Avoid deep recursion in serialization write-buffer test

The "many small structural writes exceed the write buffer" subcase built
a 1100-deep nested array and dumped it to force >1024 consecutive
single-character writes through put_char (exercising the write buffer's
flush-when-full branch). dump() recurses per nesting level, so on MSVC
debug builds (smaller default stack, larger frames) this overflowed the
stack and crashed test-serialization; Linux/macOS have enough headroom to
hide it.

Replace the nesting with a flat array of 500 empty strings. Each element
emits '"', '"', ',' via put_char, so the dump is a long run of
single-character writes (1501 bytes > the 1024-byte buffer) at nesting
depth two, hitting the same flush branch without deep recursion. Library
code is unchanged.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XAYM1qhSA2FDaDcGfPW3fG
Signed-off-by: Niels Lohmann <mail@nlohmann.me>

.github/workflows/codeql-analysis.yml

@@ -38,14 +38,14 @@ jobs:
 
     # Initializes the CodeQL tools for scanning.
     - name: Initialize CodeQL
-      uses: github/codeql-action/init@e0647621c2984b5ed2f768cb892365bf2a616ad1 # v4.37.2
+      uses: github/codeql-action/init@e4fba868fa4b1b91e1fdab776edc8cfbe6e9fb81 # v4.37.3
       with:
         languages: c-cpp
 
     # Autobuild attempts to build any compiled languages  (C/C++, C#, or Java).
     # If this step fails, then you should remove it and run the build manually (see below)
     - name: Autobuild
-      uses: github/codeql-action/autobuild@e0647621c2984b5ed2f768cb892365bf2a616ad1 # v4.37.2
+      uses: github/codeql-action/autobuild@e4fba868fa4b1b91e1fdab776edc8cfbe6e9fb81 # v4.37.3
 
     - name: Perform CodeQL Analysis
-      uses: github/codeql-action/analyze@e0647621c2984b5ed2f768cb892365bf2a616ad1 # v4.37.2
+      uses: github/codeql-action/analyze@e4fba868fa4b1b91e1fdab776edc8cfbe6e9fb81 # v4.37.3

.github/workflows/flawfinder.yml

@@ -43,6 +43,6 @@ jobs:
           output: 'flawfinder_results.sarif'
 
       - name: Upload analysis results to GitHub Security tab
-        uses: github/codeql-action/upload-sarif@e0647621c2984b5ed2f768cb892365bf2a616ad1 # v4.37.2
+        uses: github/codeql-action/upload-sarif@e4fba868fa4b1b91e1fdab776edc8cfbe6e9fb81 # v4.37.3
         with:
           sarif_file: ${{github.workspace}}/flawfinder_results.sarif

.github/workflows/scorecards.yml

@@ -76,6 +76,6 @@ jobs:
 
       # Upload the results to GitHub's code scanning dashboard.
       - name: "Upload to code-scanning"
-        uses: github/codeql-action/upload-sarif@e0647621c2984b5ed2f768cb892365bf2a616ad1 # v4.37.2
+        uses: github/codeql-action/upload-sarif@e4fba868fa4b1b91e1fdab776edc8cfbe6e9fb81 # v4.37.3
         with:
           sarif_file: results.sarif

.github/workflows/semgrep.yml

@@ -61,7 +61,7 @@ jobs:
 
       # Upload SARIF file generated in previous step
       - name: Upload SARIF file
-        uses: github/codeql-action/upload-sarif@e0647621c2984b5ed2f768cb892365bf2a616ad1 # v4.37.2
+        uses: github/codeql-action/upload-sarif@e4fba868fa4b1b91e1fdab776edc8cfbe6e9fb81 # v4.37.3
         with:
           sarif_file: semgrep.sarif
         if: always()

.github/workflows/stale.yml

@@ -20,7 +20,7 @@ jobs:
         with:
           egress-policy: audit
 
-      - uses: actions/stale@1e223db275d687790206a7acac4d1a11bd6fe629 # v10.4.0
+      - uses: actions/stale@4391f3da665fdf50b6810c1a66712fb9ba21aa93 # v11.0.0
         with:
           stale-issue-label: 'state: stale'
           stale-pr-label: 'state: stale'

.github/workflows/ubuntu.yml

@@ -142,7 +142,7 @@ jobs:
           name: code-coverage-report
           path: ${{ github.workspace }}/build/html
       - name: Publish report to Coveralls
-        uses: coverallsapp/github-action@5cbfd81b66ca5d10c19b062c04de0199c215fb6e # v2.3.7
+        uses: coverallsapp/github-action@8d6379e14d29928660c4ba802d8e85393440b329 # v2.3.8
         with:
           github-token: ${{ secrets.GITHUB_TOKEN }}
           path-to-lcov: ${{ github.workspace }}/build/json.info.filtered.noexcept

README.md

@@ -1801,13 +1801,13 @@ The library itself consists of a single header file licensed under the MIT licen
 - [**amalgamate.py - Amalgamate C source and header files**](https://github.com/edlund/amalgamate) to create a single header file
 - [**American fuzzy lop**](https://lcamtuf.coredump.cx/afl/) for fuzz testing
 - [**AppVeyor**](https://www.appveyor.com) for [continuous integration](https://ci.appveyor.com/project/nlohmann/json) on Windows
-- [**Artistic Style**](http://astyle.sourceforge.net) for automatic source code indentation
+- [**Artistic Style**](https://astyle.sourceforge.net) for automatic source code indentation
 - [**Clang**](https://clang.llvm.org) for compilation with code sanitizers
 - [**CMake**](https://cmake.org) for build automation
 - [**Codacy**](https://www.codacy.com) for further [code analysis](https://app.codacy.com/gh/nlohmann/json/dashboard)
 - [**Coveralls**](https://coveralls.io) to measure [code coverage](https://coveralls.io/github/nlohmann/json)
 - [**Coverity Scan**](https://scan.coverity.com) for [static analysis](https://scan.coverity.com/projects/nlohmann-json)
-- [**cppcheck**](http://cppcheck.sourceforge.net) for static analysis
+- [**cppcheck**](https://cppcheck.sourceforge.io) for static analysis
 - [**doctest**](https://github.com/onqtam/doctest) for the unit tests
 - [**GitHub Changelog Generator**](https://github.com/skywinder/github-changelog-generator) to generate the [ChangeLog](https://github.com/nlohmann/json/blob/develop/ChangeLog.md)
 - [**Google Benchmark**](https://github.com/google/benchmark) to implement the benchmarks
@@ -1822,6 +1822,15 @@ The library itself consists of a single header file licensed under the MIT licen
 
 ## Notes
 
+### Standards compliance
+
+The library targets strict conformance with [RFC 8259](https://tools.ietf.org/html/rfc8259.html). Both the original [JSONTestSuite](https://github.com/nst/JSONTestSuite) and its updated revision are exercised in CI; their test data is downloaded from [`nlohmann/json_test_data`](https://github.com/nlohmann/json_test_data) at configure time rather than committed to this repository (see [`tests/src/unit-testsuites.cpp`](https://github.com/nlohmann/json/blob/develop/tests/src/unit-testsuites.cpp)):
+
+- The updated revision runs all mandatory `y_` (must-accept) and `n_` (must-reject) cases through the strict [`parse()`](https://json.nlohmann.me/api/basic_json/parse/) entry point; the original suite runs its `n_` cases through `parse()` and its `y_` cases through [`operator>>`](https://json.nlohmann.me/api/operator_gtgt/).
+- The `i_` (implementation-defined) cases are, by RFC 8259, free to be accepted *or* rejected, so "passing all `i_` cases" is not a meaningful conformance metric. The library makes deliberate, documented choices there: nesting depth is not artificially limited, a leading UTF-8 byte order mark is silently ignored, [Unicode noncharacters](https://www.unicode.org/faq/private_use.html#nonchar1) are forwarded unchanged, invalid UTF-8 and lone/unpaired UTF-16 surrogates are rejected (stricter than required), and a number that cannot be stored without becoming `NaN`/`INF` raises [`out_of_range.406`](https://json.nlohmann.me/home/exceptions/#jsonexceptionout_of_range406).
+
+One behavioral nuance is worth calling out, because a superficial test often misreads it as non-compliance: [`parse()`](https://json.nlohmann.me/api/basic_json/parse/) is strict and rejects trailing data after a value, whereas [`operator>>`](https://json.nlohmann.me/api/operator_gtgt/) follows relaxed iostream semantics — it parses a single value and leaves the stream positioned right after it. Feeding "a valid document followed by trailing bytes" through `operator>>` reports success; the same input through `parse()` is rejected. This is a documented two-API design, not a conformance gap. See [**parsing**](https://json.nlohmann.me/features/parsing/) for details.
+
 ### Character encoding
 
 The library supports **Unicode input** as follows:

cmake/clang_flags.cmake

@@ -5,6 +5,9 @@
 # -Wno-extra-semi-stmt            The library uses assert which triggers this warning.
 # -Wno-padded                     We do not care about padding warnings.
 # -Wno-covered-switch-default     All switches list all cases and a default case.
+# -Wno-c2y-extensions             Clang 22.1 diagnoses __COUNTER__ as a C2y extension, also in
+#                                 C++ mode. The library does not use __COUNTER__; the warnings
+#                                 all come from vendored Doctest (SECTION/TEST_CASE macros).
 # -Wno-unsafe-buffer-usage        Pervasive: the library's own low-level numeric/buffer code
 #                                 (to_chars, serializer, lexer, binary reader/writer, input
 #                                 adapters, json_pointer) plus vendored Doctest itself (~208
@@ -20,5 +23,6 @@ set(CLANG_CXXFLAGS
     -Wno-extra-semi-stmt
     -Wno-padded
     -Wno-covered-switch-default
+    -Wno-c2y-extensions
     -Wno-unsafe-buffer-usage
 )

docs/mkdocs/docs/api/basic_json/parser_callback_t.md

@@ -29,7 +29,14 @@ Discarding a value (i.e., returning `#!cpp false`) has different effects dependi
 called:
 
 - Discarded values in structured types are skipped. That is, the parser will behave as if the discarded value was never
-  read.
+  read. This holds for every value type and for both kinds of parent: a discarded element is removed from the
+  surrounding array, and a discarded member is removed from the surrounding object together with its key.
+- Arrays and objects can be discarded either at their `parse_event_t::array_start`/`parse_event_t::object_start` event
+  or at their `parse_event_t::array_end`/`parse_event_t::object_end` event, and both remove the whole value. Discarding
+  it at the start event also means the callback is called neither for the content of the value nor for its matching end
+  event.
+- Discarding a `parse_event_t::key` event discards the whole object member. The callback is still called for the
+  associated value, but its return value has no further effect.
 - In case a value outside a structured type is skipped, it is replaced with `null`. This case happens if the top-level
   element is skipped.
 
@@ -49,7 +56,7 @@ called:
 ## Return value
 
 Whether the JSON value which called the function during parsing should be kept (`#!cpp true`) or not (`#!cpp false`). In
-the latter case, it is either skipped completely or replaced by an empty discarded object.
+the latter case, it is skipped completely, or replaced by `null` if it is the top-level value.
 
 ## Examples
 
@@ -68,6 +75,21 @@ the latter case, it is either skipped completely or replaced by an empty discard
     --8<-- "examples/parse__string__parser_callback_t.output"
     ```
 
+??? example
+
+    The example below shows where discarded values are removed. The array and the number are discarded in different
+    ways, but in each case the parse result contains neither the value nor its key.
+
+    ```cpp
+    --8<-- "examples/parser_callback_t.cpp"
+    ```
+
+    Output:
+
+    ```json
+    --8<-- "examples/parser_callback_t.output"
+    ```
+
 ## See also
 
 - [parse](parse.md) deserialize from a compatible input
@@ -76,3 +98,5 @@ the latter case, it is either skipped completely or replaced by an empty discard
 ## Version history
 
 - Added in version 1.0.0.
+- Fixed in version 3.13.0 to also remove discarded values from a parent object; before, discarding an array or a value
+  stored under an object key left a discarded member behind, which made the parse result serialize to invalid JSON.

docs/mkdocs/docs/api/macros/index.md

@@ -24,6 +24,7 @@ header. See also the [macro overview page](../../features/macros.md).
 - [**JSON_NO_IO**](json_no_io.md) - switch off functions relying on certain C++ I/O headers
 - [**JSON_SKIP_UNSUPPORTED_COMPILER_CHECK**](json_skip_unsupported_compiler_check.md) - do not warn about unsupported compilers
 - [**JSON_USE_GLOBAL_UDLS**](json_use_global_udls.md) - place user-defined string literals (UDLs) into the global namespace
+- [**JSON_USE_SIMDUTF**](json_use_simdutf.md) - use the simdutf library to accelerate UTF-8 validation
 
 ## Library version
 

docs/mkdocs/docs/api/macros/json_use_simdutf.md

@@ -0,0 +1,52 @@
+# JSON_USE_SIMDUTF
+
+```cpp
+#define JSON_USE_SIMDUTF
+```
+
+When defined, the parser validates the UTF-8 content of JSON strings that come from a **contiguous byte input**
+(`std::string`, `std::vector<char>`/`<std::uint8_t>`, string literals, `const char*` ranges, …) using the
+[simdutf](https://github.com/simdutf/simdutf) library instead of the built-in scalar validator. On text with many
+non-ASCII characters (e.g. CJK or emoji) this can validate several times faster.
+
+This is an **opt-in external dependency**. The library itself remains header-only and its behavior is unchanged: the
+same input is accepted or rejected either way, and every parse error is reported at the same position with the same
+message (simdutf is only used to fast-path *valid* runs; anything it flags falls back to the scalar path so the exact
+diagnostic is preserved). Streaming inputs (files, `std::istream`, wide strings, user-defined adapters) always use the
+scalar path.
+
+When `JSON_USE_SIMDUTF` is defined you must make the `simdutf.h` header available on the include path and link the
+simdutf library. When it is not defined, no simdutf header is included and there is no dependency.
+
+## Default definition
+
+By default, `#!cpp JSON_USE_SIMDUTF` is not defined and the portable C++11 scalar validator is used.
+
+```cpp
+#undef JSON_USE_SIMDUTF
+```
+
+## Examples
+
+??? example
+
+    The code below enables the simdutf backend for UTF-8 validation.
+
+    ```cpp
+    #define JSON_USE_SIMDUTF 1
+    #include <simdutf.h>
+    #include <nlohmann/json.hpp>
+
+    ...
+    ```
+
+    The project must also link against simdutf, e.g. with CMake:
+
+    ```cmake
+    target_compile_definitions(your_target PRIVATE JSON_USE_SIMDUTF)
+    target_link_libraries(your_target PRIVATE simdutf::simdutf)
+    ```
+
+## Version history
+
+- Added in version 3.12.1.

docs/mkdocs/docs/api/operator_gtgt.md

@@ -33,17 +33,44 @@ A UTF-8 byte order mark is silently ignored.
 Invalid Unicode escapes and unpaired surrogates in the input are reported as
 [`parse_error.101`](../home/exceptions.md#jsonexceptionparse_error101) with a detailed message.
 
-`operator>>` parses exactly one JSON value and leaves the stream positioned right after it, so it can be called
+`operator>>` parses exactly one JSON value, so it can be called repeatedly to read a sequence of concatenated JSON
-repeatedly to read a sequence of concatenated JSON values from the same stream:
+values from the same stream:
 
 ```cpp
 json j1, j2;
-input >> j1;  // parses the first value, stream now positioned right after it
+input >> j1;  // parses the first value
 input >> j2;  // parses the next value
 ```
 
-Note this does **not** work for [JSON Lines](../features/parsing/json_lines.md) (newline-delimited JSON) input --
+!!! warning "A number must be followed by whitespace"
-see that page for why and for the recommended alternative.
+
+    A number is only terminated by the character that follows it. That character is read from the stream to detect the
+    end of the number, and it is **not** put back. When a value that is a number is immediately followed by the next
+    value, the first character of that next value is lost:
+
+    ```cpp
+    std::istringstream input("1true");
+    json j1, j2;
+    input >> j1;  // j1 == 1
+    input >> j2;  // throws parse_error.101: the stream now starts at "rue"
+    ```
+
+    Separating the values with whitespace avoids this, because the character that is eaten is then the separator:
+
+    ```cpp
+    std::istringstream input("1 true");
+    json j1, j2;
+    input >> j1;  // j1 == 1
+    input >> j2;  // j2 == true
+    ```
+
+    Only numbers are affected. Values ending in a self-delimiting character do not read past themselves, so
+    `truefalse`, `[1][2]`, `{"a":1}{"b":2}`, and `"a""b"` can be read back to back without a separator.
+
+    This is tracked in [#5340](https://github.com/nlohmann/json/issues/5340).
+
+Note that reading concatenated values does **not** work for [JSON Lines](../features/parsing/json_lines.md)
+(newline-delimited JSON) input -- see that page for why and for the recommended alternative.
 
 !!! warning "Deprecation"
 

docs/mkdocs/docs/api/ordered_json.md

@@ -13,6 +13,12 @@ Therefore, adding object elements can yield a reallocation in which case all ite
 [`end()`](basic_json/end.md) iterator) and all references to the elements are invalidated. Also, any iterator or
 reference after the insertion point will point to the same index, which is now a different value.
 
+## Complexity
+
+[`ordered_map`](ordered_map.md) has no lookup index: every key-based object operation is a linear scan, so building or
+parsing an object of `n` keys costs O(n²) rather than O(n log n). See
+[`ordered_map` complexity](ordered_map.md#complexity) for the per-operation table and for measured numbers.
+
 ## Examples
 
 ??? example

docs/mkdocs/docs/api/ordered_map.md

@@ -56,6 +56,48 @@ std::equal_to<>     // since C++14
 - **find**
 - **insert**
 
+## Complexity
+
+Because the elements are stored in a `std::vector` in insertion order, there is no index to look a key up by. Every
+key-based operation performs a **linear scan** over the stored elements. With `n` denoting the number of elements in the
+container:
+
+| Operation                              | Complexity     | Note                                                     |
+|----------------------------------------|----------------|----------------------------------------------------------|
+| **emplace**                            | O(n)           | scans for an existing key, then appends (amortized O(1)) |
+| **operator\[\]**                       | O(n)           | delegates to **emplace** (non-const) or **at** (const)   |
+| **at**                                 | O(n)           | throws `#!cpp std::out_of_range` if the key is not found |
+| **find**                               | O(n)           |                                                          |
+| **count**                              | O(n)           | the result is always 0 or 1                              |
+| **erase(key)**                         | O(n)           | scan, then move the remaining elements one position down |
+| **erase(pos)**, **erase(first, last)** | O(n)           | moves all elements after the erased range                |
+| **insert(value)**                      | O(n)           | equivalent to **emplace**                                |
+| **insert(first, last)**                | O((n + m) * m) | for `m` inserted elements                                |
+
+This differs from `#!cpp std::map`, where the same operations are O(log n).
+
+!!! warning "Quadratic cost of building large objects"
+
+    Because every insertion scans all elements inserted so far, building an object of `n` distinct keys costs
+    **O(n²)** in total. This applies to filling an [`ordered_json`](ordered_json.md) object key by key as well as to
+    parsing one, since the parser inserts each key as it is read.
+
+    The cost is negligible for the object sizes typically found in configuration files or API payloads, but it grows
+    steeply for machine-generated objects with many thousands of keys. Measured with `-O2 -DNDEBUG` for parsing a flat
+    object of `n` keys, relative to `#!cpp nlohmann::json` (which uses `#!cpp std::map`):
+
+    | `n`    | `json` | `ordered_json` | factor |
+    |--------|--------|----------------|--------|
+    | 2000   | 0.7 ms | 3.6 ms         | 5×     |
+    | 4000   | 0.8 ms | 14.0 ms        | 19×    |
+    | 8000   | 1.6 ms | 67.8 ms        | 43×    |
+    | 16 000 | 3.3 ms | 181.6 ms       | 54×    |
+
+    If key order matters for objects of that size, consider a container with a lookup index, such as
+    [`tsl::ordered_map`](https://github.com/Tessil/ordered-map)
+    ([integration](https://github.com/nlohmann/json/issues/546#issuecomment-304447518)), as the object type -- see
+    [object order](../features/object_order.md).
+
 ## Examples
 
 ??? example

docs/mkdocs/docs/examples/parser_callback_t.cpp

@@ -0,0 +1,47 @@
+#include <iostream>
+#include <nlohmann/json.hpp>
+
+using json = nlohmann::json;
+
+int main()
+{
+    // a JSON text with an array and a number inside an object
+    auto text = R"({"IDs": [116, 943], "Width": 800})";
+
+    // discard the array when the parser reads its opening bracket
+    json j_array_start = json::parse(text, [](int /*depth*/, json::parse_event_t event, json & /*parsed*/)
+    {
+        return event != json::parse_event_t::array_start;
+    });
+
+    // discard the same array when the parser reads its closing bracket
+    json j_array_end = json::parse(text, [](int /*depth*/, json::parse_event_t event, json & /*parsed*/)
+    {
+        return event != json::parse_event_t::array_end;
+    });
+
+    // discard the number, but keep its key
+    json j_value = json::parse(text, [](int /*depth*/, json::parse_event_t event, json & parsed)
+    {
+        return !(event == json::parse_event_t::value && parsed == json(800));
+    });
+
+    // discard the key of the number
+    json j_key = json::parse(text, [](int /*depth*/, json::parse_event_t event, json & parsed)
+    {
+        return !(event == json::parse_event_t::key && parsed == json("Width"));
+    });
+
+    // discard the top-level object
+    json j_root = json::parse(text, [](int /*depth*/, json::parse_event_t event, json & /*parsed*/)
+    {
+        return event != json::parse_event_t::object_end;
+    });
+
+    // in every case, the discarded value is removed together with its key
+    std::cout << j_array_start << '\n'
+              << j_array_end << '\n'
+              << j_value << '\n'
+              << j_key << '\n'
+              << j_root << '\n';
+}

docs/mkdocs/docs/examples/parser_callback_t.output

@@ -0,0 +1,5 @@
+{"Width":800}
+{"Width":800}
+{"IDs":[116,943]}
+{"IDs":[116,943]}
+null

docs/mkdocs/docs/features/binary_formats/bson.md

@@ -35,6 +35,19 @@ The library uses the following mapping from JSON values types to BSON types:
     The mapping is **incomplete**, since only JSON-objects (and things contained therein) can be serialized to BSON.
     Also, keys may not contain U+0000, since they are serialized a zero-terminated c-strings.
 
+!!! warning "BSON type 0x11 interoperability"
+
+    The BSON specification defines type `0x11` as a Timestamp. This library uses marker `0x11` when serializing
+    `number_unsigned` values in the range `9223372036854775808..18446744073709551615`. Other BSON implementations may
+    therefore interpret these values as Timestamps instead of unsigned integers.
+
+!!! info "Binary values without a subtype"
+
+    BSON requires every binary value to have a subtype. If a binary value has no subtype, this library serializes it
+    with the generic subtype `0x00`. After deserialization, `has_subtype()` returns `true` and `subtype()` returns `0`.
+    As a result, serializing and deserializing a JSON object containing such a value produces a different JSON object,
+    even though the binary data is unchanged.
+
 ??? example
 
     ```cpp
@@ -82,8 +95,8 @@ The library maps BSON record types to JSON value types as follows:
 
 !!! note "Handling of BSON type 0x11"
 
-    BSON type 0x11 is used to represent uint64 numbers. This library treats these values purely as uint64 numbers 
+    This library deserializes BSON type `0x11` (Timestamp) as a `number_unsigned` value. The 64-bit value is preserved,
-    and does not parse them into date-related formats.
+    but the Timestamp type information is not.
 
 ??? example
 

docs/mkdocs/docs/features/conversions.md

@@ -66,8 +66,14 @@ auto t = j.get<std::tuple<double, std::string, int>>();  // {1.0, "hello", 42}
     std::get<1>(refs) = "world";  // modifies j[1] in place
     ```
 
-    A referenced type must be one the library actually stores (or an arithmetic type it can convert to/from);
+    A referenced element must name the type the library actually *stores* — one of [`boolean_t`](../api/basic_json/boolean_t.md),
-    otherwise this is a compile error.
+    [`number_integer_t`](../api/basic_json/number_integer_t.md), [`number_unsigned_t`](../api/basic_json/number_unsigned_t.md),
+    [`number_float_t`](../api/basic_json/number_float_t.md), [`string_t`](../api/basic_json/string_t.md),
+    [`binary_t`](../api/basic_json/binary_t.md), [`array_t`](../api/basic_json/array_t.md), or
+    [`object_t`](../api/basic_json/object_t.md). There is nothing else to refer to, so a reference to any other type is a
+    compile error even when a conversion would exist: `#!cpp std::tuple<int&>` is rejected, because the library stores a
+    `#!cpp number_integer_t` (`#!cpp std::int64_t` by default) and not an `#!cpp int`. This restriction applies only to
+    reference elements — a plain `#!cpp std::tuple<int>` converts by value as usual.
 
 ## Implicit conversions
 
@@ -116,17 +122,34 @@ which forces the explicit `get` form and can catch unintended conversions at com
     with a custom `adl_serializer<std::optional<T>>` specialization. Prefer `get<std::optional<T>>()`/`get_to()`
     over `static_cast` for optional types.
 
-!!! warning "Converting to a fixed-size `std::array` does not check length"
+!!! warning "Converting to a fixed-size destination does not check the array size"
 
-    Converting a JSON array to `#!cpp std::array<T, N>` does not check that the JSON array's size matches `N`:
+    Some destination types have a size that is fixed by their C++ type rather than by the JSON value:
-    if the JSON array is longer, the extra elements are silently dropped; if it is shorter, the remaining
+    `#!cpp std::pair<A, B>`, `#!cpp std::tuple<Ts...>`, `#!cpp std::array<T, N>`, C arrays `#!cpp T[N]`, and
-    `std::array` elements are left default-constructed. No exception is thrown in either case.
+    `#!cpp std::map`/`#!cpp std::unordered_map` with a non-string key type (which is read from an array of
+    two-element arrays). All of them read exactly as many elements as they need via
+    [`at`](../api/basic_json/at.md) and **never compare the JSON array's size to that number**. The two
+    mismatch directions therefore behave differently:
+
+    - The JSON array has **too many** elements: the surplus is **silently discarded**, and no exception is
+      thrown.
+    - The JSON array has **too few** elements: `at` throws
+      [`out_of_range.401`](../home/exceptions.md#jsonexceptionout_of_range401) for the first missing index --
+      an out-of-range error, not a [`type_error`](../home/exceptions.md#type-errors), even though the cause
+      is a shape mismatch.
 
     ```cpp
     json j = {1, 2, 3, 4, 5};
-    auto a = j.get<std::array<int, 3>>();  // {1, 2, 3} -- elements 4 and 5 silently dropped
+
+    auto a = j.get<std::array<int, 3>>();       // {1, 2, 3} -- elements 4 and 5 silently dropped
+    auto p = j.get<std::pair<int, int>>();      // (1, 2)    -- elements 3, 4, and 5 silently dropped
+
+    json k = {1};
+    auto q = k.get<std::pair<int, int>>();      // ❌ throws out_of_range.401
     ```
 
+    If a size mismatch is an error in your application, check the size yourself before converting.
+
 ## Omitting a field when serializing `std::optional`
 
 By default, `to_json` for `std::optional<T>` writes either the value or `#!json null` -- there is no built-in way

docs/mkdocs/docs/features/object_order.md

@@ -53,6 +53,12 @@ If you do want to preserve the **insertion order**, you can use the type [`nlohm
 
 Alternatively, you can use a more sophisticated ordered map like [`tsl::ordered_map`](https://github.com/Tessil/ordered-map) ([integration](https://github.com/nlohmann/json/issues/546#issuecomment-304447518)) or [`nlohmann::fifo_map`](https://github.com/nlohmann/fifo_map) ([integration](https://github.com/nlohmann/json/issues/485#issuecomment-333652309)).
 
+The [`ordered_map`](../api/ordered_map.md) behind `nlohmann::ordered_json` is deliberately minimal and has no lookup
+index, so every key access is a linear scan and building an object of `n` keys costs O(n²). This is unnoticeable at
+typical object sizes but becomes significant for objects with many thousands of keys; see
+[`ordered_map` complexity](../api/ordered_map.md#complexity). The alternatives above keep a lookup index and do not
+have this cost.
+
 ### Notes on parsing
 
 Note that you also need to call the right [`parse`](../api/basic_json/parse.md) function when reading from a file.

docs/mkdocs/docs/features/parsing/index.md

@@ -28,6 +28,22 @@ Inputs consisting of multiple values separated by newlines are handled by the [J
 By default, the library rejects comments and trailing commas. Both can be enabled with parameters of the `parse`
 function — see [comments](../comments.md) and [trailing commas](../trailing_commas.md).
 
+## Strictness and trailing data
+
+[`parse`](../../api/basic_json/parse.md) reads a single JSON value and requires the whole input to be consumed: any
+non-whitespace data after the value is reported as a parse error. Use it when you want to guarantee that an input is
+exactly one complete JSON document.
+
+[`operator>>`](../../api/operator_gtgt.md) follows relaxed `#!cpp std::istream` semantics instead: it parses one JSON
+value and leaves the stream positioned right after it, without requiring the rest of the stream to be consumed. This is
+what makes it possible to read several concatenated values from the same stream, but it also means that "a valid
+document followed by trailing bytes" is accepted rather than rejected. If you are validating conformance, or need to
+reject any input that is not exactly one JSON document, prefer `parse`.
+
+When using `operator>>` to read several concatenated values this way, a value that is a number must be followed by
+whitespace, because `operator>>` consumes the character that terminates a number — see the
+[`operator>>` notes](../../api/operator_gtgt.md#notes) for details and examples.
+
 ## SAX vs. DOM parsing
 
 The library offers two parsing models:

docs/mkdocs/docs/features/parsing/json_lines.md

@@ -49,4 +49,5 @@ JSON Lines input with more than one value is treated as invalid JSON by the [`pa
     with a JSON Lines input does not work, because the parser will try to parse one value after the last one.
 
     This is different from parsing a stream of *concatenated* (non-newline-delimited) JSON values, for which
-    `operator>>` does work -- see its [notes](../../api/operator_gtgt.md#notes) for details.
+    `operator>>` does work, provided that a value that is a number is followed by whitespace -- see its
+    [notes](../../api/operator_gtgt.md#notes) for details.

docs/mkdocs/docs/home/exceptions.md

@@ -291,9 +291,10 @@ A JSON Pointer array index must be a number.
 
 ### json.exception.parse_error.110
 
-When parsing CBOR or MessagePack, the byte vector ends before the complete value has been read.
+When parsing a [binary format](../features/binary_formats/index.md), the byte vector ends before the complete value has
+been read.
 
-!!! failure "Example message"
+!!! failure "Example messages"
 
     ```
     [json.exception.parse_error.110] parse error at byte 5: syntax error while parsing CBOR string: unexpected end of input
@@ -301,6 +302,9 @@ When parsing CBOR or MessagePack, the byte vector ends before the complete value
     ```
     [json.exception.parse_error.110] parse error at byte 2: syntax error while parsing UBJSON value: expected end of input; last byte: 0x5A
     ```
+    ```
+    [json.exception.parse_error.110] parse error at byte 8: syntax error while parsing BSON number: unexpected end of input
+    ```
 
 ### json.exception.parse_error.112
 
@@ -329,10 +333,14 @@ An unexpected byte was read in a [binary format](../features/binary_formats/inde
     ```
     [json.exception.parse_error.112] parse error at byte 9: syntax error while parsing CBOR value: negative integer overflow
     ```
+    ```
+    [json.exception.parse_error.112] parse error at byte 5: syntax error while parsing BSON document: document size 6 does not match the number of bytes read (5)
+    ```
 
 ### json.exception.parse_error.113
 
-While parsing a map key, a value that is not a string has been read.
+A string could not be read from a [binary format](../features/binary_formats/index.md): either a value that is not a
+string was read where one was required (for instance as a map key), or the string's length specification is invalid.
 
 !!! failure "Example messages"
 
@@ -345,6 +353,9 @@ While parsing a map key, a value that is not a string has been read.
     ```
     [json.exception.parse_error.113] parse error at byte 2: syntax error while parsing UBJSON char: byte after 'C' must be in range 0x00..0x7F; last byte: 0x82
     ```
+    ```
+    [json.exception.parse_error.113] parse error at byte 3: syntax error while parsing BJData string: string length must not be negative
+    ```
 
 ### json.exception.parse_error.114
 
@@ -853,13 +864,21 @@ and this exception no longer occurs.
 
 ### json.exception.out_of_range.408
 
-The size (following `#`) of an UBJSON array or object exceeds the maximal capacity.
+The size of an array or object in a [binary format](../features/binary_formats/index.md) exceeds the maximal capacity:
+the size following `#` for [UBJSON](../features/binary_formats/ubjson.md)/[BJData](../features/binary_formats/bjdata.md),
+or the encoded length for [CBOR](../features/binary_formats/cbor.md).
 
-!!! failure "Example message"
+!!! failure "Example messages"
 
     ```
     excessive array size: 8658170730974374167
     ```
+    ```
+    [json.exception.out_of_range.408] syntax error while parsing CBOR size: excessive array size
+    ```
+    ```
+    [json.exception.out_of_range.408] syntax error while parsing CBOR size: excessive map size
+    ```
 
 ### json.exception.out_of_range.409
 

docs/mkdocs/mkdocs.yml

@@ -296,6 +296,7 @@ nav:
           - 'JSON_USE_GLOBAL_UDLS': api/macros/json_use_global_udls.md
           - 'JSON_USE_IMPLICIT_CONVERSIONS': api/macros/json_use_implicit_conversions.md
           - 'JSON_USE_LEGACY_DISCARDED_VALUE_COMPARISON': api/macros/json_use_legacy_discarded_value_comparison.md
+          - 'JSON_USE_SIMDUTF': api/macros/json_use_simdutf.md
           - 'NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE, NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE_WITH_DEFAULT, NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE_ONLY_SERIALIZE, NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE, NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE_WITH_DEFAULT, NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE_ONLY_SERIALIZE': api/macros/nlohmann_define_derived_type.md
           - 'NLOHMANN_DEFINE_TYPE_INTRUSIVE, NLOHMANN_DEFINE_TYPE_INTRUSIVE_WITH_DEFAULT, NLOHMANN_DEFINE_TYPE_INTRUSIVE_ONLY_SERIALIZE': api/macros/nlohmann_define_type_intrusive.md
           - 'NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE, NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE_WITH_DEFAULT, NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE_ONLY_SERIALIZE': api/macros/nlohmann_define_type_non_intrusive.md

include/nlohmann/detail/input/binary_reader.hpp

@@ -704,13 +704,15 @@ class binary_reader
             case 0x9A: // array (four-byte uint32_t for n follow)
             {
                 std::uint32_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_array(conditional_static_cast<std::size_t>(len), tag_handler);
+                std::size_t size{};
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && get_cbor_array(size, tag_handler);
             }
 
             case 0x9B: // array (eight-byte uint64_t for n follow)
             {
                 std::uint64_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_array(conditional_static_cast<std::size_t>(len), tag_handler);
+                std::size_t size{};
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "array") && get_cbor_array(size, tag_handler);
             }
 
             case 0x9F: // array (indefinite length)
@@ -758,13 +760,15 @@ class binary_reader
             case 0xBA: // map (four-byte uint32_t for n follow)
             {
                 std::uint32_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_object(conditional_static_cast<std::size_t>(len), tag_handler);
+                std::size_t size{};
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && get_cbor_object(size, tag_handler);
             }
 
             case 0xBB: // map (eight-byte uint64_t for n follow)
             {
                 std::uint64_t len{};
-                return get_number(input_format_t::cbor, len) && get_cbor_object(conditional_static_cast<std::size_t>(len), tag_handler);
+                std::size_t size{};
+                return get_number(input_format_t::cbor, len) && get_cbor_container_size(len, size, "map") && get_cbor_object(size, tag_handler);
             }
 
             case 0xBF: // map (indefinite length)
@@ -807,25 +811,37 @@ class binary_reader
                             case 0xD8:
                             {
                                 std::uint8_t subtype_to_ignore{};
-                                get_number(input_format_t::cbor, subtype_to_ignore);
+                                if (!get_number(input_format_t::cbor, subtype_to_ignore))
+                                {
+                                    return false;
+                                }
                                 break;
                             }
                             case 0xD9:
                             {
                                 std::uint16_t subtype_to_ignore{};
-                                get_number(input_format_t::cbor, subtype_to_ignore);
+                                if (!get_number(input_format_t::cbor, subtype_to_ignore))
+                                {
+                                    return false;
+                                }
                                 break;
                             }
                             case 0xDA:
                             {
                                 std::uint32_t subtype_to_ignore{};
-                                get_number(input_format_t::cbor, subtype_to_ignore);
+                                if (!get_number(input_format_t::cbor, subtype_to_ignore))
+                                {
+                                    return false;
+                                }
                                 break;
                             }
                             case 0xDB:
                             {
                                 std::uint64_t subtype_to_ignore{};
-                                get_number(input_format_t::cbor, subtype_to_ignore);
+                                if (!get_number(input_format_t::cbor, subtype_to_ignore))
+                                {
+                                    return false;
+                                }
                                 break;
                             }
                             default:
@@ -843,28 +859,40 @@ class binary_reader
                             case 0xD8:
                             {
                                 std::uint8_t subtype{};
-                                get_number(input_format_t::cbor, subtype);
+                                if (!get_number(input_format_t::cbor, subtype))
+                                {
+                                    return false;
+                                }
                                 b.set_subtype(detail::conditional_static_cast<typename binary_t::subtype_type>(subtype));
                                 break;
                             }
                             case 0xD9:
                             {
                                 std::uint16_t subtype{};
-                                get_number(input_format_t::cbor, subtype);
+                                if (!get_number(input_format_t::cbor, subtype))
+                                {
+                                    return false;
+                                }
                                 b.set_subtype(detail::conditional_static_cast<typename binary_t::subtype_type>(subtype));
                                 break;
                             }
                             case 0xDA:
                             {
                                 std::uint32_t subtype{};
-                                get_number(input_format_t::cbor, subtype);
+                                if (!get_number(input_format_t::cbor, subtype))
+                                {
+                                    return false;
+                                }
                                 b.set_subtype(detail::conditional_static_cast<typename binary_t::subtype_type>(subtype));
                                 break;
                             }
                             case 0xDB:
                             {
                                 std::uint64_t subtype{};
-                                get_number(input_format_t::cbor, subtype);
+                                if (!get_number(input_format_t::cbor, subtype))
+                                {
+                                    return false;
+                                }
                                 b.set_subtype(detail::conditional_static_cast<typename binary_t::subtype_type>(subtype));
                                 break;
                             }
@@ -1155,6 +1183,31 @@ class binary_reader
         }
     }
 
+    /*!
+    @brief narrow a definite CBOR array/map length to std::size_t
+
+    A definite length is rejected if it does not fit in std::size_t or if it
+    equals detail::unknown_size(), which is reserved to mark an indefinite-
+    length container and would otherwise make the length read as indefinite.
+    Both cases exceed any container's max_size(), so no representable input
+    is affected.
+
+    @param[in]  len      the declared length
+    @param[out] result   the length narrowed to std::size_t
+    @param[in]  context  "array" or "map", for the error message
+    @return whether the length is usable
+    */
+    bool get_cbor_container_size(const std::uint64_t len, std::size_t& result, const char* context)
+    {
+        if (JSON_HEDLEY_UNLIKELY(!value_in_range_of<std::size_t>(len) || len == detail::unknown_size()))
+        {
+            return sax->parse_error(chars_read, get_token_string(), out_of_range::create(408,
+                                    exception_message(input_format_t::cbor, concat("excessive ", context, " size"), "size"), nullptr));
+        }
+        result = conditional_static_cast<std::size_t>(len);
+        return true;
+    }
+
     /*!
     @param[in] len  the length of the array or detail::unknown_size() for an
                     array of indefinite size
@@ -2825,7 +2878,17 @@ class binary_reader
             case token_type::value_unsigned:
                 return sax->number_unsigned(number_lexer.get_number_unsigned());
             case token_type::value_float:
-                return sax->number_float(number_lexer.get_number_float(), std::move(number_string));
+            {
+                const auto parsed_float = number_lexer.get_number_float();
+                if (JSON_HEDLEY_UNLIKELY(!std::isfinite(parsed_float)))
+                {
+                    return sax->parse_error(
+                               chars_read,
+                               number_string,
+                               out_of_range::create(406, concat("number overflow parsing '", number_string, '\''), nullptr));
+                }
+                return sax->number_float(parsed_float, std::move(number_string));
+            }
             case token_type::uninitialized:
             case token_type::literal_true:
             case token_type::literal_false:

include/nlohmann/detail/input/input_adapters.hpp

@@ -231,6 +231,33 @@ class iterator_input_adapter
         std::is_same<IteratorType, SentinelType>::value && std::is_pointer<IteratorType>::value;
 #endif
 
+  public:
+    // Whether the remaining input is a single contiguous block of 1-byte
+    // elements that the lexer can inspect directly (used for the SWAR string
+    // fast path). Restricted to same-type iterator/sentinel pairs so that plain
+    // std::distance/std::advance are well-defined in all standards.
+    static constexpr bool supports_bulk_scan =
+        iterator_is_contiguous && std::is_same<IteratorType, SentinelType>::value && sizeof(char_type) == 1;
+
+    // Pointer to the next unread element; only valid when bulk_remaining() > 0.
+    const char_type* bulk_data() const
+    {
+        return &*current;
+    }
+
+    // Number of unread elements available as one contiguous block.
+    std::size_t bulk_remaining() const
+    {
+        return static_cast<std::size_t>(std::distance(current, end));
+    }
+
+    // Consume @a n elements previously inspected via bulk_data().
+    void bulk_skip(std::size_t n)
+    {
+        std::advance(current, static_cast<typename std::iterator_traits<IteratorType>::difference_type>(n));
+    }
+
+  private:
     // contiguous fast path: bulk copy the remaining range with std::memcpy
     template<class T>
     std::size_t get_elements_impl(T* dest, std::size_t count, std::true_type /*contiguous*/)
@@ -566,6 +593,24 @@ typename iterator_input_adapter_factory<IteratorType, SentinelType>::adapter_typ
     return factory_type::create(first, last);
 }
 
+// Detect a container that stores its elements contiguously as single bytes
+// (std::string, std::vector<char/unsigned char>, std::array<char, N>,
+// std::string_view, ...). Such inputs are wrapped in a pointer-based adapter so
+// they benefit from the contiguous fast paths (bulk string scanning, memcpy for
+// binary formats) in every C++ standard - not only in C++20, where the standard
+// library iterators model std::contiguous_iterator and are detected directly.
+template<typename ContainerType, typename = void>
+struct is_contiguous_byte_container : std::false_type {};
+
+template<typename ContainerType>
+struct is_contiguous_byte_container < ContainerType, void_t <
+decltype(std::declval<const ContainerType&>().data()),
+decltype(std::declval<const ContainerType&>().size()) >>
+            : std::integral_constant < bool,
+        std::is_pointer<decltype(std::declval<const ContainerType&>().data())>::value&&
+        std::is_integral<typename std::remove_pointer<decltype(std::declval<const ContainerType&>().data())>::type>::value&&
+        sizeof(typename std::remove_pointer<decltype(std::declval<const ContainerType&>().data())>::type) == 1 > {};
+
 // Convenience shorthand from container to iterator
 // Enables ADL on begin(container) and end(container)
 // Encloses the using declarations in namespace for not to leak them to outside scope
@@ -593,12 +638,32 @@ struct container_input_adapter_factory< ContainerType,
 
 }  // namespace container_input_adapter_factory_impl
 
-template<typename ContainerType>
+// General container path (iterator-based). Contiguous single-byte containers
-typename container_input_adapter_factory_impl::container_input_adapter_factory<ContainerType>::adapter_type input_adapter(ContainerType&& container)
+// are excluded here and routed through the pointer-based overload below.
+template < typename ContainerType,
+           enable_if_t < !is_contiguous_byte_container<ContainerType>::value, int > = 0 >
+typename container_input_adapter_factory_impl::container_input_adapter_factory<ContainerType>::adapter_type input_adapter(ContainerType && container)
 {
     return container_input_adapter_factory_impl::container_input_adapter_factory<ContainerType>::create(std::forward<ContainerType>(container));
 }
 
+// Contiguous single-byte containers (std::string, std::vector<char>, ...) are
+// wrapped in a pointer-based adapter so the contiguous fast paths apply in every
+// standard. The pointer keeps the container's own element type (const char* for
+// std::string, const std::uint8_t* for std::vector<std::uint8_t>, ...), so the
+// resulting char_type - and therefore the parsing behavior - is byte-for-byte
+// identical to the iterator-based path; only the raw pointer additionally
+// enables the bulk fast paths. The container outlives the adapter for the whole
+// parse (temporaries live until the end of the full expression), exactly as the
+// iterators it replaces did.
+template < typename ContainerType,
+           enable_if_t < is_contiguous_byte_container<ContainerType>::value, int > = 0 >
+auto input_adapter(const ContainerType& container)
+-> decltype(input_adapter(container.data(), container.data() + container.size()))
+{
+    return input_adapter(container.data(), container.data() + container.size());
+}
+
 // specialization for std::string
 using string_input_adapter_type = decltype(input_adapter(std::declval<std::string>()));
 

include/nlohmann/detail/input/json_sax.hpp

@@ -626,14 +626,7 @@ class json_sax_dom_callback_parser
         if (!ref_stack.empty() && ref_stack.back() && ref_stack.back()->is_structured())
         {
             // remove discarded value
-            for (auto it = ref_stack.back()->begin(); it != ref_stack.back()->end(); ++it)
+            remove_discarded_value(*ref_stack.back());
-            {
-                if (it->is_discarded())
-                {
-                    ref_stack.back()->erase(it);
-                    break;
-                }
-            }
         }
 
         return true;
@@ -674,8 +667,9 @@ class json_sax_dom_callback_parser
     bool end_array()
     {
         bool keep = true;
+        const bool stored = ref_stack.back() != nullptr;
 
-        if (ref_stack.back())
+        if (stored)
         {
             keep = callback(static_cast<int>(ref_stack.size()) - 1, parse_event_t::array_end, *ref_stack.back());
             if (keep)
@@ -709,9 +703,19 @@ class json_sax_dom_callback_parser
         keep_stack.pop_back();
 
         // remove discarded value
-        if (!keep && !ref_stack.empty() && ref_stack.back()->is_array())
+        if (!ref_stack.empty() && ref_stack.back())
         {
-            ref_stack.back()->m_data.m_value.array->pop_back();
+            if (!keep && ref_stack.back()->is_array())
+            {
+                ref_stack.back()->m_data.m_value.array->pop_back();
+            }
+            else if ((!keep || !stored) && ref_stack.back()->is_object())
+            {
+                // the array is either still stored under its key or was never
+                // stored, leaving the placeholder key() wrote; both show up as
+                // a discarded member of the parent object
+                remove_discarded_value(*ref_stack.back());
+            }
         }
 
         return true;
@@ -801,6 +805,19 @@ class json_sax_dom_callback_parser
     }
 #endif
 
+    /// remove the discarded value the callback rejected from its parent
+    static void remove_discarded_value(BasicJsonType& parent)
+    {
+        for (auto it = parent.begin(); it != parent.end(); ++it)
+        {
+            if (it->is_discarded())
+            {
+                parent.erase(it);
+                break;
+            }
+        }
+    }
+
     /*!
     @param[in] v  value to add to the JSON value we build during parsing
     @param[in] skip_callback  whether we should skip calling the callback
@@ -841,6 +858,18 @@ class json_sax_dom_callback_parser
         // do not handle this value if we just learnt it shall be discarded
         if (!keep)
         {
+            // if the value was to become an object member, key() already
+            // stored a placeholder for it that has to be removed again
+            if (!ref_stack.empty() && ref_stack.back() && ref_stack.back()->is_object())
+            {
+                JSON_ASSERT(!key_keep_stack.empty());
+                const bool placeholder_stored = key_keep_stack.back();
+                key_keep_stack.pop_back();
+                if (placeholder_stored)
+                {
+                    remove_discarded_value(*ref_stack.back());
+                }
+            }
             return {false, nullptr};
         }
 

include/nlohmann/detail/input/lexer.hpp

@@ -19,7 +19,9 @@
 #include <vector> // vector
 
 #include <nlohmann/detail/input/input_adapters.hpp>
+#include <nlohmann/detail/input/number_parse.hpp>
 #include <nlohmann/detail/input/position_t.hpp>
+#include <nlohmann/detail/input/string_scan.hpp>
 #include <nlohmann/detail/macro_scope.hpp>
 #include <nlohmann/detail/meta/type_traits.hpp>
 
@@ -125,6 +127,25 @@ constexpr bool input_adapter_supports_seek(std::false_type /*detected*/)
     return false;
 }
 
+// Detect whether an input adapter exposes a contiguous byte block that the
+// lexer can scan directly (see iterator_input_adapter::supports_bulk_scan).
+// Adapters without the flag - file, stream, wide-string, user-defined - fall
+// back to the character-at-a-time string scanner.
+template<typename InputAdapterType>
+using detect_supports_bulk_scan = decltype(InputAdapterType::supports_bulk_scan);
+
+template<typename InputAdapterType>
+constexpr bool input_adapter_supports_bulk_scan(std::true_type /*detected*/)
+{
+    return InputAdapterType::supports_bulk_scan;
+}
+
+template<typename InputAdapterType>
+constexpr bool input_adapter_supports_bulk_scan(std::false_type /*detected*/)
+{
+    return false;
+}
+
 /*!
 @brief lexical analysis
 
@@ -146,6 +167,14 @@ class lexer : public lexer_base<BasicJsonType>
     static constexpr bool lazy_token_string =
         input_adapter_supports_seek<InputAdapterType>(is_detected<detect_supports_seek, InputAdapterType> {});
 
+    /// whether string scanning may bulk-consume runs of ordinary characters
+    /// directly from a contiguous input buffer (SWAR fast path). This requires
+    /// the token to be reconstructible lazily (lazy_token_string), so bypassing
+    /// the per-character capture in get() cannot lose error diagnostics.
+    static constexpr bool bulk_scan =
+        lazy_token_string
+        && input_adapter_supports_bulk_scan<InputAdapterType>(is_detected<detect_supports_bulk_scan, InputAdapterType> {});
+
   public:
     using token_type = typename lexer_base<BasicJsonType>::token_type;
 
@@ -265,6 +294,40 @@ class lexer : public lexer_base<BasicJsonType>
         return true;
     }
 
+    /// contiguous input: bulk-append the run of ordinary characters and complete
+    /// well-formed UTF-8 sequences starting at the current read position, leaving
+    /// the first byte that needs individual handling (the closing quote, an
+    /// escape, a control character, or an ill-formed UTF-8 byte) for get()
+    void scan_string_bulk(std::true_type /*bulk*/)
+    {
+        // a pending unget must be consumed through the normal path first
+        if (next_unget)
+        {
+            return;
+        }
+        const std::size_t remaining = ia.bulk_remaining();
+        if (remaining == 0)
+        {
+            return;
+        }
+        const auto* const data = reinterpret_cast<const unsigned char*>(ia.bulk_data());
+
+        const std::size_t pos = string_bulk_run(data, remaining);
+        if (pos == 0)
+        {
+            return;
+        }
+        token_buffer.append(reinterpret_cast<const typename string_t::value_type*>(data), pos);
+        ia.bulk_skip(pos);
+        // the run contains no newline (all bytes < 0x20 are treated as special),
+        // so only the flat character counters advance
+        position.chars_read_total += pos;
+        position.chars_read_current_line += pos;
+    }
+
+    /// streaming input: no bulk fast path
+    void scan_string_bulk(std::false_type /*bulk*/) const noexcept {}
+
     /*!
     @brief scan a string literal
 
@@ -290,6 +353,10 @@ class lexer : public lexer_base<BasicJsonType>
 
         while (true)
         {
+            // bulk-consume ordinary characters from contiguous input, then
+            // handle the next special byte through the switch below
+            scan_string_bulk(std::integral_constant<bool, bulk_scan> {});
+
             // get the next character
             switch (get())
             {
@@ -1279,45 +1346,56 @@ scan_number_done:
         // we are done scanning a number)
         unget();
 
-        char* endptr = nullptr; // NOLINT(misc-const-correctness,cppcoreguidelines-pro-type-vararg,hicpp-vararg)
+        return convert_number(number_type);
-        errno = 0;
+    }
 
-        // try to parse integers first and fall back to floats
+    /*!
+    @brief convert the number text in token_buffer to its value and token type
+
+    The digit sequence in token_buffer has already been validated (by the
+    scan_number() state machine or by the contiguous fast path) and holds the
+    locale decimal point in place of '.'. Integers are parsed first and fall
+    back to floating point on overflow. This is shared so both scanners produce
+    identical results.
+    */
+    token_type convert_number(token_type number_type)
+    {
+        const char* const num_begin = token_buffer.data();
+        const char* const num_end = num_begin + token_buffer.size();
+
+        // try to parse integers first and fall back to floats; the digit
+        // sequence has already been validated, so a dedicated parser can avoid
+        // the locale/errno overhead of strtoull
         if (number_type == token_type::value_unsigned)
         {
-            const auto x = std::strtoull(token_buffer.data(), &endptr, 10);
+            if (parse_integer_unsigned(num_begin, num_end, value_unsigned))
-
-            // we checked the number format before
-            JSON_ASSERT(endptr == token_buffer.data() + token_buffer.size());
-
-            if (errno != ERANGE)
             {
-                value_unsigned = static_cast<number_unsigned_t>(x);
+                return token_type::value_unsigned;
-                if (value_unsigned == x)
-                {
-                    return token_type::value_unsigned;
-                }
             }
         }
         else if (number_type == token_type::value_integer)
         {
-            const auto x = std::strtoll(token_buffer.data(), &endptr, 10);
+            if (parse_integer_signed(num_begin, num_end, value_integer))
-
-            // we checked the number format before
-            JSON_ASSERT(endptr == token_buffer.data() + token_buffer.size());
-
-            if (errno != ERANGE)
             {
-                value_integer = static_cast<number_integer_t>(x);
+                return token_type::value_integer;
-                if (value_integer == x)
-                {
-                    return token_type::value_integer;
-                }
             }
         }
 
         // this code is reached if we parse a floating-point number or if an
-        // integer conversion above failed
+        // integer conversion above overflowed. Prefer std::from_chars
+        // (Eisel-Lemire, locale-independent, correctly rounded) when available;
+        // otherwise the exact Clinger fast path (double only); otherwise the
+        // locale-aware strtof/strtod.
+        if (parse_float_from_chars(num_begin, num_end, value_float))
+        {
+            return token_type::value_float;
+        }
+        if (parse_float_fast(num_begin, num_end, decimal_point_char, value_float))
+        {
+            return token_type::value_float;
+        }
+
+        char* endptr = nullptr; // NOLINT(misc-const-correctness,cppcoreguidelines-pro-type-vararg,hicpp-vararg)
         strtof(value_float, token_buffer.data(), &endptr);
 
         // we checked the number format before
@@ -1326,6 +1404,130 @@ scan_number_done:
         return token_type::value_float;
     }
 
+    /*!
+    @brief contiguous fast path for scanning a number
+
+    Parses the whole number token straight from the input buffer, avoiding the
+    per-character get()/add() of scan_number(). On success it fills token_buffer
+    (with the locale decimal point substituted, as scan_number() does) and
+    returns the token type. On anything it does not fully recognize as a
+    well-formed number it makes no state change and returns
+    token_type::uninitialized, so the caller falls back to scan_number(), which
+    then produces the exact diagnostic. @a current is the first digit or the
+    leading minus (already read); the remaining bytes are taken from the adapter.
+    */
+    token_type scan_number_bulk_contiguous()
+    {
+        // a pending unget offsets the buffer position from current; fall back
+        if (next_unget)
+        {
+            return token_type::uninitialized;
+        }
+        const std::size_t rem = ia.bulk_remaining();
+        if (rem == 0)
+        {
+            // the first digit is the last input byte; let scan_number() finish
+            return token_type::uninitialized;
+        }
+        // the byte before the next unread one is current (contiguous input)
+        const char* const data = reinterpret_cast<const char*>(ia.bulk_data()) - 1;
+        const std::size_t avail = rem + 1;
+
+        // validate + classify the number extent (mirrors scan_number()'s grammar)
+        std::size_t i = 0;
+        std::size_t dot_index = std::string::npos;
+        token_type number_type = token_type::value_unsigned;
+        if (data[0] == '-')
+        {
+            number_type = token_type::value_integer;
+            i = 1;
+            if (i >= avail)
+            {
+                return token_type::uninitialized;
+            }
+        }
+        if (data[i] == '0')
+        {
+            ++i;
+        }
+        else if (data[i] >= '1' && data[i] <= '9')
+        {
+            ++i;
+            while (i < avail && data[i] >= '0' && data[i] <= '9')
+            {
+                ++i;
+            }
+        }
+        else
+        {
+            return token_type::uninitialized;
+        }
+        if (i < avail && data[i] == '.')
+        {
+            number_type = token_type::value_float;
+            dot_index = i;
+            ++i;
+            if (i >= avail || !(data[i] >= '0' && data[i] <= '9'))
+            {
+                return token_type::uninitialized;
+            }
+            while (i < avail && data[i] >= '0' && data[i] <= '9')
+            {
+                ++i;
+            }
+        }
+        if (i < avail && (data[i] == 'e' || data[i] == 'E'))
+        {
+            number_type = token_type::value_float;
+            ++i;
+            if (i < avail && (data[i] == '+' || data[i] == '-'))
+            {
+                ++i;
+            }
+            if (i >= avail || !(data[i] >= '0' && data[i] <= '9'))
+            {
+                return token_type::uninitialized;
+            }
+            while (i < avail && data[i] >= '0' && data[i] <= '9')
+            {
+                ++i;
+            }
+        }
+        const std::size_t len = i;
+
+        // materialize the token exactly as scan_number() would, substituting the
+        // locale decimal point so convert_number()'s strtof fallback stays valid.
+        // reset() already cleared token_buffer, so append() fills it (assign() is
+        // avoided because custom string_t types need not provide it)
+        reset();
+        token_buffer.append(reinterpret_cast<const typename string_t::value_type*>(data), len);
+        if (dot_index != std::string::npos)
+        {
+            token_buffer[dot_index] = static_cast<typename string_t::value_type>(decimal_point_char);
+            decimal_point_position = dot_index;
+        }
+
+        // consume the remaining bytes of the number (current was already read)
+        ia.bulk_skip(len - 1);
+        position.chars_read_total += (len - 1);
+        position.chars_read_current_line += (len - 1);
+
+        return convert_number(number_type);
+    }
+
+    /// contiguous input: try the number fast path, else the byte-path scanner
+    token_type scan_number_dispatch(std::true_type /*bulk*/)
+    {
+        const token_type t = scan_number_bulk_contiguous();
+        return (t != token_type::uninitialized) ? t : scan_number();
+    }
+
+    /// streaming input: always use the byte-path scanner
+    token_type scan_number_dispatch(std::false_type /*bulk*/)
+    {
+        return scan_number();
+    }
+
     /*!
     @param[in] literal_text  the literal text to expect
     @param[in] length        the length of the passed literal text
@@ -1680,7 +1882,7 @@ scan_number_done:
             case '7':
             case '8':
             case '9':
-                return scan_number();
+                return scan_number_dispatch(std::integral_constant<bool, bulk_scan> {});
 
             // end of input (the null byte is needed when parsing from
             // string literals)

include/nlohmann/detail/input/number_parse.hpp

@@ -0,0 +1,302 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  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
+
+#pragma once
+
+#include <array> // array
+#include <cfloat> // FLT_EVAL_METHOD
+#include <cstddef> // size_t
+#include <cstdint> // int64_t, uint64_t
+#include <limits> // numeric_limits
+
+#include <nlohmann/detail/macro_scope.hpp>
+
+// std::from_chars lives in <charconv>, but being in C++17 mode does not
+// guarantee the header exists: GCC 7 sets __cplusplus to C++17 yet ships no
+// <charconv> (added in GCC 8; floating-point support in GCC 11). Guard the
+// include with __has_include so such toolchains fall back to the scalar path.
+#if defined(JSON_HAS_CPP_17) && defined(__has_include)
+    #if __has_include(<charconv>)
+        #include <charconv> // from_chars (only used when __cpp_lib_to_chars is defined)
+        #include <system_error> // errc
+    #endif
+#endif
+
+// This file contains the value-conversion helpers used by the lexer to turn an
+// already-validated number token into a value, without the locale/errno
+// overhead of std::strtoull/std::strtod. They are free functions so the lexer
+// stays focused on scanning; see lexer::convert_number().
+
+NLOHMANN_JSON_NAMESPACE_BEGIN
+namespace detail
+{
+
+/*!
+@brief fast integer parser for an already-validated unsigned integer
+
+The number scanner has already checked that [first, last) is a valid JSON
+integer, so this only needs to accumulate the digits and detect overflow. This
+avoids the locale/errno machinery of std::strtoull, which dominates
+integer-heavy inputs.
+
+@param[in]  first   pointer to the first character (a digit)
+@param[in]  last    pointer past the last character
+@param[out] value   the parsed value on success
+@return true if the value fit into @a NumberUnsignedType; false on overflow, in
+        which case the caller falls back to floating-point parsing (matching the
+        previous std::strtoull behavior)
+*/
+template<typename NumberUnsignedType>
+bool parse_integer_unsigned(const char* first, const char* last, NumberUnsignedType& value) noexcept
+{
+    // accumulate in the widest unsigned type used by the previous strtoull
+    // path so the overflow behavior is unchanged for custom number types
+    std::uint64_t x = 0;
+    constexpr std::uint64_t cutoff = (std::numeric_limits<std::uint64_t>::max)() / 10u;
+    constexpr std::uint64_t cutlim = (std::numeric_limits<std::uint64_t>::max)() % 10u;
+    for (const char* p = first; p != last; ++p)
+    {
+        const auto digit = static_cast<std::uint64_t>(static_cast<unsigned char>(*p) - static_cast<unsigned char>('0'));
+        if (JSON_HEDLEY_UNLIKELY(x > cutoff || (x == cutoff && digit > cutlim)))
+        {
+            return false;
+        }
+        x = (x * 10u) + digit;
+    }
+    value = static_cast<NumberUnsignedType>(x);
+    // reject values that do not round-trip into a narrower NumberUnsignedType
+    return static_cast<std::uint64_t>(value) == x;
+}
+
+/*!
+@brief fast integer parser for an already-validated negative integer
+
+@param[in]  first   pointer to the leading '-'
+@param[in]  last    pointer past the last character
+@param[out] value   the parsed (negative) value on success
+@return true on success; false on overflow (caller falls back to float)
+*/
+template<typename NumberIntegerType>
+bool parse_integer_signed(const char* first, const char* last, NumberIntegerType& value) noexcept
+{
+    // the state machine only reaches the signed path via a leading '-'
+    JSON_ASSERT(first != last && *first == '-');
+    std::uint64_t magnitude = 0;
+    // |INT64_MIN| == INT64_MAX + 1; this is the largest admissible magnitude
+    constexpr std::uint64_t limit = static_cast<std::uint64_t>((std::numeric_limits<std::int64_t>::max)()) + 1u;
+    for (const char* p = first + 1; p != last; ++p)
+    {
+        const auto digit = static_cast<std::uint64_t>(static_cast<unsigned char>(*p) - static_cast<unsigned char>('0'));
+        if (JSON_HEDLEY_UNLIKELY(magnitude > (limit - digit) / 10u))
+        {
+            return false;
+        }
+        magnitude = (magnitude * 10u) + digit;
+    }
+    const std::int64_t x = (magnitude == limit)
+                           ? (std::numeric_limits<std::int64_t>::min)()
+                           : -static_cast<std::int64_t>(magnitude);
+    value = static_cast<NumberIntegerType>(x);
+    // reject values that do not round-trip into a narrower NumberIntegerType
+    return static_cast<std::int64_t>(value) == x;
+}
+
+/*!
+@brief exact fast path for parsing a `double` (Clinger's algorithm)
+
+For the common case - at most 19 significant digits, a decimal exponent in
+[-22, 22], and a significand below 2^53 - the value equals significand *
+10^exp computed in IEEE-754 double arithmetic, which is exact under
+round-to-nearest because both operands are exactly representable. This is the
+same fast path used by fast_float/simdjson; the general cases are left to
+std::strtod. The parser only activates for number_float_t == double; float and
+long double keep the std::strtof/std::strtold paths (see the templated overload
+below).
+
+@param[in]  first          pointer to the first character of the number
+@param[in]  last           pointer past the last character
+@param[in]  decimal_point  the (locale-dependent) decimal point character
+@param[out] out            the parsed value on success
+@return true if the value was parsed exactly; false to fall back to strtod
+*/
+template<typename DecimalPointType>
+bool parse_float_fast(const char* first, const char* last, DecimalPointType decimal_point, double& out) noexcept
+{
+#if defined(FLT_EVAL_METHOD) && FLT_EVAL_METHOD != 0
+    // Clinger's fast path is only exact when double operations are evaluated in
+    // true double precision. On platforms that keep intermediates in extended
+    // precision (e.g. the x87 FPU on 32-bit x86, where FLT_EVAL_METHOD == 2) the
+    // single significand * 10^scale step is double-rounded and can be 1 ULP off,
+    // so decline and let the caller fall back to the correctly-rounded
+    // std::from_chars / std::strtod path.
+    static_cast<void>(first);
+    static_cast<void>(last);
+    static_cast<void>(decimal_point);
+    static_cast<void>(out);
+    return false;
+#else
+    static const std::array<double, 23> powers_of_ten =
+    {
+        {
+            1e0, 1e1, 1e2, 1e3, 1e4, 1e5, 1e6, 1e7, 1e8, 1e9, 1e10, 1e11,
+            1e12, 1e13, 1e14, 1e15, 1e16, 1e17, 1e18, 1e19, 1e20, 1e21, 1e22
+        }
+    };
+
+    const char* p = first;
+    bool negative = false;
+    if (p != last && (*p == '-' || *p == '+'))
+    {
+        negative = (*p == '-');
+        ++p;
+    }
+
+    std::uint64_t significand = 0;
+    int num_digits = 0;
+    int fractional_digits = 0;
+    bool seen_dot = false;
+    bool any_digit = false;
+    for (; p != last; ++p)
+    {
+        const char c = *p;
+        if (c >= '0' && c <= '9')
+        {
+            any_digit = true;
+            if (JSON_HEDLEY_UNLIKELY(num_digits >= 19))
+            {
+                return false; // significand may not fit into uint64_t
+            }
+            significand = (significand * 10u) + static_cast<std::uint64_t>(c - '0');
+            ++num_digits;
+            fractional_digits += static_cast<int>(seen_dot);
+        }
+        else if (static_cast<DecimalPointType>(c) == decimal_point)
+        {
+            if (JSON_HEDLEY_UNLIKELY(seen_dot))
+            {
+                return false;
+            }
+            seen_dot = true;
+        }
+        else if (c == 'e' || c == 'E')
+        {
+            ++p;
+            break;
+        }
+        else
+        {
+            return false;
+        }
+    }
+    if (JSON_HEDLEY_UNLIKELY(!any_digit))
+    {
+        return false;
+    }
+
+    int exponent = 0;
+    if (p != last) // an exponent part remains
+    {
+        bool exp_negative = false;
+        if (p != last && (*p == '-' || *p == '+'))
+        {
+            exp_negative = (*p == '-');
+            ++p;
+        }
+        bool any_exp_digit = false;
+        for (; p != last; ++p)
+        {
+            if (JSON_HEDLEY_UNLIKELY(*p < '0' || *p > '9'))
+            {
+                return false;
+            }
+            exponent = (exponent * 10) + (*p - '0');
+            any_exp_digit = true;
+            if (JSON_HEDLEY_UNLIKELY(exponent > 9999))
+            {
+                return false;
+            }
+        }
+        if (JSON_HEDLEY_UNLIKELY(!any_exp_digit))
+        {
+            return false;
+        }
+        if (exp_negative)
+        {
+            exponent = -exponent;
+        }
+    }
+
+    const int scale = exponent - fractional_digits;
+    if (JSON_HEDLEY_UNLIKELY(significand >= (static_cast<std::uint64_t>(1) << 53)))
+    {
+        return false; // significand not exactly representable as double
+    }
+
+    auto result = static_cast<double>(significand);
+    if (scale >= 0)
+    {
+        if (JSON_HEDLEY_UNLIKELY(scale > 22))
+        {
+            return false;
+        }
+        result *= powers_of_ten[static_cast<std::size_t>(scale)];
+    }
+    else
+    {
+        if (JSON_HEDLEY_UNLIKELY(-scale > 22))
+        {
+            return false;
+        }
+        result /= powers_of_ten[static_cast<std::size_t>(-scale)];
+    }
+    out = negative ? -result : result;
+    return true;
+#endif
+}
+
+/// fast float path is only exact for `double`; decline for float/long double
+template<typename DecimalPointType, typename FloatType>
+bool parse_float_fast(const char* /*first*/, const char* /*last*/, DecimalPointType /*decimal_point*/, FloatType& /*out*/) noexcept
+{
+    return false;
+}
+
+/*!
+@brief parse a float with std::from_chars (Eisel-Lemire) when available
+
+std::from_chars is locale-independent, correctly rounded, and - via the
+Eisel-Lemire algorithm in modern standard libraries - much faster than strtod
+over the whole value range (not just the Clinger subset). It is used only when
+__cpp_lib_to_chars indicates full floating-point support and only when it
+consumes the entire token ([first, last)); a partial parse means the buffer
+uses a non-'.' locale decimal point, in which case the caller falls back to the
+locale-aware path. An under-/overflow (result_out_of_range) also declines, so
+the caller's strtod fallback supplies the well-defined ±inf/0 result the parser
+expects (side-stepping the P4168 divergence between implementations).
+
+@return true if the value was parsed exactly and fully; false to fall back
+*/
+template<typename FloatType>
+bool parse_float_from_chars(const char* first, const char* last, FloatType& out) noexcept
+{
+    // JSON_HAS_CPP_17 must gate the use as well as the <charconv> include above:
+    // some standard libraries (e.g. libstdc++ 15) define __cpp_lib_to_chars even
+    // in C++14 mode, where <charconv> is not included.
+#if defined(JSON_HAS_CPP_17) && defined(__cpp_lib_to_chars)
+    const auto result = std::from_chars(first, last, out);
+    return result.ec == std::errc() && result.ptr == last;
+#else
+    static_cast<void>(first);
+    static_cast<void>(last);
+    static_cast<void>(out);
+    return false;
+#endif
+}
+
+}  // namespace detail
+NLOHMANN_JSON_NAMESPACE_END

include/nlohmann/detail/input/string_scan.hpp

@@ -0,0 +1,289 @@
+//     __ _____ _____ _____
+//  __|  |   __|     |   | |  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
+
+#pragma once
+
+#include <cstddef> // size_t
+#include <cstdint> // uint64_t
+#include <cstring> // memcpy
+
+#if defined(JSON_USE_SIMDUTF)
+    // Optional SIMD backend for bulk UTF-8 validation. This is an opt-in
+    // external dependency: nlohmann/json itself stays header-only and the C++11
+    // scalar validator below is always available; defining JSON_USE_SIMDUTF
+    // additionally requires the simdutf headers on the include path and linking
+    // the simdutf library. See string_bulk_run().
+    #include <simdutf.h>
+#endif
+
+#include <nlohmann/detail/macro_scope.hpp>
+
+// This file contains the byte-level string-scanning helpers used by the lexer's
+// contiguous fast path. They operate purely on raw bytes (no dependency on the
+// lexer's template parameters) so they are free functions, keeping the lexer
+// itself focused on the state machine; see lexer::scan_string_bulk().
+
+NLOHMANN_JSON_NAMESPACE_BEGIN
+namespace detail
+{
+
+// classify a single byte as needing individual string handling: the closing
+// quote, an escape, a control character, or a non-ASCII (UTF-8)
+// lead/continuation byte. Ordinary bytes (0x20..0x7F except '"' and '\\') are
+// copied verbatim, which the bulk scanner does 8 bytes at a time.
+inline bool is_string_special(unsigned char c) noexcept
+{
+    return c == '\"' || c == '\\' || c < 0x20u || c >= 0x80u;
+}
+
+// SWAR helper: return a word whose high bit is set in every byte of @a v that
+// is_string_special(); zero if the 8 bytes are all ordinary.
+inline std::uint64_t swar_string_special(std::uint64_t v) noexcept
+{
+    constexpr std::uint64_t ones = 0x0101010101010101ull;
+    constexpr std::uint64_t high = 0x8080808080808080ull;
+    const std::uint64_t q = v ^ 0x2222222222222222ull; // '"'  (0x22)
+    const std::uint64_t b = v ^ 0x5C5C5C5C5C5C5C5Cull; // '\\' (0x5C)
+    const std::uint64_t has_quote     = (q - ones) & ~q & high;
+    const std::uint64_t has_backslash = (b - ones) & ~b & high;
+    const std::uint64_t has_control   = (v - 0x2020202020202020ull) & ~v & high; // < 0x20
+    const std::uint64_t has_non_ascii = v & high;                                // >= 0x80
+    return has_quote | has_backslash | has_control | has_non_ascii;
+}
+
+// return the index of the first is_string_special() byte in [data, data+n), or
+// n if every byte is ordinary; scans 8 bytes at a time
+inline std::size_t find_string_special(const unsigned char* data, std::size_t n) noexcept
+{
+    std::size_t i = 0;
+    for (; i + 8 <= n; i += 8)
+    {
+        std::uint64_t word = 0;
+        std::memcpy(&word, data + i, sizeof(word));
+        if (swar_string_special(word) != 0)
+        {
+            // a special byte is in this word; locate it (endian-agnostic)
+            for (std::size_t j = 0; j < 8; ++j)
+            {
+                if (is_string_special(data[i + j]))
+                {
+                    return i + j;
+                }
+            }
+        }
+    }
+    for (; i < n; ++i)
+    {
+        if (is_string_special(data[i]))
+        {
+            return i;
+        }
+    }
+    return n;
+}
+
+// classify a byte as one the serializer must NOT copy verbatim when
+// ensure_ascii is requested: the closing quote, an escape, a control character
+// (< 0x20), DEL (0x7F), or any non-ASCII byte (>= 0x80). Everything else -
+// printable ASCII except '"' and '\\' - is emitted unchanged. Note this differs
+// from is_string_special() only in that 0x7F is also a stop (it is escaped as
+// \u007f under ensure_ascii).
+inline bool is_ascii_copyable(unsigned char c) noexcept
+{
+    return c >= 0x20u && c < 0x7Fu && c != '\"' && c != '\\';
+}
+
+// return the index of the first byte in [data, data+n) that is NOT
+// is_ascii_copyable(), or n if every byte can be copied verbatim; scans 8 bytes
+// at a time. Used by the serializer's ensure_ascii fast path.
+inline std::size_t find_ascii_copyable_run(const unsigned char* data, std::size_t n) noexcept
+{
+    constexpr std::uint64_t ones = 0x0101010101010101ull;
+    constexpr std::uint64_t high = 0x8080808080808080ull;
+    std::size_t i = 0;
+    for (; i + 8 <= n; i += 8)
+    {
+        std::uint64_t v = 0;
+        std::memcpy(&v, data + i, sizeof(v));
+        const std::uint64_t q = v ^ 0x2222222222222222ull; // '"'  (0x22)
+        const std::uint64_t b = v ^ 0x5C5C5C5C5C5C5C5Cull; // '\\' (0x5C)
+        const std::uint64_t d = v ^ 0x7F7F7F7F7F7F7F7Full; // DEL  (0x7F)
+        const std::uint64_t stop = ((q - ones) & ~q & high)   // == '"'
+                                   | ((b - ones) & ~b & high)  // == '\\'
+                                   | ((d - ones) & ~d & high)  // == 0x7F
+                                   | ((v - 0x2020202020202020ull) & ~v & high) // < 0x20
+                                   | (v & high);               // >= 0x80
+        if (stop != 0)
+        {
+            for (std::size_t j = 0; j < 8; ++j)
+            {
+                if (!is_ascii_copyable(data[i + j]))
+                {
+                    return i + j;
+                }
+            }
+        }
+    }
+    for (; i < n; ++i)
+    {
+        if (!is_ascii_copyable(data[i]))
+        {
+            return i;
+        }
+    }
+    return n;
+}
+
+// Validate one UTF-8 sequence at the front of [data, data+avail). Returns its
+// length (2..4) only when the bytes form a *well-formed* sequence using exactly
+// the same ranges as scan_string()'s per-byte switch, so the bulk path accepts
+// precisely what the byte path accepts. Returns 0 for anything that is invalid,
+// incomplete, or that the byte path must diagnose (the caller then defers to
+// that path, keeping error messages unchanged). Lead bytes < 0x80 are handled
+// by the caller and never passed here.
+inline std::size_t validate_one_utf8(const unsigned char* data, std::size_t avail) noexcept
+{
+    const unsigned char c0 = data[0];
+    if (c0 >= 0xC2 && c0 <= 0xDF)  // U+0080..U+07FF
+    {
+        if (avail >= 2 && data[1] >= 0x80 && data[1] <= 0xBF)
+        {
+            return 2;
+        }
+    }
+    else if (c0 == 0xE0)  // U+0800..U+0FFF
+    {
+        if (avail >= 3 && data[1] >= 0xA0 && data[1] <= 0xBF && data[2] >= 0x80 && data[2] <= 0xBF)
+        {
+            return 3;
+        }
+    }
+    else if ((c0 >= 0xE1 && c0 <= 0xEC) || c0 == 0xEE || c0 == 0xEF)  // U+1000..U+CFFF, U+E000..U+FFFF
+    {
+        if (avail >= 3 && data[1] >= 0x80 && data[1] <= 0xBF && data[2] >= 0x80 && data[2] <= 0xBF)
+        {
+            return 3;
+        }
+    }
+    else if (c0 == 0xED)  // U+D000..U+D7FF (excludes surrogates)
+    {
+        if (avail >= 3 && data[1] >= 0x80 && data[1] <= 0x9F && data[2] >= 0x80 && data[2] <= 0xBF)
+        {
+            return 3;
+        }
+    }
+    else if (c0 == 0xF0)  // U+10000..U+3FFFF
+    {
+        if (avail >= 4 && data[1] >= 0x90 && data[1] <= 0xBF && data[2] >= 0x80 && data[2] <= 0xBF && data[3] >= 0x80 && data[3] <= 0xBF)
+        {
+            return 4;
+        }
+    }
+    else if (c0 >= 0xF1 && c0 <= 0xF3)  // U+40000..U+FFFFF
+    {
+        if (avail >= 4 && data[1] >= 0x80 && data[1] <= 0xBF && data[2] >= 0x80 && data[2] <= 0xBF && data[3] >= 0x80 && data[3] <= 0xBF)
+        {
+            return 4;
+        }
+    }
+    else if (c0 == 0xF4)  // U+100000..U+10FFFF
+    {
+        if (avail >= 4 && data[1] >= 0x80 && data[1] <= 0x8F && data[2] >= 0x80 && data[2] <= 0xBF && data[3] >= 0x80 && data[3] <= 0xBF)
+        {
+            return 4;
+        }
+    }
+    return 0; // invalid, incomplete, or must be diagnosed by the byte path
+}
+
+// Scalar (C++11) computation of the bulk run length: the number of leading
+// bytes in [data, data+n) that are ordinary ASCII or complete well-formed UTF-8
+// sequences, stopping before the first byte that needs individual handling (the
+// closing quote, an escape, a control character, or an ill-formed/truncated
+// sequence). ASCII is skipped 8 bytes at a time.
+inline std::size_t scalar_string_bulk_run(const unsigned char* data, std::size_t n) noexcept
+{
+    std::size_t pos = 0;
+    while (pos < n)
+    {
+        pos += find_string_special(data + pos, n - pos);
+        if (pos >= n || data[pos] < 0x80u)
+        {
+            break; // end of buffer, or a quote/escape/control byte
+        }
+        const std::size_t seq = validate_one_utf8(data + pos, n - pos);
+        if (seq == 0)
+        {
+            break; // ill-formed or truncated: let the byte path diagnose it
+        }
+        pos += seq;
+    }
+    return pos;
+}
+
+#if defined(JSON_USE_SIMDUTF)
+// Index of the first quote/escape/control byte in [data, data+n) (non-ASCII
+// bytes are *not* stops here - the whole run is handed to simdutf), or n.
+inline std::size_t find_string_delimiter(const unsigned char* data, std::size_t n) noexcept
+{
+    constexpr std::uint64_t ones = 0x0101010101010101ull;
+    constexpr std::uint64_t high = 0x8080808080808080ull;
+    std::size_t i = 0;
+    for (; i + 8 <= n; i += 8)
+    {
+        std::uint64_t v = 0;
+        std::memcpy(&v, data + i, sizeof(v));
+        const std::uint64_t q = v ^ 0x2222222222222222ull;
+        const std::uint64_t b = v ^ 0x5C5C5C5C5C5C5C5Cull;
+        const std::uint64_t hit = ((q - ones) & ~q & high)
+                                  | ((b - ones) & ~b & high)
+                                  | ((v - 0x2020202020202020ull) & ~v & high);
+        if (hit != 0)
+        {
+            for (std::size_t j = 0; j < 8; ++j)
+            {
+                const unsigned char c = data[i + j];
+                if (c == '\"' || c == '\\' || c < 0x20u)
+                {
+                    return i + j;
+                }
+            }
+        }
+    }
+    for (; i < n; ++i)
+    {
+        const unsigned char c = data[i];
+        if (c == '\"' || c == '\\' || c < 0x20u)
+        {
+            return i;
+        }
+    }
+    return n;
+}
+#endif
+
+// Backend-dispatched bulk run length. With JSON_USE_SIMDUTF the run up to the
+// next delimiter is validated in one shot by simdutf; on the rare failure the
+// scalar helper recomputes the exact valid prefix so the byte path still
+// produces the precise diagnostic. Without it, the pure scalar path is used.
+inline std::size_t string_bulk_run(const unsigned char* data, std::size_t n) noexcept
+{
+#if defined(JSON_USE_SIMDUTF)
+    const std::size_t run = find_string_delimiter(data, n);
+    if (run != 0 && simdutf::validate_utf8(reinterpret_cast<const char*>(data), run))
+    {
+        return run;
+    }
+    return scalar_string_bulk_run(data, n);
+#else
+    return scalar_string_bulk_run(data, n);
+#endif
+}
+
+}  // namespace detail
+NLOHMANN_JSON_NAMESPACE_END

include/nlohmann/detail/meta/type_traits.hpp

@@ -231,7 +231,9 @@ struct char_traits<signed char> : std::char_traits<char>
     // Redefine to_int_type function
     static int_type to_int_type(char_type c) noexcept
     {
-        return static_cast<int_type>(c);
+        // cast via unsigned char: sign-extending a negative char_type would make
+        // byte 0xFF indistinguishable from eof()
+        return static_cast<int_type>(static_cast<unsigned char>(c));
     }
 
     static char_type to_char_type(int_type i) noexcept
@@ -699,21 +701,35 @@ struct is_json_pointer_of<A, ::nlohmann::json_pointer<A>> : std::true_type {};
 template <typename A>
 struct is_json_pointer_of<A, ::nlohmann::json_pointer<A>&> : std::true_type {};
 
-// checks if A and B are comparable using Compare functor
+// checks if A and B are comparable using Compare functor, assuming that
+// neither A nor B is a json_pointer type (that case is handled by
+// is_comparable below, which never instantiates this helper otherwise)
 template<typename Compare, typename A, typename B, typename = void>
-struct is_comparable : std::false_type {};
+struct is_comparable_no_json_pointer : std::false_type {};
 
-// We exclude json_pointer here, because the checks using Compare(A, B) will
-// use json_pointer::operator string_t() which triggers a deprecation warning
-// for GCC. See https://github.com/nlohmann/json/issues/4621. The call to
-// is_json_pointer_of can be removed once the deprecated function has been
-// removed.
 template<typename Compare, typename A, typename B>
-struct is_comparable < Compare, A, B, enable_if_t < !is_json_pointer_of<A, B>::value
+struct is_comparable_no_json_pointer < Compare, A, B, enable_if_t <
-&& std::is_constructible <decltype(std::declval<Compare>()(std::declval<A>(), std::declval<B>()))>::value
+std::is_constructible <decltype(std::declval<Compare>()(std::declval<A>(), std::declval<B>()))>::value
 && std::is_constructible <decltype(std::declval<Compare>()(std::declval<B>(), std::declval<A>()))>::value
 >> : std::true_type {};
 
+// checks if A and B are comparable using Compare functor
+// We dispatch on is_json_pointer_of as a plain bool (rather than folding it
+// into a single enable_if_t condition together with the checks below) so
+// that the Compare(A, B) checks are only ever written - and thus only ever
+// instantiated - when A/B are not a json_pointer/string pair. Those checks
+// use json_pointer::operator string_t() (GCC, see #4621) resp. the
+// deprecated json_pointer/string operator== (Clang, see #5288), and merely
+// naming them as later operands of a plain && chain is not sufficient to
+// avoid their instantiation on all compilers, even when the first operand
+// is false. The dispatch on is_json_pointer_of can be removed once the
+// deprecated json_pointer comparison operators have been removed.
+template<typename Compare, typename A, typename B, bool = is_json_pointer_of<A, B>::value>
+struct is_comparable : std::false_type {};
+
+template<typename Compare, typename A, typename B>
+struct is_comparable<Compare, A, B, false> : is_comparable_no_json_pointer<Compare, A, B> {};
+
 template<typename T>
 using detect_is_transparent = typename T::is_transparent;
 

include/nlohmann/detail/output/binary_writer.hpp

@@ -1662,7 +1662,15 @@ class binary_writer
         std::size_t len = (value.at(key).empty() ? 0 : 1);
         for (const auto& el : value.at(key))
         {
-            len *= static_cast<std::size_t>(el.m_data.m_value.number_unsigned);
+            // a dimension is read as an unsigned value below, so anything that
+            // is not a non-negative integer is rejected: a non-integer entry
+            // would pun unrelated bytes as the dimension, and a negative one
+            // would wrap into a nonsensical length
+            if (!el.is_number_integer() || (!el.is_number_unsigned() && el.template get<std::int64_t>() < 0))
+            {
+                return true;
+            }
+            len *= static_cast<std::size_t>(el.template get<std::uint64_t>());
         }
 
         key = "_ArrayData_";
@@ -1671,6 +1679,24 @@ class binary_writer
             return true;
         }
 
+        // every element is written below as the number kind dtype names, so it
+        // has to actually be a number of that category: an element of any other
+        // type would reinterpret unrelated bytes, e.g. a string's heap pointer,
+        // as that number. Such an object falls back to a plain object encoding.
+        // dtype names the wire type, not the storage type: whether an integer
+        // is held as number_integer or number_unsigned depends on how the value
+        // was built (parsing stores non-negative integers as unsigned, the C++
+        // API stores int literals as signed), so both are accepted here and the
+        // writes below go through get<>, which reads the member that is active.
+        const bool ndarray_is_float = (dtype == 'd' || dtype == 'D');
+        for (const auto& el : value.at(key))
+        {
+            if (ndarray_is_float ? !el.is_number_float() : !el.is_number_integer())
+            {
+                return true;
+            }
+        }
+
         oa->write_character('[');
         oa->write_character('$');
         oa->write_character(dtype);
@@ -1684,70 +1710,70 @@ class binary_writer
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<std::uint8_t>(el.m_data.m_value.number_unsigned), true);
+                write_number(static_cast<std::uint8_t>(el.template get<std::uint64_t>()), true);
             }
         }
         else if (dtype == 'i')
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<std::int8_t>(el.m_data.m_value.number_integer), true);
+                write_number(static_cast<std::int8_t>(el.template get<std::int64_t>()), true);
             }
         }
         else if (dtype == 'u')
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<std::uint16_t>(el.m_data.m_value.number_unsigned), true);
+                write_number(static_cast<std::uint16_t>(el.template get<std::uint64_t>()), true);
             }
         }
         else if (dtype == 'I')
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<std::int16_t>(el.m_data.m_value.number_integer), true);
+                write_number(static_cast<std::int16_t>(el.template get<std::int64_t>()), true);
             }
         }
         else if (dtype == 'm')
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<std::uint32_t>(el.m_data.m_value.number_unsigned), true);
+                write_number(static_cast<std::uint32_t>(el.template get<std::uint64_t>()), true);
             }
         }
         else if (dtype == 'l')
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<std::int32_t>(el.m_data.m_value.number_integer), true);
+                write_number(static_cast<std::int32_t>(el.template get<std::int64_t>()), true);
             }
         }
         else if (dtype == 'M')
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<std::uint64_t>(el.m_data.m_value.number_unsigned), true);
+                write_number(el.template get<std::uint64_t>(), true);
             }
         }
         else if (dtype == 'L')
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<std::int64_t>(el.m_data.m_value.number_integer), true);
+                write_number(el.template get<std::int64_t>(), true);
             }
         }
         else if (dtype == 'd')
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<float>(el.m_data.m_value.number_float), true);
+                write_number(static_cast<float>(el.template get<double>()), true);
             }
         }
         else if (dtype == 'D')
         {
             for (const auto& el : value.at(key))
             {
-                write_number(static_cast<double>(el.m_data.m_value.number_float), true);
+                write_number(el.template get<double>(), true);
             }
         }
         return false;

include/nlohmann/detail/output/serializer.hpp

@@ -16,6 +16,7 @@
 #include <cstddef> // size_t, ptrdiff_t
 #include <cstdint> // uint8_t
 #include <cstdio> // snprintf
+#include <cstring> // memcpy
 #include <limits> // numeric_limits
 #include <string> // string, char_traits
 #include <iomanip> // setfill, setw
@@ -24,6 +25,7 @@
 
 #include <nlohmann/detail/conversions/to_chars.hpp>
 #include <nlohmann/detail/exceptions.hpp>
+#include <nlohmann/detail/input/string_scan.hpp>
 #include <nlohmann/detail/macro_scope.hpp>
 #include <nlohmann/detail/meta/cpp_future.hpp>
 #include <nlohmann/detail/output/binary_writer.hpp>
@@ -109,6 +111,25 @@ class serializer
               const bool ensure_ascii,
               const unsigned int indent_step,
               const unsigned int current_indent = 0)
+    {
+        dump_internal(val, pretty_print, ensure_ascii, indent_step, current_indent);
+        flush();
+    }
+
+  JSON_PRIVATE_UNLESS_TESTED:
+    /*!
+    @brief recursive worker for @ref dump
+
+    Identical in behavior to the historical @ref dump, but writes into the
+    serializer's internal @ref write_buffer instead of issuing a virtual call
+    per token. The public @ref dump wraps this and flushes the buffer once the
+    top-level value has been serialized.
+    */
+    void dump_internal(const BasicJsonType& val,
+                       const bool pretty_print,
+                       const bool ensure_ascii,
+                       const unsigned int indent_step,
+                       const unsigned int current_indent = 0)
     {
         switch (val.m_data.m_type)
         {
@@ -116,13 +137,13 @@ class serializer
             {
                 if (val.m_data.m_value.object->empty())
                 {
-                    o->write_characters("{}", 2);
+                    put_chars("{}", 2);
                     return;
                 }
 
                 if (pretty_print)
                 {
-                    o->write_characters("{\n", 2);
+                    put_chars("{\n", 2);
 
                     // variable to hold indentation for recursive calls
                     const auto new_indent = current_indent + indent_step;
@@ -135,51 +156,51 @@ class serializer
                     auto i = val.m_data.m_value.object->cbegin();
                     for (std::size_t cnt = 0; cnt < val.m_data.m_value.object->size() - 1; ++cnt, ++i)
                     {
-                        o->write_characters(indent_string.c_str(), new_indent);
+                        put_chars(indent_string.c_str(), new_indent);
-                        o->write_character('\"');
+                        put_char('\"');
                         dump_escaped(i->first, ensure_ascii);
-                        o->write_characters("\": ", 3);
+                        put_chars("\": ", 3);
-                        dump(i->second, true, ensure_ascii, indent_step, new_indent);
+                        dump_internal(i->second, true, ensure_ascii, indent_step, new_indent);
-                        o->write_characters(",\n", 2);
+                        put_chars(",\n", 2);
                     }
 
                     // last element
                     JSON_ASSERT(i != val.m_data.m_value.object->cend());
                     JSON_ASSERT(std::next(i) == val.m_data.m_value.object->cend());
-                    o->write_characters(indent_string.c_str(), new_indent);
+                    put_chars(indent_string.c_str(), new_indent);
-                    o->write_character('\"');
+                    put_char('\"');
                     dump_escaped(i->first, ensure_ascii);
-                    o->write_characters("\": ", 3);
+                    put_chars("\": ", 3);
-                    dump(i->second, true, ensure_ascii, indent_step, new_indent);
+                    dump_internal(i->second, true, ensure_ascii, indent_step, new_indent);
 
-                    o->write_character('\n');
+                    put_char('\n');
-                    o->write_characters(indent_string.c_str(), current_indent);
+                    put_chars(indent_string.c_str(), current_indent);
-                    o->write_character('}');
+                    put_char('}');
                 }
                 else
                 {
-                    o->write_character('{');
+                    put_char('{');
 
                     // first n-1 elements
                     auto i = val.m_data.m_value.object->cbegin();
                     for (std::size_t cnt = 0; cnt < val.m_data.m_value.object->size() - 1; ++cnt, ++i)
                     {
-                        o->write_character('\"');
+                        put_char('\"');
                         dump_escaped(i->first, ensure_ascii);
-                        o->write_characters("\":", 2);
+                        put_chars("\":", 2);
-                        dump(i->second, false, ensure_ascii, indent_step, current_indent);
+                        dump_internal(i->second, false, ensure_ascii, indent_step, current_indent);
-                        o->write_character(',');
+                        put_char(',');
                     }
 
                     // last element
                     JSON_ASSERT(i != val.m_data.m_value.object->cend());
                     JSON_ASSERT(std::next(i) == val.m_data.m_value.object->cend());
-                    o->write_character('\"');
+                    put_char('\"');
                     dump_escaped(i->first, ensure_ascii);
-                    o->write_characters("\":", 2);
+                    put_chars("\":", 2);
-                    dump(i->second, false, ensure_ascii, indent_step, current_indent);
+                    dump_internal(i->second, false, ensure_ascii, indent_step, current_indent);
 
-                    o->write_character('}');
+                    put_char('}');
                 }
 
                 return;
@@ -189,13 +210,13 @@ class serializer
             {
                 if (val.m_data.m_value.array->empty())
                 {
-                    o->write_characters("[]", 2);
+                    put_chars("[]", 2);
                     return;
                 }
 
                 if (pretty_print)
                 {
-                    o->write_characters("[\n", 2);
+                    put_chars("[\n", 2);
 
                     // variable to hold indentation for recursive calls
                     const auto new_indent = current_indent + indent_step;
@@ -208,37 +229,37 @@ class serializer
                     for (auto i = val.m_data.m_value.array->cbegin();
                             i != val.m_data.m_value.array->cend() - 1; ++i)
                     {
-                        o->write_characters(indent_string.c_str(), new_indent);
+                        put_chars(indent_string.c_str(), new_indent);
-                        dump(*i, true, ensure_ascii, indent_step, new_indent);
+                        dump_internal(*i, true, ensure_ascii, indent_step, new_indent);
-                        o->write_characters(",\n", 2);
+                        put_chars(",\n", 2);
                     }
 
                     // last element
                     JSON_ASSERT(!val.m_data.m_value.array->empty());
-                    o->write_characters(indent_string.c_str(), new_indent);
+                    put_chars(indent_string.c_str(), new_indent);
-                    dump(val.m_data.m_value.array->back(), true, ensure_ascii, indent_step, new_indent);
+                    dump_internal(val.m_data.m_value.array->back(), true, ensure_ascii, indent_step, new_indent);
 
-                    o->write_character('\n');
+                    put_char('\n');
-                    o->write_characters(indent_string.c_str(), current_indent);
+                    put_chars(indent_string.c_str(), current_indent);
-                    o->write_character(']');
+                    put_char(']');
                 }
                 else
                 {
-                    o->write_character('[');
+                    put_char('[');
 
                     // first n-1 elements
                     for (auto i = val.m_data.m_value.array->cbegin();
                             i != val.m_data.m_value.array->cend() - 1; ++i)
                     {
-                        dump(*i, false, ensure_ascii, indent_step, current_indent);
+                        dump_internal(*i, false, ensure_ascii, indent_step, current_indent);
-                        o->write_character(',');
+                        put_char(',');
                     }
 
                     // last element
                     JSON_ASSERT(!val.m_data.m_value.array->empty());
-                    dump(val.m_data.m_value.array->back(), false, ensure_ascii, indent_step, current_indent);
+                    dump_internal(val.m_data.m_value.array->back(), false, ensure_ascii, indent_step, current_indent);
 
-                    o->write_character(']');
+                    put_char(']');
                 }
 
                 return;
@@ -246,9 +267,9 @@ class serializer
 
             case value_t::string:
             {
-                o->write_character('\"');
+                put_char('\"');
                 dump_escaped(*val.m_data.m_value.string, ensure_ascii);
-                o->write_character('\"');
+                put_char('\"');
                 return;
             }
 
@@ -256,7 +277,7 @@ class serializer
             {
                 if (pretty_print)
                 {
-                    o->write_characters("{\n", 2);
+                    put_chars("{\n", 2);
 
                     // variable to hold indentation for recursive calls
                     const auto new_indent = current_indent + indent_step;
@@ -265,9 +286,9 @@ class serializer
                         indent_string.resize(indent_string.size() * 2, ' ');
                     }
 
-                    o->write_characters(indent_string.c_str(), new_indent);
+                    put_chars(indent_string.c_str(), new_indent);
 
-                    o->write_characters("\"bytes\": [", 10);
+                    put_chars("\"bytes\": [", 10);
 
                     if (!val.m_data.m_value.binary->empty())
                     {
@@ -275,30 +296,30 @@ class serializer
                                 i != val.m_data.m_value.binary->cend() - 1; ++i)
                         {
                             dump_integer(*i);
-                            o->write_characters(", ", 2);
+                            put_chars(", ", 2);
                         }
                         dump_integer(val.m_data.m_value.binary->back());
                     }
 
-                    o->write_characters("],\n", 3);
+                    put_chars("],\n", 3);
-                    o->write_characters(indent_string.c_str(), new_indent);
+                    put_chars(indent_string.c_str(), new_indent);
 
-                    o->write_characters("\"subtype\": ", 11);
+                    put_chars("\"subtype\": ", 11);
                     if (val.m_data.m_value.binary->has_subtype())
                     {
                         dump_integer(val.m_data.m_value.binary->subtype());
                     }
                     else
                     {
-                        o->write_characters("null", 4);
+                        put_chars("null", 4);
                     }
-                    o->write_character('\n');
+                    put_char('\n');
-                    o->write_characters(indent_string.c_str(), current_indent);
+                    put_chars(indent_string.c_str(), current_indent);
-                    o->write_character('}');
+                    put_char('}');
                 }
                 else
                 {
-                    o->write_characters("{\"bytes\":[", 10);
+                    put_chars("{\"bytes\":[", 10);
 
                     if (!val.m_data.m_value.binary->empty())
                     {
@@ -306,20 +327,20 @@ class serializer
                                 i != val.m_data.m_value.binary->cend() - 1; ++i)
                         {
                             dump_integer(*i);
-                            o->write_character(',');
+                            put_char(',');
                         }
                         dump_integer(val.m_data.m_value.binary->back());
                     }
 
-                    o->write_characters("],\"subtype\":", 12);
+                    put_chars("],\"subtype\":", 12);
                     if (val.m_data.m_value.binary->has_subtype())
                     {
                         dump_integer(val.m_data.m_value.binary->subtype());
-                        o->write_character('}');
+                        put_char('}');
                     }
                     else
                     {
-                        o->write_characters("null}", 5);
+                        put_chars("null}", 5);
                     }
                 }
                 return;
@@ -329,11 +350,11 @@ class serializer
             {
                 if (val.m_data.m_value.boolean)
                 {
-                    o->write_characters("true", 4);
+                    put_chars("true", 4);
                 }
                 else
                 {
-                    o->write_characters("false", 5);
+                    put_chars("false", 5);
                 }
                 return;
             }
@@ -358,13 +379,13 @@ class serializer
 
             case value_t::discarded:
             {
-                o->write_characters("<discarded>", 11);
+                put_chars("<discarded>", 11);
                 return;
             }
 
             case value_t::null:
             {
-                o->write_characters("null", 4);
+                put_chars("null", 4);
                 return;
             }
 
@@ -400,6 +421,45 @@ class serializer
 
         for (std::size_t i = 0; i < s.size(); ++i)
         {
+            // Fast path: at a character boundary (state == UTF8_ACCEPT),
+            // bulk-copy the longest run of bytes that need no escaping using a
+            // SWAR scanner shared with the lexer's contiguous path. The scanner
+            // stops exactly at the first byte dump_escaped would handle
+            // individually, so that byte is left to the byte-at-a-time path
+            // below, keeping escaping output and error diagnostics unchanged.
+            //
+            // - ensure_ascii == false: string_bulk_run() copies ordinary bytes
+            //   and complete well-formed UTF-8, stopping at a quote, backslash,
+            //   control character (< 0x20), or ill-formed/truncated sequence.
+            // - ensure_ascii == true: only printable ASCII may be copied
+            //   verbatim; find_ascii_copyable_run() additionally stops at 0x7F
+            //   and every non-ASCII byte (>= 0x80), which must be \u-escaped.
+            if (state == UTF8_ACCEPT)
+            {
+                const auto* const data = reinterpret_cast<const unsigned char*>(s.data());
+                const std::size_t run = ensure_ascii
+                                        ? find_ascii_copyable_run(data + i, s.size() - i)
+                                        : string_bulk_run(data + i, s.size() - i);
+                if (run != 0)
+                {
+                    // emit any bytes still pending in string_buffer first to
+                    // preserve output order, then write the run directly
+                    if (bytes != 0)
+                    {
+                        put_chars(string_buffer.data(), bytes);
+                        bytes = 0;
+                    }
+                    put_chars(s.data() + i, run);
+                    bytes_after_last_accept = 0;
+                    undumped_chars = 0;
+                    i += run;
+                    if (i >= s.size())
+                    {
+                        break;
+                    }
+                }
+            }
+
             const auto byte = static_cast<std::uint8_t>(s[i]);
 
             switch (decode(state, codepoint, byte))
@@ -488,7 +548,7 @@ class serializer
                     // written ("\uxxxx\uxxxx\0") for one code point
                     if (string_buffer.size() - bytes < 13)
                     {
-                        o->write_characters(string_buffer.data(), bytes);
+                        put_chars(string_buffer.data(), bytes);
                         bytes = 0;
                     }
 
@@ -547,7 +607,7 @@ class serializer
                                 // written ("\uxxxx\uxxxx\0") for one code point
                                 if (string_buffer.size() - bytes < 13)
                                 {
-                                    o->write_characters(string_buffer.data(), bytes);
+                                    put_chars(string_buffer.data(), bytes);
                                     bytes = 0;
                                 }
 
@@ -586,7 +646,7 @@ class serializer
             // write buffer
             if (bytes > 0)
             {
-                o->write_characters(string_buffer.data(), bytes);
+                put_chars(string_buffer.data(), bytes);
             }
         }
         else
@@ -602,22 +662,22 @@ class serializer
                 case error_handler_t::ignore:
                 {
                     // write all accepted bytes
-                    o->write_characters(string_buffer.data(), bytes_after_last_accept);
+                    put_chars(string_buffer.data(), bytes_after_last_accept);
                     break;
                 }
 
                 case error_handler_t::replace:
                 {
                     // write all accepted bytes
-                    o->write_characters(string_buffer.data(), bytes_after_last_accept);
+                    put_chars(string_buffer.data(), bytes_after_last_accept);
                     // add a replacement character
                     if (ensure_ascii)
                     {
-                        o->write_characters("\\ufffd", 6);
+                        put_chars("\\ufffd", 6);
                     }
                     else
                     {
-                        o->write_characters("\xEF\xBF\xBD", 3);
+                        put_chars("\xEF\xBF\xBD", 3);
                     }
                     break;
                 }
@@ -628,6 +688,66 @@ class serializer
         }
     }
 
+  private:
+    /*!
+    @brief append a single character to the write buffer
+
+    Structural characters ('{', '"', ',', ...) previously went straight to the
+    output adapter, one virtual call each. Buffering them and flushing in bulk
+    turns those many indirect calls into a single memcpy plus an occasional
+    flush, which dominates the cost of serializing object/array-heavy values.
+    */
+    void put_char(char c)
+    {
+        if (JSON_HEDLEY_UNLIKELY(write_buffer_pos == write_buffer.size()))
+        {
+            flush();
+        }
+        write_buffer[write_buffer_pos++] = c;
+    }
+
+    /*!
+    @brief append @a length characters to the write buffer
+
+    Runs that do not fit the buffer are written straight through the output
+    adapter (after flushing what is pending), so large string/number payloads
+    are not copied an extra time.
+    */
+    JSON_HEDLEY_NON_NULL(2)
+    void put_chars(const char* s, std::size_t length)
+    {
+        if (JSON_HEDLEY_UNLIKELY(length >= write_buffer.size()))
+        {
+            flush();
+            o->write_characters(s, length);
+            return;
+        }
+        if (JSON_HEDLEY_UNLIKELY(write_buffer_pos + length > write_buffer.size()))
+        {
+            flush();
+        }
+        std::memcpy(write_buffer.data() + write_buffer_pos, s, length);
+        write_buffer_pos += length;
+    }
+
+  JSON_PRIVATE_UNLESS_TESTED:
+    /*!
+    @brief flush the write buffer to the output adapter
+
+    Writing zero characters is a well-defined no-op for every output adapter, so
+    the buffered length is passed through unconditionally (no empty-guard branch
+    to leave uncovered).
+
+    @note dump_escaped() and dump_integer()/dump_float() write into the internal
+    write buffer; callers that invoke them directly (rather than through the
+    public dump()) must call flush() before inspecting the output.
+    */
+    void flush()
+    {
+        o->write_characters(write_buffer.data(), write_buffer_pos);
+        write_buffer_pos = 0;
+    }
+
   private:
     /*!
     @brief count digits
@@ -752,7 +872,7 @@ class serializer
         // special case for "0"
         if (x == 0)
         {
-            o->write_character('0');
+            put_char('0');
             return;
         }
 
@@ -805,7 +925,7 @@ class serializer
             *(--buffer_ptr) = static_cast<char>('0' + abs_value);
         }
 
-        o->write_characters(number_buffer.data(), n_chars);
+        put_chars(number_buffer.data(), n_chars);
     }
 
     /*!
@@ -821,7 +941,7 @@ class serializer
         // NaN / inf
         if (!std::isfinite(x))
         {
-            o->write_characters("null", 4);
+            put_chars("null", 4);
             return;
         }
 
@@ -842,7 +962,7 @@ class serializer
         auto* begin = number_buffer.data();
         auto* end = ::nlohmann::detail::to_chars(begin, begin + number_buffer.size(), x);
 
-        o->write_characters(begin, static_cast<size_t>(end - begin));
+        put_chars(begin, static_cast<size_t>(end - begin));
     }
 
     JSON_HEDLEY_NON_NULL(1)
@@ -893,7 +1013,7 @@ class serializer
             }
         }
 
-        o->write_characters(number_buffer.data(), static_cast<std::size_t>(len));
+        put_chars(number_buffer.data(), static_cast<std::size_t>(len));
 
         // determine if we need to append ".0"
         const bool value_is_int_like =
@@ -905,7 +1025,7 @@ class serializer
 
         if (value_is_int_like)
         {
-            o->write_characters(".0", 2);
+            put_chars(".0", 2);
         }
     }
 
@@ -1015,6 +1135,12 @@ class serializer
 
     /// error_handler how to react on decoding errors
     const error_handler_t error_handler;
+
+    /// buffer collecting output before it is flushed to the output adapter, so
+    /// that the many small structural writes become few bulk writes
+    std::array<char, 1024> write_buffer{{}};
+    /// number of valid bytes currently held in @ref write_buffer
+    std::size_t write_buffer_pos = 0;
 };
 
 }  // namespace detail

tests/src/unit-32bit.cpp

@@ -132,3 +132,37 @@ TEST_CASE("BJData")
         }
     }
 }
+
+TEST_CASE("CBOR")
+{
+    SECTION("parse errors")
+    {
+        SECTION("array/map size larger than std::size_t")
+        {
+            // declared lengths do not fit in a 32-bit std::size_t and must not be truncated
+            std::vector<uint8_t> const varr = {0x9B, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x05};
+            std::vector<uint8_t> const vmap = {0xBB, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x05};
+
+            json _;
+            CHECK_THROWS_WITH_AS(_ = json::from_cbor(varr), "[json.exception.out_of_range.408] syntax error while parsing CBOR size: excessive array size", json::out_of_range&);
+            CHECK(json::from_cbor(varr, true, false).is_discarded());
+
+            CHECK_THROWS_WITH_AS(_ = json::from_cbor(vmap), "[json.exception.out_of_range.408] syntax error while parsing CBOR size: excessive map size", json::out_of_range&);
+            CHECK(json::from_cbor(vmap, true, false).is_discarded());
+        }
+
+        SECTION("array/map size equal to the indefinite-length sentinel")
+        {
+            // on 32-bit platforms a four-byte length of 0xFFFFFFFF aliases unknown_size()
+            std::vector<uint8_t> const varr = {0x9A, 0xFF, 0xFF, 0xFF, 0xFF};
+            std::vector<uint8_t> const vmap = {0xBA, 0xFF, 0xFF, 0xFF, 0xFF};
+
+            json _;
+            CHECK_THROWS_WITH_AS(_ = json::from_cbor(varr), "[json.exception.out_of_range.408] syntax error while parsing CBOR size: excessive array size", json::out_of_range&);
+            CHECK(json::from_cbor(varr, true, false).is_discarded());
+
+            CHECK_THROWS_WITH_AS(_ = json::from_cbor(vmap), "[json.exception.out_of_range.408] syntax error while parsing CBOR size: excessive map size", json::out_of_range&);
+            CHECK(json::from_cbor(vmap, true, false).is_discarded());
+        }
+    }
+}

tests/src/unit-bjdata.cpp

@@ -1347,6 +1347,8 @@ TEST_CASE("BJData")
                     CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vec2), "[json.exception.parse_error.115] parse error at byte 5: syntax error while parsing BJData high-precision number: invalid number text: 1A", json::parse_error);
                     std::vector<uint8_t> const vec3 = {'H', 'i', 2, '1', '.'};
                     CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vec3), "[json.exception.parse_error.115] parse error at byte 5: syntax error while parsing BJData high-precision number: invalid number text: 1.", json::parse_error);
+                    std::vector<uint8_t> const vec_overflow = {'H', 'i', 5, '1', 'e', '4', '0', '0'};
+                    CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vec_overflow), "[json.exception.out_of_range.406] number overflow parsing '1e400'", json::out_of_range);
                     std::vector<uint8_t> const vec4 = {'H', 2, '1', '0'};
                     CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vec4), "[json.exception.parse_error.113] parse error at byte 2: syntax error while parsing BJData size: expected length type specification (U, i, u, I, m, l, M, L) after '#'; last byte: 0x02", json::parse_error);
                 }
@@ -2587,6 +2589,69 @@ TEST_CASE("BJData")
                 CHECK(json::to_bjdata(json::from_bjdata(v_B), true, true) == v_B);
             }
 
+            SECTION("ndarray with data not matching _ArrayType_ is written as an object")
+            {
+                // A JData-annotated object is only serialized as an ndarray when
+                // its _ArrayData_ elements are actually stored as the number kind
+                // named by _ArrayType_. Otherwise the writer would read the wrong
+                // union member (e.g. a std::string's heap pointer as a uint64) and
+                // emit it, so such an object falls back to a plain object encoding
+                // that still round-trips.
+
+                // string data declared as a uint64 array
+                json const j_str = json({{"_ArrayType_", "uint64"}, {"_ArraySize_", {1}}, {"_ArrayData_", {"pointer"}}});
+                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}}});
+                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}}});
+                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]})");
+                const auto out_neg = json::to_bjdata(j_neg);
+                CHECK(out_neg.at(0) == '{');
+                CHECK(json::from_bjdata(out_neg) == j_neg);
+            }
+
+            SECTION("ndarray parsed from text is written as a typed array")
+            {
+                // json::parse stores a non-negative integer as number_unsigned while
+                // 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.
+                for (const char* type :
+                        {"uint8", "int8", "uint16", "int16", "uint32", "int32", "uint64", "int64", "char", "byte"
+                        })
+                {
+                    CAPTURE(type);
+                    const std::string text = std::string(R"({"_ArrayType_":")") + type +
+                                             R"(","_ArraySize_":[2,3],"_ArrayData_":[1,2,3,4,5,6]})";
+                    const auto from_text = json::to_bjdata(json::parse(text));
+                    CHECK(from_text.at(0) == '[');
+                    CHECK(from_text == json::to_bjdata(json({{"_ArrayType_", type}, {"_ArraySize_", {2, 3}}, {"_ArrayData_", {1, 2, 3, 4, 5, 6}}})));
+                }
+
+                // 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]})"));
+                CHECK(from_neg.at(0) == '[');
+                CHECK(from_neg == json::to_bjdata(json({{"_ArrayType_", "int32"}, {"_ArraySize_", {2}}, {"_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]})"));
+                CHECK(from_float.at(0) == '[');
+                CHECK(from_float == json::to_bjdata(json({{"_ArrayType_", "double"}, {"_ArraySize_", {2}}, {"_ArrayData_", {1.5, 2.5}}})));
+            }
+
             SECTION("optimized ndarray (type and vector-size as 1D array)")
             {
                 // create vector with two elements of the same type

tests/src/unit-bson.cpp

@@ -529,6 +529,41 @@ TEST_CASE("BSON")
             CHECK(json::from_bson(result, true, false) == j);
         }
 
+        SECTION("non-empty object with binary member without subtype")
+        {
+            const size_t N = 10;
+            const auto s = std::vector<std::uint8_t>(N, 'x');
+            json const j =
+            {
+                { "entry", json::binary(s) }
+            };
+
+            CHECK(!j.at("entry").get_binary().has_subtype());
+
+            std::vector<std::uint8_t> const expected =
+            {
+                0x1B, 0x00, 0x00, 0x00, // size (little endian)
+                0x05, // entry: binary
+                'e', 'n', 't', 'r', 'y', '\x00',
+
+                0x0A, 0x00, 0x00, 0x00, // size of binary (little endian)
+                0x00, // Generic binary subtype
+                0x78, 0x78, 0x78, 0x78, 0x78, 0x78, 0x78, 0x78, 0x78, 0x78,
+
+                0x00 // end marker
+            };
+
+            const auto result = json::to_bson(j);
+            CHECK(result == expected);
+
+            // roundtrip adds the generic binary subtype
+            const auto roundtrip = json::from_bson(result);
+            CHECK(roundtrip != j);
+            CHECK(roundtrip.at("entry").get_binary().has_subtype());
+            CHECK(roundtrip.at("entry").get_binary().subtype() == 0);
+            CHECK(json::from_bson(result, true, false) == roundtrip);
+        }
+
         SECTION("non-empty object with binary member with subtype")
         {
             // an MD5 hash

tests/src/unit-cbor.cpp

@@ -1999,6 +1999,42 @@ TEST_CASE("CBOR regressions")
 }
 #endif
 
+TEST_CASE("CBOR definite length equal to the indefinite-length sentinel")
+{
+    // A definite-length array or map whose declared element count equals the
+    // reserved unknown_size() sentinel (SIZE_MAX) must be rejected. Otherwise
+    // it is read as an indefinite-length container and the following bytes are
+    // silently accepted instead of the (impossible) count being reported.
+    json _;
+
+    SECTION("array")
+    {
+        // 0x9B: array with eight-byte length; length = 0xFFFFFFFFFFFFFFFF
+        const std::vector<uint8_t> input = {0x9B, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0x01, 0x02, 0xFF};
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input), "[json.exception.out_of_range.408] syntax error while parsing CBOR size: excessive array size", json::out_of_range&);
+    }
+
+    SECTION("map")
+    {
+        // 0xBB: map with eight-byte length; length = 0xFFFFFFFFFFFFFFFF
+        const std::vector<uint8_t> input = {0xBB, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0x61, 0x61, 0x01, 0xFF};
+        CHECK_THROWS_WITH_AS(_ = json::from_cbor(input), "[json.exception.out_of_range.408] syntax error while parsing CBOR size: excessive map size", json::out_of_range&);
+    }
+
+    SECTION("indefinite-length containers are unaffected")
+    {
+        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, 0x61, 0x01, 0xFF})) == json({{"a", 1}}));
+    }
+
+    SECTION("ordinary four-byte length containers are unaffected")
+    {
+        // 0x9A/0xBA carry a four-byte length; a normal count still parses
+        CHECK(json::from_cbor(std::vector<uint8_t>({0x9A, 0x00, 0x00, 0x00, 0x02, 0x01, 0x02})) == json({1, 2}));
+        CHECK(json::from_cbor(std::vector<uint8_t>({0xBA, 0x00, 0x00, 0x00, 0x01, 0x61, 0x61, 0x01})) == json({{"a", 1}}));
+    }
+}
+
 TEST_CASE("CBOR roundtrips" * doctest::skip())
 {
     SECTION("input from flynn")

tests/src/unit-class_lexer.cpp

@@ -12,6 +12,10 @@
 #include <nlohmann/json.hpp>
 using nlohmann::json;
 
+#include <sstream> // stringstream
+#include <string> // string
+#include <vector> // vector
+
 namespace
 {
 // shortcut to scan a string literal
@@ -224,3 +228,75 @@ 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("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));
+        }
+    }
+}

tests/src/unit-class_parser.cpp

@@ -1440,6 +1440,13 @@ TEST_CASE("parser class")
             ]
         )";
 
+        const auto* structured_object = R"(
+            {
+                "foo": [1, 2],
+                "bar": 3
+            }
+        )";
+
         SECTION("filter nothing")
         {
             const json j_object = json::parse(s_object, [](int /*unused*/, json::parse_event_t /*unused*/, const json& /*unused*/) noexcept
@@ -1515,6 +1522,48 @@ TEST_CASE("parser class")
             CHECK (j_filtered2 == json({1}));
         }
 
+        SECTION("filter array in object")
+        {
+            // the array is discarded once it is already stored under its key
+            const json j_filtered1 = json::parse(structured_object, [](int /*unused*/, json::parse_event_t e, const json& /*parsed*/) noexcept
+            {
+                return e != json::parse_event_t::array_end;
+            });
+
+            CHECK (j_filtered1 == json({{"bar", 3}}));
+
+            // the array is discarded before it is stored, leaving the
+            // placeholder the key event wrote
+            const json j_filtered2 = json::parse(structured_object, [](int /*unused*/, json::parse_event_t e, const json& /*parsed*/) noexcept
+            {
+                return e != json::parse_event_t::array_start;
+            });
+
+            CHECK (j_filtered2 == json({{"bar", 3}}));
+        }
+
+        SECTION("filter value in object")
+        {
+            // the value is discarded after its key was kept, leaving the
+            // placeholder the key event wrote
+            const json j_filtered1 = json::parse(structured_object, [](int /*unused*/, json::parse_event_t e, const json & parsed) noexcept
+            {
+                return !(e == json::parse_event_t::value && parsed == json(3));
+            });
+
+            CHECK (j_filtered1 == json({{"foo", {1, 2}}}));
+
+            // the same value is discarded together with its key, so no
+            // placeholder was stored for it
+            const json j_filtered2 = json::parse(structured_object, [](int /*unused*/, json::parse_event_t e, const json & parsed) noexcept
+            {
+                return !((e == json::parse_event_t::key && parsed == json("bar")) ||
+                         (e == json::parse_event_t::value && parsed == json(3)));
+            });
+
+            CHECK (j_filtered2 == json({{"foo", {1, 2}}}));
+        }
+
         SECTION("filter specific events")
         {
             SECTION("first closing event")

tests/src/unit-convenience.cpp

@@ -100,6 +100,7 @@ void check_escaped(const char* original, const char* escaped, const bool ensure_
     std::stringstream ss;
     json::serializer s(nlohmann::detail::output_adapter<char>(ss), ' ');
     s.dump_escaped(original, ensure_ascii);
+    s.flush(); // dump_escaped writes into the serializer's internal buffer
     CHECK(ss.str() == escaped);
 }
 } // namespace

tests/src/unit-deserialization.cpp

@@ -427,6 +427,30 @@ TEST_CASE("deserialization")
                 CHECK(l.events == std::vector<std::string>({"boolean(true)"}));
             }
 
+            SECTION("from std::vector<signed char>")
+            {
+                std::vector<signed char> const v = {'t', 'r', 'u', 'e'};
+                CHECK(json::parse(v) == json(true));
+                CHECK(json::accept(v));
+
+                SaxEventLogger l;
+                CHECK(json::sax_parse(v, &l));
+                CHECK(l.events.size() == 1);
+                CHECK(l.events == std::vector<std::string>({"boolean(true)"}));
+
+                // bytes outside ASCII are negative here and must not be sign-extended;
+                // 0xC3 and 0xA9 do not fit in signed char (MSVC C4309), so spell them as negative values
+                std::vector<signed char> const umlaut = {'"', static_cast<signed char>(0xC3 - 0x100), static_cast<signed char>(0xA9 - 0x100), '"'};
+                CHECK(json::parse(umlaut) == json("\xC3\xA9"));
+                CHECK(json::accept(umlaut));
+
+                // 0xFF (spelled as -1 to stay in range) must not be reported as end of input
+                std::vector<signed char> const trailing = {'t', 'r', 'u', 'e', static_cast<signed char>(0xFF - 0x100)};
+                json _;
+                CHECK_THROWS_WITH_AS(_ = json::parse(trailing), "[json.exception.parse_error.101] parse error at line 1, column 5: syntax error while parsing value - invalid literal; last read: 'true\xFF'; expected end of input", json::parse_error&);
+                CHECK(!json::accept(trailing));
+            }
+
             SECTION("from std::array")
             {
                 std::array<uint8_t, 5> const v { {'t', 'r', 'u', 'e'} };

tests/src/unit-regression2.cpp

@@ -1530,4 +1530,29 @@ TEST_CASE("issue #4320 - custom base class must not leak nlohmann::detail into A
     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());
+        }
+    }
+}
+
 DOCTEST_CLANG_SUPPRESS_WARNING_POP

tests/src/unit-serialization.cpp

@@ -382,3 +382,91 @@ 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(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\"");
+    }
+}

tests/src/unit-type_traits.cpp

@@ -53,4 +53,34 @@ TEST_CASE("type traits")
             // NOLINTEND(hicpp-avoid-c-arrays,modernize-avoid-c-arrays,cppcoreguidelines-avoid-c-arrays)
         }
     }
+
+    SECTION("char_traits")
+    {
+        SECTION("to_int_type does not sign-extend")
+        {
+            using unsigned_traits = nlohmann::detail::char_traits<unsigned char>;
+            using signed_traits = nlohmann::detail::char_traits<signed char>;
+
+            CHECK(unsigned_traits::to_int_type(static_cast<unsigned char>(0x7F)) == 0x7F);
+            CHECK(unsigned_traits::to_int_type(static_cast<unsigned char>(0x80)) == 0x80);
+            CHECK(unsigned_traits::to_int_type(static_cast<unsigned char>(0xFF)) == 0xFF);
+
+            CHECK(signed_traits::to_int_type(static_cast<signed char>(0x7F)) == 0x7F);
+            // 0x80 and 0xFF do not fit in signed char (MSVC C4309), so spell them as negative values
+            CHECK(signed_traits::to_int_type(static_cast<signed char>(0x80 - 0x100)) == 0x80);
+            CHECK(signed_traits::to_int_type(static_cast<signed char>(0xFF - 0x100)) == 0xFF);
+        }
+
+        SECTION("no byte value collides with eof")
+        {
+            using unsigned_traits = nlohmann::detail::char_traits<unsigned char>;
+            using signed_traits = nlohmann::detail::char_traits<signed char>;
+
+            for (int i = 0; i < 256; ++i)
+            {
+                CHECK(unsigned_traits::to_int_type(static_cast<unsigned char>(i)) != unsigned_traits::eof());
+                CHECK(signed_traits::to_int_type(static_cast<signed char>(i)) != signed_traits::eof());
+            }
+        }
+    }
 }

tests/src/unit-ubjson.cpp

@@ -819,6 +819,8 @@ TEST_CASE("UBJSON")
                     CHECK_THROWS_WITH_AS(_ = json::from_ubjson(vec2), "[json.exception.parse_error.115] parse error at byte 5: syntax error while parsing UBJSON high-precision number: invalid number text: 1A", json::parse_error);
                     std::vector<uint8_t> const vec3 = {'H', 'i', 2, '1', '.'};
                     CHECK_THROWS_WITH_AS(_ = json::from_ubjson(vec3), "[json.exception.parse_error.115] parse error at byte 5: syntax error while parsing UBJSON high-precision number: invalid number text: 1.", json::parse_error);
+                    std::vector<uint8_t> const vec_overflow = {'H', 'i', 5, '1', 'e', '4', '0', '0'};
+                    CHECK_THROWS_WITH_AS(_ = json::from_ubjson(vec_overflow), "[json.exception.out_of_range.406] number overflow parsing '1e400'", json::out_of_range&);
                     std::vector<uint8_t> const vec4 = {'H', 2, '1', '0'};
                     CHECK_THROWS_WITH_AS(_ = json::from_ubjson(vec4), "[json.exception.parse_error.113] parse error at byte 2: syntax error while parsing UBJSON size: expected length type specification (U, i, I, l, L) after '#'; last byte: 0x02", json::parse_error);
                 }