NLOHMANN_JSON_SERIALIZE_ENUM uses a static std::pair
+DOCTEST_CLANG_SUPPRESS_WARNING_PUSH
+DOCTEST_CLANG_SUPPRESS_WARNING("-Wexit-time-destructors")
+/////////////////////////////////////////////////////////////////////
+// for #3077
+/////////////////////////////////////////////////////////////////////
+
+class FooAlloc
+{};
+
+class Foo
+{
+  public:
+    explicit Foo(const FooAlloc& /* unused */ = FooAlloc()) {}
+
+    bool value = false;
+};
+
+class FooBar
+{
+  public:
+    Foo foo{}; // NOLINT(readability-redundant-member-init)
+};
+
+inline void from_json(const nlohmann::json& j, FooBar& fb) // NOLINT(misc-use-internal-linkage)
+{
+    j.at("value").get_to(fb.foo.value);
+}
+
+/////////////////////////////////////////////////////////////////////
+// for #3171
+/////////////////////////////////////////////////////////////////////
+
+struct for_3171_base // NOLINT(cppcoreguidelines-special-member-functions)
+{
+    for_3171_base(const std::string& /*unused*/ = {}) {}
+    virtual ~for_3171_base();
+
+    for_3171_base(const for_3171_base& other) // NOLINT(hicpp-use-equals-default,modernize-use-equals-default)
+        : str(other.str)
+    {}
+
+    for_3171_base& operator=(const for_3171_base& other)
+    {
+        if (this != &other)
+        {
+            str = other.str;
+        }
+        return *this;
+    }
+
+    for_3171_base(for_3171_base&& other) noexcept
+        : str(std::move(other.str))
+    {}
+
+    for_3171_base& operator=(for_3171_base&& other) noexcept
+    {
+        if (this != &other)
+        {
+            str = std::move(other.str);
+        }
+        return *this;
+    }
+
+    virtual void _from_json(const json& j)
+    {
+        j.at("str").get_to(str);
+    }
+
+    std::string str{}; // NOLINT(readability-redundant-member-init)
+};
+
+for_3171_base::~for_3171_base() = default;
+
+struct for_3171_derived : public for_3171_base
+{
+    for_3171_derived() = default;
+    ~for_3171_derived() override;
+    explicit for_3171_derived(const std::string& /*unused*/) { }
+
+    for_3171_derived(const for_3171_derived& other) // NOLINT(hicpp-use-equals-default,modernize-use-equals-default)
+        : for_3171_base(other)
+    {}
+
+    for_3171_derived& operator=(const for_3171_derived& other)
+    {
+        if (this != &other)
+        {
+            for_3171_base::operator=(other); // Call base class assignment operator
+        }
+        return *this;
+    }
+
+    for_3171_derived(for_3171_derived&& other) noexcept
+        : for_3171_base(std::move(other))
+    {}
+
+    for_3171_derived& operator=(for_3171_derived&& other) noexcept
+    {
+        if (this != &other)
+        {
+            for_3171_base::operator=(std::move(other)); // Call base class move assignment operator
+        }
+        return *this;
+    }
+};
+
+for_3171_derived::~for_3171_derived() = default;
+
+inline void from_json(const json& j, for_3171_base& tb) // NOLINT(misc-use-internal-linkage)
+{
+    tb._from_json(j);
+}
+
+/////////////////////////////////////////////////////////////////////
+// for #3312
+/////////////////////////////////////////////////////////////////////
+
+#ifdef JSON_HAS_CPP_20
+struct for_3312
+{
+    std::string name;
+};
+
+inline void from_json(const json& j, for_3312& obj) // NOLINT(misc-use-internal-linkage)
+{
+    j.at("name").get_to(obj.name);
+}
+#endif
+
+/////////////////////////////////////////////////////////////////////
+// for #3204
+/////////////////////////////////////////////////////////////////////
+
+struct for_3204_foo
+{
+    for_3204_foo() = default;
+    explicit for_3204_foo(std::string /*unused*/) {} // NOLINT(performance-unnecessary-value-param)
+};
+
+struct for_3204_bar
+{
+    enum constructed_from_t // NOLINT(cppcoreguidelines-use-enum-class)
+    {
+        constructed_from_none = 0,
+        constructed_from_foo = 1,
+        constructed_from_json = 2
+    };
+
+    explicit for_3204_bar(std::function<void(for_3204_foo)> /*unused*/) noexcept // NOLINT(performance-unnecessary-value-param)
+        : constructed_from(constructed_from_foo) {}
+    explicit for_3204_bar(std::function<void(json)> /*unused*/) noexcept // NOLINT(performance-unnecessary-value-param)
+        : constructed_from(constructed_from_json) {}
+
+    constructed_from_t constructed_from = constructed_from_none;
+};
+
+/////////////////////////////////////////////////////////////////////
+// for #3333
+/////////////////////////////////////////////////////////////////////
+
+struct for_3333 final
+{
+    for_3333(int x_ = 0, int y_ = 0) : x(x_), y(y_) {}
+
+    template <class T>
+    for_3333(const T& /*unused*/)
+    {
+        CHECK(false);
+    }
+
+    int x = 0;
+    int y = 0;
+};
+
+template <>
+inline for_3333::for_3333(const json& j)
+    : for_3333(j.value("x", 0), j.value("y", 0))
+{}
+
+/////////////////////////////////////////////////////////////////////
+// for #3810
+/////////////////////////////////////////////////////////////////////
+
+struct Example_3810
+{
+    int bla{};
+
+    Example_3810() = default;
+};
+
+NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE(Example_3810, bla) // NOLINT(misc-use-internal-linkage)
+
+/////////////////////////////////////////////////////////////////////
+// for #4740
+/////////////////////////////////////////////////////////////////////
+
+#ifdef JSON_HAS_CPP_17
+struct Example_4740
+{
+    std::optional<std::string> host = std::nullopt;
+    std::optional<int> port = std::nullopt;
+    NLOHMANN_DEFINE_TYPE_INTRUSIVE_WITH_DEFAULT(Example_4740, host, port)
+};
+#endif
+
+TEST_CASE("regression tests 3")
+{
+#if JSON_HAS_FILESYSTEM || JSON_HAS_EXPERIMENTAL_FILESYSTEM
+    // JSON_HAS_CPP_17 (do not remove; see note at top of file)
+    SECTION("issue #3070 - Version 3.10.3 breaks backward-compatibility with 3.10.2 ")
+    {
+        nlohmann::detail::std_fs::path text_path("/tmp/text.txt");
+        const json j(text_path);
+
+        const auto j_path = j.get<nlohmann::detail::std_fs::path>();
+        CHECK(j_path == text_path);
+
+#if DOCTEST_CLANG || DOCTEST_GCC >= DOCTEST_COMPILER(8, 4, 0)
+        // only known to work on Clang and GCC >=8.4
+        CHECK_THROWS_WITH_AS(nlohmann::detail::std_fs::path(json(1)), "[json.exception.type_error.302] type must be string, but is number", json::type_error);
+#endif
+    }
+#endif
+
+    SECTION("issue #3077 - explicit constructor with default does not compile")
+    {
+        json j;
+        j[0]["value"] = true;
+        std::vector<FooBar> foo;
+        j.get_to(foo);
+    }
+
+    SECTION("issue #3108 - ordered_json doesn't support range based erase")
+    {
+        ordered_json j = {1, 2, 2, 4};
+
+        auto last = std::unique(j.begin(), j.end());
+        j.erase(last, j.end());
+
+        CHECK(j.dump() == "[1,2,4]");
+
+        j.erase(std::remove_if(j.begin(), j.end(), [](const ordered_json & val)
+        {
+            return val == 2;
+        }), j.end());
+
+        CHECK(j.dump() == "[1,4]");
+    }
+
+    SECTION("issue #3343 - json and ordered_json are not interchangeable")
+    {
+        json::object_t jobj({ { "product", "one" } });
+        ordered_json::object_t ojobj({{"product", "one"}});
+
+        auto jit = jobj.begin();
+        auto ojit = ojobj.begin();
+
+        CHECK(jit->first == ojit->first);
+        CHECK(jit->second.get<std::string>() == ojit->second.get<std::string>());
+    }
+
+    SECTION("issue #3171 - if class is_constructible from std::string wrong from_json overload is being selected, compilation failed")
+    {
+        const json j{{ "str", "value"}};
+
+        // failed with: error: no match for ‘operator=’ (operand types are ‘for_3171_derived’ and ‘const nlohmann::basic_json<>::string_t’
+        //                                               {aka ‘const std::__cxx11::basic_string<char>’})
+        //                  s = *j.template get_ptr<const typename BasicJsonType::string_t*>();
+        auto td = j.get<for_3171_derived>();
+
+        CHECK(td.str == "value");
+    }
+
+#ifdef JSON_HAS_CPP_20
+    SECTION("issue #3312 - Parse to custom class from unordered_json breaks on G++11.2.0 with C++20")
+    {
+        // see test for #3171
+        const ordered_json j = {{"name", "class"}};
+        for_3312 obj{};
+
+        j.get_to(obj);
+
+        CHECK(obj.name == "class");
+    }
+#endif
+
+#if defined(JSON_HAS_CPP_17) && JSON_USE_IMPLICIT_CONVERSIONS
+    SECTION("issue #3428 - Error occurred when converting nlohmann::json to std::any")
+    {
+        const json j;
+        const std::any a1 = j;
+        std::any&& a2 = j;
+
+        CHECK(a1.type() == typeid(j));
+        CHECK(a2.type() == typeid(j));
+    }
+#endif
+
+    SECTION("issue #3204 - ambiguous regression")
+    {
+        const for_3204_bar bar_from_foo([](for_3204_foo) noexcept {}); // NOLINT(performance-unnecessary-value-param)
+        const for_3204_bar bar_from_json([](json) noexcept {}); // NOLINT(performance-unnecessary-value-param)
+
+        CHECK(bar_from_foo.constructed_from == for_3204_bar::constructed_from_foo);
+        CHECK(bar_from_json.constructed_from == for_3204_bar::constructed_from_json);
+    }
+
+    SECTION("issue #3333 - Ambiguous conversion from nlohmann::basic_json<> to custom class")
+    {
+        const json j
+        {
+            {"x", 1},
+            {"y", 2}
+        };
+        const for_3333 p = j;
+
+        CHECK(p.x == 1);
+        CHECK(p.y == 2);
+    }
+
+    SECTION("issue #3810 - ordered_json doesn't support construction from C array of custom type")
+    {
+        Example_3810 states[45]; // NOLINT(cppcoreguidelines-avoid-c-arrays,hicpp-avoid-c-arrays,modernize-avoid-c-arrays)
+
+        // fix "not used" warning
+        states[0].bla = 1;
+
+        const auto* const expected = R"([{"bla":1},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0},{"bla":0}])";
+
+        // This works:
+        nlohmann::json j;
+        j["test"] = states;
+        CHECK(j["test"].dump() == expected);
+
+        // This doesn't compile:
+        nlohmann::ordered_json oj;
+        oj["test"] = states;
+        CHECK(oj["test"].dump() == expected);
+    }
+
+#ifdef JSON_HAS_CPP_17
+    SECTION("issue #4740 - build issue with std::optional")
+    {
+        const auto t1 = Example_4740();
+        const auto j1 = nlohmann::json(t1);
+        CHECK(j1.dump() == "{\"host\":null,\"port\":null}");
+        const auto t2 = j1.get<Example_4740>();
+        CHECK(!t2.host.has_value());
+        CHECK(!t2.port.has_value());
+
+        // improve coverage
+        auto t3 = Example_4740();
+        t3.port = 80;
+        t3.host = "example.com";
+        const auto j2 = nlohmann::json(t3);
+        CHECK(j2.dump() == "{\"host\":\"example.com\",\"port\":80}");
+        const auto t4 = j2.get<Example_4740>();
+        CHECK(t4.host.has_value());
+        CHECK(t4.port.has_value());
+    }
+#endif
+
+#if !defined(_MSVC_LANG)
+    // MSVC returns garbage on invalid enum values, so this test is excluded
+    // there.
+    SECTION("issue #4762 - json exception 302 with unhelpful explanation : type must be number, but is number")
+    {
+        // In #4762, the main issue was that a json object with an invalid type
+        // returned "number" as type_name(), because this was the default case.
+        // This test makes sure we now return "invalid" instead.
+        json j;
+        j.m_data.m_type = static_cast<json::value_t>(100); // NOLINT(clang-analyzer-optin.core.EnumCastOutOfRange)
+        CHECK(j.type_name() == "invalid");
+    }
+#endif
+
+#ifdef JSON_HAS_CPP_17
+    SECTION("issue #4804: from_cbor incompatible with std::vector<std::byte> as binary_t")
+    {
+        const std::vector<std::uint8_t> data = {0x80};
+        const auto decoded = json_4804::from_cbor(data);
+        CHECK((decoded == json_4804::array()));
+    }
+
+#ifndef SKIP_TESTS_FOR_ENUM_SERIALIZATION
+    SECTION("discussion #4209 - custom BinaryType direct assignment and round-tripping")
+    {
+        // Test that assigning a custom BinaryType directly creates a binary value, not an array
+        const std::vector<std::byte> original{std::byte{1}, std::byte{2}, std::byte{3}};
+        const json_4804 j = original;
+        CHECK(j.is_binary());
+        CHECK(!j.is_array());
+
+        // Test round-tripping: extracting the binary value back as the custom container type
+        const auto extracted = j.get<std::vector<std::byte>>();
+        CHECK(extracted == original);
+
+        // Test that the default json alias behavior is unchanged: std::vector<uint8_t> -> array
+        const json default_json = std::vector<std::uint8_t> {1, 2, 3};
+        CHECK(default_json.is_array());
+        CHECK(!default_json.is_binary());
+    }
+
+    SECTION("discussion #4209 - custom BinaryType extraction from parsed array")
+    {
+        // Test that extracting a custom BinaryType from a parsed JSON array still works
+        // (not just from a binary-typed node)
+        const auto j = json_4804::parse("[1,2,3]");
+        CHECK(j.is_array());
+        CHECK(!j.is_binary());
+
+        // Extracting as custom BinaryType should work from arrays
+        const auto extracted = j.get<std::vector<std::byte>>();
+        CHECK(extracted.size() == 3);
+        CHECK(extracted[0] == std::byte{1});
+        CHECK(extracted[1] == std::byte{2});
+        CHECK(extracted[2] == std::byte{3});
+    }
+#endif
+
+    SECTION("issue #5046 - implicit conversion of return json to std::optional no longer implicit")
+    {
+        const json jval{};
+        auto GetValue = [](const json & valRoot) -> std::optional<json>
+        {
+            if (valRoot.contains("default"))
+            {
+                return valRoot.at("default");
+            }
+            return std::nullopt;
+        };
+        auto result = GetValue(jval);
+        CHECK(!result.has_value());
+    }
+#endif
+
+#if JSON_HAS_RANGES == 1
+    SECTION("issue #4440 - assert when using std::views::filter and GCC 10")
+    {
+        auto noOpFilter = std::views::filter([](auto&&) noexcept
+        {
+            return true;
+        });
+        json j = {1, 2, 3};
+        auto filtered = j | noOpFilter;
+        CHECK(*filtered.begin() == 1);
+    }
+#endif
+
+#if JSON_HAS_RANGES && !defined(__MINGW32__)
+    SECTION("issue #4916 - constructing array from C++20 ranges view does not work")
+    {
+        std::vector<int> nums{1, 2, 37, 42, 21};
+        auto filteredNums = nums | std::views::filter([](int i)
+        {
+            return i > 10;
+        });
+        json const j(filteredNums);
+        CHECK(j.type() == json::value_t::array);
+        CHECK(j == json({37, 42, 21}));
+    }
+#endif
+
+    // owning_view is not available in libstdc++ < 12
+#if JSON_HAS_RANGES && !defined(__MINGW32__) && !(defined(__GLIBCXX__) && _GLIBCXX_RELEASE < 12)
+    SECTION("issue #4916 - constructing array from prvalue C++20 ranges view (owning_view)")
+    {
+        json const j(std::vector<int> {1, 2, 37, 42, 21} | std::views::filter([](int i)
+        {
+            return i > 10;
+        }));
+        CHECK(j.type() == json::value_t::array);
+        CHECK(j == json({37, 42, 21}));
+    }
+#endif
+
+#if JSON_HAS_RANGES && !defined(__MINGW32__)
+    SECTION("issue #4916 - constructing array from C++20 transform view (prvalue elements)")
+    {
+        std::vector<int> nums{1, 2, 3};
+        auto t = nums | std::views::transform([](int i) noexcept
+        {
+            return i * 2;
+        });
+        json const j(t);
+        CHECK(j.type() == json::value_t::array);
+        CHECK(j == json({2, 4, 6}));
+    }
+#endif
+}
+
+TEST_CASE_TEMPLATE("issue #4798 - nlohmann::json::to_msgpack() encode float NaN as double", T, double, float) // NOLINT(readability-math-missing-parentheses, bugprone-throwing-static-initialization)
+{
+    // With issue #4798, we encode NaN, infinity, and -infinity as float instead
+    // of double to allow for smaller encodings.
+    const json jx = std::numeric_limits<T>::quiet_NaN();
+    const json jy = std::numeric_limits<T>::infinity();
+    const json jz = -std::numeric_limits<T>::infinity();
+
+    /////////////////////////////////////////////////////////////////////////
+    // MessagePack
+    /////////////////////////////////////////////////////////////////////////
+
+    // expected MessagePack values
+    const std::vector<std::uint8_t> msgpack_x = {{0xCA, 0x7F, 0xC0, 0x00, 0x00}};
+    const std::vector<std::uint8_t> msgpack_y = {{0xCA, 0x7F, 0x80, 0x00, 0x00}};
+    const std::vector<std::uint8_t> msgpack_z = {{0xCA, 0xFF, 0x80, 0x00, 0x00}};
+
+    CHECK(json::to_msgpack(jx) == msgpack_x);
+    CHECK(json::to_msgpack(jy) == msgpack_y);
+    CHECK(json::to_msgpack(jz) == msgpack_z);
+
+    CHECK(std::isnan(json::from_msgpack(msgpack_x).get<T>()));
+    CHECK(json::from_msgpack(msgpack_y).get<T>() == std::numeric_limits<T>::infinity());
+    CHECK(json::from_msgpack(msgpack_z).get<T>() == -std::numeric_limits<T>::infinity());
+
+    // Make sure the other MessagePakc encodings for NaN, infinity, and
+    // -infinity are still supported.
+    const std::vector<std::uint8_t> msgpack_x_2 = {{0xCB, 0x7F, 0xF8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
+    const std::vector<std::uint8_t> msgpack_y_2 = {{0xCB, 0x7F, 0xF0, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
+    const std::vector<std::uint8_t> msgpack_z_2 = {{0xCB, 0xFF, 0xF0, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
+    CHECK(std::isnan(json::from_msgpack(msgpack_x_2).get<T>()));
+    CHECK(json::from_msgpack(msgpack_y_2).get<T>() == std::numeric_limits<T>::infinity());
+    CHECK(json::from_msgpack(msgpack_z_2).get<T>() == -std::numeric_limits<T>::infinity());
+
+    /////////////////////////////////////////////////////////////////////////
+    // CBOR
+    /////////////////////////////////////////////////////////////////////////
+
+    // expected CBOR values
+    const std::vector<std::uint8_t> cbor_x = {{0xF9, 0x7E, 0x00}};
+    const std::vector<std::uint8_t> cbor_y = {{0xF9, 0x7C, 0x00}};
+    const std::vector<std::uint8_t> cbor_z = {{0xF9, 0xfC, 0x00}};
+
+    CHECK(json::to_cbor(jx) == cbor_x);
+    CHECK(json::to_cbor(jy) == cbor_y);
+    CHECK(json::to_cbor(jz) == cbor_z);
+
+    CHECK(std::isnan(json::from_cbor(cbor_x).get<T>()));
+    CHECK(json::from_cbor(cbor_y).get<T>() == std::numeric_limits<T>::infinity());
+    CHECK(json::from_cbor(cbor_z).get<T>() == -std::numeric_limits<T>::infinity());
+
+    // Make sure the other CBOR encodings for NaN, infinity, and -infinity are
+    // still supported.
+    const std::vector<std::uint8_t> cbor_x_2 = {{0xFA, 0x7F, 0xC0, 0x00, 0x00}};
+    const std::vector<std::uint8_t> cbor_y_2 = {{0xFA, 0x7F, 0x80, 0x00, 0x00}};
+    const std::vector<std::uint8_t> cbor_z_2 = {{0xFA, 0xFF, 0x80, 0x00, 0x00}};
+    const std::vector<std::uint8_t> cbor_x_3 = {{0xFB, 0x7F, 0xF8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
+    const std::vector<std::uint8_t> cbor_y_3 = {{0xFB, 0x7F, 0xF0, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
+    const std::vector<std::uint8_t> cbor_z_3 = {{0xFB, 0xFF, 0xF0, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}};
+    CHECK(std::isnan(json::from_cbor(cbor_x_2).get<T>()));
+    CHECK(json::from_cbor(cbor_y_2).get<T>() == std::numeric_limits<T>::infinity());
+    CHECK(json::from_cbor(cbor_z_2).get<T>() == -std::numeric_limits<T>::infinity());
+    CHECK(std::isnan(json::from_cbor(cbor_x_3).get<T>()));
+    CHECK(json::from_cbor(cbor_y_3).get<T>() == std::numeric_limits<T>::infinity());
+    CHECK(json::from_cbor(cbor_z_3).get<T>() == -std::numeric_limits<T>::infinity());
+}
+
+TEST_CASE("regression test #5074 - portable workaround for single-element brace init")
+{
+    json const j_obj = {{"key", "value"}};
+
+    json const j = json::array({j_obj});
+    CHECK(j.is_array());
+    CHECK(j.size() == 1);
+    CHECK(j[0] == j_obj);
+}
+
+struct Example_5122
+{
+    float b = 2;
+    nlohmann::ordered_map<std::string, std::string> c{}; // NOLINT(readability-redundant-member-init): needed for GCC -Weffc++
+    int a = 1;
+    NLOHMANN_DEFINE_TYPE_INTRUSIVE_WITH_DEFAULT(Example_5122, b, c, a)
+};
+
+TEST_CASE("regression test #5122 - from_json into types holding nlohmann::ordered_map")
+{
+    Example_5122 src;
+    src.c.emplace("first", "1");
+    src.c.emplace("second", "2");
+
+    ordered_json const j = src;
+    Example_5122 const dst = j.get<Example_5122>();
+
+    CHECK(dst.b == src.b);
+    CHECK(dst.a == src.a);
+    REQUIRE(dst.c.size() == src.c.size());
+    auto src_it = src.c.begin();
+    auto dst_it = dst.c.begin();
+    for (; src_it != src.c.end(); ++src_it, ++dst_it)
+    {
+        CHECK(dst_it->first == src_it->first);
+        CHECK(dst_it->second == src_it->second);
+    }
+}
+
+// -Wself-assign-overloaded was introduced in Clang 7. Gate the pragma on
+// __has_warning so older Clang versions do not error with "unknown warning
+// group". The __has_warning check has to stay inside the __clang__ branch
+// because GCC does not provide it and would tokenize-error on the argument.
+#if defined(__clang__) && defined(__has_warning)
+    #if __has_warning("-Wself-assign-overloaded")
+        DOCTEST_CLANG_SUPPRESS_WARNING_PUSH
+        DOCTEST_CLANG_SUPPRESS_WARNING("-Wself-assign-overloaded")
+    #endif
+#endif
+
+TEST_CASE("regression test #5122 - nlohmann::ordered_map copy-assignment is self-assignment safe")
+{
+    nlohmann::ordered_map<std::string, std::string> m;
+    m.emplace("first", "1");
+    m.emplace("second", "2");
+
+    // Insertion order is preserved by ordered_map, so we can check it directly.
+    m = m;
+
+    REQUIRE(m.size() == 2);
+    auto it = m.begin();
+    CHECK(it->first == "first");
+    CHECK(it->second == "1");
+    ++it;
+    CHECK(it->first == "second");
+    CHECK(it->second == "2");
+}
+
+#if defined(__clang__) && defined(__has_warning)
+    #if __has_warning("-Wself-assign-overloaded")
+        DOCTEST_CLANG_SUPPRESS_WARNING_POP
+    #endif
+#endif
+
+TEST_CASE("regression test #5122 - nlohmann::ordered_map move-assignment transfers contents")
+{
+    nlohmann::ordered_map<std::string, std::string> src;
+    src.emplace("first", "1");
+    src.emplace("second", "2");
+
+    nlohmann::ordered_map<std::string, std::string> dst;
+    dst.emplace("stale", "x");
+    dst = std::move(src);
+
+    REQUIRE(dst.size() == 2);
+    auto it = dst.begin();
+    CHECK(it->first == "first");
+    CHECK(it->second == "1");
+    ++it;
+    CHECK(it->first == "second");
+    CHECK(it->second == "2");
+
+    // Re-assigning into the moved-from object must leave it in a usable state.
+    src = nlohmann::ordered_map<std::string, std::string> {};
+    src.emplace("after-move", "3");
+    REQUIRE(src.size() == 1);
+    CHECK(src.begin()->first == "after-move");
+}
+
+// Stand-in for a third-party library (e.g., Eigen as of 3.4, which added
+// STL-compatible begin()/end() to its vector types), living in its own
+// namespace with its own to_json overload for its vector type.
+namespace issue_4320_eigen
+{
+// "array-compatible" from the library's point of view (it has begin()/end()),
+// but for which this (fake) third-party namespace provides its own to_json.
+struct vector3
+{
+    double v[3]; // NOLINT(cppcoreguidelines-avoid-c-arrays,hicpp-avoid-c-arrays,modernize-avoid-c-arrays,cppcoreguidelines-use-default-member-init,modernize-use-default-member-init)
+    vector3(double x, double y, double z) : v{x, y, z} {} // NOLINT(hicpp-member-init,cppcoreguidelines-pro-type-member-init)
+    double x() const
+    {
+        return v[0];
+    }
+    double y() const
+    {
+        return v[1];
+    }
+    double z() const
+    {
+        return v[2];
+    }
+    double* begin()
+    {
+        return v;
+    }
+    double* end()
+    {
+        return v + 3;
+    }
+    const double* begin() const
+    {
+        return v;
+    }
+    const double* end() const
+    {
+        return v + 3;
+    }
+};
+
+inline void to_json(json& j, const vector3& v) // NOLINT(misc-use-internal-linkage)
+{
+    j = {{"x", v.x()}, {"y", v.y()}, {"z", v.z()}};
+}
+} // namespace issue_4320_eigen
+
+// The user's own namespace, using the (fake) Eigen type as an implementation
+// detail behind a payload type that has nothing to do with vectors/arrays.
+namespace issue_4320
+{
+// Publicly derives from issue_4320_eigen::vector3 but does *not* define its
+// own to_json - it is only ever used as a temporary to reach the base
+// class's to_json via ADL.
+struct vector3_wrapper : issue_4320_eigen::vector3
+{
+    using issue_4320_eigen::vector3::vector3;
+};
+
+struct payload
+{
+    double x, y, z;
+};
+
+inline vector3_wrapper to_eigen(const payload& p) // NOLINT(misc-use-internal-linkage)
+{
+    return {p.x, p.y, p.z};
+}
+
+inline void to_json(json& j, const payload& p) // NOLINT(misc-use-internal-linkage)
+{
+    // Unqualified call, passing a *derived* vector3_wrapper: relies on ADL
+    // finding issue_4320_eigen::to_json(json&, const vector3&) through the
+    // vector3 base class, via a derived-to-base conversion. Must NOT resolve
+    // to the library's own generic array-compatible to_json (an exact-match
+    // template for vector3_wrapper, since it also has begin()/end()), which
+    // would serialize this as [x, y, z] instead of {"x":x, "y":y, "z":z}.
+    to_json(j, to_eigen(p));
+}
+} // namespace issue_4320
+
+TEST_CASE("issue #4320 - custom base class must not leak nlohmann::detail into ADL")
+{
+    // Before the fix, basic_json unconditionally derived from a type living in
+    // nlohmann::detail (json_default_base), which made nlohmann::detail an
+    // associated namespace of every basic_json for ADL purposes. That leaked
+    // the library's internal generic-array to_json overload into unqualified
+    // to_json() calls made from user code, silently bypassing user-defined
+    // to_json overloads reached via a derived-to-base conversion.
+    const issue_4320::payload p{1.0, 2.0, 3.0};
+
+    json j;
+    to_json(j, p);
+    CHECK(j == json({{"x", 1.0}, {"y", 2.0}, {"z", 3.0}}));
+}
+
+TEST_CASE("issue #5338 - truncated CBOR tagged binary subtype is rejected")
+{
+    const std::vector<std::vector<std::uint8_t>> truncated_tags =
+    {
+        {0xD8},
+        {0xD9, 0x00},
+        {0xDA, 0x00, 0x00, 0x00},
+        {0xDB, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}
+    };
+
+    for (const auto& data : truncated_tags)
+    {
+        CAPTURE(data);
+        for (const auto tag_handler :
+                {
+                    json::cbor_tag_handler_t::ignore, json::cbor_tag_handler_t::store
+                })
+        {
+            CAPTURE(tag_handler);
+            const auto result = json::from_cbor(data, true, false, tag_handler);
+            CHECK(result.is_discarded());
+        }
+    }
+}
+
+TEST_CASE("issue #5402 - update(merge_objects=true) overwrites a primitive with an object")
+{
+    json t = {{"k", 1}};
+    t.update(json{{"k", {{"x", 2}}}}, true);
+    CHECK(t == json({{"k", {{"x", 2}}}}));
+
+    json mixed = {{"keep", {{"a", 1}}}, {"replace", 1}};
+    mixed.update(json{{"keep", {{"b", 2}}}, {"replace", {{"x", 2}}}}, true);
+    CHECK(mixed == json({{"keep", {{"a", 1}, {"b", 2}}}, {"replace", {{"x", 2}}}}));
+}
+
+
+TEST_CASE("regression test #5476 - array type without reserve()")
+{
+    // the capacity reserved for definite-length arrays must not require the
+    // array type to have a reserve() member function
+    using deque_json = nlohmann::basic_json<std::map, std::deque>;
+
+    SECTION("std::deque")
+    {
+        const auto j = deque_json::parse(R"({"a":[1,[2,3]],"b":[]})");
+        CHECK(j.dump() == R"({"a":[1,[2,3]],"b":[]})");
+
+        // the binary formats pass a definite length to start_array()
+        CHECK(deque_json::from_cbor(deque_json::to_cbor(j)) == j);
+        CHECK(deque_json::from_msgpack(deque_json::to_msgpack(j)) == j);
+        CHECK(deque_json::from_bon8(deque_json::to_bon8(j)) == j);
+
+        // parse() instantiates the callback parser as well, which reserves too
+        const auto with_callback = deque_json::parse(R"([1,2,3])", [](int /*depth*/, deque_json::parse_event_t /*event*/, deque_json& /*parsed*/) noexcept
+        {
+            return true;
+        });
+        CHECK(with_callback == deque_json({1, 2, 3}));
+    }
+
+    SECTION("std::vector still reserves")
+    {
+        json array = json::array();
+        for (int i = 0; i < 100; ++i)
+        {
+            array.push_back(i);
+        }
+
+        const auto j = json::from_cbor(json::to_cbor(array));
+        CHECK(j == array);
+        CHECK(j.get_ref<const json::array_t&>().capacity() >= 100);
+    }
+
+    SECTION("the reservation stays capped")
+    {
+        // CBOR array announcing 2^32-1 elements, but truncated right after the
+        // header: the input must be rejected without reserving that capacity
+        const std::vector<std::uint8_t> truncated = {0x9A, 0xFF, 0xFF, 0xFF, 0xFF};
+        CHECK(json::from_cbor(truncated, true, false).is_discarded());
+    }
+}
+
+DOCTEST_CLANG_SUPPRESS_WARNING_POP

tests/src/unit-serialization.cpp

@@ -15,6 +15,8 @@ using nlohmann::json;
 #include <sstream>
 #include <iomanip>
 
+#include "test_utils.hpp"
+
 TEST_CASE("serialization")
 {
     SECTION("operator<<")
@@ -84,19 +86,31 @@ TEST_CASE("serialization")
         {
             const json j = "ä\xA9ü";
 
-            CHECK_THROWS_WITH_AS(j.dump(), "[json.exception.type_error.316] invalid UTF-8 byte at index 2: 0xA9", json::type_error&);
-            CHECK_THROWS_WITH_AS(j.dump(1, ' ', false, json::error_handler_t::strict), "[json.exception.type_error.316] invalid UTF-8 byte at index 2: 0xA9", json::type_error&);
+            // dump() is nodiscard; the exception is thrown by dump() itself before it would return
+            CHECK_THROWS_WITH_AS(utils::ignore_return_value(j.dump()), "[json.exception.type_error.316] invalid UTF-8 byte at index 2: 0xA9", json::type_error&);
+            CHECK_THROWS_WITH_AS(utils::ignore_return_value(j.dump(1, ' ', false, json::error_handler_t::strict)), "[json.exception.type_error.316] invalid UTF-8 byte at index 2: 0xA9", json::type_error&);
             CHECK(j.dump(-1, ' ', false, json::error_handler_t::ignore) == "\"äü\"");
             CHECK(j.dump(-1, ' ', false, json::error_handler_t::replace) == "\"ä\xEF\xBF\xBDü\"");
             CHECK(j.dump(-1, ' ', true, json::error_handler_t::replace) == "\"\\u00e4\\ufffd\\u00fc\"");
         }
 
+        SECTION("invalid character (regression guard for shared UTF-8 decoder, see #5529)")
+        {
+            // dump_escaped_impl() now calls the UTF-8 decoder shared with the
+            // binary readers (detail::decode() in string_utils.hpp) instead
+            // of a private copy; the exact type_error.316 message/behavior
+            // must stay byte-for-byte the same as before that extraction
+            const json j = "ä\xA9ü";
+            CHECK_THROWS_WITH_AS(utils::ignore_return_value(j.dump()), "[json.exception.type_error.316] invalid UTF-8 byte at index 2: 0xA9", json::type_error&);
+        }
+
         SECTION("ending with incomplete character")
         {
             const json j = "123\xC2";
 
-            CHECK_THROWS_WITH_AS(j.dump(), "[json.exception.type_error.316] incomplete UTF-8 string; last byte: 0xC2", json::type_error&);
-            CHECK_THROWS_AS(j.dump(1, ' ', false, json::error_handler_t::strict), json::type_error&);
+            // dump() is nodiscard; the exception is thrown by dump() itself before it would return
+            CHECK_THROWS_WITH_AS(utils::ignore_return_value(j.dump()), "[json.exception.type_error.316] incomplete UTF-8 string; last byte: 0xC2", json::type_error&);
+            CHECK_THROWS_AS(utils::ignore_return_value(j.dump(1, ' ', false, json::error_handler_t::strict)), json::type_error&);
             CHECK(j.dump(-1, ' ', false, json::error_handler_t::ignore) == "\"123\"");
             CHECK(j.dump(-1, ' ', false, json::error_handler_t::replace) == "\"123\xEF\xBF\xBD\"");
             CHECK(j.dump(-1, ' ', true, json::error_handler_t::replace) == "\"123\\ufffd\"");
@@ -106,8 +120,9 @@ TEST_CASE("serialization")
         {
             const json j = "123\xF1\xB0\x34\x35\x36";
 
-            CHECK_THROWS_WITH_AS(j.dump(), "[json.exception.type_error.316] invalid UTF-8 byte at index 5: 0x34", json::type_error&);
-            CHECK_THROWS_AS(j.dump(1, ' ', false, json::error_handler_t::strict), json::type_error&);
+            // dump() is nodiscard; the exception is thrown by dump() itself before it would return
+            CHECK_THROWS_WITH_AS(utils::ignore_return_value(j.dump()), "[json.exception.type_error.316] invalid UTF-8 byte at index 5: 0x34", json::type_error&);
+            CHECK_THROWS_AS(utils::ignore_return_value(j.dump(1, ' ', false, json::error_handler_t::strict)), json::type_error&);
             CHECK(j.dump(-1, ' ', false, json::error_handler_t::ignore) == "\"123456\"");
             CHECK(j.dump(-1, ' ', false, json::error_handler_t::replace) == "\"123\xEF\xBF\xBD\x34\x35\x36\"");
             CHECK(j.dump(-1, ' ', true, json::error_handler_t::replace) == "\"123\\ufffd456\"");
@@ -382,3 +397,405 @@ TEST_CASE("dump for basic_json with long double number_float_t")
         check_same(100.0L,  100.0);
     }
 }
+
+TEST_CASE("serialization of strings (bulk fast path)")
+{
+    // These cases exercise the SWAR bulk-copy fast path in dump_escaped and the
+    // internal write buffer: long runs, escapes interrupting runs, 0x7F/DEL,
+    // multibyte UTF-8 under both ensure_ascii settings, and payloads larger than
+    // the write buffer.
+
+    SECTION("long unescaped ASCII exceeds the write buffer")
+    {
+        const std::string big(3000, 'a');
+        const json j = big;
+        CHECK(j.dump() == '"' + big + '"');
+        CHECK(j.dump(-1, ' ', true) == '"' + big + '"');
+        // round-trips
+        CHECK(json::parse(j.dump()) == j);
+    }
+
+    SECTION("runs interrupted by escapes")
+    {
+        const json j = std::string(500, 'x') + "\n\"\\" + std::string(500, 'y');
+        const std::string out = j.dump();
+        CHECK(out == '"' + std::string(500, 'x') + "\\n\\\"\\\\" + std::string(500, 'y') + '"');
+        CHECK(json::parse(out) == j);
+    }
+
+    SECTION("DEL (0x7F) depends on ensure_ascii")
+    {
+        const json j = std::string("a\x7f" "b");
+        CHECK(j.dump(-1, ' ', false) == "\"a\x7f" "b\"");   // copied verbatim
+        CHECK(j.dump(-1, ' ', true) == "\"a\\u007fb\"");    // escaped
+    }
+
+    SECTION("multibyte UTF-8 under both ensure_ascii settings")
+    {
+        const json j = std::string("A\xc3\xa9\xe4\xbd\xa0\xf0\x9f\x98\x80Z"); // A é 你 😀 Z
+        // not escaping non-ASCII: bytes are copied through the bulk validator
+        CHECK(j.dump(-1, ' ', false) == "\"A\xc3\xa9\xe4\xbd\xa0\xf0\x9f\x98\x80Z\"");
+        // ensure_ascii: escaped (with a surrogate pair for the emoji)
+        CHECK(j.dump(-1, ' ', true) == "\"A\\u00e9\\u4f60\\ud83d\\ude00Z\"");
+        CHECK(json::parse(j.dump(-1, ' ', true)) == j);
+    }
+
+    SECTION("many small structural writes exceed the write buffer")
+    {
+        json arr = json::array();
+        for (int i = 0; i < 2000; ++i)
+        {
+            arr.push_back(i);
+        }
+        const std::string out = arr.dump();
+        CHECK(out.front() == '[');
+        CHECK(out.back() == ']');
+        CHECK(json::parse(out) == arr);
+
+        json obj = json::object();
+        for (int i = 0; i < 500; ++i)
+        {
+            obj["key" + std::to_string(i)] = i;
+        }
+        CHECK(json::parse(obj.dump()) == obj);
+        CHECK(json::parse(obj.dump(2)) == obj);
+
+        // an array of many empty strings emits a long run of single-character
+        // writes ('"', '"', ',') at shallow nesting depth, so the write buffer
+        // fills and flushes mid-run without the deep recursion that would
+        // overflow the stack on some debug builds
+        json many_empty = json::array();
+        for (int i = 0; i < 500; ++i)
+        {
+            many_empty.push_back("");
+        }
+        const std::string out2 = many_empty.dump();
+        CHECK(out2.size() > 1024); // spans multiple write-buffer flushes
+        CHECK(out2.front() == '[');
+        CHECK(out2.back() == ']');
+        CHECK(json::parse(out2) == many_empty);
+    }
+
+    SECTION("invalid UTF-8 handling is unaffected by the fast path")
+    {
+        const json j = std::string("valid\xff" "more");
+        CHECK_THROWS_WITH_AS(utils::ignore_return_value(j.dump()), "[json.exception.type_error.316] invalid UTF-8 byte at index 5: 0xFF", json::type_error&);
+        CHECK(j.dump(-1, ' ', false, json::error_handler_t::replace) == "\"valid\xef\xbf\xbd" "more\"");
+        CHECK(j.dump(-1, ' ', true, json::error_handler_t::replace) == "\"valid\\ufffdmore\"");
+        CHECK(j.dump(-1, ' ', false, json::error_handler_t::ignore) == "\"validmore\"");
+    }
+}
+
+TEST_CASE("indentation is written straight into the write buffer")
+{
+    // put_indent() memsets the indentation into the write buffer instead of
+    // copying it out of a pre-grown indentation string. These cases cover an
+    // indentation wider than the buffer, a non-space indentation character, and
+    // nesting deep enough that the accumulated indentation spans several
+    // buffer-fulls - the situations the old grow-a-string approach got wrong.
+
+    SECTION("indent_step wider than the write buffer")
+    {
+        const json j = {{"a", 1}};
+        // 2000 > the 1024-byte write buffer, and > the 512 the indentation
+        // string used to start at
+        CHECK(j.dump(2000) == "{\n" + std::string(2000, ' ') + "\"a\": 1\n}");
+        // several whole buffer-fulls, so the buffer is refilled once and then
+        // flushed repeatedly
+        CHECK(j.dump(5000) == "{\n" + std::string(5000, ' ') + "\"a\": 1\n}");
+        CHECK(j.dump(5000, '\t') == "{\n" + std::string(5000, '\t') + "\"a\": 1\n}");
+        // an exact multiple of the buffer size
+        CHECK(j.dump(4096) == "{\n" + std::string(4096, ' ') + "\"a\": 1\n}");
+    }
+
+    SECTION("a non-space indentation character is used throughout")
+    {
+        const json j = {{"a", 1}};
+        // 600 is past the point where the indentation used to be grown, which
+        // is where a hard-coded space would have shown up
+        CHECK(j.dump(600, '\t') == "{\n" + std::string(600, '\t') + "\"a\": 1\n}");
+        CHECK(j.dump(3, '.') == "{\n...\"a\": 1\n}");
+    }
+
+    SECTION("accumulated indentation spans several buffer-fulls")
+    {
+        // five levels deep at 400 per level: the innermost value is indented by
+        // 2000 characters, reached in steps that each straddle the buffer end
+        json j = json::array({1});
+        for (int i = 0; i < 4; ++i)
+        {
+            j = json::array({j});
+        }
+
+        const std::string out = j.dump(400);
+        CHECK(out.find(std::string("\n") + std::string(2000, ' ') + "1\n") != std::string::npos);
+        CHECK(json::parse(out) == j);
+    }
+
+    SECTION("binary values are indented the same way")
+    {
+        // a binary value is serialized as an object with "bytes" and
+        // "subtype" keys; the byte array itself is always written compactly
+        // (see dump_byte()), so only the surrounding object's indentation
+        // goes through put_indent()
+        const json j = json::binary({1, 2, 3}, 128);
+        CHECK(j.dump(2000) == "{\n" + std::string(2000, ' ') + "\"bytes\": [1, 2, 3],\n"
+              + std::string(2000, ' ') + "\"subtype\": 128\n}");
+        CHECK(j.dump(2000, '\t') == "{\n" + std::string(2000, '\t') + "\"bytes\": [1, 2, 3],\n"
+              + std::string(2000, '\t') + "\"subtype\": 128\n}");
+    }
+
+    SECTION("indentation is unchanged for ordinary widths")
+    {
+        const json j = {{"a", {1, 2}}, {"b", nullptr}};
+        CHECK(j.dump(2) == "{\n  \"a\": [\n    1,\n    2\n  ],\n  \"b\": null\n}");
+        CHECK(j.dump(0) == "{\n\"a\": [\n1,\n2\n],\n\"b\": null\n}");
+    }
+}
+
+TEST_CASE("serialization of deeply nested values")
+{
+    // dump() descends into a bounded number of levels and writes out whatever
+    // is nested deeper than that without the call stack; see
+    // https://github.com/nlohmann/json/issues/5387
+
+    SECTION("nested deeper than the call stack could follow")
+    {
+        // parsing is iterative, so building these costs little
+        const std::size_t depth = 100000;
+
+        const std::string array_text = std::string(depth, '[') + '0' + std::string(depth, ']');
+        CHECK(json::parse(array_text).dump() == array_text);
+
+        std::string object_text;
+        object_text.reserve((6 * depth) + 1);
+        for (std::size_t i = 0; i < depth; ++i)
+        {
+            object_text += "{\"a\":";
+        }
+        object_text += '1';
+        object_text.append(depth, '}');
+        CHECK(json::parse(object_text).dump() == object_text);
+    }
+
+    SECTION("depths around the bound of the descent")
+    {
+        // Cover every depth around the bound, so that the two ways of writing a
+        // value are known to meet cleanly - wherever the bound is set.
+        for (std::size_t d = 1; d <= 300; ++d)
+        {
+            CAPTURE(d);
+
+            const std::string array_text = std::string(d, '[') + '7' + std::string(d, ']');
+            CHECK(json::parse(array_text).dump() == array_text);
+
+            std::string object_text;
+            for (std::size_t i = 0; i < d; ++i)
+            {
+                object_text += "{\"k\":";
+            }
+            object_text += '7';
+            object_text.append(d, '}');
+            CHECK(json::parse(object_text).dump() == object_text);
+        }
+    }
+
+    SECTION("pretty-printing across the bound")
+    {
+        for (std::size_t d = 120; d <= 140; ++d)
+        {
+            CAPTURE(d);
+
+            const json j = json::parse(std::string(d, '[') + '7' + std::string(d, ']'));
+
+            std::string expected;
+            for (std::size_t i = 0; i < d; ++i)
+            {
+                expected += std::string(2 * i, ' ') + "[\n";
+            }
+            expected += std::string(2 * d, ' ') + '7';
+            for (std::size_t i = d; i > 0; --i)
+            {
+                expected += '\n' + std::string(2 * (i - 1), ' ') + ']';
+            }
+
+            CHECK(j.dump(2) == expected);
+        }
+    }
+
+    SECTION("an empty container below the bound")
+    {
+        // an empty container is written out in full and never descended into,
+        // so it must not gain a newline when it is reached iteratively
+        for (std::size_t d = 125; d <= 135; ++d)
+        {
+            CAPTURE(d);
+
+            const std::string compact = std::string(d, '[') + "[]" + std::string(d, ']');
+            CHECK(json::parse(compact).dump() == compact);
+
+            const std::string with_object = std::string(d, '[') + "{}" + std::string(d, ']');
+            CHECK(json::parse(with_object).dump() == with_object);
+        }
+    }
+}
+
+namespace
+{
+// wraps @a inner into @a depth single-element arrays
+json wrap_in_arrays(const json& inner, const std::size_t depth)
+{
+    json j = inner;
+    for (std::size_t i = 0; i < depth; ++i)
+    {
+        j = json::array({std::move(j)});
+    }
+    return j;
+}
+
+// what wrap_in_arrays(inner, depth).dump(2) is expected to be: the arrays
+// around inner.dump(2), with inner's own lines indented by the depth
+std::string expected_pretty_in_arrays(const json& inner, const std::size_t depth)
+{
+    std::string expected;
+    for (std::size_t i = 0; i < depth; ++i)
+    {
+        expected += std::string(2 * i, ' ') + "[\n";
+    }
+
+    const std::string indent(2 * depth, ' ');
+    expected += indent;
+    for (const char c : inner.dump(2))
+    {
+        expected += c;
+        if (c == '\n')
+        {
+            expected += indent;
+        }
+    }
+
+    for (std::size_t i = depth; i > 0; --i)
+    {
+        expected += '\n' + std::string(2 * (i - 1), ' ') + ']';
+    }
+    return expected;
+}
+} // namespace
+
+TEST_CASE("serialization of every kind of value below the bound of the descent")
+{
+    // Values nested deeper than the bound are written without the call stack,
+    // by code of their own; each kind of value must come out the same there as
+    // it does at the top level, compact and pretty-printed.
+    std::vector<json> values =
+    {
+        json::parse(R"({"a": 1, "b": [1, 2, {"c": "x"}], "d": {}, "e": []})"),
+        json::parse(R"([1, [2, 3], {"k": null}, "s"])"),
+        json::object(),
+        json::array(),
+        json::binary({1, 2, 3}, 42),
+        json::binary({1, 2, 3}),
+        json::binary({}, 7),
+        json::binary({}),
+        "a string with \"escapes\"\n",
+        true,
+        false,
+        -42,
+        42u,
+        1.5,
+        nullptr,
+        json(json::value_t::discarded),
+    };
+    // a pretty-printed object whose members are themselves deep
+    values.push_back({{"x", wrap_in_arrays(1, 5)}, {"y", {{"z", 2}}}});
+
+    for (const std::size_t depth : std::vector<std::size_t> {1, 200})
+    {
+        CAPTURE(depth);
+        for (const auto& inner : values)
+        {
+            CAPTURE(inner.dump());
+            const json j = wrap_in_arrays(inner, depth);
+            CHECK(j.dump() == std::string(depth, '[') + inner.dump() + std::string(depth, ']'));
+            CHECK(j.dump(2) == expected_pretty_in_arrays(inner, depth));
+        }
+    }
+
+    SECTION("pretty-printed objects across the bound")
+    {
+        for (std::size_t d = 120; d <= 140; ++d)
+        {
+            CAPTURE(d);
+
+            // built from the inside out: {"k": <level below>, "n": <level>}
+            json j = 7;
+            std::string expected = "7";
+            for (std::size_t i = d; i > 0; --i)
+            {
+                j = json({{"k", std::move(j)}, {"n", i}});
+
+                const std::string indent(2 * i, ' ');
+                const std::string outer_indent(2 * (i - 1), ' ');
+                std::string next = "{\n";
+                next += indent;
+                next += "\"k\": ";
+                next += expected;
+                next += ",\n";
+                next += indent;
+                next += "\"n\": ";
+                next += std::to_string(i);
+                next += '\n';
+                next += outer_indent;
+                next += '}';
+                expected = std::move(next);
+            }
+
+            CHECK(j.dump(2) == expected);
+            CHECK(json::parse(j.dump(2)) == j);
+            CHECK(json::parse(j.dump()) == j);
+        }
+    }
+}
+
+TEST_CASE("serializer buffers are flushed mid-string and mid-binary")
+{
+    SECTION("a long run of escaped characters")
+    {
+        // each character is escaped on its own, so the escape buffer fills up
+        const json newlines = std::string(600, '\n');
+        std::string expected = "\"";
+        for (int i = 0; i < 600; ++i)
+        {
+            expected += "\\n";
+        }
+        expected += '"';
+        CHECK(newlines.dump() == expected);
+
+        // every character is \u-escaped under ensure_ascii
+        std::string umlauts;
+        std::string escaped_umlauts = "\"";
+        for (int i = 0; i < 300; ++i)
+        {
+            umlauts += "\xC3\xA4";
+            escaped_umlauts += "\\u00e4";
+        }
+        escaped_umlauts += '"';
+        CHECK(json(umlauts).dump(-1, ' ', true) == escaped_umlauts);
+    }
+
+    SECTION("a large binary value")
+    {
+        std::vector<std::uint8_t> bytes(3000);
+        std::string expected_bytes;
+        std::string expected_pretty_bytes;
+        for (std::size_t i = 0; i < bytes.size(); ++i)
+        {
+            bytes[i] = static_cast<std::uint8_t>(i % 256);
+            expected_bytes += (i == 0 ? "" : ",") + std::to_string(i % 256);
+            expected_pretty_bytes += (i == 0 ? "" : ", ") + std::to_string(i % 256);
+        }
+        const json j = json::binary(bytes);
+        CHECK(j.dump() == "{\"bytes\":[" + expected_bytes + "],\"subtype\":null}");
+        CHECK(j.dump(2) == "{\n  \"bytes\": [" + expected_pretty_bytes + "],\n  \"subtype\": null\n}");
+    }
+}

tests/src/unit-std-format.cpp

@@ -17,6 +17,7 @@
 
 #include <nlohmann/json.hpp>
 using json = nlohmann::json;
+using ordered_json = nlohmann::ordered_json;
 
 // JSON_HAS_CPP_20 (do not remove; see note at top of file)
 #if JSON_HAS_STD_FORMAT
@@ -52,6 +53,23 @@ TEST_CASE("std::formatter<nlohmann::json>")
         CHECK(std::format("{:2}", j) == j.dump(2));
         CHECK(std::format("{:#2}", j) == j.dump(2));
         CHECK(std::format("{:8}", j) == j.dump(8));
+        // multi-digit widths must accumulate every digit, not just the first
+        CHECK(std::format("{:12}", j) == j.dump(12));
+        CHECK(std::format("{:#12}", j) == j.dump(12));
+        CHECK(std::format("{:10}", j) == j.dump(10));
+    }
+
+    SECTION("bare alignment with no fill character defaults to a space indent character")
+    {
+        const json j = {{"foo", 1}, {"bar", {1, 2, 3}}};
+        // without a preceding fill character, the alignment character itself must not
+        // be mistaken for the indent character -- the default space is kept
+        CHECK(std::format("{:<}", j) == j.dump());
+        CHECK(std::format("{:>}", j) == j.dump());
+        CHECK(std::format("{:^}", j) == j.dump());
+        CHECK(std::format("{:<3}", j) == j.dump(3, ' '));
+        CHECK(std::format("{:>3}", j) == j.dump(3, ' '));
+        CHECK(std::format("{:^3}", j) == j.dump(3, ' '));
     }
 
     SECTION("fill-and-align sets the indent character, like dump(indent, indent_char)")
@@ -84,6 +102,29 @@ TEST_CASE("std::formatter<nlohmann::json>")
         CHECK_THROWS_AS(std::vformat("{:{}}", std::make_format_args(j, dynamic_width)), std::format_error); // dynamic width
     }
 
+    SECTION("a format spec may run to the end of the parse context")
+    {
+        // std::format always hands parse() a range that still holds the closing
+        // '}', but a parse context may also end right after the spec
+        const auto parse = [](const char* spec)
+        {
+            std::format_parse_context ctx(spec);
+            std::formatter<json> f;
+            CHECK(f.parse(ctx) == ctx.end());
+            return f;
+        };
+
+        CHECK(parse("").indent == -1);
+        CHECK(parse(">").indent == -1);
+        CHECK(parse("#").indent == 4);
+        CHECK(parse("3").indent == 3);
+        CHECK(parse("#12").indent == 12);
+
+        const auto f = parse(".>");
+        CHECK(f.indent == -1);
+        CHECK(f.indent_char == '.');
+    }
+
     SECTION("std::format_to writes through an arbitrary output iterator")
     {
         const json j = {{"foo", 1}, {"bar", {1, 2, 3}}};
@@ -93,4 +134,16 @@ TEST_CASE("std::formatter<nlohmann::json>")
     }
 }
 
+TEST_CASE("std::formatter<nlohmann::ordered_json>")
+{
+    // spot-check a non-default basic_json instantiation, since the formatter
+    // is written against the generic NLOHMANN_BASIC_JSON_TPL_DECLARATION
+    // template and must actually instantiate (and behave correctly) for
+    // template arguments other than nlohmann::json
+    const ordered_json j = {{"foo", 1}, {"bar", {1, 2, 3}}};
+    CHECK(std::format("{}", j) == j.dump());
+    CHECK(std::format("{:#}", j) == j.dump(4));
+    CHECK(std::format("{:2}", j) == j.dump(2));
+}
+
 #endif

tests/src/unit-ubjson.cpp

@@ -15,6 +15,7 @@ using nlohmann::json;
 #include <fstream>
 #include <set>
 #include "make_test_data_available.hpp"
+#include "round_trip_corpus.hpp"
 #include "test_utils.hpp"
 
 namespace
@@ -264,7 +265,7 @@ TEST_CASE("UBJSON")
 
                 SECTION("-32768..-129 (int16)")
                 {
-                    for (int32_t i = -32768; i <= -129; ++i)
+                    for (int32_t i = -32768; i <= -129; i = utils::next_integer_sample(i, -129, 7))
                     {
                         CAPTURE(i)
 
@@ -424,7 +425,7 @@ TEST_CASE("UBJSON")
 
                 SECTION("256..32767 (int16)")
                 {
-                    for (size_t i = 256; i <= 32767; ++i)
+                    for (size_t i = 256; i <= 32767; i = utils::next_integer_sample(i, static_cast<size_t>(32767), static_cast<size_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -630,7 +631,7 @@ TEST_CASE("UBJSON")
 
                 SECTION("256..32767 (int16)")
                 {
-                    for (size_t i = 256; i <= 32767; ++i)
+                    for (size_t i = 256; i <= 32767; i = utils::next_integer_sample(i, static_cast<size_t>(32767), static_cast<size_t>(7)))
                     {
                         CAPTURE(i)
 
@@ -1639,6 +1640,29 @@ TEST_CASE("UBJSON")
                 });
                 CHECK_THROWS_AS(_ = json::sax_parse(v_ubjson, &scp, json::input_format_t::ubjson), json::out_of_range&);
             }
+
+            SECTION("array with a known size, read with a callback")
+            {
+                // a sized array announces its length to start_array()
+                std::vector<uint8_t> const v_ubjson = {'[', '#', 'i', 2, 'i', 1, 'i', 2};
+                json j;
+                nlohmann::detail::json_sax_dom_callback_parser<json, decltype(nlohmann::detail::input_adapter(v_ubjson))> scp(j, [](int /*unused*/, json::parse_event_t /*unused*/, const json& /*unused*/) noexcept
+                {
+                    return true;
+                });
+                CHECK(json::sax_parse(v_ubjson, &scp, json::input_format_t::ubjson));
+                CHECK(j == json({1, 2}));
+
+                // the readers reject a size this large before they announce
+                // it, so it can only reach start_array() directly (the largest
+                // value stands for an unknown size and is never checked)
+                json k;
+                nlohmann::detail::json_sax_dom_callback_parser<json, decltype(nlohmann::detail::input_adapter(v_ubjson))> scp2(k, [](int /*unused*/, json::parse_event_t /*unused*/, const json& /*unused*/) noexcept
+                {
+                    return true;
+                });
+                CHECK_THROWS_AS(scp2.start_array((std::numeric_limits<std::size_t>::max)() - 1), json::out_of_range&);
+            }
         }
     }
 
@@ -1713,6 +1737,44 @@ TEST_CASE("UBJSON")
             CHECK(json::to_ubjson(json::from_ubjson(s_L)) == s_i);
         }
 
+        SECTION("no-op markers")
+        {
+            // A no-op ('N') is valid wherever a value may start; it is consumed
+            // by get_ignore_noop() before the value is read. It is not valid
+            // where a string length type specification is expected.
+
+            SECTION("accepted where a value may start")
+            {
+                // at top level, also repeated
+                CHECK(json::from_ubjson(std::vector<uint8_t>({'N', 'i', 1})) == json(1));
+                CHECK(json::from_ubjson(std::vector<uint8_t>({'N', 'N', 'N', 'i', 1})) == json(1));
+
+                // inside an array of unknown size, before and after an element
+                CHECK(json::from_ubjson(std::vector<uint8_t>({'[', 'N', 'i', 1, ']'})) == json({1}));
+                CHECK(json::from_ubjson(std::vector<uint8_t>({'[', 'i', 1, 'N', ']'})) == json({1}));
+
+                // inside an object of unknown size: before a key, between key
+                // and value, and before the closing '}'
+                CHECK(json::from_ubjson(std::vector<uint8_t>({'{', 'N', 'U', 1, 'a', 'i', 1, '}'})) == json({{"a", 1}}));
+                CHECK(json::from_ubjson(std::vector<uint8_t>({'{', 'U', 1, 'a', 'N', 'i', 1, '}'})) == json({{"a", 1}}));
+                CHECK(json::from_ubjson(std::vector<uint8_t>({'{', 'U', 1, 'a', 'i', 1, 'N', '}'})) == json({{"a", 1}}));
+            }
+
+            SECTION("rejected where a length type specification is expected")
+            {
+                json _;
+
+                // after the 'S' marker of a string value
+                std::vector<uint8_t> const v_S = {'S', 'N', 'U', 1, 'a'};
+                CHECK_THROWS_WITH_AS(_ = json::from_ubjson(v_S), "[json.exception.parse_error.113] parse error at byte 2: syntax error while parsing UBJSON string: expected length type specification (U, i, I, l, L); last byte: 0x4E", json::parse_error&);
+
+                // as the key length of an object with a known size, where
+                // no-ops are not permitted in the first place
+                std::vector<uint8_t> const v_key = {'{', '#', 'i', 1, 'N', 'U', 1, 'a', 'i', 1};
+                CHECK_THROWS_WITH_AS(_ = json::from_ubjson(v_key), "[json.exception.parse_error.113] parse error at byte 5: syntax error while parsing UBJSON string: expected length type specification (U, i, I, l, L); last byte: 0x4E", json::parse_error&);
+            }
+        }
+
         SECTION("number")
         {
             SECTION("float")
@@ -2111,6 +2173,320 @@ TEST_CASE("UBJSON")
     }
 }
 
+TEST_CASE("UBJSON nesting does not consume the call stack")
+{
+    // Containers used to be read by calling back into the value reader once
+    // per element, so the native call stack grew with the nesting depth of the
+    // input. '[' alone opens a container, so a payload of repeated '[' crashed
+    // the process (#5104), as did the optimized forms, which reach the same
+    // path through a type or size annotation. The containers are kept on a
+    // heap stack now.
+    //
+    // Deeply nested values must not be compared, copied or dumped here: those
+    // operations are still recursive and would reintroduce the crash.
+    json _;
+
+    SECTION("containers that end at a marker")
+    {
+        const std::vector<uint8_t> input(500000, '[');
+        CHECK_THROWS_WITH_AS(_ = json::from_ubjson(input), "[json.exception.parse_error.110] parse error at byte 500001: syntax error while parsing UBJSON value: unexpected end of input", json::parse_error&);
+        CHECK(json::from_ubjson(input, true, false).is_discarded());
+    }
+
+    SECTION("containers with a size")
+    {
+        std::vector<uint8_t> input;
+        for (std::size_t i = 0; i < 100000; ++i)
+        {
+            input.push_back('[');
+            input.push_back('#');
+            input.push_back('i');
+            input.push_back(1);
+        }
+        CHECK_THROWS_AS(_ = json::from_ubjson(input), json::parse_error&);
+        CHECK(json::from_ubjson(input, true, false).is_discarded());
+    }
+
+    SECTION("containers with a type and a size")
+    {
+        // '[' is a permitted optimized type in UBJSON, so each element of such
+        // a container is itself a container, read without a marker of its own
+        std::vector<uint8_t> input;
+        for (std::size_t i = 0; i < 100000; ++i)
+        {
+            const std::vector<uint8_t> level = {'[', '$', '[', '#', 'i', 1};
+            input.insert(input.end(), level.begin(), level.end());
+        }
+        CHECK_THROWS_AS(_ = json::from_ubjson(input), json::parse_error&);
+        CHECK(json::from_ubjson(input, true, false).is_discarded());
+    }
+
+    SECTION("a well-formed deep value is read through the SAX interface")
+    {
+        std::vector<uint8_t> input(100000, '[');
+        input.insert(input.end(), 100000, ']');
+
+        SaxCountdown accept_all(1000000);
+        CHECK(json::sax_parse(input, &accept_all, json::input_format_t::ubjson));
+    }
+
+    SECTION("a well-formed deep value is read into a value")
+    {
+        const std::size_t depth = 10000;
+        std::vector<uint8_t> input(depth, '[');
+        input.insert(input.end(), depth, ']');
+
+        json j = json::from_ubjson(input);
+
+        std::size_t measured = 0;
+        const json* p = &j;
+        while (p->is_array() && !p->empty())
+        {
+            p = &p->front();
+            ++measured;
+        }
+        // the innermost array is empty, so the descent stops one level short
+        CHECK(measured == depth - 1);
+    }
+
+    SECTION("containers are still read the same way")
+    {
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', ']'})) == json::array());
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'{', '}'})) == json::object());
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '#', 'i', 0})) == json::array());
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'{', '#', 'i', 0})) == json::object());
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', 'i', '#', 'i', 2, 1, 2})) == json({1, 2}));
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '#', 'i', 2, 'i', 1, 'i', 2})) == json({1, 2}));
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'{', '$', 'i', '#', 'i', 1, 'i', 1, 'a', 1})) == json({{"a", 1}}));
+        // a no-op is not a value, so a container of them holds none
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', 'N', '#', 'i', 2})) == json::array());
+        // sized and unsized forms nested inside one another
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '[', '#', 'i', 2, 'i', 1, 'i', 2, ']'})) == json({{1, 2}}));
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '#', 'i', 1, '[', 'i', 1, ']'})) == json({{1}}));
+        // an optimized container of containers
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', '[', '#', 'i', 2, 'i', 1, ']', 'i', 2, ']'})) == json({{1}, {2}}));
+    }
+
+    SECTION("BJData containers are still read the same way")
+    {
+        // the ND-array wrapper and the binary shortcut are complete values,
+        // not containers the reader descends into
+        CHECK(json::from_bjdata(std::vector<uint8_t>({'[', '$', 'U', '#', '[', '$', 'i', '#', 'i', 2, 2, 3, 1, 2, 3, 4, 5, 6})) ==
+        json({{"_ArrayType_", "uint8"}, {"_ArraySize_", {2, 3}}, {"_ArrayData_", {1, 2, 3, 4, 5, 6}}}));
+        CHECK(json::from_bjdata(std::vector<uint8_t>({'[', '$', 'i', '#', 'i', 2, 1, 2})) == json({1, 2}));
+        CHECK(json::from_bjdata(std::vector<uint8_t>({'[', '[', 'i', 1, ']', ']'})) == json({{1}}));
+    }
+}
+
+TEST_CASE("UBJSON input that cannot be read is discarded by every overload")
+{
+    std::vector<std::uint8_t> input = json::to_ubjson(json({{"a", {1, 2}}}));
+    input.pop_back();
+
+    json _;
+    CHECK_THROWS_AS(_ = json::from_ubjson(input.begin(), input.end()), json::parse_error&);
+    CHECK(json::from_ubjson(input, true, false).is_discarded());
+    CHECK(json::from_ubjson(input.begin(), input.end(), true, false).is_discarded());
+    CHECK(json::from_ubjson(input.data(), input.size(), true, false).is_discarded());
+    CHECK(json::from_ubjson({input.data(), input.size()}, true, false).is_discarded());
+}
+
+TEST_CASE("UBJSON SAX parsing stops at every event")
+{
+    // Containers are opened and closed by the loop that reads them; a SAX
+    // handler that rejects any event - including the end of a nested
+    // container - must stop the parse right there.
+    const auto count_events = [](const std::vector<std::uint8_t>& input)
+    {
+        int events = 0;
+        while (true)
+        {
+            SaxCountdown scp(events);
+            if (json::sax_parse(input, &scp, json::input_format_t::ubjson))
+            {
+                return events;
+            }
+            ++events;
+            REQUIRE(events < 1000);
+        }
+    };
+
+    // 20 events: every container kind closes inside another one
+    const json j = json::parse(R"({"a": [1, {"b": []}], "c": {"d": [[2]]}})");
+    CHECK(count_events(json::to_ubjson(j)) == 20);
+    CHECK(count_events(json::to_ubjson(j, true)) == 20);
+    CHECK(count_events(json::to_ubjson(j, true, true)) == 20);
+}
+
+TEST_CASE("UBJSON optimized arrays of a valueless type are bounded")
+{
+    // An element of type 'Z', 'T' or 'F' is encoded by its marker alone, so an
+    // optimized array of one of those has no payload and the declared count is
+    // the only thing deciding how much is allocated. Ten bytes used to produce
+    // billions of values (#2793); every other type costs at least one byte per
+    // element and is bounded by the end of the input.
+    json _;
+
+    SECTION("an excessive count is rejected")
+    {
+        // 'l' is a big-endian int32: 0x7FFFFFFF elements, about 34 GB of value;
+        // OSS-Fuzz reported this shape as a parse_ubjson_fuzzer timeout
+        // (testcase 6347769435193344, no issue filed)
+        for (const auto marker :
+                {'Z', 'T', 'F'
+                })
+        {
+            const std::vector<uint8_t> input = {'[', '$', static_cast<uint8_t>(marker), '#', 'l', 0x7F, 0xFF, 0xFF, 0xFF};
+            CHECK_THROWS_WITH_AS(_ = json::from_ubjson(input), "[json.exception.out_of_range.408] syntax error while parsing UBJSON size: excessive array size", json::out_of_range&);
+            CHECK(json::from_ubjson(input, true, false).is_discarded());
+        }
+    }
+
+    SECTION("ordinary counts are unaffected")
+    {
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', 'Z', '#', 'i', 3})) == json({nullptr, nullptr, nullptr}));
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', 'T', '#', 'i', 2})) == json({true, true}));
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', 'F', '#', 'i', 2})) == json({false, false}));
+        // 'N' is a no-op rather than a value, and still yields an empty array
+        CHECK(json::from_ubjson(std::vector<uint8_t>({'[', '$', 'N', '#', 'i', 2})) == json::array());
+    }
+
+    SECTION("a type with a payload is unaffected")
+    {
+        // A count past the limit is not rejected for 'U', which costs a byte
+        // per element and is bounded by the end of the input instead. The
+        // count is kept just past the limit rather than made huge, because a
+        // count that also exceeds the array's max_size() is reported as
+        // out_of_range before the input runs out, and max_size() depends on
+        // the width of std::size_t.
+        const std::vector<uint8_t> input = {'[', '$', 'U', '#', 'l', 0x00, 0x10, 0x00, 0x01};
+        CHECK_THROWS_WITH_AS(_ = json::from_ubjson(input), "[json.exception.parse_error.110] parse error at byte 10: syntax error while parsing UBJSON number: unexpected end of input", json::parse_error&);
+        CHECK(json::from_ubjson(input, true, false).is_discarded());
+    }
+
+    SECTION("the writer stays within what the reader accepts")
+    {
+        // below the limit the optimized form is used and is tiny; above it the
+        // writer falls back so that the result can still be read back
+        json const at_limit(1048576, nullptr);
+        const auto v_at_limit = json::to_ubjson(at_limit, true, true);
+        CHECK(v_at_limit.size() == 9);
+        CHECK(v_at_limit.at(1) == '$');
+        CHECK(json::from_ubjson(v_at_limit) == at_limit);
+
+        json const above_limit(1048577, nullptr);
+        const auto v_above_limit = json::to_ubjson(above_limit, true, true);
+        CHECK(v_above_limit.at(1) != '$');
+        CHECK(json::from_ubjson(v_above_limit) == above_limit);
+    }
+}
+
+TEST_CASE("issue #5405 - array reserve for definite-length UBJSON arrays")
+{
+#if !defined(JSON_NOEXCEPTION)
+    // this SECTION relies on catching a thrown exception to distinguish
+    // which of two acceptable, bounded rejections a hostile header took;
+    // under JSON_NOEXCEPTION, JSON_THROW never produces a catchable C++
+    // exception (it aborts instead), so this cannot be tested that way here
+    SECTION("a huge claimed length with no element data must not over-allocate")
+    {
+        // optimized form [$type#count: type 'i' (int8), count as a four-byte
+        // 'l' (int32) of 0x7FFFFFFF (2147483647), but no element data at all.
+        // max_size() for a std::vector is far larger than this count, so it
+        // does not reject the header outright; the (capped) reservation must
+        // not attempt to allocate space for billions of elements before the
+        // missing data is detected.
+        json _;
+        const std::vector<uint8_t> input = {'[', '$', 'i', '#', 'l', 0x7F, 0xFF, 0xFF, 0xFF};
+        // On a platform where std::vector<json>::max_size() is smaller than
+        // the claimed count (e.g. 32-bit, where max_size() is bounded by a
+        // 32-bit SIZE_MAX divided by sizeof(json)), the SAX consumer's own
+        // check rejects the header outright (out_of_range.408, with the
+        // claimed count in the message) instead of accepting it and only
+        // finding it short of data once the (capped) reservation looks for
+        // element bytes that were never provided (parse_error.110). Either
+        // is an acceptable, bounded rejection of the hostile header -- the
+        // property under test is that no path attempts to allocate space
+        // for billions of elements.
+        bool threw = false;
+        try
+        {
+            _ = json::from_ubjson(input);
+        }
+        catch (const json::parse_error& e)
+        {
+            threw = true;
+            CHECK(e.id == 110);
+            CHECK(std::string(e.what()) == "[json.exception.parse_error.110] parse error at byte 10: syntax error while parsing UBJSON number: unexpected end of input");
+        }
+        catch (const json::out_of_range& e)
+        {
+            threw = true;
+            CHECK(e.id == 408);
+            CHECK(std::string(e.what()).find("excessive array size") != std::string::npos);
+        }
+        CHECK(threw);
+
+        // json_sax_dom_parser::start_array()'s max_size() check (unlike the
+        // scanner's own parse_error path) throws unconditionally via
+        // JSON_THROW rather than going through sax->parse_error(), so it is
+        // not gated by allow_exceptions=false on a platform where this
+        // header hits that check (e.g. 32-bit, see above) -- allow either
+        // a discarded result or the same out_of_range it throws with
+        // exceptions enabled.
+        try
+        {
+            CHECK(json::from_ubjson(input, true, false).is_discarded());
+        }
+        catch (const json::out_of_range& e)
+        {
+            CHECK(e.id == 408);
+        }
+    }
+#endif
+
+    SECTION("arrays of various sizes decode to the same value as before the reserve optimization")
+    {
+        for (const auto size :
+                {
+                    std::size_t{0}, std::size_t{1}, std::size_t{5}, // small
+                    std::size_t{16384},                             // exactly at the reserve cap
+                    std::size_t{20000}                              // above the reserve cap
+                })
+        {
+            CAPTURE(size)
+            json j = json::array();
+            for (std::size_t i = 0; i < size; ++i)
+            {
+                j.push_back(static_cast<int>(i % 1000));
+            }
+
+            // exercise both the plain and the optimized [$type#count encoding
+            const auto packed_plain = json::to_ubjson(j);
+            CHECK(json::from_ubjson(packed_plain) == j);
+
+            const auto packed_optimized = json::to_ubjson(j, true, true);
+            CHECK(json::from_ubjson(packed_optimized) == j);
+        }
+    }
+
+    SECTION("a user-defined SAX consumer is unaffected by the internal DOM reserve optimization")
+    {
+        // the reserve() call is local to json_sax_dom_parser / json_sax_dom_callback_parser;
+        // a custom SAX consumer that does not touch a DOM array sees identical events
+        json j = json::array();
+        for (int i = 0; i < 100; ++i)
+        {
+            j.push_back(i);
+        }
+        const auto packed = json::to_ubjson(j, true, true);
+
+        SaxCountdown scp(1000000); // large enough to never trigger an abort
+        CHECK(json::sax_parse(packed, &scp, json::input_format_t::ubjson));
+    }
+}
+
+
 TEST_CASE("Universal Binary JSON Specification Examples 1")
 {
     SECTION("Null Value")
@@ -2465,6 +2841,93 @@ TEST_CASE("all UBJSON first bytes")
 }
 #endif
 
+TEST_CASE("UBJSON use_type requires use_size")
+{
+    SECTION("non-empty array throws other_error.502")
+    {
+        const json j = {1, 2, 3};
+        CHECK_THROWS_WITH_AS(json::to_ubjson(j, false, true),
+                             "[json.exception.other_error.502] use_type requires use_size = true",
+                             json::other_error&);
+    }
+
+    SECTION("non-empty object throws other_error.502")
+    {
+        const json j = {{"a", 1}, {"b", 2}};
+        CHECK_THROWS_WITH_AS(json::to_ubjson(j, false, true),
+                             "[json.exception.other_error.502] use_type requires use_size = true",
+                             json::other_error&);
+    }
+
+    SECTION("scalars do not throw with use_type=true, use_count=false")
+    {
+        CHECK_NOTHROW(json::to_ubjson(42, false, true));
+        CHECK_NOTHROW(json::to_ubjson(3.14, false, true));
+        CHECK_NOTHROW(json::to_ubjson("hello", false, true));
+        CHECK_NOTHROW(json::to_ubjson(true, false, true));
+        CHECK_NOTHROW(json::to_ubjson(nullptr, false, true));
+    }
+
+    SECTION("empty containers do not throw with use_type=true, use_count=false")
+    {
+        CHECK_NOTHROW(json::to_ubjson(json::array(), false, true));
+        CHECK_NOTHROW(json::to_ubjson(json::object(), false, true));
+    }
+
+    SECTION("valid combinations on non-empty containers")
+    {
+        const json j = {1, 2, 3};
+        CHECK_NOTHROW(json::to_ubjson(j, false, false));
+        CHECK_NOTHROW(json::to_ubjson(j, true, false));
+        CHECK_NOTHROW(json::to_ubjson(j, true, true));
+    }
+}
+
+TEST_CASE("UBJSON round-trip invariants")
+{
+    // This checks what the parse_ubjson_fuzzer driver checks (see
+    // tests/src/fuzzer-parse_ubjson.cpp), so that a regression shows up in CI
+    // rather than as an OSS-Fuzz report: every value from_ubjson() returns
+    // (j1) can be serialized with any combination of options, the result can
+    // be parsed back (j2), and serializing j2 again with the same options
+    // reproduces the exact bytes. Beyond the driver, this also checks that j2
+    // equals j1. Values are compared with dump() rather than operator==,
+    // because a NaN never compares equal to itself.
+    struct options
+    {
+        bool use_size;
+        bool use_type;
+    };
+    const std::vector<options> all_options =
+    {
+        {false, false},
+        {true, false},
+        {true, true},
+    };
+
+    for (const auto& j0 : utils::round_trip_corpus::values())
+    {
+        // turn the corpus value into a value as from_ubjson() returns it; this
+        // has no binary values, as UBJSON writes them as arrays of integers
+        for (const auto& initial : all_options)
+        {
+            const json j1 = json::from_ubjson(json::to_ubjson(j0, initial.use_size, initial.use_type));
+
+            for (const auto& o : all_options)
+            {
+                INFO("j1 = " << j1.dump() << ", use_size = " << o.use_size << ", use_type = " << o.use_type);
+
+                const std::vector<std::uint8_t> vec = json::to_ubjson(j1, o.use_size, o.use_type);
+                json j2;
+                // anything the library writes must be parsable by the library
+                REQUIRE_NOTHROW(j2 = json::from_ubjson(vec));
+                CHECK(j2.dump() == j1.dump());
+                CHECK(json::to_ubjson(j2, o.use_size, o.use_type) == vec);
+            }
+        }
+    }
+}
+
 TEST_CASE("UBJSON roundtrips" * doctest::skip())
 {
     SECTION("input from self-generated UBJSON files")
@@ -2517,60 +2980,34 @@ TEST_CASE("UBJSON roundtrips" * doctest::skip())
         {
             CAPTURE(filename)
 
+            std::ifstream f_json(filename);
+            json const j1 = json::parse(f_json);
+            auto const packed = utils::read_binary_file(filename + ".ubjson");
+
             {
                 INFO_WITH_TEMP(filename + ": std::vector<uint8_t>");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                json const j1 = json::parse(f_json);
-
-                // parse UBJSON file
-                auto const packed = utils::read_binary_file(filename + ".ubjson");
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_ubjson(packed));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": std::ifstream");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                json const j1 = json::parse(f_json);
-
-                // parse UBJSON file
                 std::ifstream f_ubjson(filename + ".ubjson", std::ios::binary);
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_ubjson(f_ubjson));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": uint8_t* and size");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                const json j1 = json::parse(f_json);
-
-                // parse UBJSON file
-                auto const packed = utils::read_binary_file(filename + ".ubjson");
                 json j2;
                 CHECK_NOTHROW(j2 = json::from_ubjson({packed.data(), packed.size()}));
-
-                // compare parsed JSON values
                 CHECK(j1 == j2);
             }
 
             {
                 INFO_WITH_TEMP(filename + ": output to output adapters");
-                // parse JSON file
-                std::ifstream f_json(filename);
-                json const j1 = json::parse(f_json);
-
-                // parse UBJSON file
-                auto const packed = utils::read_binary_file(filename + ".ubjson");
-
                 {
                     INFO_WITH_TEMP(filename + ": output adapters: std::vector<uint8_t>");
                     std::vector<uint8_t> vec;
@@ -2581,3 +3018,18 @@ TEST_CASE("UBJSON roundtrips" * doctest::skip())
         }
     }
 }
+
+TEST_CASE("UBJSON optimized array of unsigned integers beyond int64")
+{
+    // UBJSON has no unsigned 64-bit type, so such values are written as
+    // high-precision numbers - also as the type of an optimized container
+    const json j = {18446744073709551615ULL, 9223372036854775808ULL};
+    const std::vector<std::uint8_t> expected =
+    {
+        '[', '$', 'H', '#', 'i', 2,
+        'i', 20, '1', '8', '4', '4', '6', '7', '4', '4', '0', '7', '3', '7', '0', '9', '5', '5', '1', '6', '1', '5',
+        'i', 19, '9', '2', '2', '3', '3', '7', '2', '0', '3', '6', '8', '5', '4', '7', '7', '5', '8', '0', '8'
+    };
+    CHECK(json::to_ubjson(j, true, true) == expected);
+    CHECK(json::from_ubjson(expected) == j);
+}

tests/src/unit-udt_macro.cpp

@@ -778,6 +778,193 @@ class derived_person_only_serialize_private_3 : person_without_default_construct
     NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE_ONLY_SERIALIZE_WITH_NAMES(derived_person_only_serialize_private_3, person_without_default_constructor_3, "json_hair_color", hair_color)
 };
 
+// Zero-member types for issue #4041: NLOHMANN_DEFINE_TYPE_* and
+// NLOHMANN_DEFINE_DERIVED_TYPE_* must compile and produce a valid (empty)
+// JSON object when no member arguments are given.
+class empty_intrusive
+{
+  public:
+    bool operator==(const empty_intrusive& /*rhs*/) const
+    {
+        return true;
+    }
+    NLOHMANN_DEFINE_TYPE_INTRUSIVE(empty_intrusive)
+};
+
+class empty_intrusive_with_default
+{
+  public:
+    bool operator==(const empty_intrusive_with_default& /*rhs*/) const
+    {
+        return true;
+    }
+    NLOHMANN_DEFINE_TYPE_INTRUSIVE_WITH_DEFAULT(empty_intrusive_with_default)
+};
+
+class empty_intrusive_only_serialize
+{
+  public:
+    NLOHMANN_DEFINE_TYPE_INTRUSIVE_ONLY_SERIALIZE(empty_intrusive_only_serialize)
+};
+
+class empty_non_intrusive
+{
+  public:
+    bool operator==(const empty_non_intrusive& /*rhs*/) const
+    {
+        return true;
+    }
+};
+// NOLINTNEXTLINE(misc-use-internal-linkage)
+NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE(empty_non_intrusive)
+
+class empty_non_intrusive_with_default
+{
+  public:
+    bool operator==(const empty_non_intrusive_with_default& /*rhs*/) const
+    {
+        return true;
+    }
+};
+// NOLINTNEXTLINE(misc-use-internal-linkage)
+NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE_WITH_DEFAULT(empty_non_intrusive_with_default)
+
+class empty_non_intrusive_only_serialize {};
+// NOLINTNEXTLINE(misc-use-internal-linkage)
+NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE_ONLY_SERIALIZE(empty_non_intrusive_only_serialize)
+
+class empty_derived_intrusive : public person_with_private_data
+{
+  public:
+    empty_derived_intrusive() = default;
+    empty_derived_intrusive(std::string name_, int age_, json metadata_)
+        : person_with_private_data(std::move(name_), age_, std::move(metadata_))
+    {}
+    NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE(empty_derived_intrusive, person_with_private_data)
+};
+
+class empty_derived_intrusive_with_default : public person_with_private_data
+{
+  public:
+    empty_derived_intrusive_with_default() = default;
+    empty_derived_intrusive_with_default(std::string name_, int age_, json metadata_)
+        : person_with_private_data(std::move(name_), age_, std::move(metadata_))
+    {}
+    NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE_WITH_DEFAULT(empty_derived_intrusive_with_default, person_with_private_data)
+};
+
+class empty_derived_intrusive_only_serialize : public person_with_private_data
+{
+  public:
+    empty_derived_intrusive_only_serialize() = default;
+    empty_derived_intrusive_only_serialize(std::string name_, int age_, json metadata_)
+        : person_with_private_data(std::move(name_), age_, std::move(metadata_))
+    {}
+    NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE_ONLY_SERIALIZE(empty_derived_intrusive_only_serialize, person_with_private_data)
+};
+
+class empty_derived_non_intrusive : public person_with_private_data
+{
+  public:
+    empty_derived_non_intrusive() = default;
+    empty_derived_non_intrusive(std::string name_, int age_, json metadata_)
+        : person_with_private_data(std::move(name_), age_, std::move(metadata_))
+    {}
+};
+// NOLINTNEXTLINE(misc-use-internal-linkage)
+NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE(empty_derived_non_intrusive, person_with_private_data)
+
+class empty_derived_non_intrusive_with_default : public person_with_private_data
+{
+  public:
+    empty_derived_non_intrusive_with_default() = default;
+    empty_derived_non_intrusive_with_default(std::string name_, int age_, json metadata_)
+        : person_with_private_data(std::move(name_), age_, std::move(metadata_))
+    {}
+};
+// NOLINTNEXTLINE(misc-use-internal-linkage)
+NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE_WITH_DEFAULT(empty_derived_non_intrusive_with_default, person_with_private_data)
+
+class empty_derived_non_intrusive_only_serialize : public person_with_private_data
+{
+  public:
+    empty_derived_non_intrusive_only_serialize() = default;
+    empty_derived_non_intrusive_only_serialize(std::string name_, int age_, json metadata_)
+        : person_with_private_data(std::move(name_), age_, std::move(metadata_))
+    {}
+};
+// NOLINTNEXTLINE(misc-use-internal-linkage)
+NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE_ONLY_SERIALIZE(empty_derived_non_intrusive_only_serialize, person_with_private_data)
+
+// Types at the documented maximum member count (63) for issue #4041's
+// argument-count dispatch. The derived-type macros carry a two-token
+// Type,BaseType prefix, so they reach two slots further into
+// NLOHMANN_JSON_GET_MACRO than the non-derived ones and are the first to break
+// if the tag dispatch runs out of positional slots.
+class max_members
+{
+  public:
+    int m1{}, m2{}, m3{}, m4{}, m5{}, m6{}, m7{}, m8{}, m9{}, m10{}, m11{}, m12{}, m13{}, m14{}, m15{}, m16{}, m17{}, m18{}, m19{}, m20{}, m21{}, m22{}, m23{}, m24{}, m25{}, m26{}, m27{}, m28{}, m29{}, m30{}, m31{}, m32{}, m33{}, m34{}, m35{}, m36{}, m37{}, m38{}, m39{}, m40{}, m41{}, m42{}, m43{}, m44{}, m45{}, m46{}, m47{}, m48{}, m49{}, m50{}, m51{}, m52{}, m53{}, m54{}, m55{}, m56{}, m57{}, m58{}, m59{}, m60{}, m61{}, m62{}, m63{};
+
+    NLOHMANN_DEFINE_TYPE_INTRUSIVE(max_members, m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45, m46, m47, m48, m49, m50, m51, m52, m53, m54, m55, m56, m57, m58, m59, m60, m61, m62, m63)
+};
+
+class max_members_base
+{
+  public:
+    int base_value = 0;
+
+    NLOHMANN_DEFINE_TYPE_INTRUSIVE(max_members_base, base_value)
+};
+
+class max_members_derived : public max_members_base
+{
+  public:
+    int m1{}, m2{}, m3{}, m4{}, m5{}, m6{}, m7{}, m8{}, m9{}, m10{}, m11{}, m12{}, m13{}, m14{}, m15{}, m16{}, m17{}, m18{}, m19{}, m20{}, m21{}, m22{}, m23{}, m24{}, m25{}, m26{}, m27{}, m28{}, m29{}, m30{}, m31{}, m32{}, m33{}, m34{}, m35{}, m36{}, m37{}, m38{}, m39{}, m40{}, m41{}, m42{}, m43{}, m44{}, m45{}, m46{}, m47{}, m48{}, m49{}, m50{}, m51{}, m52{}, m53{}, m54{}, m55{}, m56{}, m57{}, m58{}, m59{}, m60{}, m61{}, m62{}, m63{};
+
+    NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE(max_members_derived, max_members_base, m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45, m46, m47, m48, m49, m50, m51, m52, m53, m54, m55, m56, m57, m58, m59, m60, m61, m62, m63)
+};
+
+// User macros named like the dispatch suffixes (EMPTY is a common empty-macro
+// idiom) must not leak into the NLOHMANN_DEFINE_TYPE_* dispatch.
+#define EMPTY
+#define MEMBERS clobbered_by_user_macro
+
+class dispatch_with_user_macros_empty
+{
+  public:
+    NLOHMANN_DEFINE_TYPE_INTRUSIVE(dispatch_with_user_macros_empty)
+};
+
+class dispatch_with_user_macros_members
+{
+  public:
+    int value = 0;
+
+    NLOHMANN_DEFINE_TYPE_INTRUSIVE(dispatch_with_user_macros_members, value)
+};
+
+class dispatch_with_user_macros_derived_empty : public dispatch_with_user_macros_members
+{
+};
+// NOLINTNEXTLINE(misc-use-internal-linkage)
+NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE(dispatch_with_user_macros_derived_empty, dispatch_with_user_macros_members)
+
+class dispatch_with_user_macros_derived_members : public dispatch_with_user_macros_members
+{
+  public:
+    int own = 0;
+};
+// NOLINTNEXTLINE(misc-use-internal-linkage)
+NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE(dispatch_with_user_macros_derived_members, dispatch_with_user_macros_members, own)
+
+// testing for the macros also keeps -Wunused-macros from rejecting them
+#if !defined(EMPTY) || !defined(MEMBERS)
+    #error "EMPTY and MEMBERS must stay defined for the tests above"
+#endif
+#undef EMPTY
+#undef MEMBERS
+
 } // namespace persons
 
 TEST_CASE_TEMPLATE("Serialization/deserialization via NLOHMANN_DEFINE_TYPE_INTRUSIVE and NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE", Pair, // NOLINT(readability-math-missing-parentheses, bugprone-throwing-static-initialization)
@@ -1191,3 +1378,183 @@ TEST_CASE_TEMPLATE("Serialization of non-default-constructible classes via NLOHM
         }
     }
 }
+
+// Regression tests for issue #4041: NLOHMANN_DEFINE_TYPE_* and
+// NLOHMANN_DEFINE_DERIVED_TYPE_* macros must compile and produce valid
+// (empty, or base-only for the derived case) JSON objects when no member
+// arguments are given, on every supported C++ standard.
+TEST_CASE_TEMPLATE("Serialization/deserialization of zero-member types via NLOHMANN_DEFINE_TYPE_* (issue #4041)", Json, // NOLINT(readability-math-missing-parentheses, bugprone-throwing-static-initialization)
+                   nlohmann::json, nlohmann::ordered_json)
+{
+    constexpr bool is_ordered = std::is_same<Json, nlohmann::ordered_json>::value;
+    const char* const derived_dump = is_ordered
+                                     ? R"({"age":1,"name":"Erik","metadata":null})"
+                                     : R"({"age":1,"metadata":null,"name":"Erik"})";
+
+    SECTION("NLOHMANN_DEFINE_TYPE_INTRUSIVE with zero members")
+    {
+        persons::empty_intrusive obj{};
+        Json j = obj;
+        CHECK(j.dump() == "{}");
+        CHECK(j.template get<persons::empty_intrusive>() == obj);
+    }
+
+    SECTION("NLOHMANN_DEFINE_TYPE_INTRUSIVE_WITH_DEFAULT with zero members")
+    {
+        persons::empty_intrusive_with_default obj{};
+        Json j = obj;
+        CHECK(j.dump() == "{}");
+        CHECK(j.template get<persons::empty_intrusive_with_default>() == obj);
+    }
+
+    SECTION("NLOHMANN_DEFINE_TYPE_INTRUSIVE_ONLY_SERIALIZE with zero members")
+    {
+        const persons::empty_intrusive_only_serialize obj{};
+        Json j = obj;
+        CHECK(j.dump() == "{}");
+    }
+
+    SECTION("NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE with zero members")
+    {
+        persons::empty_non_intrusive obj{};
+        Json j = obj;
+        CHECK(j.dump() == "{}");
+        CHECK(j.template get<persons::empty_non_intrusive>() == obj);
+    }
+
+    SECTION("NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE_WITH_DEFAULT with zero members")
+    {
+        persons::empty_non_intrusive_with_default obj{};
+        Json j = obj;
+        CHECK(j.dump() == "{}");
+        CHECK(j.template get<persons::empty_non_intrusive_with_default>() == obj);
+    }
+
+    SECTION("NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE_ONLY_SERIALIZE with zero members")
+    {
+        const persons::empty_non_intrusive_only_serialize obj{};
+        Json j = obj;
+        CHECK(j.dump() == "{}");
+    }
+
+    SECTION("NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE with zero own members")
+    {
+        persons::empty_derived_intrusive obj{"Erik", 1, nullptr};
+        Json j = obj;
+        CHECK(j.dump() == derived_dump);
+        CHECK(j.template get<persons::empty_derived_intrusive>() == obj);
+    }
+
+    SECTION("NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE_WITH_DEFAULT with zero own members")
+    {
+        persons::empty_derived_intrusive_with_default obj{"Erik", 1, nullptr};
+        Json j = obj;
+        CHECK(j.dump() == derived_dump);
+        CHECK(j.template get<persons::empty_derived_intrusive_with_default>() == obj);
+    }
+
+    SECTION("NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE_ONLY_SERIALIZE with zero own members")
+    {
+        const persons::empty_derived_intrusive_only_serialize obj{"Erik", 1, nullptr};
+        Json j = obj;
+        CHECK(j.dump() == derived_dump);
+    }
+
+    SECTION("NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE with zero own members")
+    {
+        persons::empty_derived_non_intrusive obj{"Erik", 1, nullptr};
+        Json j = obj;
+        CHECK(j.dump() == derived_dump);
+        CHECK(j.template get<persons::empty_derived_non_intrusive>() == obj);
+    }
+
+    SECTION("NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE_WITH_DEFAULT with zero own members")
+    {
+        persons::empty_derived_non_intrusive_with_default obj{"Erik", 1, nullptr};
+        Json j = obj;
+        CHECK(j.dump() == derived_dump);
+        CHECK(j.template get<persons::empty_derived_non_intrusive_with_default>() == obj);
+    }
+
+    SECTION("NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE_ONLY_SERIALIZE with zero own members")
+    {
+        const persons::empty_derived_non_intrusive_only_serialize obj{"Erik", 1, nullptr};
+        Json j = obj;
+        CHECK(j.dump() == derived_dump);
+    }
+}
+
+// Regression test for the argument-count dispatch added for issue #4041: the
+// documented maximum of 63 members must keep working, including for the
+// derived-type macros whose Type,BaseType prefix consumes two dispatch slots.
+TEST_CASE_TEMPLATE("Serialization/deserialization of maximum-member-count types via NLOHMANN_DEFINE_TYPE_*", Json, // NOLINT(readability-math-missing-parentheses, bugprone-throwing-static-initialization)
+                   nlohmann::json, nlohmann::ordered_json)
+{
+    SECTION("NLOHMANN_DEFINE_TYPE_INTRUSIVE with 63 members")
+    {
+        persons::max_members obj{};
+        obj.m1 = 1;
+        obj.m63 = 63;
+        Json j = obj;
+        CHECK(j.size() == 63);
+        const auto obj2 = j.template get<persons::max_members>();
+        CHECK(obj2.m1 == 1);
+        CHECK(obj2.m63 == 63);
+    }
+
+    SECTION("NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE with 63 own members")
+    {
+        persons::max_members_derived obj{};
+        obj.base_value = 7;
+        obj.m1 = 1;
+        obj.m63 = 63;
+        Json j = obj;
+        CHECK(j.size() == 64);
+        const auto obj2 = j.template get<persons::max_members_derived>();
+        CHECK(obj2.base_value == 7);
+        CHECK(obj2.m1 == 1);
+        CHECK(obj2.m63 == 63);
+    }
+}
+
+TEST_CASE_TEMPLATE("NLOHMANN_DEFINE_TYPE_* dispatch is unaffected by user macros named EMPTY or MEMBERS", Json, // NOLINT(readability-math-missing-parentheses, bugprone-throwing-static-initialization)
+                   nlohmann::json, nlohmann::ordered_json)
+{
+    SECTION("zero members")
+    {
+        const persons::dispatch_with_user_macros_empty obj{};
+        const Json j = obj;
+        CHECK(j == Json::object());
+        CHECK_NOTHROW(j.template get<persons::dispatch_with_user_macros_empty>());
+    }
+
+    SECTION("one member")
+    {
+        persons::dispatch_with_user_macros_members obj{};
+        obj.value = 42;
+        const Json j = obj;
+        CHECK(j == Json({{"value", 42}}));
+        CHECK(j.template get<persons::dispatch_with_user_macros_members>().value == 42);
+    }
+
+    SECTION("derived with zero own members")
+    {
+        persons::dispatch_with_user_macros_derived_empty obj{};
+        obj.value = 42;
+        const Json j = obj;
+        CHECK(j == Json({{"value", 42}}));
+        CHECK(j.template get<persons::dispatch_with_user_macros_derived_empty>().value == 42);
+    }
+
+    SECTION("derived with own members")
+    {
+        persons::dispatch_with_user_macros_derived_members obj{};
+        obj.value = 42;
+        obj.own = 7;
+        const Json j = obj;
+        CHECK(j == Json({{"value", 42}, {"own", 7}}));
+        const auto obj2 = j.template get<persons::dispatch_with_user_macros_derived_members>();
+        CHECK(obj2.value == 42);
+        CHECK(obj2.own == 7);
+    }
+}

tests/src/unit-unicode1.cpp

@@ -17,6 +17,7 @@ using nlohmann::json;
 #include <sstream>
 #include <iomanip>
 #include "make_test_data_available.hpp"
+#include "test_utils.hpp"
 
 TEST_CASE("Unicode (1/5)" * doctest::skip())
 {
@@ -240,7 +241,8 @@ void roundtrip(bool success_expected, const std::string& s)
     if (success_expected)
     {
         // serialization succeeds
-        CHECK_NOTHROW(j.dump());
+        // dump() is nodiscard; this only checks that dumping does not throw
+        CHECK_NOTHROW(utils::ignore_return_value(j.dump()));
 
         // exclude parse test for U+0000
         if (s[0] != '\0')
@@ -259,7 +261,8 @@ void roundtrip(bool success_expected, const std::string& s)
     else
     {
         // serialization fails
-        CHECK_THROWS_AS(j.dump(), json::type_error&);
+        // dump() is nodiscard; the exception is thrown by dump() itself before it would return
+        CHECK_THROWS_AS(utils::ignore_return_value(j.dump()), json::type_error&);
 
         // parsing JSON text fails
         CHECK_THROWS_AS(_ = json::parse(ps), json::parse_error&);

tests/src/unit-unicode2.cpp

@@ -19,6 +19,7 @@ using nlohmann::json;
 #include <iostream>
 #include <iomanip>
 #include "make_test_data_available.hpp"
+#include "test_utils.hpp"
 
 // this test suite uses static variables with non-trivial destructors
 DOCTEST_CLANG_SUPPRESS_WARNING_PUSH
@@ -97,7 +98,8 @@ void check_utf8dump(bool success_expected, int byte1, int byte2 = -1, int byte3
     else
     {
         // strict mode must throw if success is not expected
-        CHECK_THROWS_AS(j.dump(), json::type_error&);
+        // dump() is nodiscard; the exception is thrown by dump() itself before it would return
+        CHECK_THROWS_AS(utils::ignore_return_value(j.dump()), json::type_error&);
         // ignore and replace must create different dumps
         CHECK(s_ignored != s_replaced);
 

tests/src/unit-unicode3.cpp

@@ -19,6 +19,7 @@ using nlohmann::json;
 #include <iostream>
 #include <iomanip>
 #include "make_test_data_available.hpp"
+#include "test_utils.hpp"
 
 // this test suite uses static variables with non-trivial destructors
 DOCTEST_CLANG_SUPPRESS_WARNING_PUSH
@@ -97,7 +98,8 @@ void check_utf8dump(bool success_expected, int byte1, int byte2 = -1, int byte3
     else
     {
         // strict mode must throw if success is not expected
-        CHECK_THROWS_AS(j.dump(), json::type_error&);
+        // dump() is nodiscard; the exception is thrown by dump() itself before it would return
+        CHECK_THROWS_AS(utils::ignore_return_value(j.dump()), json::type_error&);
         // ignore and replace must create different dumps
         CHECK(s_ignored != s_replaced);
 
@@ -304,8 +306,8 @@ TEST_CASE("Unicode (3/5)" * doctest::skip())
                         {
                             for (int byte4 = 0x00; byte4 <= 0xFF; ++byte4)
                             {
-                                // skip fourth second byte
-                                if (0x80 <= byte3 && byte3 <= 0xBF)
+                                // skip correct fourth byte
+                                if (0x80 <= byte4 && byte4 <= 0xBF)
                                 {
                                     continue;
                                 }

tests/src/unit-unicode4.cpp

@@ -19,6 +19,7 @@ using nlohmann::json;
 #include <iostream>
 #include <iomanip>
 #include "make_test_data_available.hpp"
+#include "test_utils.hpp"
 
 // this test suite uses static variables with non-trivial destructors
 DOCTEST_CLANG_SUPPRESS_WARNING_PUSH
@@ -97,7 +98,8 @@ void check_utf8dump(bool success_expected, int byte1, int byte2 = -1, int byte3
     else
     {
         // strict mode must throw if success is not expected
-        CHECK_THROWS_AS(j.dump(), json::type_error&);
+        // dump() is nodiscard; the exception is thrown by dump() itself before it would return
+        CHECK_THROWS_AS(utils::ignore_return_value(j.dump()), json::type_error&);
         // ignore and replace must create different dumps
         CHECK(s_ignored != s_replaced);
 
@@ -305,7 +307,7 @@ TEST_CASE("Unicode (4/5)" * doctest::skip())
                             for (int byte4 = 0x00; byte4 <= 0xFF; ++byte4)
                             {
                                 // skip correct fourth byte
-                                if (0x80 <= byte3 && byte3 <= 0xBF)
+                                if (0x80 <= byte4 && byte4 <= 0xBF)
                                 {
                                     continue;
                                 }

tests/src/unit-unicode5.cpp

@@ -19,6 +19,7 @@ using nlohmann::json;
 #include <iostream>
 #include <iomanip>
 #include "make_test_data_available.hpp"
+#include "test_utils.hpp"
 
 // this test suite uses static variables with non-trivial destructors
 DOCTEST_CLANG_SUPPRESS_WARNING_PUSH
@@ -97,7 +98,8 @@ void check_utf8dump(bool success_expected, int byte1, int byte2 = -1, int byte3
     else
     {
         // strict mode must throw if success is not expected
-        CHECK_THROWS_AS(j.dump(), json::type_error&);
+        // dump() is nodiscard; the exception is thrown by dump() itself before it would return
+        CHECK_THROWS_AS(utils::ignore_return_value(j.dump()), json::type_error&);
         // ignore and replace must create different dumps
         CHECK(s_ignored != s_replaced);
 
@@ -305,7 +307,7 @@ TEST_CASE("Unicode (5/5)" * doctest::skip())
                             for (int byte4 = 0x00; byte4 <= 0xFF; ++byte4)
                             {
                                 // skip correct fourth byte
-                                if (0x80 <= byte3 && byte3 <= 0xBF)
+                                if (0x80 <= byte4 && byte4 <= 0xBF)
                                 {
                                     continue;
                                 }

tests/src/unit-user_defined_input.cpp

@@ -18,7 +18,12 @@
 #include <nlohmann/json.hpp>
 using nlohmann::json;
 
+#include <array> // array
+#include <cstddef> // size_t
+#include <cstdint> // uint8_t
 #include <list>
+#include <string> // string
+#include <vector> // vector
 
 #if defined(__cpp_lib_concepts) && defined(JSON_HAS_CPP_20)
     #include <iterator>
@@ -212,6 +217,66 @@ TEST_CASE("Parse with heterogeneous iterator and sentinel types")
     CHECK(j2.at(0) == 1);
 }
 
+// A type whose data() hands out raw bytes but whose size() counts something
+// else - here fixed-size records. Reading [data(), data() + size()) as bytes
+// would silently truncate the input, so data() and size() alone must not be
+// taken as evidence of contiguous byte storage.
+struct record_buffer
+{
+    using value_type = std::array<char, 4>;
+
+    std::string bytes;
+
+    const char* data() const noexcept
+    {
+        return bytes.data();
+    }
+    std::size_t size() const noexcept
+    {
+        return bytes.size() / sizeof(value_type);
+    }
+    const char* begin() const noexcept
+    {
+        return bytes.data();
+    }
+    const char* end() const noexcept
+    {
+        return bytes.data() + bytes.size();
+    }
+};
+
+TEST_CASE("Contiguous byte containers take the pointer adapter")
+{
+    // Containers with contiguous single-byte storage are routed through the
+    // pointer-based adapter so the bulk fast paths apply in every standard, not
+    // only in C++20 where the library iterators model std::contiguous_iterator.
+    CHECK(nlohmann::detail::is_contiguous_byte_container<std::string>::value);
+    CHECK(nlohmann::detail::is_contiguous_byte_container<std::vector<char>>::value);
+    CHECK(nlohmann::detail::is_contiguous_byte_container<std::vector<std::uint8_t>>::value);
+    CHECK(nlohmann::detail::is_contiguous_byte_container<std::array<char, 4>>::value);
+
+    // input_adapter() takes its container by forwarding reference, so the trait
+    // is also asked about reference types
+    CHECK(nlohmann::detail::is_contiguous_byte_container<std::string&>::value);
+    CHECK(nlohmann::detail::is_contiguous_byte_container<const std::string&>::value);
+
+    // everything else keeps the iterator-based adapter
+    CHECK_FALSE(nlohmann::detail::is_contiguous_byte_container<std::list<char>>::value);
+    CHECK_FALSE(nlohmann::detail::is_contiguous_byte_container<std::vector<int>>::value);
+    CHECK_FALSE(nlohmann::detail::is_contiguous_byte_container<const char*>::value);
+
+    // including a type that has data() and size() but whose size() does not
+    // count the units data() points at: its value_type says so
+    CHECK_FALSE(nlohmann::detail::is_contiguous_byte_container<record_buffer>::value);
+
+    // and such a container still parses through its iterators, in full - taking
+    // it for a byte container would stop after data() + size() bytes
+    const record_buffer buffer{"[1,2,3,4,5]"};
+    CHECK(buffer.data() == buffer.bytes.data());
+    CHECK(buffer.size() * sizeof(record_buffer::value_type) < buffer.bytes.size());
+    CHECK(json::parse(buffer) == json({1, 2, 3, 4, 5}));
+}
+
 #if defined(__cpp_lib_concepts) && defined(JSON_HAS_CPP_20)
 // JSON_HAS_CPP_20 (do not remove; see note at top of file)
 TEST_CASE("Parse with std::counted_iterator and std::default_sentinel_t")
@@ -228,6 +293,180 @@ TEST_CASE("Parse with std::counted_iterator and std::default_sentinel_t")
     const std::counted_iterator<iterator_type> first2(json_str.begin(), len);
     CHECK(json::accept(first2, std::default_sentinel));
 }
+
+TEST_CASE("std::counted_iterator reaches the contiguous fast paths")
+{
+    // A sized sentinel makes the remaining element count computable in O(1), so
+    // std::counted_iterator over a contiguous iterator must reach the same bulk
+    // string/number scanners as a plain pointer - not just the byte-at-a-time
+    // fallback (see #5268 for the equivalent memcpy fast path).
+#if JSON_HAS_RANGES
+    // JSON_HAS_RANGES is 0 on standard libraries with an incomplete <ranges>
+    // (libstdc++ < 11, libc++ < 16), where the adapter deliberately falls back
+    // to the byte-at-a-time scanner; everything below still has to work there.
+    using adapter_type = nlohmann::detail::iterator_input_adapter<std::counted_iterator<const char*>, std::default_sentinel_t>;
+    CHECK(adapter_type::supports_bulk_scan);
+    CHECK(adapter_type::supports_seek);
+#endif
+
+    // exercise every fast path: long ASCII run, multibyte UTF-8, escapes, and
+    // integer/floating-point numbers
+    const std::string json_str =
+        R"({"ascii":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",)"
+        "\"utf8\":\"\xe4\xb8\xad\xe6\x96\x87\xf0\x9f\x98\x80\xc3\xa9\","
+        R"("escaped":"aéb\n\\","ints":[0,-1,18446744073709551615,-9223372036854775808],)"
+        R"("floats":[1.5,-2.25e3,0.30000000000000004]})";
+    const auto len = static_cast<std::iter_difference_t<const char*>>(json_str.size());
+
+    const std::counted_iterator<const char*> first(json_str.data(), len);
+    const json j = json::parse(first, std::default_sentinel);
+
+    // parsing through the pointer adapter must give exactly the same result
+    CHECK(j == json::parse(json_str));
+
+#if !defined(JSON_NOEXCEPTION)
+    // Diagnostics that quote the offending token are reconstructed from the
+    // already-consumed input (supports_seek), a path a sized sentinel only
+    // reaches now; check a few that include the "last read" text. Parsing
+    // invalid input aborts when exceptions are off, hence the guard.
+    // Raw strings and explicit bytes: an escaped literal and two literals
+    // written next to each other both read as mistakes to static analysis.
+    const auto byte = [](int value)
+    {
+        return std::string(1, static_cast<char>(value));
+    };
+    const std::vector<std::string> diagnostic_docs =
+    {
+        "1\nx",
+        "truX",
+        "[tru]",
+        R"("abc)",
+        R"(["\ud834"])",
+        R"(["a)" + byte(0x01) + R"(b"])",
+        R"([")" + byte(0xC3) + byte(0x28) + R"("])",
+        "[1e]",
+        R"(["aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaX)"
+    };
+
+    for (const auto& text : diagnostic_docs)
+    {
+        CAPTURE(text);
+        const std::counted_iterator<const char*> it(text.data(), static_cast<std::iter_difference_t<const char*>>(text.size()));
+        std::string counted_message;
+        std::string string_message;
+        try
+        {
+            const json counted_result = json::parse(it, std::default_sentinel);
+            static_cast<void>(counted_result);
+        }
+        catch (const json::parse_error& e)
+        {
+            counted_message = e.what();
+        }
+        try
+        {
+            const json string_result = json::parse(text);
+            static_cast<void>(string_result);
+        }
+        catch (const json::parse_error& e)
+        {
+            string_message = e.what();
+        }
+        CHECK_FALSE(counted_message.empty());
+        CHECK(counted_message == string_message);
+    }
+
+    // and errors must still be reported identically
+    const std::string bad = "[01\n]";
+    const std::counted_iterator<const char*> bad_first(bad.data(), static_cast<std::iter_difference_t<const char*>>(bad.size()));
+    std::string counted_what;
+    std::string string_what;
+    try
+    {
+        const json counted_result = json::parse(bad_first, std::default_sentinel);
+        static_cast<void>(counted_result);
+    }
+    catch (const json::parse_error& e)
+    {
+        counted_what = e.what();
+    }
+    try
+    {
+        const json string_result = json::parse(bad);
+        static_cast<void>(string_result);
+    }
+    catch (const json::parse_error& e)
+    {
+        string_what = e.what();
+    }
+    CHECK_FALSE(counted_what.empty());
+    CHECK(counted_what == string_what);
+#endif
+}
+
+#if !defined(JSON_NOEXCEPTION)
+// several cases below are truncated on purpose, and parsing invalid input
+// aborts when exceptions are off
+TEST_CASE("std::counted_iterator bulk scanning stops at the counted end")
+{
+    // The count, not the size of the underlying buffer, is the end of the
+    // input: the bulk scanners must never look at the bytes behind it, even
+    // though they are readable. Each case is compared against parsing the
+    // equivalent prefix as a std::string.
+    const auto via_counted = [](const std::string & buf, std::size_t n) -> std::string
+    {
+        const std::counted_iterator<const char*> first(buf.data(), static_cast<std::iter_difference_t<const char*>>(n));
+        try
+        {
+            const json j = json::parse(first, std::default_sentinel);
+            return "OK|" + j.dump();
+        }
+        catch (const json::parse_error& e)
+        {
+            return {e.what()};
+        }
+    };
+    const auto via_prefix = [](const std::string & buf, std::size_t n) -> std::string
+    {
+        try
+        {
+            const json j = json::parse(buf.substr(0, n));
+            return "OK|" + j.dump();
+        }
+        catch (const json::parse_error& e)
+        {
+            return {e.what()};
+        }
+    };
+
+    struct testcase // NOLINT(cppcoreguidelines-pro-type-member-init,hicpp-member-init)
+    {
+        const char* buffer;
+        std::size_t count;
+    };
+    const std::vector<testcase> cases =
+    {
+        {"[\"abc\"]____TRAILING____", 7},                    // exact fit, tail hidden
+        {"[\"abcdefghijklmnop\"]____", 8},                   // cut inside a string
+        {"[\"abc\"]____", 6},                                // cut just before the closing quote
+        {"[12345]xxxxx", 4},                                 // cut inside a number
+        {"[123]999999", 5},                                  // number ends exactly at the count
+        {"[\"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaa\"]", 12},        // closing quote only behind the count
+        {"[\"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa\"]", 19},      // cut inside an 8-byte SWAR stride
+        {"[\"\xe4\xb8\xad\xe6\x96\x87\"]", 5},               // cut inside a UTF-8 sequence
+        {"[\"\xe4\xb8\xad\xe6\x96\x87\"]____", 10},          // complete UTF-8, tail hidden
+        {"[1.25e3]TRAILINGDIGITS999", 7},                    // number token reaches the count
+    };
+
+    for (const auto& tc : cases)
+    {
+        CAPTURE(tc.buffer);
+        CAPTURE(tc.count);
+        const std::string buffer = tc.buffer;
+        CHECK(via_counted(buffer, tc.count) == via_prefix(buffer, tc.count));
+    }
+}
+#endif
 #endif
 
 } // namespace

tests/src/unit-wstring.cpp

@@ -70,6 +70,8 @@ TEST_CASE("wide strings")
             CHECK_THROWS_WITH_AS(_ = json::parse(std::wstring{L'"', static_cast<wchar_t>(0xDC00), L'"'}), error_low_surrogate, json::parse_error&);
             // a high surrogate followed by a non-low-surrogate unit is invalid
             CHECK_THROWS_WITH_AS(_ = json::parse(std::wstring{L'"', static_cast<wchar_t>(0xD800), L'a', L'"'}), error_high_surrogate, json::parse_error&);
+            // ... also when the unit is above the low surrogates
+            CHECK_THROWS_WITH_AS(_ = json::parse(std::wstring{L'"', static_cast<wchar_t>(0xD800), static_cast<wchar_t>(0xE000), L'"'}), error_high_surrogate, json::parse_error&);
             // a lone low surrogate must not swallow the following unit: pairing
             // it with any second unit would produce valid UTF-8, so the error
             // has to report an ill-formed byte at the surrogate's own position
@@ -99,6 +101,8 @@ TEST_CASE("wide strings")
             CHECK_THROWS_WITH_AS(_ = json::parse(std::u16string{u'"', 0xDC00, u'"'}), "[json.exception.parse_error.101] parse error at line 1, column 2: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"<U+0000>'", json::parse_error&);
             // a high surrogate followed by a non-low-surrogate unit is invalid
             CHECK_THROWS_WITH_AS(_ = json::parse(std::u16string{u'"', 0xD800, u'a', u'"'}), "[json.exception.parse_error.101] parse error at line 1, column 2: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"<U+0000>'", json::parse_error&);
+            // ... also when the unit is above the low surrogates
+            CHECK_THROWS_WITH_AS(_ = json::parse(std::u16string{u'"', 0xD800, 0xE000, u'"'}), "[json.exception.parse_error.101] parse error at line 1, column 2: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"<U+0000>'", json::parse_error&);
             // a lone low surrogate must not swallow the following unit: pairing
             // it with any second unit would produce valid UTF-8, so the error
             // has to report an ill-formed byte at the surrogate's own position
@@ -125,6 +129,16 @@ TEST_CASE("wide strings")
             std::u32string const w = U"\"\x110000";
             json _;
             CHECK_THROWS_AS(_ = json::parse(w), json::parse_error&);
+
+            // a code unit above U+10FFFF must not be narrowed onto the EOF
+            // sentinel: 0xFFFFFFFF would otherwise end the document silently and
+            // let everything following it pass the strict end-of-input check
+            std::u32string const trailing{U'[', U'1', U']', static_cast<char32_t>(0xFFFFFFFF), U'x'};
+            CHECK_THROWS_WITH_AS(_ = json::parse(trailing), "[json.exception.parse_error.101] parse error at line 1, column 4: syntax error while parsing value - invalid literal; last read: '1]\xFF'; expected end of input", json::parse_error&);
+            CHECK(!json::accept(trailing));
+
+            // the same unit inside a string is reported as an ill-formed byte
+            CHECK_THROWS_WITH_AS(_ = json::parse(std::u32string{U'"', static_cast<char32_t>(0xFFFFFFFF), U'"'}), "[json.exception.parse_error.101] parse error at line 1, column 2: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"\xFF'", json::parse_error&);
         }
     }
 }