Use the shared descent bookkeeping rather than a second set

Comparing kept a thread_local count, a limit and a guard of its own beside the ones copying already had, all three the same thing under a different name. They are gone; the shared count, limit and guard do the work. The guard grows a second constructor here, because the comparison operators are written as a macro and a macro cannot use the preprocessor: it cannot look the count up behind an #ifdef the way copy_structured does, so the guard looks it up for it. nesting_depth_exhausted() arrives for the same reason - whether an operator descends at all is a constant at every call site, and testing it there is what MSVC reports as C4127. Also say in compare_leaves what happens to a pair that is an array on one side and an object on the other, since the answer is not obvious from the code: an operator only descends into two values of the same type, so such a pair is told apart by its types alone - unequal, and ordered the way the types are - exactly as it is above the bound. And record what the explicit stack costs: the comparison operators are noexcept and the container comparison this replaces allocated nothing, so running out of memory here ends the process instead of throwing. It takes a value nested past the bound and an exhausted heap to reach, and the same comparison used to exhaust the call stack, but it is a new way to fail. Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Parenthesise the reserve() computation in the comparison test
2026-08-21 16:43:17 +00:00 · 2026-08-21 12:30:56 +02:00 · 2026-08-21 12:25:05 +02:00 · 2026-08-21 12:25:05 +02:00 · 2026-08-21 12:25:05 +02:00 · 2026-08-21 12:25:05 +02:00
29 changed files with 3019 additions and 1253 deletions
@@ -108,7 +108,9 @@ The tests are located in [`tests/src/unit-*.cpp`](https://github.com/nlohmann/js
 are structured along the features of the library or the nature of the tests. Usually, it should be clear from the
 context which existing file needs to be extended, and only very few cases require creating new test files.

-When fixing a bug, edit `unit-regression2.cpp` and add a section referencing the fixed issue.
+When fixing a bug, edit `unit-regression3.cpp` and add a test case referencing the fixed issue. Its predecessors
+`unit-regression1.cpp` and `unit-regression2.cpp` stay as they are: the MinGW linker fails on the object a file this
+size produces, which is why the tests are spread over several files in the first place.

 #### Exceptions

@@ -11,7 +11,7 @@ jobs:
    runs-on: ubuntu-latest
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -34,7 +34,7 @@ jobs:

    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -9,7 +9,7 @@ jobs:
    runs-on: ubuntu-22.04
    steps:
    - name: Harden Runner
-      uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+      uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
      with:
        egress-policy: audit

@@ -27,7 +27,7 @@ jobs:

    steps:
    - name: Harden Runner
-      uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+      uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
      with:
        egress-policy: audit

@@ -19,7 +19,7 @@ jobs:
      pull-requests: write
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -17,7 +17,7 @@ jobs:
    runs-on: ubuntu-latest
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -27,7 +27,7 @@ jobs:
      security-events: write
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -17,7 +17,7 @@ jobs:

    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -26,7 +26,7 @@ jobs:
    runs-on: ubuntu-22.04
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -36,7 +36,7 @@ jobs:

    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -32,7 +32,7 @@ jobs:
    runs-on: ubuntu-latest
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -16,7 +16,7 @@ jobs:

    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -35,7 +35,7 @@ jobs:
    runs-on: ubuntu-latest
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -60,7 +60,7 @@ jobs:
        target: [ci_test_amalgamation, ci_test_single_header, ci_cppcheck, ci_cpplint, ci_reproducible_tests, ci_non_git_tests, ci_offline_testdata, ci_reuse_compliance, ci_test_valgrind]
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -100,7 +100,7 @@ jobs:
    container: ubuntu:focal
    strategy:
      matrix:
-        target: [ci_cmake_flags, ci_test_diagnostics, ci_test_diagnostic_positions, ci_test_noexceptions, ci_test_noimplicitconversions, ci_test_legacycomparison, ci_test_noglobaludls]
+        target: [ci_cmake_flags, ci_test_diagnostics, ci_test_diagnostic_positions, ci_test_noexceptions, ci_test_noimplicitconversions, ci_test_legacycomparison, ci_test_noglobaludls, ci_test_no_thread_local]
    steps:
      - name: Install build-essential
        run: apt-get update ; apt-get install -y build-essential unzip wget git libssl-dev
@@ -118,7 +118,7 @@ jobs:
    runs-on: ubuntu-latest
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -369,7 +369,7 @@ jobs:
    runs-on: ubuntu-latest
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -392,7 +392,7 @@ jobs:
        target: [ci_test_examples, ci_test_build_documentation]
    steps:
      - name: Harden Runner
-        uses: step-security/harden-runner@05e31511f85b41b11d1cf0ef85d0992719546e2c # v2.21.0
+        uses: step-security/harden-runner@b09bb98e06d4d774595224525879c09bc6e98c40 # v2.20.1
        with:
          egress-policy: audit

@@ -158,6 +158,10 @@ jobs:
      # to fit: IMAGE_REL_AMD64_SECREL against `.debug_line'" because the
      # MinGW linker cannot relocate the debug sections this test produces.
      # The tests are only built and run here, so the debug info is not used.
+      # Do not add -O1 here to shrink the objects further: it does make them
+      # link, but the binaries clang 11.0.1 and clang 18.1.8 then produce crash
+      # before doctest prints its first line - 39 of 102 tests on clang 18.
+      # Keep the objects small by splitting the test files instead.
      - name: Run CMake
        run: cmake -S . -B build ^
             -DCMAKE_CXX_COMPILER="C:/Program Files/LLVM/bin/clang++.exe" ^
@@ -242,6 +242,25 @@ add_custom_target(ci_test_noglobaludls
    COMMENT "Compile and test with global UDLs disabled"
 )

+###############################################################################
+# Disable thread-local storage.
+###############################################################################
+
+# Without thread-local storage, copying and comparing cannot bound their
+# descent and handle every object and array without the call stack. Those paths
+# are otherwise only reached by values nested deeper than the bound, so this
+# target is what runs the whole test suite through them.
+add_custom_target(ci_test_no_thread_local
+    COMMAND ${CMAKE_COMMAND}
+    -DCMAKE_BUILD_TYPE=Debug -GNinja
+    -DJSON_BuildTests=ON
+    -DCMAKE_CXX_FLAGS=-DJSON_NO_THREAD_LOCAL
+    -S${PROJECT_SOURCE_DIR} -B${PROJECT_BINARY_DIR}/build_no_thread_local
+    COMMAND ${CMAKE_COMMAND} --build ${PROJECT_BINARY_DIR}/build_no_thread_local
+    COMMAND cd ${PROJECT_BINARY_DIR}/build_no_thread_local && ${CMAKE_CTEST_COMMAND} --parallel ${N} --output-on-failure
+    COMMENT "Compile and test without thread-local storage"
+)
+
 ###############################################################################
 # Coverage.
 ###############################################################################
@@ -22,6 +22,7 @@ header. See also the [macro overview page](../../features/macros.md).
 - [**JSON_HAS_STD_FORMAT**](json_has_std_format.md) - control `std::format`/`std::formatter` support
 - [**JSON_HAS_THREE_WAY_COMPARISON**](json_has_three_way_comparison.md) - control 3-way comparison support
 - [**JSON_NO_IO**](json_no_io.md) - switch off functions relying on certain C++ I/O headers
+- [**JSON_NO_THREAD_LOCAL**](json_no_thread_local.md) - switch off the use of `thread_local` storage
 - [**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

@@ -0,0 +1,48 @@
+# JSON_NO_THREAD_LOCAL
+
+```cpp
+#define JSON_NO_THREAD_LOCAL
+```
+
+When defined, the library does not use `#!cpp thread_local` storage. This is relevant for the few environments whose
+toolchain does not support it.
+
+Copying a value and comparing two values both descend into the first levels by letting the containers copy or compare
+themselves, and finish whatever is nested deeper than that without the call stack, so that neither can exhaust the stack
+however deeply the values are nested. Each counts the levels it has descended into in a `#!cpp thread_local` variable, as
+a counter shared between threads would be raced.
+
+Without those counters, no descent can be bounded safely, so objects and arrays are copied and compared without the call
+stack right away. Both keep working exactly as they do otherwise - the same values come out, the same comparisons hold,
+and deeply nested values are handled just as safely - but both are slower, because the containers no longer copy or
+compare themselves. Copying the benchmark documents takes 9% (`canada.json`) to 34% (`twitter.json`) longer, and
+comparing two equal ones 10% (`citm_catalog.json`) to 90% (`canada.json`) longer.
+
+## Default definition
+
+By default, `#!cpp JSON_NO_THREAD_LOCAL` is not defined.
+
+```cpp
+#undef JSON_NO_THREAD_LOCAL
+```
+
+The library defines it by itself for Clang targeting MinGW, which does not survive the `#!cpp thread_local` storage:
+copying a value segfaults there, with both old and current Clang versions, while GCC targeting MinGW is unaffected.
+Copying and comparing fall back to working without the call stack there, as they do whenever the macro is defined.
+
+## Examples
+
+??? example
+
+    The code below forces the library not to use `#!cpp thread_local` storage.
+
+    ```cpp
+    #define JSON_NO_THREAD_LOCAL 1
+    #include <nlohmann/json.hpp>
+
+    ...
+    ```
+
+## Version history
+
+- Added in version 3.12.1.
@@ -91,6 +91,14 @@ security reasons (e.g., Intel Software Guard Extensions (SGX)).

 See [full documentation of `JSON_NO_IO`](../api/macros/json_no_io.md).

+## `JSON_NO_THREAD_LOCAL`
+
+When defined, the library does not use `#!cpp thread_local` storage. Copying a value and comparing two values then
+always avoid the call stack rather than descending into a bounded number of levels first, which is slower but yields the
+same values and the same comparisons.
+
+See [full documentation of `JSON_NO_THREAD_LOCAL`](../api/macros/json_no_thread_local.md).
+
 ## `JSON_SKIP_LIBRARY_VERSION_CHECK`

 When defined, the library will not create a compiler warning when a different version of the library was already
@@ -291,6 +291,7 @@ nav:
          - 'JSON_HAS_THREE_WAY_COMPARISON': api/macros/json_has_three_way_comparison.md
          - 'JSON_NOEXCEPTION': api/macros/json_noexception.md
          - 'JSON_NO_IO': api/macros/json_no_io.md
+          - 'JSON_NO_THREAD_LOCAL': api/macros/json_no_thread_local.md
          - 'JSON_SKIP_LIBRARY_VERSION_CHECK': api/macros/json_skip_library_version_check.md
          - 'JSON_SKIP_UNSUPPORTED_COMPILER_CHECK': api/macros/json_skip_unsupported_compiler_check.md
          - 'JSON_USE_GLOBAL_UDLS': api/macros/json_use_global_udls.md
@@ -186,6 +186,15 @@
    #define JSON_NO_UNIQUE_ADDRESS
 #endif

+// Clang targeting MinGW does not survive the thread_local storage the copy
+// constructor uses to bound its descent: every test that copies a value
+// segfaults with clang 11.0.1 and clang 18.1.8, while the same tests pass with
+// GCC targeting MinGW and with every other toolchain the library is tested on.
+// Copying works the same way without the counter, only more slowly.
+#if !defined(JSON_NO_THREAD_LOCAL) && defined(__clang__) && defined(__MINGW32__)
+    #define JSON_NO_THREAD_LOCAL 1
+#endif
+
 // disable documentation warnings on clang
 #if defined(__clang__)
    #pragma clang diagnostic push
@@ -28,14 +28,14 @@
    #pragma GCC diagnostic ignored "-Wignored-attributes"
 #endif

-#include <algorithm> // all_of, find, for_each
+#include <algorithm> // all_of, find, for_each, none_of
 #include <cstddef> // nullptr_t, ptrdiff_t, size_t
 #include <functional> // hash, less
 #include <initializer_list> // initializer_list
 #ifndef JSON_NO_IO
    #include <iosfwd> // istream, ostream
 #endif  // JSON_NO_IO
-#include <iterator> // random_access_iterator_tag
+#include <iterator> // make_move_iterator, random_access_iterator_tag
 #include <memory> // unique_ptr
 #include <string> // string, stoi, to_string
 #include <utility> // declval, forward, move, pair, swap
@@ -821,6 +821,626 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        return j;
    }

+    /// the number of levels an operation descends into before it finishes the
+    /// value below it without the call stack
+    static constexpr std::size_t nesting_depth_limit()
+    {
+        return 128;
+    }
+
+#ifndef JSON_NO_THREAD_LOCAL
+    /*!
+    @brief how many levels the operation going on in this thread has descended into
+
+    Copying a value and comparing two values share this count. The library never
+    nests one inside the other - copying a value does not compare one, and
+    comparing two values does not copy them - and where user code nests them
+    anyway, sharing the count only ends a descent sooner than it had to, which
+    costs a little speed and is never wrong.
+
+    A byte is enough: the count never exceeds the limit by more than the single
+    level that notices the limit has been reached.
+    */
+    static std::size_t& nesting_depth() noexcept
+    {
+        static thread_local std::size_t depth = 0; // NOLINT(misc-use-internal-linkage)
+        return depth;
+    }
+#endif
+
+    /*!
+    @brief whether a descent must stop here and finish without the call stack
+
+    @a may_descend says whether the operator descends at all; it is a constant
+    at every call site, and is passed rather than tested by the caller so that
+    the test does not become a constant condition there, which MSVC reports as
+    C4127.
+    */
+    static bool nesting_depth_exhausted(bool may_descend = true) noexcept
+    {
+#ifdef JSON_NO_THREAD_LOCAL
+        // without a count of its own per thread, a descent cannot be bounded
+        // without racing another one, so none is made
+        static_cast<void>(may_descend);
+        return true;
+#else
+        return !may_descend || nesting_depth() >= nesting_depth_limit();
+#endif
+    }
+
+    /*!
+    @brief counts one level of a bounded descent for as long as it runs
+
+    The constructor taking the count is for callers that have looked it up
+    already to test it: reaching thread-local storage is not free, and the path
+    that is taken almost every time should reach it once rather than twice. The
+    other is for callers that cannot look it up - the comparison operators are
+    written as a macro, and a macro cannot use the preprocessor.
+    */
+    class nesting_depth_guard
+    {
+      public:
+        explicit nesting_depth_guard(std::size_t& depth) noexcept
+            : m_depth(&depth)
+        {
+            ++*m_depth;
+        }
+
+        nesting_depth_guard() noexcept
+            : m_depth(countable())
+        {
+            if (m_depth != nullptr)
+            {
+                ++*m_depth;
+            }
+        }
+
+        ~nesting_depth_guard()
+        {
+            if (m_depth != nullptr)
+            {
+                --*m_depth;
+            }
+        }
+
+        nesting_depth_guard(const nesting_depth_guard&) = delete;
+        nesting_depth_guard& operator=(const nesting_depth_guard&) = delete;
+        nesting_depth_guard(nesting_depth_guard&&) = delete;
+        nesting_depth_guard& operator=(nesting_depth_guard&&) = delete;
+
+      private:
+        /// @brief the count to keep, or nullptr where there is none to keep
+        static std::size_t* countable() noexcept
+        {
+#ifdef JSON_NO_THREAD_LOCAL
+            return nullptr;
+#else
+            return &nesting_depth();
+#endif
+        }
+
+        std::size_t* m_depth;
+    };
+
+    /// an entry of the iterative deep copy's worklist: a structured value and
+    /// the value that is to become its copy
+    using copy_worklist_t = std::vector<std::pair<const basic_json*, basic_json*>>;
+
+    /// scratch space to build the key skeleton of an object copy in one go
+    using copy_scratch_t = std::vector<std::pair<typename object_t::key_type, basic_json>>;
+
+    /// @brief copy everything of @a src into @a dst but its type and value
+    static void copy_metadata(const basic_json& src, basic_json& dst)
+    {
+        // a custom base class is only required to be copy-constructible and
+        // move-assignable, so the copy has to go through a temporary
+        static_cast<json_base_class_t&>(dst) = json_base_class_t(static_cast<const json_base_class_t&>(src));
+
+#if JSON_DIAGNOSTIC_POSITIONS
+        dst.start_position = src.start_position;
+        dst.end_position = src.end_position;
+#else
+        static_cast<void>(src);
+        static_cast<void>(dst);
+#endif
+    }
+
+    /*!
+    @brief copy the value of @a src into @a dst, which must not be structured
+
+    Objects and arrays are left alone: creating those is the one thing the copy
+    constructor and @ref copy_shallow do differently from one another, and it is
+    the reason copying a value can descend at all.
+    */
+    /// @note inlined on purpose: both callers have already told an object or an
+    ///       array apart from the rest, and letting the compiler fold that test
+    ///       into this switch is worth a few percent when copying a value made
+    ///       mostly of numbers
+    JSON_HEDLEY_ALWAYS_INLINE
+    static void copy_leaf_value(const basic_json& src, basic_json& dst)
+    {
+        switch (src.m_data.m_type)
+        {
+            case value_t::string:
+            {
+                dst.m_data.m_value = *src.m_data.m_value.string;
+                break;
+            }
+
+            case value_t::binary:
+            {
+                dst.m_data.m_value = *src.m_data.m_value.binary;
+                break;
+            }
+
+            case value_t::boolean:
+            {
+                dst.m_data.m_value = src.m_data.m_value.boolean;
+                break;
+            }
+
+            case value_t::number_integer:
+            {
+                dst.m_data.m_value = src.m_data.m_value.number_integer;
+                break;
+            }
+
+            case value_t::number_unsigned:
+            {
+                dst.m_data.m_value = src.m_data.m_value.number_unsigned;
+                break;
+            }
+
+            case value_t::number_float:
+            {
+                dst.m_data.m_value = src.m_data.m_value.number_float;
+                break;
+            }
+
+            case value_t::object:
+            case value_t::array:
+            case value_t::null:
+            case value_t::discarded:
+            default:
+                break;
+        }
+    }
+
+    /*!
+    @brief copy everything of @a src into the null value @a dst but the children
+
+    Objects and arrays are not copied here; they are appended to @a worklist to
+    be created later by @ref copy_iteratively. Until that happens, @a dst remains
+    a null value, so that a partially built copy can be destroyed at any point
+    without ever violating the class invariants.
+    */
+    static void copy_shallow(const basic_json& src, basic_json& dst, copy_worklist_t& worklist)
+    {
+        copy_metadata(src, dst);
+
+        if (src.m_data.m_type == value_t::object || src.m_data.m_type == value_t::array)
+        {
+            // defer: dst stays a null value until its container exists
+            worklist.emplace_back(&src, &dst);
+            return;
+        }
+
+        copy_leaf_value(src, dst);
+
+        // only now that the value exists may the type be set: had the creation
+        // of the value thrown, dst would have been left as a valid null value
+        dst.m_data.m_type = src.m_data.m_type;
+    }
+
+    /// @brief create the copy of the array @a src in @a dst
+    /// @note structured elements are appended to @a worklist instead
+    static void copy_array_level(const basic_json& src, basic_json& dst, copy_worklist_t& worklist)
+    {
+        const array_t& src_array = *src.m_data.m_value.array;
+
+        // create all elements up front: growing the array afterwards could
+        // invalidate the pointers that are handed to the worklist
+        dst.m_data.m_value.array = create<array_t>(src_array.size(), basic_json());
+
+        auto dst_it = dst.m_data.m_value.array->begin();
+        for (auto src_it = src_array.cbegin(); src_it != src_array.cend(); ++src_it, ++dst_it)
+        {
+            copy_shallow(*src_it, *dst_it, worklist);
+        }
+    }
+
+    /// @brief create the copy of the object @a src in @a dst
+    /// @note structured values are appended to @a worklist instead
+    static void copy_object_level(const basic_json& src, basic_json& dst,
+                                  copy_worklist_t& worklist, copy_scratch_t& scratch)
+    {
+        const object_t& src_object = *src.m_data.m_value.object;
+
+        // build the complete key skeleton and hand it to the object's range
+        // constructor: adding the keys one by one would be quadratic for object
+        // types that are backed by a vector, such as nlohmann::ordered_map
+        scratch.clear();
+        scratch.reserve(src_object.size());
+        for (const auto& element : src_object)
+        {
+            scratch.emplace_back(element.first, basic_json());
+        }
+
+        dst.m_data.m_value.object = create<object_t>(std::make_move_iterator(scratch.begin()),
+                                    std::make_move_iterator(scratch.end()));
+        scratch.clear();
+
+        // pair every value of the copy with its counterpart in the original;
+        // both are enumerated in the same order for every object type with a
+        // deterministic order, so the lookup is only needed for exotic ones
+        auto src_it = src_object.cbegin();
+        for (auto& element : *dst.m_data.m_value.object)
+        {
+            if (JSON_HEDLEY_LIKELY(src_it != src_object.cend() && src_it->first == element.first))
+            {
+                copy_shallow(src_it->second, element.second, worklist);
+                ++src_it;
+            }
+            else
+            {
+                const auto found = src_object.find(element.first);
+                JSON_ASSERT(found != src_object.cend());
+                copy_shallow(found->second, element.second, worklist);
+            }
+        }
+    }
+
+    /*!
+    @brief deep-copy the object or array @a src into this value without recursing
+
+    The values whose copy has not been created yet are kept on an explicit
+    worklist rather than on the call stack. This is only reached for values
+    nested deeper than @ref nesting_depth_limit levels, which is why it copies
+    every container by hand instead of letting the container do it: the fast
+    ways of doing so would descend into the elements and defeat the purpose.
+    */
+    void copy_iteratively(const basic_json& src)
+    {
+        copy_worklist_t worklist;
+        copy_scratch_t scratch;
+
+        const basic_json* src_value = &src;
+        basic_json* dst_value = this;
+
+        for (;;)
+        {
+            if (src_value->m_data.m_type == value_t::array)
+            {
+                copy_array_level(*src_value, *dst_value, worklist);
+            }
+            else
+            {
+                copy_object_level(*src_value, *dst_value, worklist, scratch);
+            }
+
+            // the container is complete and will not be modified again
+            dst_value->set_parents();
+
+            if (worklist.empty())
+            {
+                break;
+            }
+
+            const auto& next = worklist.back();
+            src_value = next.first;
+            dst_value = next.second;
+            worklist.pop_back();
+
+            // the value stops being a null value exactly here
+            dst_value->m_data.m_type = src_value->m_data.m_type;
+        }
+    }
+
+    /*!
+    @brief copy one level of the object or array @a src into this value
+
+    The container copies its own elements, which is the fastest way to fill it.
+    Every element that is structured itself comes back to @ref copy_structured.
+    */
+    void copy_level(const basic_json& src)
+    {
+        if (m_data.m_type == value_t::object)
+        {
+            m_data.m_value = *src.m_data.m_value.object;
+        }
+        else
+        {
+            m_data.m_value = *src.m_data.m_value.array;
+        }
+
+        set_parents();
+    }
+
+    /*!
+    @brief deep-copy the object or array @a src into this value
+
+    Copying a container copies its elements, so a value nested deeply enough
+    used to exhaust the call stack. The descent is bounded here: the first
+    @ref nesting_depth_limit levels are copied by the containers themselves, just
+    as they always were, and anything below that is copied without the call
+    stack by @ref copy_iteratively. Copying a value can therefore no longer
+    exhaust the stack, however deeply it is nested, just like destroying one
+    cannot since #1436.
+
+    Nothing has to be scanned or built by hand to reach that: a value that is
+    not nested deeper than the limit - all but a vanishing minority - is copied
+    exactly as it was before, and this whole detour costs it one counter.
+
+    @sa https://github.com/nlohmann/json/issues/5387
+    */
+    void copy_structured(const basic_json& src)
+    {
+#ifndef JSON_NO_THREAD_LOCAL
+        std::size_t& depth = nesting_depth();
+
+        if (JSON_HEDLEY_LIKELY(depth < nesting_depth_limit()))
+        {
+            const nesting_depth_guard guard(depth);
+            copy_level(src);
+            return;
+        }
+#endif
+
+        // Finish this value without descending any further. It is completed
+        // before this returns, so a copy made by a custom base class - or by
+        // anything else that runs while a copy is going on - is unaffected by
+        // the copy it is nested in.
+        copy_iteratively(src);
+    }
+
+
+    /// the result of comparing two values, including values that cannot be
+    /// ordered at all, such as a discarded value or a NaN
+    enum class compare_result { less, equal, greater, unordered };
+
+#if JSON_HAS_THREE_WAY_COMPARISON
+    /// @brief the ordering that @a result stands for
+    static std::partial_ordering to_partial_ordering(compare_result result) noexcept // *NOPAD*
+    {
+        switch (result)
+        {
+            case compare_result::less:
+                return std::partial_ordering::less;
+            case compare_result::greater:
+                return std::partial_ordering::greater;
+            case compare_result::equal:
+                return std::partial_ordering::equivalent;
+            case compare_result::unordered:
+            default:
+                return std::partial_ordering::unordered;
+        }
+    }
+#endif
+
+    /*!
+    @brief compare two values that are not both an array or both an object
+
+    Such a pair is compared by the operators themselves, which cannot descend
+    into it and therefore cannot recurse.
+
+    That holds for a pair whose types differ as much as for a pair of leaves: an
+    array and an object are told apart by their types alone, because an operator
+    only ever descends into two values of the same type. So `==` reports them as
+    unequal without looking inside either, and an ordering falls back to the
+    order of the types - an object sorts before an array - exactly as it does
+    for a value that is not nested deeply enough to get here.
+    */
+    template<bool Ordered>
+    static compare_result compare_leaves(const_reference lhs, const_reference rhs) noexcept
+    {
+        if (lhs == rhs)
+        {
+            return compare_result::equal;
+        }
+
+        return order_leaves(lhs, rhs, std::integral_constant<bool, Ordered> {});
+    }
+
+    /*!
+    @brief compare two object keys
+
+    An object compares its entries as pairs of a key and a value, so its keys
+    are compared exactly as std::pair compares them: with < where the objects
+    are being ordered, and with == where they are only checked for equality.
+    Note that this is not the object's own comparator, which for a vector-backed
+    object type such as nlohmann::ordered_map tells equality rather than order.
+    */
+    static compare_result compare_keys(const typename object_t::key_type& lhs,
+                                       const typename object_t::key_type& rhs,
+                                       std::true_type /*ordered*/)
+    {
+        if (lhs < rhs)
+        {
+            return compare_result::less;
+        }
+
+        if (rhs < lhs)
+        {
+            return compare_result::greater;
+        }
+
+        return compare_result::equal;
+    }
+
+    /// @brief check two object keys for equality
+    static compare_result compare_keys(const typename object_t::key_type& lhs,
+                                       const typename object_t::key_type& rhs,
+                                       std::false_type /*ordered*/)
+    {
+        return lhs == rhs ? compare_result::equal : compare_result::unordered;
+    }
+
+    /// @brief tell apart two values that are not equal
+    /// @note only instantiated where the values are being ordered, as a key or
+    ///       string type is not required to be ordered to be compared for equality
+    static compare_result order_leaves(const_reference lhs, const_reference rhs, std::true_type /*ordered*/) noexcept
+    {
+        if (lhs < rhs)
+        {
+            return compare_result::less;
+        }
+
+        if (rhs < lhs)
+        {
+            return compare_result::greater;
+        }
+
+        return compare_result::unordered;
+    }
+
+    /// @brief report two values as not equal without ordering them
+    static compare_result order_leaves(const_reference /*lhs*/, const_reference /*rhs*/, std::false_type /*ordered*/) noexcept
+    {
+        return compare_result::unordered;
+    }
+
+    /*!
+    @brief compare @a lhs and @a rhs without descending into them
+
+    Reached once a comparison has descended @ref nesting_depth_limit levels, so
+    that comparing values cannot exhaust the call stack however deeply they are
+    nested. The two values are walked in lockstep on an explicit stack and
+    compared lexicographically, element by element in the order the containers
+    enumerate them - which is how the container types this library ships compare
+    themselves: a std::map enumerates its entries in key order, and
+    nlohmann::ordered_map in insertion order. An object type that enumerates its
+    entries in an unspecified order, such as std::unordered_map, compares them
+    pairwise instead; the difference could only ever show below the bound.
+
+    Note that the stack this walks with is allocated, while the comparison
+    operators are noexcept and the container comparison this replaces allocated
+    nothing. Failing that allocation therefore ends the process rather than
+    throwing. It only arises for values nested past the bound, and only when
+    memory has run out - where the same comparison used to exhaust the call
+    stack instead - but it is a way to fail that the operators did not have.
+    */
+    template<bool Ordered>
+    static compare_result compare_iteratively(const_reference lhs, const_reference rhs,
+            const bool unordered_compares_equal) noexcept
+    {
+        /// a pair of containers being compared in lockstep
+        struct frame
+        {
+            const basic_json* lhs_value{nullptr};
+            const basic_json* rhs_value{nullptr};
+            typename array_t::const_iterator lhs_array_it{};
+            typename array_t::const_iterator rhs_array_it{};
+            typename object_t::const_iterator lhs_object_it{};
+            typename object_t::const_iterator rhs_object_it{};
+        };
+
+        std::vector<frame> stack;
+        const basic_json* left = &lhs;
+        const basic_json* right = &rhs;
+
+        for (;;)
+        {
+            const auto type = left->m_data.m_type;
+
+            if (type == right->m_data.m_type && (type == value_t::array || type == value_t::object))
+            {
+                // descend: the elements decide, and are compared further down
+                stack.emplace_back();
+                frame& pushed = stack.back();
+                pushed.lhs_value = left;
+                pushed.rhs_value = right;
+
+                if (type == value_t::array)
+                {
+                    pushed.lhs_array_it = left->m_data.m_value.array->cbegin();
+                    pushed.rhs_array_it = right->m_data.m_value.array->cbegin();
+                }
+                else
+                {
+                    pushed.lhs_object_it = left->m_data.m_value.object->cbegin();
+                    pushed.rhs_object_it = right->m_data.m_value.object->cbegin();
+                }
+            }
+            else
+            {
+                const compare_result result = compare_leaves<Ordered>(*left, *right);
+
+                // Values that cannot be ordered - a NaN, say - end an ordered
+                // comparison for std::lexicographical_compare_three_way, but
+                // std::lexicographical_compare treats them as equivalent and
+                // carries on with the next element. Both are reproduced here,
+                // so that a value nested too deeply to descend into compares
+                // exactly as one that is not.
+                if (result != compare_result::equal &&
+                        !(unordered_compares_equal && result == compare_result::unordered))
+                {
+                    return result;
+                }
+            }
+
+            // walk back up past the containers that are exhausted, then take the
+            // next pair of elements from the innermost one that is not
+            for (;;)
+            {
+                if (stack.empty())
+                {
+                    return compare_result::equal;
+                }
+
+                frame& current = stack.back();
+                const bool is_object = current.lhs_value->m_data.m_type == value_t::object;
+
+                const bool lhs_done = is_object
+                                      ? current.lhs_object_it == current.lhs_value->m_data.m_value.object->cend()
+                                      : current.lhs_array_it == current.lhs_value->m_data.m_value.array->cend();
+                const bool rhs_done = is_object
+                                      ? current.rhs_object_it == current.rhs_value->m_data.m_value.object->cend()
+                                      : current.rhs_array_it == current.rhs_value->m_data.m_value.array->cend();
+
+                if (lhs_done || rhs_done)
+                {
+                    // whichever ran out first holds the smaller container; if
+                    // both did, they are equal and the container above decides
+                    if (lhs_done != rhs_done)
+                    {
+                        return lhs_done ? compare_result::less : compare_result::greater;
+                    }
+
+                    stack.pop_back();
+                    continue;
+                }
+
+                if (is_object)
+                {
+                    // an entry is a key and a value, and the key decides first
+                    const compare_result key_result =
+                        compare_keys(current.lhs_object_it->first, current.rhs_object_it->first,
+                                     std::integral_constant<bool, Ordered> {});
+
+                    if (key_result != compare_result::equal)
+                    {
+                        return key_result;
+                    }
+
+                    left = &(current.lhs_object_it->second);
+                    right = &(current.rhs_object_it->second);
+                    ++current.lhs_object_it;
+                    ++current.rhs_object_it;
+                }
+                else
+                {
+                    left = &(*current.lhs_array_it);
+                    right = &(*current.rhs_array_it);
+                    ++current.lhs_array_it;
+                    ++current.rhs_array_it;
+                }
+
+                break;
+            }
+        }
+    }
+
+
  public:
    //////////////////////////
    // JSON parser callback //
@@ -1200,60 +1820,15 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        // check of passed value is valid
        other.assert_invariant();

-        switch (m_data.m_type)
+        if (m_data.m_type == value_t::object || m_data.m_type == value_t::array)
        {
-            case value_t::object:
-            {
-                m_data.m_value = *other.m_data.m_value.object;
-                break;
-            }
-
-            case value_t::array:
-            {
-                m_data.m_value = *other.m_data.m_value.array;
-                break;
-            }
-
-            case value_t::string:
-            {
-                m_data.m_value = *other.m_data.m_value.string;
-                break;
-            }
-
-            case value_t::boolean:
-            {
-                m_data.m_value = other.m_data.m_value.boolean;
-                break;
-            }
-
-            case value_t::number_integer:
-            {
-                m_data.m_value = other.m_data.m_value.number_integer;
-                break;
-            }
-
-            case value_t::number_unsigned:
-            {
-                m_data.m_value = other.m_data.m_value.number_unsigned;
-                break;
-            }
-
-            case value_t::number_float:
-            {
-                m_data.m_value = other.m_data.m_value.number_float;
-                break;
-            }
-
-            case value_t::binary:
-            {
-                m_data.m_value = *other.m_data.m_value.binary;
-                break;
-            }
-
-            case value_t::null:
-            case value_t::discarded:
-            default:
-                break;
+            // copying the container directly would call this constructor again
+            // for every element, once per nesting level
+            copy_structured(other);
+        }
+        else
+        {
+            copy_leaf_value(other, *this);
        }

        set_parents();
@@ -3658,7 +4233,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
    // because any negative signed value is smaller than any unsigned value.
    // Otherwise, the non-negative signed value is cast to unsigned before the
    // comparison to avoid wraparound.
-#define JSON_IMPLEMENT_OPERATOR(op, null_result, unordered_result, default_result)                       \
+#define JSON_IMPLEMENT_OPERATOR(op, null_result, unordered_result, default_result, deep_result, may_descend) \
    const auto lhs_type = lhs.type();                                                                    \
    const auto rhs_type = rhs.type();                                                                    \
    \
@@ -3667,11 +4242,25 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        switch (lhs_type)                                                                                \
        {                                                                                                \
            case value_t::array:                                                                         \
+            {                                                                                            \
+                if (JSON_HEDLEY_UNLIKELY(nesting_depth_exhausted(may_descend)))                        \
+                {                                                                                        \
+                    return (deep_result);                                                                \
+                }                                                                                        \
+                const nesting_depth_guard guard;                                                         \
                return (*lhs.m_data.m_value.array) op (*rhs.m_data.m_value.array);                                     \
-                \
+            }                                                                                            \
+            \
            case value_t::object:                                                                        \
+            {                                                                                            \
+                if (JSON_HEDLEY_UNLIKELY(nesting_depth_exhausted(may_descend)))                        \
+                {                                                                                        \
+                    return (deep_result);                                                                \
+                }                                                                                        \
+                const nesting_depth_guard guard;                                                         \
                return (*lhs.m_data.m_value.object) op (*rhs.m_data.m_value.object);                                   \
-                \
+            }                                                                                            \
+            \
            case value_t::null:                                                                          \
                return (null_result);                                                                    \
                \
@@ -3771,7 +4360,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
 #pragma GCC diagnostic ignored "-Wfloat-equal"
 #endif
        const_reference lhs = *this;
-        JSON_IMPLEMENT_OPERATOR( ==, true, false, false)
+        JSON_IMPLEMENT_OPERATOR( ==, true, false, false,
+                                 compare_iteratively<false>(lhs, rhs, false) == compare_result::equal, true)
 #ifdef __GNUC__
 #pragma GCC diagnostic pop
 #endif
@@ -3796,7 +4386,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        JSON_IMPLEMENT_OPERATOR(<=>, // *NOPAD*
                                std::partial_ordering::equivalent,
                                std::partial_ordering::unordered,
-                                lhs_type <=> rhs_type) // *NOPAD*
+                                lhs_type <=> rhs_type, // *NOPAD*
+                                to_partial_ordering(compare_iteratively<true>(lhs, rhs, false)), true)
    }

    /// @brief comparison: 3-way
@@ -3863,7 +4454,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
 #pragma GCC diagnostic push
 #pragma GCC diagnostic ignored "-Wfloat-equal"
 #endif
-        JSON_IMPLEMENT_OPERATOR( ==, true, false, false)
+        JSON_IMPLEMENT_OPERATOR( ==, true, false, false,
+                                 compare_iteratively<false>(lhs, rhs, false) == compare_result::equal, true)
 #ifdef __GNUC__
 #pragma GCC diagnostic pop
 #endif
@@ -3919,7 +4511,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        // default_result is used if we cannot compare values. In that case,
        // we compare types. Note we have to call the operator explicitly,
        // because MSVC has problems otherwise.
-        JSON_IMPLEMENT_OPERATOR( <, false, false, operator<(lhs_type, rhs_type))
+        JSON_IMPLEMENT_OPERATOR( <, false, false, operator<(lhs_type, rhs_type),
+                                 compare_iteratively<true>(lhs, rhs, true) == compare_result::less, false)
    }

    /// @brief comparison: less than
@@ -28,14 +28,14 @@
    #pragma GCC diagnostic ignored "-Wignored-attributes"
 #endif

-#include <algorithm> // all_of, find, for_each
+#include <algorithm> // all_of, find, for_each, none_of
 #include <cstddef> // nullptr_t, ptrdiff_t, size_t
 #include <functional> // hash, less
 #include <initializer_list> // initializer_list
 #ifndef JSON_NO_IO
    #include <iosfwd> // istream, ostream
 #endif  // JSON_NO_IO
-#include <iterator> // random_access_iterator_tag
+#include <iterator> // make_move_iterator, random_access_iterator_tag
 #include <memory> // unique_ptr
 #include <string> // string, stoi, to_string
 #include <utility> // declval, forward, move, pair, swap
@@ -2561,6 +2561,15 @@ JSON_HEDLEY_DIAGNOSTIC_POP
    #define JSON_NO_UNIQUE_ADDRESS
 #endif

+// Clang targeting MinGW does not survive the thread_local storage the copy
+// constructor uses to bound its descent: every test that copies a value
+// segfaults with clang 11.0.1 and clang 18.1.8, while the same tests pass with
+// GCC targeting MinGW and with every other toolchain the library is tested on.
+// Copying works the same way without the counter, only more slowly.
+#if !defined(JSON_NO_THREAD_LOCAL) && defined(__clang__) && defined(__MINGW32__)
+    #define JSON_NO_THREAD_LOCAL 1
+#endif
+
 // disable documentation warnings on clang
 #if defined(__clang__)
    #pragma clang diagnostic push
@@ -22166,6 +22175,626 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        return j;
    }

+    /// the number of levels an operation descends into before it finishes the
+    /// value below it without the call stack
+    static constexpr std::size_t nesting_depth_limit()
+    {
+        return 128;
+    }
+
+#ifndef JSON_NO_THREAD_LOCAL
+    /*!
+    @brief how many levels the operation going on in this thread has descended into
+
+    Copying a value and comparing two values share this count. The library never
+    nests one inside the other - copying a value does not compare one, and
+    comparing two values does not copy them - and where user code nests them
+    anyway, sharing the count only ends a descent sooner than it had to, which
+    costs a little speed and is never wrong.
+
+    A byte is enough: the count never exceeds the limit by more than the single
+    level that notices the limit has been reached.
+    */
+    static std::size_t& nesting_depth() noexcept
+    {
+        static thread_local std::size_t depth = 0; // NOLINT(misc-use-internal-linkage)
+        return depth;
+    }
+#endif
+
+    /*!
+    @brief whether a descent must stop here and finish without the call stack
+
+    @a may_descend says whether the operator descends at all; it is a constant
+    at every call site, and is passed rather than tested by the caller so that
+    the test does not become a constant condition there, which MSVC reports as
+    C4127.
+    */
+    static bool nesting_depth_exhausted(bool may_descend = true) noexcept
+    {
+#ifdef JSON_NO_THREAD_LOCAL
+        // without a count of its own per thread, a descent cannot be bounded
+        // without racing another one, so none is made
+        static_cast<void>(may_descend);
+        return true;
+#else
+        return !may_descend || nesting_depth() >= nesting_depth_limit();
+#endif
+    }
+
+    /*!
+    @brief counts one level of a bounded descent for as long as it runs
+
+    The constructor taking the count is for callers that have looked it up
+    already to test it: reaching thread-local storage is not free, and the path
+    that is taken almost every time should reach it once rather than twice. The
+    other is for callers that cannot look it up - the comparison operators are
+    written as a macro, and a macro cannot use the preprocessor.
+    */
+    class nesting_depth_guard
+    {
+      public:
+        explicit nesting_depth_guard(std::size_t& depth) noexcept
+            : m_depth(&depth)
+        {
+            ++*m_depth;
+        }
+
+        nesting_depth_guard() noexcept
+            : m_depth(countable())
+        {
+            if (m_depth != nullptr)
+            {
+                ++*m_depth;
+            }
+        }
+
+        ~nesting_depth_guard()
+        {
+            if (m_depth != nullptr)
+            {
+                --*m_depth;
+            }
+        }
+
+        nesting_depth_guard(const nesting_depth_guard&) = delete;
+        nesting_depth_guard& operator=(const nesting_depth_guard&) = delete;
+        nesting_depth_guard(nesting_depth_guard&&) = delete;
+        nesting_depth_guard& operator=(nesting_depth_guard&&) = delete;
+
+      private:
+        /// @brief the count to keep, or nullptr where there is none to keep
+        static std::size_t* countable() noexcept
+        {
+#ifdef JSON_NO_THREAD_LOCAL
+            return nullptr;
+#else
+            return &nesting_depth();
+#endif
+        }
+
+        std::size_t* m_depth;
+    };
+
+    /// an entry of the iterative deep copy's worklist: a structured value and
+    /// the value that is to become its copy
+    using copy_worklist_t = std::vector<std::pair<const basic_json*, basic_json*>>;
+
+    /// scratch space to build the key skeleton of an object copy in one go
+    using copy_scratch_t = std::vector<std::pair<typename object_t::key_type, basic_json>>;
+
+    /// @brief copy everything of @a src into @a dst but its type and value
+    static void copy_metadata(const basic_json& src, basic_json& dst)
+    {
+        // a custom base class is only required to be copy-constructible and
+        // move-assignable, so the copy has to go through a temporary
+        static_cast<json_base_class_t&>(dst) = json_base_class_t(static_cast<const json_base_class_t&>(src));
+
+#if JSON_DIAGNOSTIC_POSITIONS
+        dst.start_position = src.start_position;
+        dst.end_position = src.end_position;
+#else
+        static_cast<void>(src);
+        static_cast<void>(dst);
+#endif
+    }
+
+    /*!
+    @brief copy the value of @a src into @a dst, which must not be structured
+
+    Objects and arrays are left alone: creating those is the one thing the copy
+    constructor and @ref copy_shallow do differently from one another, and it is
+    the reason copying a value can descend at all.
+    */
+    /// @note inlined on purpose: both callers have already told an object or an
+    ///       array apart from the rest, and letting the compiler fold that test
+    ///       into this switch is worth a few percent when copying a value made
+    ///       mostly of numbers
+    JSON_HEDLEY_ALWAYS_INLINE
+    static void copy_leaf_value(const basic_json& src, basic_json& dst)
+    {
+        switch (src.m_data.m_type)
+        {
+            case value_t::string:
+            {
+                dst.m_data.m_value = *src.m_data.m_value.string;
+                break;
+            }
+
+            case value_t::binary:
+            {
+                dst.m_data.m_value = *src.m_data.m_value.binary;
+                break;
+            }
+
+            case value_t::boolean:
+            {
+                dst.m_data.m_value = src.m_data.m_value.boolean;
+                break;
+            }
+
+            case value_t::number_integer:
+            {
+                dst.m_data.m_value = src.m_data.m_value.number_integer;
+                break;
+            }
+
+            case value_t::number_unsigned:
+            {
+                dst.m_data.m_value = src.m_data.m_value.number_unsigned;
+                break;
+            }
+
+            case value_t::number_float:
+            {
+                dst.m_data.m_value = src.m_data.m_value.number_float;
+                break;
+            }
+
+            case value_t::object:
+            case value_t::array:
+            case value_t::null:
+            case value_t::discarded:
+            default:
+                break;
+        }
+    }
+
+    /*!
+    @brief copy everything of @a src into the null value @a dst but the children
+
+    Objects and arrays are not copied here; they are appended to @a worklist to
+    be created later by @ref copy_iteratively. Until that happens, @a dst remains
+    a null value, so that a partially built copy can be destroyed at any point
+    without ever violating the class invariants.
+    */
+    static void copy_shallow(const basic_json& src, basic_json& dst, copy_worklist_t& worklist)
+    {
+        copy_metadata(src, dst);
+
+        if (src.m_data.m_type == value_t::object || src.m_data.m_type == value_t::array)
+        {
+            // defer: dst stays a null value until its container exists
+            worklist.emplace_back(&src, &dst);
+            return;
+        }
+
+        copy_leaf_value(src, dst);
+
+        // only now that the value exists may the type be set: had the creation
+        // of the value thrown, dst would have been left as a valid null value
+        dst.m_data.m_type = src.m_data.m_type;
+    }
+
+    /// @brief create the copy of the array @a src in @a dst
+    /// @note structured elements are appended to @a worklist instead
+    static void copy_array_level(const basic_json& src, basic_json& dst, copy_worklist_t& worklist)
+    {
+        const array_t& src_array = *src.m_data.m_value.array;
+
+        // create all elements up front: growing the array afterwards could
+        // invalidate the pointers that are handed to the worklist
+        dst.m_data.m_value.array = create<array_t>(src_array.size(), basic_json());
+
+        auto dst_it = dst.m_data.m_value.array->begin();
+        for (auto src_it = src_array.cbegin(); src_it != src_array.cend(); ++src_it, ++dst_it)
+        {
+            copy_shallow(*src_it, *dst_it, worklist);
+        }
+    }
+
+    /// @brief create the copy of the object @a src in @a dst
+    /// @note structured values are appended to @a worklist instead
+    static void copy_object_level(const basic_json& src, basic_json& dst,
+                                  copy_worklist_t& worklist, copy_scratch_t& scratch)
+    {
+        const object_t& src_object = *src.m_data.m_value.object;
+
+        // build the complete key skeleton and hand it to the object's range
+        // constructor: adding the keys one by one would be quadratic for object
+        // types that are backed by a vector, such as nlohmann::ordered_map
+        scratch.clear();
+        scratch.reserve(src_object.size());
+        for (const auto& element : src_object)
+        {
+            scratch.emplace_back(element.first, basic_json());
+        }
+
+        dst.m_data.m_value.object = create<object_t>(std::make_move_iterator(scratch.begin()),
+                                    std::make_move_iterator(scratch.end()));
+        scratch.clear();
+
+        // pair every value of the copy with its counterpart in the original;
+        // both are enumerated in the same order for every object type with a
+        // deterministic order, so the lookup is only needed for exotic ones
+        auto src_it = src_object.cbegin();
+        for (auto& element : *dst.m_data.m_value.object)
+        {
+            if (JSON_HEDLEY_LIKELY(src_it != src_object.cend() && src_it->first == element.first))
+            {
+                copy_shallow(src_it->second, element.second, worklist);
+                ++src_it;
+            }
+            else
+            {
+                const auto found = src_object.find(element.first);
+                JSON_ASSERT(found != src_object.cend());
+                copy_shallow(found->second, element.second, worklist);
+            }
+        }
+    }
+
+    /*!
+    @brief deep-copy the object or array @a src into this value without recursing
+
+    The values whose copy has not been created yet are kept on an explicit
+    worklist rather than on the call stack. This is only reached for values
+    nested deeper than @ref nesting_depth_limit levels, which is why it copies
+    every container by hand instead of letting the container do it: the fast
+    ways of doing so would descend into the elements and defeat the purpose.
+    */
+    void copy_iteratively(const basic_json& src)
+    {
+        copy_worklist_t worklist;
+        copy_scratch_t scratch;
+
+        const basic_json* src_value = &src;
+        basic_json* dst_value = this;
+
+        for (;;)
+        {
+            if (src_value->m_data.m_type == value_t::array)
+            {
+                copy_array_level(*src_value, *dst_value, worklist);
+            }
+            else
+            {
+                copy_object_level(*src_value, *dst_value, worklist, scratch);
+            }
+
+            // the container is complete and will not be modified again
+            dst_value->set_parents();
+
+            if (worklist.empty())
+            {
+                break;
+            }
+
+            const auto& next = worklist.back();
+            src_value = next.first;
+            dst_value = next.second;
+            worklist.pop_back();
+
+            // the value stops being a null value exactly here
+            dst_value->m_data.m_type = src_value->m_data.m_type;
+        }
+    }
+
+    /*!
+    @brief copy one level of the object or array @a src into this value
+
+    The container copies its own elements, which is the fastest way to fill it.
+    Every element that is structured itself comes back to @ref copy_structured.
+    */
+    void copy_level(const basic_json& src)
+    {
+        if (m_data.m_type == value_t::object)
+        {
+            m_data.m_value = *src.m_data.m_value.object;
+        }
+        else
+        {
+            m_data.m_value = *src.m_data.m_value.array;
+        }
+
+        set_parents();
+    }
+
+    /*!
+    @brief deep-copy the object or array @a src into this value
+
+    Copying a container copies its elements, so a value nested deeply enough
+    used to exhaust the call stack. The descent is bounded here: the first
+    @ref nesting_depth_limit levels are copied by the containers themselves, just
+    as they always were, and anything below that is copied without the call
+    stack by @ref copy_iteratively. Copying a value can therefore no longer
+    exhaust the stack, however deeply it is nested, just like destroying one
+    cannot since #1436.
+
+    Nothing has to be scanned or built by hand to reach that: a value that is
+    not nested deeper than the limit - all but a vanishing minority - is copied
+    exactly as it was before, and this whole detour costs it one counter.
+
+    @sa https://github.com/nlohmann/json/issues/5387
+    */
+    void copy_structured(const basic_json& src)
+    {
+#ifndef JSON_NO_THREAD_LOCAL
+        std::size_t& depth = nesting_depth();
+
+        if (JSON_HEDLEY_LIKELY(depth < nesting_depth_limit()))
+        {
+            const nesting_depth_guard guard(depth);
+            copy_level(src);
+            return;
+        }
+#endif
+
+        // Finish this value without descending any further. It is completed
+        // before this returns, so a copy made by a custom base class - or by
+        // anything else that runs while a copy is going on - is unaffected by
+        // the copy it is nested in.
+        copy_iteratively(src);
+    }
+
+
+    /// the result of comparing two values, including values that cannot be
+    /// ordered at all, such as a discarded value or a NaN
+    enum class compare_result { less, equal, greater, unordered };
+
+#if JSON_HAS_THREE_WAY_COMPARISON
+    /// @brief the ordering that @a result stands for
+    static std::partial_ordering to_partial_ordering(compare_result result) noexcept // *NOPAD*
+    {
+        switch (result)
+        {
+            case compare_result::less:
+                return std::partial_ordering::less;
+            case compare_result::greater:
+                return std::partial_ordering::greater;
+            case compare_result::equal:
+                return std::partial_ordering::equivalent;
+            case compare_result::unordered:
+            default:
+                return std::partial_ordering::unordered;
+        }
+    }
+#endif
+
+    /*!
+    @brief compare two values that are not both an array or both an object
+
+    Such a pair is compared by the operators themselves, which cannot descend
+    into it and therefore cannot recurse.
+
+    That holds for a pair whose types differ as much as for a pair of leaves: an
+    array and an object are told apart by their types alone, because an operator
+    only ever descends into two values of the same type. So `==` reports them as
+    unequal without looking inside either, and an ordering falls back to the
+    order of the types - an object sorts before an array - exactly as it does
+    for a value that is not nested deeply enough to get here.
+    */
+    template<bool Ordered>
+    static compare_result compare_leaves(const_reference lhs, const_reference rhs) noexcept
+    {
+        if (lhs == rhs)
+        {
+            return compare_result::equal;
+        }
+
+        return order_leaves(lhs, rhs, std::integral_constant<bool, Ordered> {});
+    }
+
+    /*!
+    @brief compare two object keys
+
+    An object compares its entries as pairs of a key and a value, so its keys
+    are compared exactly as std::pair compares them: with < where the objects
+    are being ordered, and with == where they are only checked for equality.
+    Note that this is not the object's own comparator, which for a vector-backed
+    object type such as nlohmann::ordered_map tells equality rather than order.
+    */
+    static compare_result compare_keys(const typename object_t::key_type& lhs,
+                                       const typename object_t::key_type& rhs,
+                                       std::true_type /*ordered*/)
+    {
+        if (lhs < rhs)
+        {
+            return compare_result::less;
+        }
+
+        if (rhs < lhs)
+        {
+            return compare_result::greater;
+        }
+
+        return compare_result::equal;
+    }
+
+    /// @brief check two object keys for equality
+    static compare_result compare_keys(const typename object_t::key_type& lhs,
+                                       const typename object_t::key_type& rhs,
+                                       std::false_type /*ordered*/)
+    {
+        return lhs == rhs ? compare_result::equal : compare_result::unordered;
+    }
+
+    /// @brief tell apart two values that are not equal
+    /// @note only instantiated where the values are being ordered, as a key or
+    ///       string type is not required to be ordered to be compared for equality
+    static compare_result order_leaves(const_reference lhs, const_reference rhs, std::true_type /*ordered*/) noexcept
+    {
+        if (lhs < rhs)
+        {
+            return compare_result::less;
+        }
+
+        if (rhs < lhs)
+        {
+            return compare_result::greater;
+        }
+
+        return compare_result::unordered;
+    }
+
+    /// @brief report two values as not equal without ordering them
+    static compare_result order_leaves(const_reference /*lhs*/, const_reference /*rhs*/, std::false_type /*ordered*/) noexcept
+    {
+        return compare_result::unordered;
+    }
+
+    /*!
+    @brief compare @a lhs and @a rhs without descending into them
+
+    Reached once a comparison has descended @ref nesting_depth_limit levels, so
+    that comparing values cannot exhaust the call stack however deeply they are
+    nested. The two values are walked in lockstep on an explicit stack and
+    compared lexicographically, element by element in the order the containers
+    enumerate them - which is how the container types this library ships compare
+    themselves: a std::map enumerates its entries in key order, and
+    nlohmann::ordered_map in insertion order. An object type that enumerates its
+    entries in an unspecified order, such as std::unordered_map, compares them
+    pairwise instead; the difference could only ever show below the bound.
+
+    Note that the stack this walks with is allocated, while the comparison
+    operators are noexcept and the container comparison this replaces allocated
+    nothing. Failing that allocation therefore ends the process rather than
+    throwing. It only arises for values nested past the bound, and only when
+    memory has run out - where the same comparison used to exhaust the call
+    stack instead - but it is a way to fail that the operators did not have.
+    */
+    template<bool Ordered>
+    static compare_result compare_iteratively(const_reference lhs, const_reference rhs,
+            const bool unordered_compares_equal) noexcept
+    {
+        /// a pair of containers being compared in lockstep
+        struct frame
+        {
+            const basic_json* lhs_value{nullptr};
+            const basic_json* rhs_value{nullptr};
+            typename array_t::const_iterator lhs_array_it{};
+            typename array_t::const_iterator rhs_array_it{};
+            typename object_t::const_iterator lhs_object_it{};
+            typename object_t::const_iterator rhs_object_it{};
+        };
+
+        std::vector<frame> stack;
+        const basic_json* left = &lhs;
+        const basic_json* right = &rhs;
+
+        for (;;)
+        {
+            const auto type = left->m_data.m_type;
+
+            if (type == right->m_data.m_type && (type == value_t::array || type == value_t::object))
+            {
+                // descend: the elements decide, and are compared further down
+                stack.emplace_back();
+                frame& pushed = stack.back();
+                pushed.lhs_value = left;
+                pushed.rhs_value = right;
+
+                if (type == value_t::array)
+                {
+                    pushed.lhs_array_it = left->m_data.m_value.array->cbegin();
+                    pushed.rhs_array_it = right->m_data.m_value.array->cbegin();
+                }
+                else
+                {
+                    pushed.lhs_object_it = left->m_data.m_value.object->cbegin();
+                    pushed.rhs_object_it = right->m_data.m_value.object->cbegin();
+                }
+            }
+            else
+            {
+                const compare_result result = compare_leaves<Ordered>(*left, *right);
+
+                // Values that cannot be ordered - a NaN, say - end an ordered
+                // comparison for std::lexicographical_compare_three_way, but
+                // std::lexicographical_compare treats them as equivalent and
+                // carries on with the next element. Both are reproduced here,
+                // so that a value nested too deeply to descend into compares
+                // exactly as one that is not.
+                if (result != compare_result::equal &&
+                        !(unordered_compares_equal && result == compare_result::unordered))
+                {
+                    return result;
+                }
+            }
+
+            // walk back up past the containers that are exhausted, then take the
+            // next pair of elements from the innermost one that is not
+            for (;;)
+            {
+                if (stack.empty())
+                {
+                    return compare_result::equal;
+                }
+
+                frame& current = stack.back();
+                const bool is_object = current.lhs_value->m_data.m_type == value_t::object;
+
+                const bool lhs_done = is_object
+                                      ? current.lhs_object_it == current.lhs_value->m_data.m_value.object->cend()
+                                      : current.lhs_array_it == current.lhs_value->m_data.m_value.array->cend();
+                const bool rhs_done = is_object
+                                      ? current.rhs_object_it == current.rhs_value->m_data.m_value.object->cend()
+                                      : current.rhs_array_it == current.rhs_value->m_data.m_value.array->cend();
+
+                if (lhs_done || rhs_done)
+                {
+                    // whichever ran out first holds the smaller container; if
+                    // both did, they are equal and the container above decides
+                    if (lhs_done != rhs_done)
+                    {
+                        return lhs_done ? compare_result::less : compare_result::greater;
+                    }
+
+                    stack.pop_back();
+                    continue;
+                }
+
+                if (is_object)
+                {
+                    // an entry is a key and a value, and the key decides first
+                    const compare_result key_result =
+                        compare_keys(current.lhs_object_it->first, current.rhs_object_it->first,
+                                     std::integral_constant<bool, Ordered> {});
+
+                    if (key_result != compare_result::equal)
+                    {
+                        return key_result;
+                    }
+
+                    left = &(current.lhs_object_it->second);
+                    right = &(current.rhs_object_it->second);
+                    ++current.lhs_object_it;
+                    ++current.rhs_object_it;
+                }
+                else
+                {
+                    left = &(*current.lhs_array_it);
+                    right = &(*current.rhs_array_it);
+                    ++current.lhs_array_it;
+                    ++current.rhs_array_it;
+                }
+
+                break;
+            }
+        }
+    }
+
+
  public:
    //////////////////////////
    // JSON parser callback //
@@ -22545,60 +23174,15 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        // check of passed value is valid
        other.assert_invariant();

-        switch (m_data.m_type)
+        if (m_data.m_type == value_t::object || m_data.m_type == value_t::array)
        {
-            case value_t::object:
-            {
-                m_data.m_value = *other.m_data.m_value.object;
-                break;
-            }
-
-            case value_t::array:
-            {
-                m_data.m_value = *other.m_data.m_value.array;
-                break;
-            }
-
-            case value_t::string:
-            {
-                m_data.m_value = *other.m_data.m_value.string;
-                break;
-            }
-
-            case value_t::boolean:
-            {
-                m_data.m_value = other.m_data.m_value.boolean;
-                break;
-            }
-
-            case value_t::number_integer:
-            {
-                m_data.m_value = other.m_data.m_value.number_integer;
-                break;
-            }
-
-            case value_t::number_unsigned:
-            {
-                m_data.m_value = other.m_data.m_value.number_unsigned;
-                break;
-            }
-
-            case value_t::number_float:
-            {
-                m_data.m_value = other.m_data.m_value.number_float;
-                break;
-            }
-
-            case value_t::binary:
-            {
-                m_data.m_value = *other.m_data.m_value.binary;
-                break;
-            }
-
-            case value_t::null:
-            case value_t::discarded:
-            default:
-                break;
+            // copying the container directly would call this constructor again
+            // for every element, once per nesting level
+            copy_structured(other);
+        }
+        else
+        {
+            copy_leaf_value(other, *this);
        }

        set_parents();
@@ -25003,7 +25587,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
    // because any negative signed value is smaller than any unsigned value.
    // Otherwise, the non-negative signed value is cast to unsigned before the
    // comparison to avoid wraparound.
-#define JSON_IMPLEMENT_OPERATOR(op, null_result, unordered_result, default_result)                       \
+#define JSON_IMPLEMENT_OPERATOR(op, null_result, unordered_result, default_result, deep_result, may_descend) \
    const auto lhs_type = lhs.type();                                                                    \
    const auto rhs_type = rhs.type();                                                                    \
    \
@@ -25012,11 +25596,25 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        switch (lhs_type)                                                                                \
        {                                                                                                \
            case value_t::array:                                                                         \
+            {                                                                                            \
+                if (JSON_HEDLEY_UNLIKELY(nesting_depth_exhausted(may_descend)))                        \
+                {                                                                                        \
+                    return (deep_result);                                                                \
+                }                                                                                        \
+                const nesting_depth_guard guard;                                                         \
                return (*lhs.m_data.m_value.array) op (*rhs.m_data.m_value.array);                                     \
-                \
+            }                                                                                            \
+            \
            case value_t::object:                                                                        \
+            {                                                                                            \
+                if (JSON_HEDLEY_UNLIKELY(nesting_depth_exhausted(may_descend)))                        \
+                {                                                                                        \
+                    return (deep_result);                                                                \
+                }                                                                                        \
+                const nesting_depth_guard guard;                                                         \
                return (*lhs.m_data.m_value.object) op (*rhs.m_data.m_value.object);                                   \
-                \
+            }                                                                                            \
+            \
            case value_t::null:                                                                          \
                return (null_result);                                                                    \
                \
@@ -25116,7 +25714,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
 #pragma GCC diagnostic ignored "-Wfloat-equal"
 #endif
        const_reference lhs = *this;
-        JSON_IMPLEMENT_OPERATOR( ==, true, false, false)
+        JSON_IMPLEMENT_OPERATOR( ==, true, false, false,
+                                 compare_iteratively<false>(lhs, rhs, false) == compare_result::equal, true)
 #ifdef __GNUC__
 #pragma GCC diagnostic pop
 #endif
@@ -25141,7 +25740,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        JSON_IMPLEMENT_OPERATOR(<=>, // *NOPAD*
                                std::partial_ordering::equivalent,
                                std::partial_ordering::unordered,
-                                lhs_type <=> rhs_type) // *NOPAD*
+                                lhs_type <=> rhs_type, // *NOPAD*
+                                to_partial_ordering(compare_iteratively<true>(lhs, rhs, false)), true)
    }

    /// @brief comparison: 3-way
@@ -25208,7 +25808,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
 #pragma GCC diagnostic push
 #pragma GCC diagnostic ignored "-Wfloat-equal"
 #endif
-        JSON_IMPLEMENT_OPERATOR( ==, true, false, false)
+        JSON_IMPLEMENT_OPERATOR( ==, true, false, false,
+                                 compare_iteratively<false>(lhs, rhs, false) == compare_result::equal, true)
 #ifdef __GNUC__
 #pragma GCC diagnostic pop
 #endif
@@ -25264,7 +25865,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
        // default_result is used if we cannot compare values. In that case,
        // we compare types. Note we have to call the operator explicitly,
        // because MSVC has problems otherwise.
-        JSON_IMPLEMENT_OPERATOR( <, false, false, operator<(lhs_type, rhs_type))
+        JSON_IMPLEMENT_OPERATOR( <, false, false, operator<(lhs_type, rhs_type),
+                                 compare_iteratively<true>(lhs, rhs, true) == compare_result::less, false)
    }

    /// @brief comparison: less than
@@ -216,6 +216,57 @@ TEST_CASE("controlled bad_alloc")
            CHECK_THROWS_AS(my_json(s), std::bad_alloc&);
            next_construct_fails = false;
        }
+
+        SECTION("basic_json(const basic_json&) of a deeply nested value (#5387)")
+        {
+            // Copying a value nested deeper than the descent bound builds the
+            // copy from the top down: every value whose own copy has not been
+            // made yet stays a null value until it is. Failing an allocation
+            // part-way through is what proves such a half-built copy can still
+            // be destroyed.
+            //
+            // Which path the failure lands in depends on the build: the first
+            // allocation of a copy belongs to the outermost level, so here it
+            // is the descending one. Built with JSON_NO_THREAD_LOCAL - as the
+            // ci_test_no_thread_local target builds the whole suite - no
+            // descent is made at all and the very same failure lands in the
+            // iterative path instead, part-way through its worklist.
+            const auto check_deep_copy = [](bool objects)
+            {
+                CAPTURE(objects);
+
+                next_construct_fails = false;
+
+                // deeper than the 128 levels the copy constructor descends into
+                const std::size_t depth = 300;
+
+                my_json j = 1;
+                for (std::size_t i = 0; i < depth; ++i)
+                {
+                    if (objects)
+                    {
+                        my_json wrapper = my_json::object();
+                        wrapper["a"] = std::move(j);
+                        j = std::move(wrapper);
+                    }
+                    else
+                    {
+                        j = my_json::array({std::move(j)});
+                    }
+                }
+
+                // NOLINTNEXTLINE(performance-unnecessary-copy-initialization): the copy is what is tested
+                CHECK_NOTHROW(my_json(j));
+
+                next_construct_fails = true;
+                // NOLINTNEXTLINE(performance-unnecessary-copy-initialization): the copy is what is tested
+                CHECK_THROWS_AS(my_json(j), std::bad_alloc&);
+                next_construct_fails = false;
+            };
+
+            check_deep_copy(false);
+            check_deep_copy(true);
+        }
    }
 }

@@ -68,6 +68,72 @@ TEST_CASE("Better diagnostics with positions")
        CHECK(j.end_pos() == root.size());
    }

+    SECTION("copying keeps the positions of nested values (#5387)")
+    {
+        // Values nested deeper than the copy constructor's descent bound are
+        // copied without the call stack, on a path that has to carry the
+        // positions over itself; shallower ones copy their containers, which
+        // bring the positions along. Both sides of the bound are checked here.
+        const auto check_copy = [](std::size_t depth, bool objects)
+        {
+            CAPTURE(depth)
+            CAPTURE(objects)
+
+            const std::string opening = objects ? R"({"a":)" : "[";
+            const std::string closing = objects ? "}" : "]";
+
+            std::string text;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                text += opening;
+            }
+            text += "12";
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                text += closing;
+            }
+
+            const json original = json::parse(text);
+            const json copy(original); // NOLINT(performance-unnecessary-copy-initialization)
+
+            const json* o = &original;
+            const json* c = &copy;
+            for (std::size_t level = 0; level <= depth; ++level)
+            {
+                CAPTURE(level)
+                REQUIRE(c->start_pos() == o->start_pos());
+                REQUIRE(c->end_pos() == o->end_pos());
+
+                if (level < depth)
+                {
+                    o = objects ? &o->at("a") : &o->at(0);
+                    c = objects ? &c->at("a") : &c->at(0);
+                }
+            }
+        };
+
+        const auto check_arrays = [&check_copy](std::size_t depth)
+        {
+            check_copy(depth, false);
+        };
+        const auto check_objects = [&check_copy](std::size_t depth)
+        {
+            check_copy(depth, true);
+        };
+
+        check_arrays(1);
+        check_arrays(127);
+        check_arrays(128);
+        check_arrays(129);
+        check_arrays(300);
+
+        check_objects(1);
+        check_objects(127);
+        check_objects(128);
+        check_objects(129);
+        check_objects(300);
+    }
+
    SECTION("JSON patch add to primitive parent (#4292)")
    {
        // the JSON Patch "add" target /foo/bar/baz has a string parent
@@ -273,5 +273,62 @@ TEST_CASE("Regression tests for extended diagnostics")
        CHECK(j1["numbers"]["two"] == 2);
        CHECK(j1["string"] == "t");
    }
+
+    SECTION("Regression test for issue #5387 - copying keeps the parents of nested values")
+    {
+        // A value nested deeper than the copy constructor's descent bound is
+        // copied without the call stack. Every container that path creates has
+        // to have the parents of its children set, or the JSON Pointer in the
+        // diagnostic is cut short.
+        const std::size_t depth = 300;
+
+        SECTION("objects")
+        {
+            json j = "not a number";
+            std::string pointer;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                j = json{{"a", j}};
+                pointer += "/a";
+            }
+
+            json const copy(j); // NOLINT(performance-unnecessary-copy-initialization)
+
+            const json* inner = &copy;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                inner = &inner->at("a");
+            }
+
+            std::string const expected = "[json.exception.type_error.302] (" + pointer + ") type must be number, but is string";
+            int i = 0;
+            CHECK_THROWS_WITH_AS(i = inner->get<int>(), expected.c_str(), json::type_error);
+            CHECK(i == 0);
+        }
+
+        SECTION("arrays")
+        {
+            json j = "not a number";
+            std::string pointer;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                j = json::array({j});
+                pointer += "/0";
+            }
+
+            json const copy(j); // NOLINT(performance-unnecessary-copy-initialization)
+
+            const json* inner = &copy;
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                inner = &inner->at(0);
+            }
+
+            std::string const expected = "[json.exception.type_error.302] (" + pointer + ") type must be number, but is string";
+            int i = 0;
+            CHECK_THROWS_WITH_AS(i = inner->get<int>(), expected.c_str(), json::type_error);
+            CHECK(i == 0);
+        }
+    }
 }

@@ -12,6 +12,7 @@
 using nlohmann::json;

 #include <algorithm>
+#include <string>

 TEST_CASE("tests on very large JSONs")
 {
@@ -27,3 +28,201 @@ TEST_CASE("tests on very large JSONs")
    }
 }

+namespace
+{
+
+// Descend a chain of single-element containers and return the value at its end,
+// reporting the number of levels traversed in @a depth.
+//
+// The values in the test case below are nested far deeper than the call stack
+// can follow, so they must not be inspected with operator== or dump(): both are
+// still recursive and would overflow the stack themselves.
+const json* innermost_value(const json& j, std::size_t& depth)
+{
+    const json* current = &j;
+    depth = 0;
+
+    while ((current->is_array() || current->is_object()) && !current->empty())
+    {
+        current = current->is_array()
+                  ? &current->front()
+                  : &current->begin().value();
+        ++depth;
+    }
+
+    return current;
+}
+
+} // namespace
+
+TEST_CASE("tests on deeply nested JSONs")
+{
+    // deep enough to exhaust the call stack, but small enough to stay cheap:
+    // parsing is iterative, so building the values below costs little
+    const std::size_t depth = 100000;
+
+    SECTION("issue #5387 - stack overflow in the copy constructor")
+    {
+        SECTION("array")
+        {
+            const json j = json::parse(std::string(depth, '[') + '0' + std::string(depth, ']'));
+
+            const json copy(j); // NOLINT(performance-unnecessary-copy-initialization): the copy is what is tested
+
+            std::size_t copy_depth = 0;
+            CHECK(*innermost_value(copy, copy_depth) == 0);
+            CHECK(copy_depth == depth);
+        }
+
+        SECTION("object")
+        {
+            std::string s;
+            s.reserve((6 * depth) + 1);
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                s += "{\"a\":";
+            }
+            s += '1';
+            s.append(depth, '}');
+
+            const json j = json::parse(s);
+
+            const json copy(j); // NOLINT(performance-unnecessary-copy-initialization): the copy is what is tested
+
+            std::size_t copy_depth = 0;
+            CHECK(*innermost_value(copy, copy_depth) == 1);
+            CHECK(copy_depth == depth);
+        }
+
+        SECTION("copy assignment")
+        {
+            // operator=(basic_json) takes its argument by value, so the deep
+            // copy happens in the copy constructor
+            const json j = json::parse(std::string(depth, '[') + '0' + std::string(depth, ']'));
+
+            json target;
+            target = j;
+
+            std::size_t target_depth = 0;
+            CHECK(*innermost_value(target, target_depth) == 0);
+            CHECK(target_depth == depth);
+        }
+
+        SECTION("depths around the bound of the recursive descent")
+        {
+            // The copy constructor descends into a bounded number of levels and
+            // completes whatever is below that without the call stack. Cover
+            // every depth around that bound, so that the two ways of copying
+            // are known to meet cleanly - wherever the bound is set.
+            for (std::size_t d = 1; d <= 300; ++d)
+            {
+                CAPTURE(d);
+
+                const json array = json::parse(std::string(d, '[') + '0' + std::string(d, ']'));
+                const json array_copy(array); // NOLINT(performance-unnecessary-copy-initialization): the copy is what is tested
+                std::size_t array_depth = 0;
+                CHECK(*innermost_value(array_copy, array_depth) == 0);
+                CHECK(array_depth == d);
+
+                std::string object_text;
+                for (std::size_t i = 0; i < d; ++i)
+                {
+                    object_text += "{\"a\":";
+                }
+                object_text += '1';
+                object_text.append(d, '}');
+
+                const json object = json::parse(object_text);
+                const json object_copy(object); // NOLINT(performance-unnecessary-copy-initialization): the copy is what is tested
+                std::size_t object_depth = 0;
+                CHECK(*innermost_value(object_copy, object_depth) == 1);
+                CHECK(object_depth == d);
+            }
+        }
+
+        SECTION("a value that is deep in one place only")
+        {
+            json j = json::object();
+            j["shallow"] = 1;
+            j["deep"] = json::parse(std::string(depth, '[') + '0' + std::string(depth, ']'));
+            j["also_shallow"] = json::array({1, 2, 3});
+
+            const json copy(j);
+
+            CHECK(copy["shallow"] == 1);
+            CHECK(copy["also_shallow"] == json::array({1, 2, 3}));
+
+            std::size_t deep_depth = 0;
+            CHECK(*innermost_value(copy["deep"], deep_depth) == 0);
+            CHECK(deep_depth == depth);
+        }
+
+        SECTION("comparing")
+        {
+            // Comparing used to descend once per level, and an ordered
+            // comparison used to compare every pair of elements twice, once in
+            // each direction, which took exponentially long in the nesting
+            // depth. Both are gone: these finish in milliseconds, where the
+            // second used to take longer than anyone would wait even for a
+            // value nested only a few dozen levels deep.
+            const std::string text = std::string(depth, '[') + '0' + std::string(depth, ']');
+            const json j = json::parse(text);
+            const json same = json::parse(text);
+            const json larger = json::parse(std::string(depth, '[') + '1' + std::string(depth, ']'));
+
+            CHECK(j == same);
+            CHECK_FALSE(j == larger);
+            CHECK(j != larger);
+
+            CHECK(j < larger);
+            CHECK_FALSE(larger < j);
+            CHECK(larger > j);
+            CHECK(j <= same);
+            CHECK(j >= same);
+
+            // a value that ends earlier is the smaller one
+            const json shorter = json::parse(std::string(depth - 1, '[') + '0' + std::string(depth - 1, ']'));
+            CHECK_FALSE(j == shorter);
+        }
+
+        SECTION("comparing objects")
+        {
+            std::string text;
+            text.reserve((6 * depth) + 1);
+            for (std::size_t i = 0; i < depth; ++i)
+            {
+                text += "{\"a\":";
+            }
+            text += '1';
+            text.append(depth, '}');
+
+            const json j = json::parse(text);
+            const json same = json::parse(text);
+
+            CHECK(j == same);
+            CHECK_FALSE(j != same);
+            CHECK(j <= same);
+            CHECK(j >= same);
+        }
+
+        SECTION("the copy is independent of the original")
+        {
+            const json j = json::parse(std::string(depth, '[') + '0' + std::string(depth, ']'));
+
+            json copy(j);
+
+            // reach the innermost value without recursing and replace it
+            json* current = &copy;
+            while (current->is_array() && !current->empty())
+            {
+                current = &current->front();
+            }
+            *current = 42;
+
+            std::size_t unused = 0;
+            CHECK(*innermost_value(copy, unused) == 42);
+            CHECK(*innermost_value(j, unused) == 0);
+        }
+    }
+}
+
@@ -81,3 +81,37 @@ TEST_CASE("regression test for issue #3732 - iteration_proxy_value<iter_impl<ord
    };
    static_cast<void>(fn);
 }
+
+TEST_CASE("copying an ordered_json with nested values")
+{
+    // ordered_map is backed by a vector, so copying an object that has
+    // structured values takes a different route than copying a std::map-backed
+    // one; see https://github.com/nlohmann/json/issues/5387
+    ordered_json oj;
+    oj["z"] = 1;
+    oj["a"]["y"] = 2;
+    oj["a"]["b"]["x"] = 3;
+    oj["m"] = {1, 2, {{"w", 4}}};
+
+    const ordered_json copy(oj);
+
+    SECTION("the copy is equal to the original")
+    {
+        CHECK(copy == oj);
+        CHECK(copy.dump() == oj.dump());
+    }
+
+    SECTION("the key order is preserved at every level")
+    {
+        CHECK(copy.dump() == R"({"z":1,"a":{"y":2,"b":{"x":3}},"m":[1,2,{"w":4}]})");
+    }
+
+    SECTION("the copy is independent of the original")
+    {
+        ordered_json mutated(oj);
+        mutated["a"]["b"]["x"] = 99;
+
+        CHECK(oj["a"]["b"]["x"] == 3);
+        CHECK(mutated["a"]["b"]["x"] == 99);
+    }
+}
Author	SHA1	Message	Date
Niels Lohmann	596e33c319	Use the shared descent bookkeeping rather than a second set Comparing kept a thread_local count, a limit and a guard of its own beside the ones copying already had, all three the same thing under a different name. They are gone; the shared count, limit and guard do the work. The guard grows a second constructor here, because the comparison operators are written as a macro and a macro cannot use the preprocessor: it cannot look the count up behind an #ifdef the way copy_structured does, so the guard looks it up for it. nesting_depth_exhausted() arrives for the same reason - whether an operator descends at all is a constant at every call site, and testing it there is what MSVC reports as C4127. Also say in compare_leaves what happens to a pair that is an array on one side and an object on the other, since the answer is not obvious from the code: an operator only descends into two values of the same type, so such a pair is told apart by its types alone - unequal, and ordered the way the types are - exactly as it is above the bound. And record what the explicit stack costs: the comparison operators are noexcept and the container comparison this replaces allocated nothing, so running out of memory here ends the process instead of throwing. It takes a value nested past the bound and an exhausted heap to reach, and the same comparison used to exhaust the call stack, but it is a new way to fail. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 12:30:56 +02:00
Niels Lohmann	5fe4cd288a	Parenthesise the reserve() computation in the comparison test clang-tidy reports the mixed * and + as readability-math-missing- parentheses, as it does for the identical line in the copy test. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 12:25:05 +02:00
Niels Lohmann	616b395f71	Note the comparison fallback in the no-thread-local documentation The macro page describes what the library defines JSON_NO_THREAD_LOCAL for by itself in terms of copying alone; comparing falls back the same way. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 12:25:05 +02:00
Niels Lohmann	6392d99a73	Take the descent flag as an argument rather than testing it MSVC reports the test of a constant as C4127 ("conditional expression is constant"), which the Windows builds treat as an error: may_descend is false for operator<, so the operand short-circuits the whole condition. Passing it to compare_descent_exhausted() puts the test where the value is an ordinary parameter, and leaves the call sites with no condition of their own. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 12:25:05 +02:00
Niels Lohmann	0e9754d772	Describe comparison in the no-thread-local docs and CI target Comparing two values now bounds its descent with a thread_local counter just as copying does, so the JSON_NO_THREAD_LOCAL page, the macro overview and the ci_test_no_thread_local target cover both rather than copying alone. Also record what switching the macro on costs a comparison: on the benchmark documents, comparing two equal values takes 10% to 90% longer. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 12:25:05 +02:00
Niels Lohmann	e8b9ec8ff8	Compare values without recursing, and without comparing them twice Comparing two values compared their containers, which compare their elements, which brought the comparison back once per nesting level. Two values nested deeply enough exhausted the call stack and terminated the process with a segmentation fault - the same bug as #5387, in the last operation that still had it. Worse, an ordered comparison took exponentially long in the nesting depth before C++20. std::vector's operator< is a lexicographical comparison, which asks whether an element is less than its counterpart and then whether the counterpart is less than it - two full comparisons of everything below that element, at every level. Comparing two equal values nested 30 levels deep, which is nothing unusual, took 3.8 seconds; 40 levels would have taken an hour, and nothing about the value has to be pathological to get there. C++20 is unaffected: std::lexicographical_compare_three_way asks once. Compare a value that is nested too deeply to descend into on an explicit stack instead, in a single pass that yields less, equal, greater or unordered at once. Equality and the three-way comparison descend as they always did for the first 128 levels, which nothing measurable costs them; an ordered comparison no longer descends at all, which is what takes the exponent out of it. Objects and arrays that are not nested deeply are otherwise compared exactly as before. The results are unchanged for every pair of values: 68121 comparisons of a corpus that covers NaN, discarded values, mixed number types, binary values, empty containers and both object types are identical to develop, in C++11, C++17 and C++20, with and without thread_local storage and legacy discarded comparison. Reproducing that meant reproducing two subtleties: a lexicographic comparison steps over a pair it cannot order, where a three-way comparison stops at it, and an object compares its keys with < where its entries are ordered but with == where they are only checked for equality - not with the object's own comparator, which for nlohmann::ordered_map tells equality. Equality needs no ordering, so it no longer asks for any: a key or string type that can only be compared for equality still works. Measured (medians of 7 interleaved runs, clang -O3, C++11): comparing two equal values nested 30 levels deep 3778 ms -> 0.002 ms; ordering flat objects -33.6%; ordering flat arrays of numbers +27.3%, the one shape that pays for the single pass; equality unchanged throughout. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 12:25:05 +02:00
Niels Lohmann	bc01db0c14	Name the test's locals so Flawfinder stops matching them The code scanning job reports CWE-362 - "check when opening files" - for a test that opens no files: Flawfinder matched a local variable called open. Rename it and its partner. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 12:24:30 +02:00
Niels Lohmann	eae15bbeb1	Check that an abandoned copy can still be destroyed Copying a value without the call stack builds the copy from the top down, and every value whose own copy has not been made yet stays a null value until it is. That is what lets a copy be abandoned half-built: the destructor finds nothing but complete values and null ones. Nothing tested it. Failing an allocation part-way through a copy of a deeply nested value does, with the allocator the file already has for exactly this kind of test. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 12:24:30 +02:00
Niels Lohmann	cb1073469e	Keep the descent bookkeeping in one place Copying carried a depth count, a depth limit and a guard of its own, and the comparison in the follow-up added a second set beside them. Neither operation needs its own: they are never nested inside one another by the library - copying a value does not compare one, and comparing two values does not copy them - and where user code nests them anyway, sharing the count only ends a descent sooner than it had to. So there is now one nesting_depth(), one nesting_depth_limit() and one nesting_depth_guard, which the follow-up uses instead of adding its own. Inverting the test in copy_structured leaves the too-deep case and the no-thread-local case as the same code. The guard takes the count rather than looking it up, because the caller has looked it up already to test it against the limit, and reaching thread-local storage twice on the path that is taken almost every time is worth avoiding. The switch that copies the value of anything that is not an object or an array was written twice - once in the copy constructor, once in copy_shallow - so that adding a value_t meant editing both, and missing one would have been silent. It is copy_leaf_value now, and inlined: both callers have already sorted the containers out, and folding that test into the switch is what keeps a value made mostly of numbers copying as fast as it did. Copying canada.json, citm_catalog.json and twitter.json is within 0.6% of what it was before, measured as a paired ratio over 18 interleaved rounds against a run-to-run spread of 0.3%. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 12:24:30 +02:00
Niels Lohmann	e66a55066e	Include <span> where the split moved its only use The #2546 test case guards itself with __has_include(<span>), but the include itself sat in unit-regression2.cpp's preamble and stayed behind, so the section compiled without a declaration wherever the guard passed - which nvhpc reported and libc++ builds do not, as they skip the section altogether. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 03:19:53 +02:00
Niels Lohmann	02013c7f0d	Move the #4804 alias to the file that uses it The split left the json_4804 alias behind in unit-regression2.cpp while the test case that uses it went to unit-regression3.cpp, which does not build for C++17 and C++20 as a result. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 03:00:18 +02:00
Niels Lohmann	cdf24bccde	Split the regression tests far enough to leave room The first split left unit-regression2.cpp 0.7% below the size develop links at, which the comparison change in the follow-up immediately used up: the MinGW linker fails on test-regression2_cpp20 again, naming copy_shallow and to_partial_ordering among the relocations it cannot fit. Move the sections from "issue #2067" on, and the helper types they use, so that the file stops being the one that decides whether the tests can be linked at all. At -O0 and C++20, unit-regression2.cpp is now 2,964,944 bytes against develop's 4,708,248, and 3,070,568 bytes with the follow-up applied - roughly a third smaller either way, rather than a fraction of a percent larger. The 135 assertions are the same ones as before, now spread over three test cases in two files. Also silence the clang-tidy findings the deep-nesting tests draw: the copies they make are what is being tested, and the reserve() computation gets its parentheses. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 02:53:10 +02:00
Niels Lohmann	9d75c87de5	Check both shapes without a C-style array clang-tidy rejects the array the two shapes were iterated over (cppcoreguidelines-avoid-c-arrays). The array only existed because astyle reformats a range-for over a braced initializer list into something unreadable; naming the two cases avoids both. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 01:41:05 +02:00
Niels Lohmann	4fd940f91b	Balance the warning suppression the split separated unit-regression2.cpp opens a DOCTEST_CLANG_SUPPRESS_WARNING_PUSH block at the top and closed it at the very bottom, which the split moved into unit-regression3.cpp: one file was left with a push and no pop, the other with a pop and no push, which clang reports as an error. Give each file the pair it needs. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 01:01:57 +02:00
Niels Lohmann	540f6d11cc	Do not use thread_local storage with Clang targeting MinGW Every test that copies a value segfaults there - 42 of 105 on clang 11.0.1, 39 of 102 on clang 18.1.8 - while the same tests pass with GCC targeting MinGW, with Clang targeting MSVC, and with every other toolchain the library is tested on. The counter that bounds the copy constructor's descent is the library's first use of thread_local, so that job had never exercised it before. JSON_NO_THREAD_LOCAL already covers toolchains without thread_local storage, and copying yields the same values with it, only more slowly. Define it for this one automatically. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 00:48:42 +02:00
Niels Lohmann	6305cc9f72	Split the regression tests so that they keep linking Linking test-regression2 fails with "relocation truncated to fit: IMAGE_REL_AMD64_REL32 against `.rdata'" once its object grows past what the MinGW linker copes with, and the copy constructor's helpers push it over: the object grows by 6.3%, from 4,654,128 to 4,944,920 bytes at -O0, and develop links at the smaller of the two. Building the tests optimized shrinks the object enough to link, but the binaries clang 11.0.1 and clang 18.1.8 then produce crash before doctest prints its first line - 39 of 102 tests on clang 18 - so the objects have to become smaller rather than denser. Moving the test cases that follow "regression tests 2" into a file of their own brings that object to 4,687,888 bytes, which is 0.7% above the size that links today rather than 6.3%. Both files still build for C++11, C++17 and C++20, and run the same 9 test cases and 135 assertions as before, now spread over two binaries. New regression tests belong in unit-regression3.cpp from here on, which is what CONTRIBUTING.md now says. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 00:36:05 +02:00
Niels Lohmann	fa9b76283a	Test the copy constructor's iterative path in CI The copy constructor descends into 128 levels before it finishes a value without the call stack, so the iterative path is otherwise only reached by the few tests that nest deeper than that. JSON_NO_THREAD_LOCAL switches the descent off, which sends every value down that path. Running the whole test suite that way covers it with every object type, string type, allocator, and base class the suite already exercises. The new ci_test_no_thread_local target does that; the macro had no build coverage at all before. Copying a nested value also has to carry over what the element-wise copy constructor would have copied: the parents that JSON_DIAGNOSTICS relies on, and the positions that JSON_DIAGNOSTIC_POSITIONS reports. Both are now checked on either side of the descent bound, for objects and arrays. Neither was tested before, and dropping either one makes the new tests fail. Also quantify what JSON_NO_THREAD_LOCAL costs a copy instead of calling it "measurably slower". Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-21 00:35:45 +02:00
Niels Lohmann	e486005583	Bound the descent of the copy constructor basic_json's copy constructor copied objects and arrays by handing the container to its own copy constructor, which copy-constructs every element and so reaches this constructor again, once per nesting level. A value nested deeply enough exhausted the call stack and terminated the process with a segmentation fault - no exception, nothing the caller could catch. Parsing such a value works, as the parser is iterative, and so does destroying one, as #1436 made destruction iterative. Bound how far the copy descends rather than take the call stack away from it. The first levels are copied exactly as they were - the containers copy their own elements, which is by far the fastest way to fill them - and only once the copy has descended 128 levels is the value below it finished without the call stack, through an explicit worklist. Copying can therefore no longer exhaust the stack, however deeply a value is nested, while a value nested less deeply than the bound - all but a vanishing minority - is copied by the very same code as before and pays only for one counter. That counter lives in thread_local storage, as one shared between threads would be raced. JSON_NO_THREAD_LOCAL switches it off for toolchains without thread_local; copying then goes through the worklist right away, which yields the same values but is measurably slower. The deferred values are completed before the copy they belong to returns, so a value copied while another copy is going on - by a custom base class, say - is unaffected by the copy it is nested in. operator= takes its argument by value, so copy assignment is fixed as well. Copying is as fast as it was, within measurement noise (medians of 9 interleaved runs, clang -O3): -1.3% for an array of strings, +0.0% for a flat object, +0.1% for a flat array of numbers, +0.3% for nested arrays, +0.6% for nested objects and +1.2% for a twitter-like document. Copying a three-key object costs about ten nanoseconds more, the counter. Deferring every level instead, rather than only those below the bound, measured between 3% and 9% slower depending on the shape of the value. This fixes #5387 for the copy constructor. dump() is still recursive. Signed-off-by: Niels Lohmann <mail@nlohmann.me>	2026-08-20 19:22:16 +02:00