Add test coverage for documented lenient BSON input handling

Issue #5333 documented three intentionally-lenient behaviors of the BSON
reader (any non-zero byte accepted as a boolean `true`, BSON array element
keys not validated against the required decimal sequence, and the payload
of binary subtype 0x02 "old binary" returned as-is including its inner
length prefix), but none of them was pinned by a test, so a future change
could silently regress the documented behavior.

Also add coverage for the out_of_range.412 length-overflow check
(shared by binary, string, and (sub-)document BSON length fields) for
the string and document cases; only the binary case was previously
tested.

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

Makefile

@@ -1,4 +1,4 @@
-.PHONY: pretty clean ChangeLog.md release update_hedley update_hedley_undef
+.PHONY: pretty clean ChangeLog.md release
 
 ##########################################################################
 # configuration
@@ -41,8 +41,6 @@ all:
 	@echo "fuzz_testing_ubjson - prepare fuzz testing of the UBJSON parser"
 	@echo "pretty - beautify code with Artistic Style"
 	@echo "run_benchmarks - build and run benchmarks"
-	@echo "update_hedley - download Hedley and regenerate hedley.hpp / hedley_undef.hpp"
-	@echo "update_hedley_undef - rebuild hedley_undef.hpp from the JSON_HEDLEY_* #define names in hedley.hpp"
 
 
 ##########################################################################
@@ -243,24 +241,11 @@ update_hedley:
 	rm -f include/nlohmann/thirdparty/hedley/hedley.hpp include/nlohmann/thirdparty/hedley/hedley_undef.hpp
 	curl https://raw.githubusercontent.com/nemequ/hedley/master/hedley.h -o include/nlohmann/thirdparty/hedley/hedley.hpp
 	$(SED) -i 's/HEDLEY_/JSON_HEDLEY_/g' include/nlohmann/thirdparty/hedley/hedley.hpp
+	grep "[[:blank:]]*#[[:blank:]]*undef" include/nlohmann/thirdparty/hedley/hedley.hpp | grep -v "__" | sort | uniq | $(SED) 's/ //g' | $(SED) 's/undef/undef /g' > include/nlohmann/thirdparty/hedley/hedley_undef.hpp
 	$(SED) -i '1s/^/#pragma once\n\n/' include/nlohmann/thirdparty/hedley/hedley.hpp
-	$(MAKE) update_hedley_undef
+	$(SED) -i '1s/^/#pragma once\n\n/' include/nlohmann/thirdparty/hedley/hedley_undef.hpp
 	$(MAKE) amalgamate
 
-# Rebuild hedley_undef.hpp from every JSON_HEDLEY_* name that hedley.hpp
-# #defines. Hedley does not #undef all of its public macros internally (see
-# #5408), so grepping those #undef lines misses names such as
-# JSON_HEDLEY_PRAGMA. cmake/scripts/gen_hedley_undef_check.cmake is the
-# single source of truth for this extraction (tests/CMakeLists.txt uses the
-# same script, in MODE=checks, to generate the matching leak-check test), so
-# the vendored header, the generated #undef list, and the regression test
-# cannot drift apart.
-update_hedley_undef:
-	cmake -DHEDLEY_HPP=include/nlohmann/thirdparty/hedley/hedley.hpp \
-	      -DOUTPUT=include/nlohmann/thirdparty/hedley/hedley_undef.hpp \
-	      -DMODE=undef \
-	      -P cmake/scripts/gen_hedley_undef_check.cmake
-
 ##########################################################################
 # serve_header.py
 ##########################################################################

cmake/scripts/gen_hedley_undef_check.cmake

@@ -1,105 +0,0 @@
-# Shared extractor for the JSON_HEDLEY_* macro names defined in hedley.hpp.
-#
-# Every macro that hedley.hpp #defines must be #undef-ed again once json.hpp
-# has been fully processed (see include/nlohmann/detail/macro_unscope.hpp
-# and https://github.com/nlohmann/json/issues/5408). Deriving the macro list
-# straight from hedley.hpp here -- instead of hand-maintaining it in two
-# places -- means hedley_undef.hpp and the regression test that checks for
-# leaked macros can never drift apart, even after a future `make
-# update_hedley` pulls in new macros from upstream Hedley.
-#
-#   MODE=undef  (default): write hedley_undef.hpp (SPDX header, #pragma once,
-#               one #undef per macro name) -- used by `make update_hedley_undef`
-#   MODE=checks:            write one #ifdef/FAIL_CHECK/#endif per macro name,
-#               meant to be #include-d inside a TEST_CASE -- used by
-#               tests/CMakeLists.txt to (re)generate the include for
-#               tests/src/unit-no-macro-leak.cpp
-#
-# Required variables:
-#   HEDLEY_HPP  path to include/nlohmann/thirdparty/hedley/hedley.hpp
-#   OUTPUT      path of the file to (over)write
-# Optional:
-#   MODE        "undef" (default) or "checks"
-
-if(NOT DEFINED HEDLEY_HPP OR NOT DEFINED OUTPUT)
-    message(FATAL_ERROR "HEDLEY_HPP and OUTPUT must be set")
-endif()
-
-if(NOT EXISTS "${HEDLEY_HPP}")
-    message(FATAL_ERROR "Hedley header not found: ${HEDLEY_HPP}")
-endif()
-
-if(NOT DEFINED MODE)
-    set(MODE undef)
-endif()
-
-if(NOT MODE STREQUAL "undef" AND NOT MODE STREQUAL "checks")
-    message(FATAL_ERROR "MODE must be undef or checks, got: ${MODE}")
-endif()
-
-# Line-anchored, like `grep -oE "^[[:blank:]]*#[[:blank:]]*define[[:blank:]]+JSON_HEDLEY_[A-Za-z0-9_]+"`.
-# Unanchored matching would also pick up JSON_HEDLEY_* mentions inside
-# comments or string literals elsewhere in the file, which must not turn
-# into #undef lines.
-file(STRINGS "${HEDLEY_HPP}" hedley_lines)
-set(macro_names)
-foreach(line IN LISTS hedley_lines)
-    if("${line}" MATCHES "^[ \t]*#[ \t]*define[ \t]+(JSON_HEDLEY_[A-Za-z0-9_]+)")
-        list(APPEND macro_names "${CMAKE_MATCH_1}")
-    endif()
-endforeach()
-
-if(NOT macro_names)
-    message(FATAL_ERROR "No JSON_HEDLEY_* macros found in ${HEDLEY_HPP}")
-endif()
-
-list(REMOVE_DUPLICATES macro_names)
-# Lexicographic, locale-independent (ASCII-only names) -- matches `LC_ALL=C sort`.
-list(SORT macro_names COMPARE STRING)
-list(LENGTH macro_names macro_count)
-
-set(generated "")
-if(MODE STREQUAL "undef")
-    # Same banner `make update_hedley_undef` would stamp by hand, so the
-    # recipe is self-contained and its output is byte-stable across reruns.
-    string(APPEND generated "//     __ _____ _____ _____\n")
-    string(APPEND generated "//  __|  |   __|     |   | |  JSON for Modern C++\n")
-    string(APPEND generated "// |  |  |__   |  |  | | | |  version 3.12.0\n")
-    string(APPEND generated "// |_____|_____|_____|_|___|  https://github.com/nlohmann/json\n")
-    string(APPEND generated "//\n")
-    string(APPEND generated "// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>\n")
-    string(APPEND generated "// SPDX-License-Identifier: MIT\n")
-    string(APPEND generated "\n")
-    string(APPEND generated "#pragma once\n")
-    string(APPEND generated "\n")
-    foreach(name IN LISTS macro_names)
-        string(APPEND generated "#undef ${name}\n")
-    endforeach()
-else()
-    string(APPEND generated "// This file is generated by cmake/scripts/gen_hedley_undef_check.cmake\n")
-    string(APPEND generated "// from include/nlohmann/thirdparty/hedley/hedley.hpp. Do not edit it by\n")
-    string(APPEND generated "// hand -- it is regenerated on every build. ${macro_count} macros checked.\n\n")
-    foreach(name IN LISTS macro_names)
-        string(APPEND generated "#ifdef ${name}\n")
-        string(APPEND generated "    FAIL_CHECK(\"${name} leaked after including nlohmann/json.hpp\");\n")
-        string(APPEND generated "#endif\n")
-    endforeach()
-endif()
-
-get_filename_component(output_dir "${OUTPUT}" DIRECTORY)
-if(output_dir)
-    file(MAKE_DIRECTORY "${output_dir}")
-endif()
-
-# Avoid rewriting the file (and busting downstream incremental rebuilds)
-# when the content has not actually changed.
-set(write_output TRUE)
-if(EXISTS "${OUTPUT}")
-    file(READ "${OUTPUT}" existing_content)
-    if(existing_content STREQUAL generated)
-        set(write_output FALSE)
-    endif()
-endif()
-if(write_output)
-    file(WRITE "${OUTPUT}" "${generated}")
-endif()

include/nlohmann/thirdparty/hedley/hedley_undef.hpp

@@ -17,7 +17,7 @@
 #undef JSON_HEDLEY_CLANG_HAS_ATTRIBUTE
 #undef JSON_HEDLEY_CLANG_HAS_BUILTIN
 #undef JSON_HEDLEY_CLANG_HAS_CPP_ATTRIBUTE
-#undef JSON_HEDLEY_CLANG_HAS_DECLSPEC_ATTRIBUTE
+#undef JSON_HEDLEY_CLANG_HAS_DECLSPEC_DECLSPEC_ATTRIBUTE
 #undef JSON_HEDLEY_CLANG_HAS_EXTENSION
 #undef JSON_HEDLEY_CLANG_HAS_FEATURE
 #undef JSON_HEDLEY_CLANG_HAS_WARNING
@@ -108,10 +108,7 @@
 #undef JSON_HEDLEY_PELLES_VERSION_CHECK
 #undef JSON_HEDLEY_PGI_VERSION
 #undef JSON_HEDLEY_PGI_VERSION_CHECK
-#undef JSON_HEDLEY_PRAGMA
 #undef JSON_HEDLEY_PREDICT
-#undef JSON_HEDLEY_PREDICT_FALSE
-#undef JSON_HEDLEY_PREDICT_TRUE
 #undef JSON_HEDLEY_PRINTF_FORMAT
 #undef JSON_HEDLEY_PRIVATE
 #undef JSON_HEDLEY_PUBLIC

single_include/nlohmann/json.hpp

@@ -26965,7 +26965,7 @@ struct formatter<nlohmann::NLOHMANN_BASIC_JSON_TPL, char> // NOLINT(cert-dcl58-c
 #undef JSON_HEDLEY_CLANG_HAS_ATTRIBUTE
 #undef JSON_HEDLEY_CLANG_HAS_BUILTIN
 #undef JSON_HEDLEY_CLANG_HAS_CPP_ATTRIBUTE
-#undef JSON_HEDLEY_CLANG_HAS_DECLSPEC_ATTRIBUTE
+#undef JSON_HEDLEY_CLANG_HAS_DECLSPEC_DECLSPEC_ATTRIBUTE
 #undef JSON_HEDLEY_CLANG_HAS_EXTENSION
 #undef JSON_HEDLEY_CLANG_HAS_FEATURE
 #undef JSON_HEDLEY_CLANG_HAS_WARNING
@@ -27056,10 +27056,7 @@ struct formatter<nlohmann::NLOHMANN_BASIC_JSON_TPL, char> // NOLINT(cert-dcl58-c
 #undef JSON_HEDLEY_PELLES_VERSION_CHECK
 #undef JSON_HEDLEY_PGI_VERSION
 #undef JSON_HEDLEY_PGI_VERSION_CHECK
-#undef JSON_HEDLEY_PRAGMA
 #undef JSON_HEDLEY_PREDICT
-#undef JSON_HEDLEY_PREDICT_FALSE
-#undef JSON_HEDLEY_PREDICT_TRUE
 #undef JSON_HEDLEY_PRINTF_FORMAT
 #undef JSON_HEDLEY_PRIVATE
 #undef JSON_HEDLEY_PUBLIC

tests/CMakeLists.txt

@@ -125,51 +125,6 @@ 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)
@@ -208,14 +163,6 @@ 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()

tests/src/unit-bson.cpp

@@ -38,6 +38,26 @@ 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 reports 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
+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)
+
+    size_type size() const noexcept // NOLINT(readability-convert-member-functions-to-static)
+    {
+        // one byte more than the BSON length field can represent
+        return static_cast<size_type>((std::numeric_limits<std::int32_t>::max)()) + 1;
+    }
+};
+
+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 +125,36 @@ 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_json::string_t("value");
+
+            CHECK_THROWS_WITH_AS(huge_string_json::to_bson(j), "[json.exception.out_of_range.412] BSON length 4294967308 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_json::string_t("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 6442450963 exceeds maximum of 2147483647", huge_string_json::out_of_range&);
+        }
     }
 
     SECTION("string length must be at least 1")
@@ -193,6 +239,23 @@ 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 +562,29 @@ 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 +680,31 @@ 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 =

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

@@ -1,34 +0,0 @@
-//     __ _____ _____ _____
-//  __|  |   __|     |   | |  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
-}