Reserve output capacity up front for binary serialization

The vector-returning to_cbor/to_msgpack/to_ubjson/to_bjdata/to_bson grew
the output buffer purely by geometric reallocation. Reserving an estimate
up front avoids the early reallocations, which is the dominant per-byte
cost for array/object-heavy output.

The estimate (binary_reserve_hint) is deliberately conservative and safe
against untrusted input: it consults only the top-level element count
(O(1), no walk of the DOM), guards the multiplication against overflow,
and clamps the result to a fixed 1 MiB ceiling, so a large or hostile DOM
can never force an oversized allocation here. The buffer still grows
geometrically past the hint, so an underestimate only costs a few later
reallocations; scalars/strings/binary are written in one shot and get no
hint. Reserving capacity does not change the bytes produced.

Throughput (g++/clang -O3, vs the previous commit):
  cbor int array     +10% / +13%
  cbor object array  +20% / +38%

Output is byte-for-byte identical to develop across the binary
differential corpus.

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@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+      uses: github/codeql-action/init@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0
       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@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+      uses: github/codeql-action/autobuild@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0
 
     - name: Perform CodeQL Analysis
-      uses: github/codeql-action/analyze@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+      uses: github/codeql-action/analyze@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0

.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@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+        uses: github/codeql-action/upload-sarif@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0
         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@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+        uses: github/codeql-action/upload-sarif@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0
         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@54f647b7e1bb85c95cddabcd46b0c578ec92bc1a # v4.36.3
+        uses: github/codeql-action/upload-sarif@99df26d4f13ea111d4ec1a7dddef6063f76b97e9 # v4.37.0
         with:
           sarif_file: semgrep.sarif
         if: always()

.github/workflows/stale.yml

@@ -20,7 +20,7 @@ jobs:
         with:
           egress-policy: audit
 
-      - uses: actions/stale@eb5cf3af3ac0a1aa4c9c45633dd1ae542a27a899 # v10.3.0
+      - uses: actions/stale@1e223db275d687790206a7acac4d1a11bd6fe629 # v10.4.0
         with:
           stale-issue-label: 'state: stale'
           stale-pr-label: 'state: stale'

BUILD.bazel

@@ -21,7 +21,6 @@ cc_library(
         "include/nlohmann/adl_serializer.hpp",
         "include/nlohmann/byte_container_with_subtype.hpp",
         "include/nlohmann/detail/abi_macros.hpp",
-        "include/nlohmann/detail/conversions/from_chars.hpp",
         "include/nlohmann/detail/conversions/from_json.hpp",
         "include/nlohmann/detail/conversions/to_chars.hpp",
         "include/nlohmann/detail/conversions/to_json.hpp",

include/nlohmann/detail/conversions/from_chars.hpp

@@ -1,636 +0,0 @@
-//     __ _____ _____ _____
-//  __|  |   __|     |   | |  JSON for Modern C++
-// |  |  |__   |  |  | | | |  version 3.12.0
-// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
-//
-// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
-// SPDX-License-Identifier: MIT
-
-#pragma once
-
-#include <nlohmann/detail/macro_scope.hpp>
-
-#if JSON_HAS_FROM_CHARS
-    #include <algorithm> // find, find_if
-    #include <charconv> // from_chars, from_chars_result
-    #include <cmath> // isnan
-#else
-    #include <cctype> // isspace
-    #include <cerrno> // ERANGE
-    #include <clocale> // localeconv, newlocale, freelocale
-    #include <cstdlib> // strto*
-    #include <memory> // unique_ptr
-    #include <type_traits> // enable_if, is_unsigned, remove_pointer
-    #include <utility> // declval
-
-    #ifdef __has_include
-        #if __has_include(<xlocale.h>)
-            #include <xlocale.h> // strto*_l, isspace_l
-        #endif
-    #endif
-#endif
-#include <limits> // numeric_limits<T>::[has_]infinity, quiet_NaN
-#include <system_error> // errc
-
-NLOHMANN_JSON_NAMESPACE_BEGIN
-namespace detail
-{
-
-constexpr char get_locale_independent_decimal_point() noexcept
-{
-    return '.';
-}
-
-#if JSON_HAS_FROM_CHARS
-
-//////////////////////////////////////////////
-// Delegate to std::from_chars if supported //
-//////////////////////////////////////////////
-
-/*!
-@brief Number parsing implementation details
-
-A traits class template that allows users of nlohmann::detail::from_chars to
-query details of the used implementation.
-
-@tparam T numeric type to parse, matching the `value` argument of from_chars
-*/
-template <typename T>
-struct from_chars_traits
-{
-    /// whether the from_chars in unaffected by the current global C locale
-    static constexpr bool is_locale_independent = true;
-
-    /// getter for the character used as decimal point by from_chars
-    static constexpr char(*get_decimal_point)() = &get_locale_independent_decimal_point;
-};
-
-using std::from_chars_result;
-
-/*!
-@brief Parse integer/floating-point numbers
-
-Parses the string [first, last) into the numeric type T.
-
-This implementation merely delegates to `std::from_chars` with one noteworthy
-difference: Different C++ standard library implementations do not agree on
-whether `value` should be set in the out-of-range case for floating-point
-numbers. Presently, libstdc++ leaves `value` unchanged whereas libc++ and MS STL
-set `value` to +/-0 or +/-inf which allows its users to distinguish between the
-various cases of over- and underflow. The C++ proposal [P4168][1] details this
-discrepancy and advocates to standardize the latter behavior.
-
-Since we need to be able to distinguish absolute underflow (+/-0) from absolute
-overflow (+/-inf), this implementation "corrects" the out-of-range behavior
-accordingly by estimating the effective exponent through manual string parsing.
-
-[1]: https://isocpp.org/files/papers/P4168R0.html
-
-@tparam T numeric type to parse
-@param[in] first points to the beginning of the parsed string (inclusive)
-@param[in] end points to the end of the parsed string (exclusive)
-@param[out] value references the variable to set the parsed number
-@return structure consisting of `ptr` and `ec` such that:
-        - If the string starting at `first` matches a number, `ptr` points to
-          the address immediately after the last character that matches.
-          * If the matched number can be represented by `T`, `value` is set and
-            `ec` is value-initialized.
-          * If the matched number cannot be represented by `T`, `ec` is set to
-            `std::errc::result_out_of_range` and depending on the type of `T`:
-            + `value` is left unchanged for integral `T`,
-            + `value` is set to +/-0 or +/-inf for floating-point `T`.
-        - If the string starting at `first` doesn't match a number, `ptr`
-          points to `first`, `ec` is `std::errc::invalid_argument`, and `value`
-          is left unchanged.
-*/
-template <typename T>
-JSON_HEDLEY_NON_NULL(1, 2)
-inline from_chars_result from_chars(const char* first, const char* last, T& value)
-{
-    // No implementation of std::from_chars will set its value argument to NaN,
-    // so by using NaN as the initial value, we can tell if it changed.
-    T inner_value = std::numeric_limits<T>::quiet_NaN();
-    auto res = std::from_chars(first, last, inner_value);
-
-    if constexpr (std::numeric_limits<T>::has_infinity)
-    {
-        if (res.ec == std::errc::result_out_of_range)
-        {
-            if (std::isnan(inner_value)) // inner_value was not set
-            {
-                // [first, res.ptr) matches a valid floating-point number that's
-                // out-of-range and our std::from_chars did not set `value`, so we
-                // need to distinguish the type of under-/overflow ourselves.
-                const char* mantissa_begin = *first == '-' ? first + 1 : first;
-                const char* mantissa_end = std::find_if(mantissa_begin, res.ptr, [](char c)
-                {
-                    return c == 'e' || c == 'E';
-                });
-                const char* decimal_point = std::find(mantissa_begin, mantissa_end, '.');
-                const char* significant_digit = std::find_if(mantissa_begin, mantissa_end, [](char d)
-                {
-                    return d >= '1' && d <= '9';
-                });
-
-                // Position of the significant digit relative to the decimal gives
-                // the order of magnitude of the mantissa.
-                auto effective_exponent = static_cast<int>(decimal_point - significant_digit);
-
-                // If the number includes an explicit exponent, add that to the the
-                // total exponent.
-                if (mantissa_end != res.ptr)
-                {
-                    const char* exponent_begin = *(mantissa_end + 1) == '+'
-                                                 ? mantissa_end + 2 // skip the exponent's + sign
-                                                 : mantissa_end + 1;
-                    int exponent = 0;
-                    auto exp_res = std::from_chars(exponent_begin, res.ptr, exponent);
-                    JSON_ASSERT(exp_res.ptr == res.ptr);
-
-                    if (exp_res.ec == std::errc::result_out_of_range)
-                    {
-                        // NOTE: Parentheses around function names mitigate min/max macro collision on Windows
-                        effective_exponent = *exponent_begin == '-'
-                                             ? (std::numeric_limits<int>::min)()
-                                             : (std::numeric_limits<int>::max)();
-                    }
-                    else
-                    {
-                        JSON_ASSERT(exp_res.ec == std::errc{});
-                        effective_exponent += exponent;
-                    }
-                }
-
-                // Set `value` to the sign-correct non-finite value based on the
-                // effective exponent.
-                value = (*first == '-' ? -1 : 1) * (effective_exponent < 0
-                                                    ? T{}
-                                                    : std::numeric_limits<T>::infinity());
-            }
-            else
-            {
-                value = inner_value;
-            }
-        }
-    }
-
-    if (res.ec == std::errc{})
-    {
-        value = inner_value;
-    }
-
-    return res;
-}
-
-#else
-
-//////////////////////////////////////////////
-// Fallback implementation using strto*[_l] //
-//////////////////////////////////////////////
-
-#if defined(_WIN32)
-
-/*!
-@brief Extended locale functions on Windows
-
-A trait object which provides wrappers for isspace and the strto* family of
-functions, based on the platform's support for extended locale APIs.
-
-This is the Windows implementation. The primary class template definition
-falls back to the standard locale-dependent functions, which is used on
-MinGW, whose C runtime only provides an incomplete subset of the
-`_<funcname>_l` family (e.g., missing `_strtof_l`/`_strtold_l` depending on
-the toolchain version). A specialization for genuine MSVC (including
-clang-cl, which mimics the MSVC ABI and CRT) is used instead, as the full
-`_<funcname>_l` family has reliably been available there since at least
-Visual Studio 2005.
-*/
-template <typename = const char*, typename = float>
-struct extended_locale_traits
-{
-    static constexpr bool is_locale_independent = false;
-
-    JSON_HEDLEY_PURE
-    static char get_decimal_point() noexcept
-    {
-        const auto* loc = localeconv();
-        JSON_ASSERT(loc != nullptr);
-        return (loc->decimal_point == nullptr) ? '.' : *(loc->decimal_point);
-    }
-
-    static int isspace(int c) noexcept
-    {
-        return std::isspace(c);
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static long long strtoll(const char* str, char** endptr) noexcept
-    {
-        return std::strtoll(str, endptr, 10);
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static unsigned long long strtoull(const char* str, char** endptr) noexcept
-    {
-        return std::strtoull(str, endptr, 10);
-    }
-
-    static constexpr float(&strtof)(const char*, char**) = std::strtof;
-    static constexpr double(&strtod)(const char*, char**) = std::strtod;
-    static constexpr long double(&strtold)(const char*, char**) = std::strtold;
-};
-
-#if defined(_MSC_VER)
-
-/*!
-@brief Get a pointer to a globally shared instance of the "C" locale on Windows
-
-This function uses _create_locale/_free_locale to allocate a "C" locale instance
-which can be passed to extended locale APIs. This instance is created on first
-use and has static storage duration, such that it can be used for the duration
-of the program.
-*/
-inline _locale_t get_c_locale_t() noexcept
-{
-    struct LocaleTDeleter
-    {
-        void operator()(_locale_t loc) noexcept
-        {
-            ::_free_locale(loc);
-        }
-    };
-    using LocaleUPtr = std::unique_ptr<std::remove_pointer<_locale_t>::type, LocaleTDeleter>;
-    static const LocaleUPtr c_locale = LocaleUPtr{::_create_locale(LC_ALL, "C"), LocaleTDeleter{}};
-    return c_locale.get();
-}
-
-template <typename T>
-struct extended_locale_traits<T, float>
-{
-    static constexpr bool is_locale_independent = true;
-    static constexpr char(&get_decimal_point)() = get_locale_independent_decimal_point;
-
-    static int isspace(int c) noexcept
-    {
-        return _isspace_l(c, get_c_locale_t());
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static long long strtoll(const char* str, char** endptr) noexcept
-    {
-        return _strtoll_l(str, endptr, 10, get_c_locale_t());
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static unsigned long long strtoull(const char* str, char** endptr) noexcept
-    {
-        return _strtoull_l(str, endptr, 10, get_c_locale_t());
-    }
-
-    static float strtof(T str, char** str_end)
-    {
-        return _strtof_l(str, str_end, get_c_locale_t());
-    }
-
-    static double strtod(T str, char** str_end)
-    {
-        return _strtod_l(str, str_end, get_c_locale_t());
-    }
-
-    static long double strtold(T str, char** str_end)
-    {
-        return _strtold_l(str, str_end, get_c_locale_t());
-    }
-};
-
-#endif
-
-#else
-
-/*!
-@brief Get a pointer to a globally shared instance of the "C" locale on POSIX
-
-This function uses newlocale/freelocale to allocate a "C" locale instance
-which can be passed to extended locale APIs. This instance is created on first
-use and has static storage duration, such that it can be used for the duration
-of the program.
-
-newlocale/freelocale themselves are in POSIX and are available independent of
-the platform's support for extended locale APIs.
-*/
-inline locale_t get_c_locale_t() noexcept
-{
-    struct LocaleTDeleter
-    {
-        void operator()(locale_t loc) noexcept
-        {
-            ::freelocale(loc);
-        }
-    };
-    using LocaleUPtr = std::unique_ptr<std::remove_pointer<locale_t>::type, LocaleTDeleter>;
-
-#if defined(__clang__)
-#pragma clang diagnostic push
-#pragma clang diagnostic ignored "-Wexit-time-destructors"
-#endif
-    static const LocaleUPtr c_locale = LocaleUPtr {::newlocale(LC_ALL_MASK, "C", nullptr), LocaleTDeleter{}};
-#if defined(__clang__)
-#pragma clang diagnostic pop
-#endif
-
-    return c_locale.get();
-}
-
-/*!
-@brief Extended locale functions on POSIX
-
-A trait object which provides wrappers for isspace and the strto* family of
-functions, based on the platform's support for extended locale APIs.
-
-This is the POSIX implementation where extended locale support might not be
-available. The primary class template definition falls back to the standard
-locale-dependent functions. A partial specialization uses SFINAE to detect the
-availability of `strtof_l` (as a proxy for the whole `<funcname>_l` family) and
-provides access to locale-independent functions.
-*/
-template <typename = const char*, typename = float>
-struct extended_locale_traits
-{
-    static constexpr bool is_locale_independent = false;
-
-    /*!
-    @brief Query the decimal point used by the current C locale
-
-    Note that calling this function while switching the global C locale from
-    another thread is undefined behavior.
-    */
-    JSON_HEDLEY_PURE
-    static char get_decimal_point() noexcept
-    {
-        const auto* loc = localeconv();
-        JSON_ASSERT(loc != nullptr);
-        return (loc->decimal_point == nullptr) ? '.' : *(loc->decimal_point);
-    }
-
-    static int isspace(int c) noexcept
-    {
-        return std::isspace(c);
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static long long strtoll(const char* str, char** endptr) noexcept
-    {
-        return std::strtoll(str, endptr, 10);
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static unsigned long long strtoull(const char* str, char** endptr) noexcept
-    {
-        return std::strtoull(str, endptr, 10);
-    }
-
-    static constexpr float(&strtof)(const char*, char**) = std::strtof;
-    static constexpr double(&strtod)(const char*, char**) = std::strtod;
-    static constexpr long double(&strtold)(const char*, char**) = std::strtold;
-};
-
-template <typename T>
-struct extended_locale_traits<T, decltype(strtof_l(
-            std::declval<T>(),
-            std::declval<char**>(),
-            std::declval<locale_t>()))>
-{
-    static constexpr bool is_locale_independent = true;
-    static constexpr char(&get_decimal_point)() = get_locale_independent_decimal_point;
-
-    static int isspace(int c) noexcept
-    {
-        return isspace_l(c, get_c_locale_t());
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static long long strtoll(const char* str, char** endptr) noexcept
-    {
-        return ::strtoll_l(str, endptr, 10, get_c_locale_t());
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static unsigned long long strtoull(const char* str, char** endptr) noexcept
-    {
-        return ::strtoull_l(str, endptr, 10, get_c_locale_t());
-    }
-
-    static float strtof(T str, char** str_end)
-    {
-        return ::strtof_l(str, str_end, get_c_locale_t());
-    }
-
-    static double strtod(T str, char** str_end)
-    {
-        return ::strtod_l(str, str_end, get_c_locale_t());
-    }
-
-    static long double strtold(T str, char** str_end)
-    {
-        return ::strtold_l(str, str_end, get_c_locale_t());
-    }
-};
-
-#endif
-
-/*!
-@brief Number parsing implementation details
-
-A traits class template that allows users of nlohmann::detail::from_chars to
-query details of the used implementation.
-
-Each specialization provides the following static members:
-- is_locale_independent: whether the from_chars in unaffected by the current
-                         global C locale
-- get_decimal_point: getter for the character used as decimal point by from_chars
-- strto: dispatches to the C standard library strto* function for type T, or to
-         the corresponding extended locale function, if available.
-
-@tparam T numeric type to parse, matching the `value` argument of from_chars
-*/
-template <typename T>
-struct from_chars_traits;
-
-template <>
-struct from_chars_traits<long long> // NOLINT(runtime/int)
-{
-    static constexpr bool is_locale_independent = extended_locale_traits<>::is_locale_independent;
-
-    static char get_decimal_point()
-    {
-        return extended_locale_traits<>::get_decimal_point();
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static constexpr long long(*strto)(const char*, char**) = &extended_locale_traits<>::strtoll;
-};
-
-template <>
-struct from_chars_traits<unsigned long long> // NOLINT(runtime/int)
-{
-    static constexpr bool is_locale_independent = extended_locale_traits<>::is_locale_independent;
-
-    static char get_decimal_point()
-    {
-        return extended_locale_traits<>::get_decimal_point();
-    }
-
-    // NOLINTNEXTLINE(runtime/int)
-    static constexpr unsigned long long(*strto)(const char*, char**) = &extended_locale_traits<>::strtoull;
-};
-
-template <>
-struct from_chars_traits<float>
-{
-    static constexpr bool is_locale_independent = extended_locale_traits<>::is_locale_independent;
-
-    static char get_decimal_point()
-    {
-        return extended_locale_traits<>::get_decimal_point();
-    }
-
-    static constexpr float(*strto)(const char*, char**) = &extended_locale_traits<>::strtof;
-};
-
-template <>
-struct from_chars_traits<double>
-{
-    static constexpr bool is_locale_independent = extended_locale_traits<>::is_locale_independent;
-
-    static char get_decimal_point()
-    {
-        return extended_locale_traits<>::get_decimal_point();
-    }
-
-    static constexpr double(*strto)(const char*, char**) = &extended_locale_traits<>::strtod;
-};
-
-template <>
-struct from_chars_traits<long double>
-{
-    static constexpr bool is_locale_independent = extended_locale_traits<>::is_locale_independent;
-
-    static char get_decimal_point()
-    {
-        return extended_locale_traits<>::get_decimal_point();
-    }
-
-    static constexpr long double(*strto)(const char*, char**) = &extended_locale_traits<>::strtold;
-};
-
-template <typename T>
-constexpr typename std::enable_if<std::numeric_limits<T>::has_infinity, bool>::type is_out_of_range_value(T value) noexcept
-{
-#ifdef __GNUC__
-#pragma GCC diagnostic push
-#pragma GCC diagnostic ignored "-Wfloat-equal"
-#endif
-    return value == T {} || value == std::numeric_limits<T>::infinity() || value == -std::numeric_limits<T>::infinity();
-#ifdef __GNUC__
-#pragma GCC diagnostic pop
-#endif
-}
-
-template <typename T>
-constexpr typename std::enable_if < !std::numeric_limits<T>::has_infinity, bool >::type is_out_of_range_value(T value) noexcept
-{
-    // NOTE: Parentheses around function names mitigate min/max macro collision on Windows
-    return value == (std::numeric_limits<T>::max)() || value == (std::numeric_limits<T>::min)();
-}
-
-struct from_chars_result
-{
-    const char* ptr;
-    std::errc ec;
-};
-
-/*!
-@brief Parse integer/floating-point numbers
-
-Parses the string [first, last) into the numeric type T.
-
-This implementation uses the C standard library functions strto*, or their
-extended locale counterparts strto*_l, to emulate the behavior of
-`std::from_chars` on platforms where it is not (fully) supported.
-
-Regarding the under-/overflow behavior, this implementation adopts the behavior
-proposed by [P4168][1] and implemented by libc++ and MS STL of setting `value`
-to the corresponding non-finite floating-point value in case of an out-of-range
-result.
-
-[1]: https://isocpp.org/files/papers/P4168R0.html
-
-@pre Unlike std::from_chars, this implementation requires the string to be
-     null-terminated, such that `last` dereferences into a NUL byte.
-
-@tparam T numeric type to parse
-@param[in] first points to the beginning of the parsed string (inclusive)
-@param[in] end points to the end of the parsed string (exclusive)
-@param[out] value references the variable to set the parsed number
-@return structure consisting of `ptr` and `ec` such that:
-        - If the string starting at `first` matches a number, `ptr` points to
-          the address immediately after the last character that matches.
-          * If the matched number can be represented by `T`, `value` is set and
-            `ec` is value-initialized.
-          * If the matched number cannot be represented by `T`, `ec` is set to
-            `std::errc::result_out_of_range` and depending on the type of `T`:
-            + `value` is left unchanged for integral `T`,
-            + `value` is set to +/-0 or +/-inf for floating-point `T`.
-        - If the string starting at `first` doesn't match a number, `ptr`
-          points to `first`, `ec` is `std::errc::invalid_argument`, and `value`
-          is left unchanged.
-*/
-template <typename T>
-JSON_HEDLEY_NON_NULL(1, 2)
-inline from_chars_result from_chars(const char* first, const char* last, T& value)
-{
-    JSON_ASSERT(*last == '\0');
-
-    // Unlike strto*, from_chars does not accept leading whitespace or + signs
-    if (first == last || *first == '+'
-            || (std::is_unsigned<T>::value && *first == '-')
-            || extended_locale_traits<>::isspace(*first) != 0)
-    {
-        return {first, std::errc::invalid_argument};
-    }
-
-    errno = 0;
-    char* ptr = nullptr; // NOLINT(misc-const-correctness)
-    T result = from_chars_traits<T>::strto(first, &ptr);
-
-    if (ptr == first)
-    {
-        return {ptr, std::errc::invalid_argument};
-    }
-
-    // Upon under-/overflow, strto* returns a marginal value and sets errno.
-    // Note that it is NOT sufficient to just check errno: strto* only clears
-    // errno if the parsed string actually parses into 0/[U]LLONG_MIN/-_MAX
-    // and this result was returned without indicating an under-/overflow.
-    // Otherwise, errno may not be relied upon to indicate the *absence* of
-    // an out-of-range error.
-    if (is_out_of_range_value(result) && errno == ERANGE)
-    {
-        if (std::numeric_limits<T>::has_infinity)
-        {
-            // ONLY for floating-point types, set `value` to the same non-finite
-            // result returned by strto* to allow users to distinguish between
-            // different types of under-/overflow.
-            value = result;
-        }
-        return {ptr, std::errc::result_out_of_range};
-    }
-
-    value = result;
-    return {ptr, std::errc{}};
-}
-
-#endif
-
-}  // namespace detail
-NLOHMANN_JSON_NAMESPACE_END

include/nlohmann/detail/input/binary_reader.hpp

@@ -1846,6 +1846,29 @@ class binary_reader
         return get_ubjson_value(get_char ? get_ignore_noop() : current);
     }
 
+    /*!
+    @brief reject a negative UBJSON/BJData string length
+
+    String and key lengths are written with signed integer markers (i, I, l,
+    L). A negative value is malformed; without this check get_string() would
+    silently treat it as an empty string and leave the following bytes to be
+    misread as the next value. This mirrors the non-negative check the
+    optimized-container count path already performs in get_ubjson_size_value.
+
+    @param[in] len  the string length read from the input
+    @return whether the length is valid (non-negative)
+    */
+    template<typename NumberType>
+    bool check_ubjson_string_length(const NumberType len)
+    {
+        if (JSON_HEDLEY_UNLIKELY(len < 0))
+        {
+            return sax->parse_error(chars_read, get_token_string(), parse_error::create(113, chars_read,
+                                    exception_message(input_format, "string length must not be negative", "string"), nullptr));
+        }
+        return true;
+    }
+
     /*!
     @brief reads a UBJSON string
 
@@ -1883,25 +1906,25 @@ class binary_reader
             case 'i':
             {
                 std::int8_t len{};
-                return get_number(input_format, len) && get_string(input_format, len, result);
+                return get_number(input_format, len) && check_ubjson_string_length(len) && get_string(input_format, len, result);
             }
 
             case 'I':
             {
                 std::int16_t len{};
-                return get_number(input_format, len) && get_string(input_format, len, result);
+                return get_number(input_format, len) && check_ubjson_string_length(len) && get_string(input_format, len, result);
             }
 
             case 'l':
             {
                 std::int32_t len{};
-                return get_number(input_format, len) && get_string(input_format, len, result);
+                return get_number(input_format, len) && check_ubjson_string_length(len) && get_string(input_format, len, result);
             }
 
             case 'L':
             {
                 std::int64_t len{};
-                return get_number(input_format, len) && get_string(input_format, len, result);
+                return get_number(input_format, len) && check_ubjson_string_length(len) && get_string(input_format, len, result);
             }
 
             case 'u':

include/nlohmann/detail/input/input_adapters.hpp

@@ -395,17 +395,30 @@ struct wide_string_input_helper<BaseInputAdapter, 2>
             }
             else
             {
-                if (JSON_HEDLEY_UNLIKELY(!input.empty()))
+                // A supplementary code point is a high surrogate (0xD800..0xDBFF)
+                // followed by a low surrogate (0xDC00..0xDFFF). A lone low
+                // surrogate, a high surrogate at the end of the input, or a high
+                // surrogate followed by any other unit is malformed UTF-16. In
+                // that case the offending unit is passed through unchanged so the
+                // UTF-8 decoder rejects it, matching how \uXXXX surrogate escapes
+                // are handled in the lexer.
+                bool valid_pair = false;
+                if (wc <= 0xDBFF && JSON_HEDLEY_UNLIKELY(!input.empty()))
                 {
                     const auto wc2 = static_cast<unsigned int>(input.get_character());
-                    const auto charcode = 0x10000u + (((static_cast<unsigned int>(wc) & 0x3FFu) << 10u) | (wc2 & 0x3FFu));
-                    utf8_bytes[0] = static_cast<std::char_traits<char>::int_type>(0xF0u | (charcode >> 18u));
-                    utf8_bytes[1] = static_cast<std::char_traits<char>::int_type>(0x80u | ((charcode >> 12u) & 0x3Fu));
-                    utf8_bytes[2] = static_cast<std::char_traits<char>::int_type>(0x80u | ((charcode >> 6u) & 0x3Fu));
-                    utf8_bytes[3] = static_cast<std::char_traits<char>::int_type>(0x80u | (charcode & 0x3Fu));
-                    utf8_bytes_filled = 4;
+                    if (0xDC00 <= wc2 && wc2 <= 0xDFFF)
+                    {
+                        const auto charcode = 0x10000u + (((static_cast<unsigned int>(wc) & 0x3FFu) << 10u) | (wc2 & 0x3FFu));
+                        utf8_bytes[0] = static_cast<std::char_traits<char>::int_type>(0xF0u | (charcode >> 18u));
+                        utf8_bytes[1] = static_cast<std::char_traits<char>::int_type>(0x80u | ((charcode >> 12u) & 0x3Fu));
+                        utf8_bytes[2] = static_cast<std::char_traits<char>::int_type>(0x80u | ((charcode >> 6u) & 0x3Fu));
+                        utf8_bytes[3] = static_cast<std::char_traits<char>::int_type>(0x80u | (charcode & 0x3Fu));
+                        utf8_bytes_filled = 4;
+                        valid_pair = true;
+                    }
                 }
-                else
+
+                if (!valid_pair)
                 {
                     utf8_bytes[0] = static_cast<std::char_traits<char>::int_type>(wc);
                     utf8_bytes_filled = 1;

include/nlohmann/detail/input/lexer.hpp

@@ -9,15 +9,15 @@
 #pragma once
 
 #include <array> // array
+#include <clocale> // localeconv
 #include <cstddef> // size_t
 #include <cstdio> // snprintf
+#include <cstdlib> // strtof, strtod, strtold, strtoll, strtoull
 #include <initializer_list> // initializer_list
 #include <string> // char_traits, string
-#include <system_error> // errc
 #include <utility> // move
 #include <vector> // vector
 
-#include <nlohmann/detail/conversions/from_chars.hpp>
 #include <nlohmann/detail/input/input_adapters.hpp>
 #include <nlohmann/detail/input/position_t.hpp>
 #include <nlohmann/detail/macro_scope.hpp>
@@ -152,7 +152,7 @@ class lexer : public lexer_base<BasicJsonType>
     explicit lexer(InputAdapterType&& adapter, bool ignore_comments_ = false) noexcept
         : ia(std::move(adapter))
         , ignore_comments(ignore_comments_)
-        , decimal_point_char(static_cast<char_int_type>(from_chars_traits<number_float_t>::get_decimal_point()))
+        , decimal_point_char(static_cast<char_int_type>(get_decimal_point()))
     {}
 
     // deleted because of pointer members
@@ -163,6 +163,19 @@ class lexer : public lexer_base<BasicJsonType>
     ~lexer() = default;
 
   private:
+    /////////////////////
+    // locales
+    /////////////////////
+
+    /// return the locale-dependent decimal point
+    JSON_HEDLEY_PURE
+    static char get_decimal_point() noexcept
+    {
+        const auto* loc = localeconv();
+        JSON_ASSERT(loc != nullptr);
+        return (loc->decimal_point == nullptr) ? '.' : *(loc->decimal_point);
+    }
+
     /////////////////////
     // scan functions
     /////////////////////
@@ -928,6 +941,24 @@ class lexer : public lexer_base<BasicJsonType>
         }
     }
 
+    JSON_HEDLEY_NON_NULL(2)
+    static void strtof(float& f, const char* str, char** endptr) noexcept
+    {
+        f = std::strtof(str, endptr);
+    }
+
+    JSON_HEDLEY_NON_NULL(2)
+    static void strtof(double& f, const char* str, char** endptr) noexcept
+    {
+        f = std::strtod(str, endptr);
+    }
+
+    JSON_HEDLEY_NON_NULL(2)
+    static void strtof(long double& f, const char* str, char** endptr) noexcept
+    {
+        f = std::strtold(str, endptr);
+    }
+
     /*!
     @brief scan a number literal
 
@@ -1248,18 +1279,19 @@ scan_number_done:
         // we are done scanning a number)
         unget();
 
+        char* endptr = nullptr; // NOLINT(misc-const-correctness,cppcoreguidelines-pro-type-vararg,hicpp-vararg)
+        errno = 0;
+
         // try to parse integers first and fall back to floats
         if (number_type == token_type::value_unsigned)
         {
-            unsigned long long x{}; // NOLINT(runtime/int)
-            const auto res = ::nlohmann::detail::from_chars(token_buffer.data(), token_buffer.data() + token_buffer.size(), x);
+            const auto x = std::strtoull(token_buffer.data(), &endptr, 10);
 
             // we checked the number format before
-            JSON_ASSERT(res.ptr == token_buffer.data() + token_buffer.size());
+            JSON_ASSERT(endptr == token_buffer.data() + token_buffer.size());
 
-            if (res.ec != std::errc::result_out_of_range)
+            if (errno != ERANGE)
             {
-                JSON_ASSERT(res.ec == std::errc{});
                 value_unsigned = static_cast<number_unsigned_t>(x);
                 if (value_unsigned == x)
                 {
@@ -1269,15 +1301,13 @@ scan_number_done:
         }
         else if (number_type == token_type::value_integer)
         {
-            long long x{}; // NOLINT(runtime/int)
-            const auto res = ::nlohmann::detail::from_chars(token_buffer.data(), token_buffer.data() + token_buffer.size(), x);
+            const auto x = std::strtoll(token_buffer.data(), &endptr, 10);
 
             // we checked the number format before
-            JSON_ASSERT(res.ptr == token_buffer.data() + token_buffer.size());
+            JSON_ASSERT(endptr == token_buffer.data() + token_buffer.size());
 
-            if (res.ec != std::errc::result_out_of_range)
+            if (errno != ERANGE)
             {
-                JSON_ASSERT(res.ec == std::errc{});
                 value_integer = static_cast<number_integer_t>(x);
                 if (value_integer == x)
                 {
@@ -1288,11 +1318,10 @@ scan_number_done:
 
         // this code is reached if we parse a floating-point number or if an
         // integer conversion above failed
-        const auto res = ::nlohmann::detail::from_chars(token_buffer.data(), token_buffer.data() + token_buffer.size(), value_float);
+        strtof(value_float, token_buffer.data(), &endptr);
 
         // we checked the number format before
-        JSON_ASSERT(res.ptr == token_buffer.data() + token_buffer.size());
-        JSON_ASSERT(res.ec != std::errc::invalid_argument);
+        JSON_ASSERT(endptr == token_buffer.data() + token_buffer.size());
 
         return token_type::value_float;
     }

include/nlohmann/detail/macro_scope.hpp

@@ -124,20 +124,6 @@
     #define JSON_HAS_FILESYSTEM 0
 #endif
 
-#ifndef JSON_HAS_FROM_CHARS
-    #if defined(JSON_HAS_CPP_17) && defined(__cpp_lib_to_chars)
-        // The std::from_chars<float> implementation in libstdc++ 11 gets some of the corner cases wrong;
-        // starting with libstdc++ 12, these are fixed by switching the implementation to fast_float.
-        #if defined(_GLIBCXX_RELEASE) && _GLIBCXX_RELEASE < 12
-            #define JSON_HAS_FROM_CHARS 0
-        #else
-            #define JSON_HAS_FROM_CHARS 1
-        #endif
-    #else
-        #define JSON_HAS_FROM_CHARS 0
-    #endif
-#endif
-
 #ifndef JSON_HAS_THREE_WAY_COMPARISON
     #if defined(__cpp_impl_three_way_comparison) && __cpp_impl_three_way_comparison >= 201907L \
         && defined(__cpp_lib_three_way_comparison) && __cpp_lib_three_way_comparison >= 201907L

include/nlohmann/detail/macro_unscope.hpp

@@ -39,7 +39,6 @@
     #undef JSON_HAS_CPP_26
     #undef JSON_HAS_FILESYSTEM
     #undef JSON_HAS_EXPERIMENTAL_FILESYSTEM
-    #undef JSON_HAS_FROM_CHARS
     #undef JSON_HAS_THREE_WAY_COMPARISON
     #undef JSON_HAS_RANGES
     #undef JSON_HAS_STD_FORMAT

include/nlohmann/detail/output/output_adapters.hpp

@@ -13,6 +13,7 @@
 #include <iterator> // back_inserter
 #include <memory> // shared_ptr, make_shared
 #include <string> // basic_string
+#include <utility> // move
 #include <vector> // vector
 
 #ifndef JSON_NO_IO
@@ -118,6 +119,72 @@ class output_string_adapter : public output_adapter_protocol<CharType>
     StringType& str;
 };
 
+/// @brief non-virtual output sink writing into a std::vector
+///
+/// Unlike output_vector_adapter, this sink is not part of the virtual
+/// output_adapter_protocol hierarchy: it is passed to binary_writer by value as
+/// a template parameter, so write_character()/write_characters() are ordinary
+/// (inlinable) calls with no vtable lookup and no shared_ptr. It is used for the
+/// common `to_cbor`/`to_msgpack`/... into a std::vector.
+template<typename CharType, typename AllocatorType = std::allocator<CharType>>
+class output_vector_sink
+{
+  public:
+    explicit output_vector_sink(std::vector<CharType, AllocatorType>& vec) noexcept
+        : v(vec)
+    {}
+
+    void write_character(CharType c)
+    {
+        v.push_back(c);
+    }
+
+    // no JSON_HEDLEY_NON_NULL here: binary_writer legitimately passes a null
+    // pointer with length 0 for empty strings/binary values. Appending an empty
+    // range is a no-op; the type-erased path tolerates this via the (unattributed)
+    // virtual base, and the concrete sink must do the same.
+    void write_characters(const CharType* s, std::size_t length)
+    {
+        v.insert(v.end(), s, s + length);
+    }
+
+  private:
+    std::vector<CharType, AllocatorType>& v;
+};
+
+/// @brief output sink forwarding to a type-erased output adapter
+///
+/// Wraps the polymorphic output_adapter_t so the same binary_writer template can
+/// also target arbitrary adapters (output streams, strings, user-provided
+/// adapters) via the `output_adapter`-based overloads. Each write still goes
+/// through one virtual call, exactly as before; only the concrete sinks above
+/// avoid it.
+template<typename CharType>
+class output_adapter_sink
+{
+  public:
+    explicit output_adapter_sink(output_adapter_t<CharType> adapter)
+        : oa(std::move(adapter))
+    {
+        JSON_ASSERT(oa);
+    }
+
+    void write_character(CharType c)
+    {
+        oa->write_character(c);
+    }
+
+    // no JSON_HEDLEY_NON_NULL: forwards (null, 0) for empty payloads, exactly as
+    // the type-erased path already did before this sink existed
+    void write_characters(const CharType* s, std::size_t length)
+    {
+        oa->write_characters(s, length);
+    }
+
+  private:
+    output_adapter_t<CharType> oa = nullptr;
+};
+
 template<typename CharType, typename StringType = std::basic_string<CharType>>
 class output_adapter
 {

include/nlohmann/json.hpp

@@ -140,7 +140,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     friend ::nlohmann::detail::serializer<basic_json>;
     template<typename BasicJsonType>
     friend class ::nlohmann::detail::iter_impl;
-    template<typename BasicJsonType, typename CharType>
+    template<typename BasicJsonType, typename CharType, typename OutputSinkType>
     friend class ::nlohmann::detail::binary_writer;
     template<typename BasicJsonType, typename InputType, typename SAX>
     friend class ::nlohmann::detail::binary_reader;
@@ -4327,7 +4327,9 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     static std::vector<std::uint8_t> to_cbor(const basic_json& j)
     {
         std::vector<std::uint8_t> result;
-        to_cbor(j, result);
+        result.reserve(detail::binary_reserve_hint(j));
+        detail::binary_writer<basic_json, std::uint8_t, detail::output_vector_sink<std::uint8_t>>(
+            detail::output_vector_sink<std::uint8_t>(result)).write_cbor(j);
         return result;
     }
 
@@ -4350,7 +4352,9 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     static std::vector<std::uint8_t> to_msgpack(const basic_json& j)
     {
         std::vector<std::uint8_t> result;
-        to_msgpack(j, result);
+        result.reserve(detail::binary_reserve_hint(j));
+        detail::binary_writer<basic_json, std::uint8_t, detail::output_vector_sink<std::uint8_t>>(
+            detail::output_vector_sink<std::uint8_t>(result)).write_msgpack(j);
         return result;
     }
 
@@ -4375,7 +4379,9 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
             const bool use_type = false)
     {
         std::vector<std::uint8_t> result;
-        to_ubjson(j, result, use_size, use_type);
+        result.reserve(detail::binary_reserve_hint(j));
+        detail::binary_writer<basic_json, std::uint8_t, detail::output_vector_sink<std::uint8_t>>(
+            detail::output_vector_sink<std::uint8_t>(result)).write_ubjson(j, use_size, use_type);
         return result;
     }
 
@@ -4403,7 +4409,9 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
             const bjdata_version_t version = bjdata_version_t::draft2)
     {
         std::vector<std::uint8_t> result;
-        to_bjdata(j, result, use_size, use_type, version);
+        result.reserve(detail::binary_reserve_hint(j));
+        detail::binary_writer<basic_json, std::uint8_t, detail::output_vector_sink<std::uint8_t>>(
+            detail::output_vector_sink<std::uint8_t>(result)).write_ubjson(j, use_size, use_type, true, true, version);
         return result;
     }
 
@@ -4430,7 +4438,9 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     static std::vector<std::uint8_t> to_bson(const basic_json& j)
     {
         std::vector<std::uint8_t> result;
-        to_bson(j, result);
+        result.reserve(detail::binary_reserve_hint(j));
+        detail::binary_writer<basic_json, std::uint8_t, detail::output_vector_sink<std::uint8_t>>(
+            detail::output_vector_sink<std::uint8_t>(result)).write_bson(j);
         return result;
     }
 

tests/CMakeLists.txt

@@ -121,10 +121,6 @@ json_test_set_test_options(test-disabled_exceptions
 # raise timeout of expensive Unicode test
 json_test_set_test_options(test-unicode4 TEST_PROPERTIES TIMEOUT 3000)
 
-# link pthreads to tests that need it
-find_package(Threads REQUIRED)
-json_test_set_test_options(test-regression2 LINK_LIBRARIES Threads::Threads)
-
 #############################################################################
 # add unit tests
 #############################################################################

tests/src/unit-bjdata.cpp

@@ -2721,6 +2721,19 @@ TEST_CASE("BJData")
                 CHECK_THROWS_WITH_AS(_ = json::from_bjdata(v), "[json.exception.parse_error.113] parse error at byte 2: syntax error while parsing BJData string: expected length type specification (U, i, u, I, m, l, M, L); last byte: 0x31", json::parse_error&);
             }
 
+            SECTION("negative length")
+            {
+                json _;
+
+                std::vector<uint8_t> const vi = {'S', 'i', 0xFF};
+                CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vi), "[json.exception.parse_error.113] parse error at byte 3: syntax error while parsing BJData string: string length must not be negative", json::parse_error&);
+                CHECK(json::from_bjdata(vi, true, false).is_discarded());
+
+                std::vector<uint8_t> const vl = {'S', 'l', 0xFF, 0xFF, 0xFF, 0xFF};
+                CHECK_THROWS_WITH_AS(_ = json::from_bjdata(vl), "[json.exception.parse_error.113] parse error at byte 6: syntax error while parsing BJData string: string length must not be negative", json::parse_error&);
+                CHECK(json::from_bjdata(vl, true, false).is_discarded());
+            }
+
             SECTION("parse bjdata markers in ubjson")
             {
                 // create a single-character string for all number types

tests/src/unit-class_lexer.cpp

@@ -55,8 +55,6 @@ TEST_CASE("lexer class")
 
         SECTION("numbers")
         {
-            // Number parsing implementation uses std::from_chars where available,
-            // so run this test suite with JSON_HAS_CPP_17 as well.
             CHECK((scan_string("0") == json::lexer::token_type::value_unsigned));
             CHECK((scan_string("1") == json::lexer::token_type::value_unsigned));
             CHECK((scan_string("2") == json::lexer::token_type::value_unsigned));
@@ -67,20 +65,13 @@ TEST_CASE("lexer class")
             CHECK((scan_string("7") == json::lexer::token_type::value_unsigned));
             CHECK((scan_string("8") == json::lexer::token_type::value_unsigned));
             CHECK((scan_string("9") == json::lexer::token_type::value_unsigned));
-            CHECK((scan_string("18446744073709551615") == json::lexer::token_type::value_unsigned));
 
             CHECK((scan_string("-0") == json::lexer::token_type::value_integer));
             CHECK((scan_string("-1") == json::lexer::token_type::value_integer));
-            CHECK((scan_string("-9223372036854775808") == json::lexer::token_type::value_integer));
 
             CHECK((scan_string("1.1") == json::lexer::token_type::value_float));
             CHECK((scan_string("-1.1") == json::lexer::token_type::value_float));
             CHECK((scan_string("1E10") == json::lexer::token_type::value_float));
-
-            // out-of-range integers/floats are treated as value_float tokens
-            CHECK((scan_string("18446744073709551616") == json::lexer::token_type::value_float));
-            CHECK((scan_string("-9223372036854775809") == json::lexer::token_type::value_float));
-            CHECK((scan_string("1E400") == json::lexer::token_type::value_float));
         }
 
         SECTION("whitespace")

tests/src/unit-from_chars.cpp

@@ -1,278 +0,0 @@
-//     __ _____ _____ _____
-//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
-// |  |  |__   |  |  | | | |  version 3.12.0
-// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
-//
-// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
-// SPDX-License-Identifier: MIT
-
-#include "doctest_compatibility.h"
-
-#include <nlohmann/json.hpp>
-
-#include <algorithm>
-#include <cmath>
-#include <cstddef>
-#include <limits>
-#include <string>
-#include <system_error>
-
-// nlohmann::details::from_chars is a wrapper around std::from_chars when it is
-// available (__cpp_lib_to_chars is defined), otherwise our fallback
-// implementation kicks in.
-//
-// Due to the incomplete implementation of this C++17 standard library feature
-// (for example, lack of support for long double even in current libc++),
-// JSON_HAS_CPP_17 does not guarantee that std::from_chars is available.
-// However, the reverse holds: If the test is compiled in C++11 mode, the
-// fallback implementation will be used for sure.
-// By mentioning the JSON_HAS_CPP_17 macro in this here comment, the test will
-// be compiled both using our fallback and, at least on some platforms,
-// the C++17 standard library version, as a way of ensuring that the
-// expectations formulated in this test align with `std::from_chars`.
-
-namespace
-{
-struct init_val_t {};
-constexpr init_val_t init_val{};
-
-template <typename T>
-void check_result(std::string& str, T& value, std::ptrdiff_t expected_ptr_offset, std::errc expected_ec)
-{
-    // On platforms where neither `std::from_chars`, nor extended locale support (`strtof_l`) are
-    // available, the `strtof`-based implementation is in fact locale-dependent and might use a
-    // different decimal separator, so we need to adapt the test expectations accordingly.
-    std::replace(str.begin(), str.end(), '.', nlohmann::detail::from_chars_traits<T>::get_decimal_point());
-
-    auto res = nlohmann::detail::from_chars(str.data(), str.data() + str.size(), value);
-    CHECK_MESSAGE(res.ec == expected_ec, "Error code mismatch while parsing: \"", str,
-                  "\": Actual: ", make_error_code(res.ec).message(),
-                  "; expected: ", make_error_code(expected_ec).message());
-    CHECK_MESSAGE(res.ptr - str.data() == expected_ptr_offset, "Ptr offset mismatch while parsing: \"", str, "\"");
-}
-
-template <typename T>
-typename std::enable_if<std::is_integral<T>::value, void>::type
-check(std::string str, T expected_value, std::ptrdiff_t expected_ptr_offset, std::errc expected_ec)
-{
-    T value = 42;
-    check_result(str, value, expected_ptr_offset, expected_ec);
-    CHECK_MESSAGE(value == expected_value, "while parsing: \"", str, "\"");
-}
-
-template <typename T>
-typename std::enable_if<std::is_floating_point<T>::value, void>::type
-check(std::string str, T expected_value, std::ptrdiff_t expected_ptr_offset, std::errc expected_ec)
-{
-    {
-        T value = 42;
-        check_result(str, value, expected_ptr_offset, expected_ec);
-        CHECK_MESSAGE(value == expected_value, "while parsing: \"", str, "\"");
-        CHECK_MESSAGE(std::signbit(value) == std::signbit(expected_value), "while parsing: \"", str, "\"");
-    }
-    {
-        T value = std::numeric_limits<T>::quiet_NaN();
-        check_result(str, value, expected_ptr_offset, expected_ec);
-        CHECK_MESSAGE(value == expected_value, "while parsing: \"", str, "\"");
-    }
-}
-
-template <typename T>
-typename std::enable_if<std::is_integral<T>::value, void>::type
-check(std::string str, init_val_t /*unused*/, std::ptrdiff_t expected_ptr_offset, std::errc expected_ec)
-{
-    T const init_value = 42;
-    T value = init_value;
-    check_result(str, value, expected_ptr_offset, expected_ec);
-    CHECK_MESSAGE(value == init_value, "while parsing: \"", str, "\"");
-}
-
-template <typename T>
-typename std::enable_if<std::is_floating_point<T>::value, void>::type
-check(std::string str, init_val_t /*unused*/, std::ptrdiff_t expected_ptr_offset, std::errc expected_ec)
-{
-    {
-        T const init_value = 42;
-        T value = init_value;
-        check_result(str, value, expected_ptr_offset, expected_ec);
-        CHECK_MESSAGE(value == init_value, "while parsing: \"", str, "\"");
-    }
-    {
-        T value = std::numeric_limits<T>::quiet_NaN();
-        check_result(str, value, expected_ptr_offset, expected_ec);
-        CHECK_MESSAGE(doctest::IsNaN<T>(value), "while parsing: \"", str, "\"");
-    }
-}
-
-}  // namespace
-
-TEST_CASE("integral results consistent with std::from_chars")
-{
-    SECTION("unsigned long long")
-    {
-        check<unsigned long long>("", init_val, 0, std::errc::invalid_argument);
-        check<unsigned long long>("0", 0ULL, 1, std::errc{});
-        check<unsigned long long>(" 123", init_val, 0, std::errc::invalid_argument);
-        check<unsigned long long>("123 ", 123ULL, 3, std::errc{});
-        check<unsigned long long>("+123", init_val, 0, std::errc::invalid_argument);
-        check<unsigned long long>("-123", init_val, 0, std::errc::invalid_argument);
-        check<unsigned long long>("123", 123ULL, 3, std::errc{});
-        check<unsigned long long>("123e10", 123ULL, 3, std::errc{});
-        check<unsigned long long>("18446744073709551615", 18446744073709551615ULL, 20, std::errc{});
-        check<unsigned long long>("18446744073709551616", init_val, 20, std::errc::result_out_of_range);
-    }
-
-    SECTION("long long")
-    {
-        check<long long>("", init_val, 0, std::errc::invalid_argument);
-        check<long long>("0", 0LL, 1, std::errc{});
-        check<long long>(" 123", init_val, 0, std::errc::invalid_argument);
-        check<long long>("123 ", 123LL, 3, std::errc{});
-        check<long long>("+123", init_val, 0, std::errc::invalid_argument);
-        check<long long>("-123", -123LL, 4, std::errc{});
-        check<long long>("123", 123LL, 3, std::errc{});
-        check<long long>("123e10", 123LL, 3, std::errc{});
-        check<long long>("9223372036854775807", 9223372036854775807LL, 19, std::errc{});
-        check<long long>("9223372036854775808", init_val, 19, std::errc::result_out_of_range);
-        check<long long>("-9223372036854775808", -9223372036854775807LL - 1, 20, std::errc{});
-        check<long long>("-9223372036854775809", init_val, 20, std::errc::result_out_of_range);
-    }
-}
-
-TEST_CASE("floating point results consistent with std::from_chars")
-{
-    INFO("Using decimal point '",
-         std::string{1, nlohmann::detail::from_chars_traits<float>::get_decimal_point()},
-         "' to match our possibly-locale-dependent from_chars fallback implementation");
-
-    SECTION("single precision")
-    {
-        check<float>("", init_val, 0, std::errc::invalid_argument);
-        check<float>(" 123", init_val, 0, std::errc::invalid_argument);
-        check<float>("123 ", 123.0f, 3, std::errc{});
-        check<float>("+123", init_val, 0, std::errc::invalid_argument);
-        check<float>("-123", -123.0f, 4, std::errc{});
-        check<float>("123", 123.0f, 3, std::errc{});
-        check<float>("123e10", 123e10f, 6, std::errc{});
-        check<float>("123e+10", 123e10f, 7, std::errc{});
-        check<float>("123e-10", 123e-10f, 7, std::errc{});
-        check<float>("123.456", 123.456f, 7, std::errc{});
-        check<float>("123;456", 123.0f, 3, std::errc{});
-        check<float>("123.456 ", 123.456f, 7, std::errc{});
-        check<float>("123;456 ", 123.0f, 3, std::errc{});
-        check<float>("123456789.123456789", 123456789.123456789f, 19, std::errc{});
-        check<float>("1e40", std::numeric_limits<float>::infinity(), 4, std::errc::result_out_of_range);
-        check<float>("-1e40", -std::numeric_limits<float>::infinity(), 5, std::errc::result_out_of_range);
-        check<float>("2e308", std::numeric_limits<float>::infinity(), 5, std::errc::result_out_of_range);
-        check<float>("-2e308", -std::numeric_limits<float>::infinity(), 6, std::errc::result_out_of_range);
-        check<float>("123.456e-789", 0.0f, 12, std::errc::result_out_of_range);
-        check<float>("-123.456e-789", -0.0f, 13, std::errc::result_out_of_range);
-        check<float>("1e-45", 1e-45f, 5, std::errc{});
-        check<float>("1e-46", 0.0f, 5, std::errc::result_out_of_range);
-        check<float>("1E-45", 1e-45f, 5, std::errc{});
-        check<float>("1E-46", 0.0f, 5, std::errc::result_out_of_range);
-        check<float>("10e-46", 1e-45f, 6, std::errc{});
-        check<float>("0.1e-45", 0.0f, 7, std::errc::result_out_of_range);
-        check<float>("100000000000000000000000000000000000000", 1e38f, 39, std::errc{});
-        check<float>("100000000000000000000000000000000000000.0", 1e38f, 41, std::errc{});
-        check<float>("1000000000000000000000000000000000000000e-1", 1e38f, 43, std::errc{});
-        check<float>("0.0000000000000000000000000000000000000000000001e84", 1e38f, 51, std::errc{});
-        check<float>("0.000000000000000000000000000000000000000000001", 1e-45f, 47, std::errc{});
-        check<float>("00.000000000000000000000000000000000000000000001", 1e-45f, 48, std::errc{});
-        check<float>("0.0000000000000000000000000000000000000000000001e1", 1e-45f, 50, std::errc{});
-        check<float>("1000000000000000000000000000000000000000e-84", 1e-45f, 44, std::errc{});
-        check<float>("1000000000000000000000000000000000000000", std::numeric_limits<float>::infinity(), 40, std::errc::result_out_of_range);
-        check<float>("1000000000000000000000000000000000000000.0", std::numeric_limits<float>::infinity(), 42, std::errc::result_out_of_range);
-        check<float>("100000000000000000000000000000000000000e1", std::numeric_limits<float>::infinity(), 41, std::errc::result_out_of_range);
-        check<float>("0.0000000000000000000000000000000000000000000001e85", std::numeric_limits<float>::infinity(), 51, std::errc::result_out_of_range);
-        check<float>("1000000000000000000000000000000000000000e-85", 0.0f, 44, std::errc::result_out_of_range);
-        check<float>("0.0000000000000000000000000000000000000000000001", 0.0f, 48, std::errc::result_out_of_range);
-        check<float>("0.000000000000000000000000000000000000000000001e-1", 0.0f, 50, std::errc::result_out_of_range);
-        check<float>("1e-99999999999999999999", 0.0f, 23, std::errc::result_out_of_range);
-        check<float>("1e+99999999999999999999", std::numeric_limits<float>::infinity(), 23, std::errc::result_out_of_range);
-        check<float>("-1e-99999999999999999999", -0.0f, 24, std::errc::result_out_of_range);
-        check<float>("-1e+99999999999999999999", -std::numeric_limits<float>::infinity(), 24, std::errc::result_out_of_range);
-    }
-
-    SECTION("double precision")
-    {
-        check<double>("", init_val, 0, std::errc::invalid_argument);
-        check<double>(" 123", init_val, 0, std::errc::invalid_argument);
-        check<double>("123 ", 123.0, 3, std::errc{});
-        check<double>("+123", init_val, 0, std::errc::invalid_argument);
-        check<double>("-123", -123.0, 4, std::errc{});
-        check<double>("123", 123.0, 3, std::errc{});
-        check<double>("123e10", 123e10, 6, std::errc{});
-        check<double>("123e+10", 123e10, 7, std::errc{});
-        check<double>("123e-10", 123e-10, 7, std::errc{});
-        check<double>("123.456", 123.456, 7, std::errc{});
-        check<double>("123;456", 123.0, 3, std::errc{});
-        check<double>("123.456 ", 123.456, 7, std::errc{});
-        check<double>("123;456 ", 123.0, 3, std::errc{});
-        check<double>("123456789.123456789", 123456789.123456789, 19, std::errc{});
-        check<double>("1e40", 1e40, 4, std::errc{});
-        check<double>("-1e40", -1e40, 5, std::errc{});
-        check<double>("2e308", std::numeric_limits<double>::infinity(), 5, std::errc::result_out_of_range);
-        check<double>("-2e308", -std::numeric_limits<double>::infinity(), 6, std::errc::result_out_of_range);
-        check<double>("2e-324", 0.0, 6, std::errc::result_out_of_range);
-        check<double>("-2e-324", -0.0, 7, std::errc::result_out_of_range);
-        check<double>("123.456e-789", 0.0, 12, std::errc::result_out_of_range);
-        check<double>("-123.456e-789", -0.0, 13, std::errc::result_out_of_range);
-        check<double>("1e-99999999999999999999", 0.0, 23, std::errc::result_out_of_range);
-        check<double>("1e+99999999999999999999", std::numeric_limits<double>::infinity(), 23, std::errc::result_out_of_range);
-        check<double>("-1e-99999999999999999999", -0.0, 24, std::errc::result_out_of_range);
-        check<double>("-1e+99999999999999999999", -std::numeric_limits<double>::infinity(), 24, std::errc::result_out_of_range);
-    }
-
-    SECTION("long double precision")
-    {
-        check<long double>("", init_val, 0, std::errc::invalid_argument);
-        check<long double>(" 123", init_val, 0, std::errc::invalid_argument);
-        check<long double>("123 ", 123.0L, 3, std::errc{});
-        check<long double>("+123", init_val, 0, std::errc::invalid_argument);
-        check<long double>("-123", -123.0L, 4, std::errc{});
-        check<long double>("123", 123.0L, 3, std::errc{});
-        check<long double>("123e10", 123e10L, 6, std::errc{});
-        check<long double>("123e+10", 123e10L, 7, std::errc{});
-        check<long double>("123e-10", 123e-10L, 7, std::errc{});
-        check<long double>("123.456", 123.456L, 7, std::errc{});
-        check<long double>("123;456", 123.0L, 3, std::errc{});
-        check<long double>("123.456 ", 123.456L, 7, std::errc{});
-        check<long double>("123;456 ", 123.0L, 3, std::errc{});
-        check<long double>("123456789.123456789", 123456789.123456789L, 19, std::errc{});
-        check<long double>("1e40", 1e40L, 4, std::errc{});
-        check<long double>("-1e40", -1e40L, 5, std::errc{});
-#if defined(_MSC_VER)
-#pragma warning(push)
-#pragma warning(disable: 4127)
-#endif
-        if (sizeof(long double) > 8)
-#if defined(_MSC_VER)
-#pragma warning(pop)
-#endif
-        {
-            // Right-hand-side is calculated to avoid warning about literal
-            // exceeding range on platforms where this branch is NOT taken.
-            check<long double>("2e308", 1e308L * 2, 5, std::errc{});
-            check<long double>("-2e308", -1e308L * 2, 6, std::errc{});
-            check<long double>("2e-324", 8e-324L / 4, 6, std::errc{});
-            check<long double>("-2e-324", -8e-324L / 4, 7, std::errc{});
-        }
-        else
-        {
-            check<long double>("2e308", std::numeric_limits<long double>::infinity(), 5, std::errc::result_out_of_range);
-            check<long double>("-2e308", -std::numeric_limits<long double>::infinity(), 6, std::errc::result_out_of_range);
-            check<long double>("2e-324", 0.0L, 6, std::errc::result_out_of_range);
-            check<long double>("-2e-324", -0.0L, 7, std::errc::result_out_of_range);
-        }
-        check<long double>("1e5000", std::numeric_limits<long double>::infinity(), 6, std::errc::result_out_of_range);
-        check<long double>("-1e5000", -std::numeric_limits<long double>::infinity(), 7, std::errc::result_out_of_range);
-        check<long double>("123.456e-7890", 0.0L, 13, std::errc::result_out_of_range);
-        check<long double>("-123.456e-7890", -0.0L, 14, std::errc::result_out_of_range);
-        check<long double>("1e-99999999999999999999", 0.0L, 23, std::errc::result_out_of_range);
-        check<long double>("1e+99999999999999999999", std::numeric_limits<long double>::infinity(), 23, std::errc::result_out_of_range);
-        check<long double>("-1e-99999999999999999999", -0.0L, 24, std::errc::result_out_of_range);
-        check<long double>("-1e+99999999999999999999", -std::numeric_limits<long double>::infinity(), 24, std::errc::result_out_of_range);
-    }
-}
-

tests/src/unit-regression2.cpp

@@ -26,11 +26,8 @@ using ordered_json = nlohmann::ordered_json;
     using namespace nlohmann::literals; // NOLINT(google-build-using-namespace)
 #endif
 
-#include <atomic>
-#include <clocale>
 #include <cstdio>
 #include <list>
-#include <thread>
 #include <type_traits>
 #include <utility>
 
@@ -1242,44 +1239,6 @@ TEST_CASE("regression tests 2")
         CHECK(j == json({2, 4, 6}));
     }
 #endif
-
-    SECTION("issue #5198 - TOCTOU race between lexer construction and locale changes causes float truncation")
-    {
-        std::atomic<bool> stop_requested{};
-        std::thread switcher([&stop_requested]
-        {
-            bool german = true;
-            const std::string initial_locale = std::setlocale(LC_NUMERIC, nullptr);
-
-            while (!stop_requested)
-            {
-                const bool setlocale_res = std::setlocale(LC_NUMERIC, german ? "de_DE.UTF-8" : "C");
-                WARN_MESSAGE(setlocale_res,
-                             "Setting locale failed, probably because de_DE.UTF-8 is not installed. "
-                             "This test was skipped as it would be inconclusive.");
-                if (!setlocale_res)
-                {
-                    stop_requested = true;
-                }
-                german = !german;
-            }
-
-            (void)std::setlocale(LC_NUMERIC, initial_locale.c_str()); // restore original locale
-        });
-
-        for (std::size_t i = 0; i < 10000 && !stop_requested; ++i)
-        {
-            const json j = json::parse("{\"val\": 99.123456789}");
-            const double parsed = j.value("val", 0.0);
-            if (!CHECK(parsed == 99.123456789))
-            {
-                break;
-            }
-        }
-
-        stop_requested = true;
-        switcher.join();
-    }
 }
 
 TEST_CASE_TEMPLATE("issue #4798 - nlohmann::json::to_msgpack() encode float NaN as double", T, double, float) // NOLINT(readability-math-missing-parentheses, bugprone-throwing-static-initialization)

tests/src/unit-ubjson.cpp

@@ -1862,6 +1862,31 @@ TEST_CASE("UBJSON")
                 json _;
                 CHECK_THROWS_WITH_AS(_ = json::from_ubjson(v), "[json.exception.parse_error.113] parse error at byte 2: syntax error while parsing UBJSON string: expected length type specification (U, i, I, l, L); last byte: 0x31", json::parse_error&);
             }
+
+            SECTION("negative length")
+            {
+                json _;
+
+                std::vector<uint8_t> const vi = {'S', 'i', 0xFF};
+                CHECK_THROWS_WITH_AS(_ = json::from_ubjson(vi), "[json.exception.parse_error.113] parse error at byte 3: syntax error while parsing UBJSON string: string length must not be negative", json::parse_error&);
+                CHECK(json::from_ubjson(vi, true, false).is_discarded());
+
+                std::vector<uint8_t> const vI = {'S', 'I', 0xFF, 0xFF};
+                CHECK_THROWS_WITH_AS(_ = json::from_ubjson(vI), "[json.exception.parse_error.113] parse error at byte 4: syntax error while parsing UBJSON string: string length must not be negative", json::parse_error&);
+                CHECK(json::from_ubjson(vI, true, false).is_discarded());
+
+                std::vector<uint8_t> const vl = {'S', 'l', 0xFF, 0xFF, 0xFF, 0xFF};
+                CHECK_THROWS_WITH_AS(_ = json::from_ubjson(vl), "[json.exception.parse_error.113] parse error at byte 6: syntax error while parsing UBJSON string: string length must not be negative", json::parse_error&);
+                CHECK(json::from_ubjson(vl, true, false).is_discarded());
+
+                std::vector<uint8_t> const vL = {'S', 'L', 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF};
+                CHECK_THROWS_WITH_AS(_ = json::from_ubjson(vL), "[json.exception.parse_error.113] parse error at byte 10: syntax error while parsing UBJSON string: string length must not be negative", json::parse_error&);
+                CHECK(json::from_ubjson(vL, true, false).is_discarded());
+
+                // a length of zero remains valid and yields an empty string
+                std::vector<uint8_t> const v0 = {'S', 'i', 0};
+                CHECK(json::from_ubjson(v0) == json(""));
+            }
         }
 
         SECTION("array")

tests/src/unit-wstring.cpp

@@ -53,6 +53,27 @@ TEST_CASE("wide strings")
             std::wstring const w = L"\"\xDBFF";
             json _;
             CHECK_THROWS_AS(_ = json::parse(w), json::parse_error&);
+
+            // the exact message depends on the width of wchar_t: a 16-bit
+            // wchar_t passes the lone surrogate to the UTF-8 decoder unchanged
+            // (rejected as a single ill-formed byte at column 2), while a
+            // 32-bit wchar_t first encodes it as an ill-formed three-byte
+            // sequence (rejected one byte later, at column 3)
+            const char* const error_low_surrogate = sizeof(wchar_t) == 2
+                                                    ? "[json.exception.parse_error.101] parse error at line 1, column 2: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"<U+0000>'"
+                                                    : "[json.exception.parse_error.101] parse error at line 1, column 3: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"\xED\xB0'";
+            const char* const error_high_surrogate = sizeof(wchar_t) == 2
+                ? "[json.exception.parse_error.101] parse error at line 1, column 2: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"<U+0000>'"
+                : "[json.exception.parse_error.101] parse error at line 1, column 3: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"\xED\xA0'";
+
+            // a lone low surrogate cannot start a pair
+            CHECK_THROWS_WITH_AS(_ = json::parse(std::wstring{L'"', static_cast<wchar_t>(0xDC00), L'"'}), error_low_surrogate, json::parse_error&);
+            // a high surrogate followed by a non-low-surrogate unit is invalid
+            CHECK_THROWS_WITH_AS(_ = json::parse(std::wstring{L'"', static_cast<wchar_t>(0xD800), L'a', L'"'}), error_high_surrogate, json::parse_error&);
+            // a lone low surrogate must not swallow the following unit: pairing
+            // it with any second unit would produce valid UTF-8, so the error
+            // has to report an ill-formed byte at the surrogate's own position
+            CHECK_THROWS_WITH_AS(_ = json::parse(std::wstring{L'"', static_cast<wchar_t>(0xDC00), L'a', L'"'}), error_low_surrogate, json::parse_error&);
         }
     }
 
@@ -68,11 +89,22 @@ TEST_CASE("wide strings")
 
     SECTION("invalid std::u16string")
     {
-        if (wstring_is_utf16())
+        if (u16string_is_utf16())
         {
             std::u16string const w = u"\"\xDBFF";
             json _;
             CHECK_THROWS_AS(_ = json::parse(w), json::parse_error&);
+
+            // a lone low surrogate cannot start a pair
+            CHECK_THROWS_WITH_AS(_ = json::parse(std::u16string{u'"', 0xDC00, u'"'}), "[json.exception.parse_error.101] parse error at line 1, column 2: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"<U+0000>'", json::parse_error&);
+            // a high surrogate followed by a non-low-surrogate unit is invalid
+            CHECK_THROWS_WITH_AS(_ = json::parse(std::u16string{u'"', 0xD800, u'a', u'"'}), "[json.exception.parse_error.101] parse error at line 1, column 2: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"<U+0000>'", json::parse_error&);
+            // a lone low surrogate must not swallow the following unit: pairing
+            // it with any second unit would produce valid UTF-8, so the error
+            // has to report an ill-formed byte at the surrogate's own position
+            CHECK_THROWS_WITH_AS(_ = json::parse(std::u16string{u'"', 0xDC00, u'a', u'"'}), "[json.exception.parse_error.101] parse error at line 1, column 2: syntax error while parsing value - invalid string: ill-formed UTF-8 byte; last read: '\"<U+0000>'", json::parse_error&);
+            // a valid surrogate pair is still decoded (U+1F600)
+            CHECK(json::parse(std::u16string{u'"', 0xD83D, 0xDE00, u'"'}).get<std::string>() == "\xF0\x9F\x98\x80");
         }
     }