Restore a duplicate key's prior value when the callback rejects its new value

json_sax_dom_callback_parser::key() unconditionally overwrote the object
slot for a key with a `discarded` placeholder as soon as the key was
accepted by the parser callback. For a duplicate key (legal JSON), this
destroyed the pre-existing value from an earlier occurrence of the same
key before the new value was even parsed. If the new value was then
rejected by the callback, remove_discarded_value() erased the member
entirely instead of leaving the original value in place, contradicting
the documented behavior that a discarded value behaves as if it was
never read.

Add a small stash of (slot pointer, previous value) pairs so that when
key() overwrites an existing member with the discarded placeholder, the
previous value can be restored later if the corresponding value (scalar,
object, or array) is rejected, instead of being erased. The stash entry
is dropped without restoring once the new value is definitively
accepted (in handle_value() for scalars, end_object()/end_array() for
containers), so a duplicate key whose new value is accepted still keeps
the last value as before. Non-duplicate keys are unaffected: rejecting
their value still removes the member entirely, since there is nothing
to restore.

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

include/nlohmann/detail/input/json_sax.hpp

@@ -8,10 +8,11 @@
 
 #pragma once
 
+#include <algorithm> // find_if
 #include <cstddef>
 #include <string> // string
 #include <type_traits> // enable_if_t
-#include <utility> // move
+#include <utility> // move, pair
 #include <vector> // vector
 
 #include <nlohmann/detail/exceptions.hpp>
@@ -601,7 +602,17 @@ class json_sax_dom_callback_parser
         // add discarded value at the given key and store the reference for later
         if (keep && ref_stack.back())
         {
-            object_element = &(ref_stack.back()->m_data.m_value.object->operator[](val) = discarded);
+            auto& obj = *ref_stack.back()->m_data.m_value.object;
+            const auto it = obj.find(val);
+            if (it != obj.end())
+            {
+                // this is a duplicate key (legal in JSON); remember its
+                // current value so it can be restored later if the new
+                // value is rejected by the callback, instead of being
+                // erased together with the discarded placeholder
+                duplicate_key_stash.emplace_back(&(it->second), it->second);
+            }
+            object_element = &(obj[val] = discarded);
         }
 
         return true;
@@ -613,7 +624,11 @@ class json_sax_dom_callback_parser
         {
             if (!callback(static_cast<int>(ref_stack.size()) - 1, parse_event_t::object_end, *ref_stack.back()))
             {
-                // discard object
+                // discard object, unless this slot holds a duplicate key's
+                // previous value pending restoration, in which case that
+                // value is restored instead of being discarded
+                if (!resolve_duplicate_key_stash(ref_stack.back(), true))
+                {
                     *ref_stack.back() = discarded;
 
 #if JSON_DIAGNOSTIC_POSITIONS
@@ -621,6 +636,7 @@ class json_sax_dom_callback_parser
                     handle_diagnostic_positions_for_json_value(*ref_stack.back());
 #endif
                 }
+            }
             else
             {
 
@@ -633,6 +649,10 @@ class json_sax_dom_callback_parser
 #endif
 
                 ref_stack.back()->set_parents();
+                // this object is finally, definitively kept; drop any
+                // pending duplicate-key stash entry for its slot since it
+                // can no longer be restored
+                resolve_duplicate_key_stash(ref_stack.back(), false);
             }
         }
 
@@ -708,10 +728,18 @@ class json_sax_dom_callback_parser
 #endif
 
                 ref_stack.back()->set_parents();
+                // this array is finally, definitively kept; drop any
+                // pending duplicate-key stash entry for its slot since it
+                // can no longer be restored
+                resolve_duplicate_key_stash(ref_stack.back(), false);
             }
             else
             {
-                // discard array
+                // discard array, unless this slot holds a duplicate key's
+                // previous value pending restoration, in which case that
+                // value is restored instead of being discarded
+                if (!resolve_duplicate_key_stash(ref_stack.back(), true))
+                {
                     *ref_stack.back() = discarded;
 
 #if JSON_DIAGNOSTIC_POSITIONS
@@ -720,6 +748,7 @@ class json_sax_dom_callback_parser
 #endif
                 }
             }
+        }
 
         JSON_ASSERT(!ref_stack.empty());
         JSON_ASSERT(!keep_stack.empty());
@@ -851,6 +880,35 @@ class json_sax_dom_callback_parser
         return string_t{};
     }
 
+    /// if there is a pending duplicate-key stash entry for this exact slot,
+    /// remove it from the stash; if restore_value is true, the stashed
+    /// previous value is moved back into the slot first (use this when the
+    /// new value at that slot was rejected); otherwise the stash entry is
+    /// simply dropped (use this when the new value was accepted, so it
+    /// correctly supersedes the old one and no restore should ever happen
+    /// for this slot again)
+    /// @return whether a matching stash entry was found (and processed)
+    bool resolve_duplicate_key_stash(BasicJsonType* slot, bool restore_value)
+    {
+        const auto it = std::find_if(duplicate_key_stash.begin(), duplicate_key_stash.end(),
+                                     [slot](const std::pair<BasicJsonType*, BasicJsonType>& entry)
+        {
+            return entry.first == slot;
+        });
+
+        if (it == duplicate_key_stash.end())
+        {
+            return false;
+        }
+
+        if (restore_value)
+        {
+            *slot = std::move(it->second);
+        }
+        duplicate_key_stash.erase(it);
+        return true;
+    }
+
     /*!
     @brief remove the discarded value the callback rejected from its parent
 
@@ -862,12 +920,14 @@ class json_sax_dom_callback_parser
 
     Finding no discarded value there means none was stored in the first place -
     the callback rejected the value before it reached its parent - so there is
-    nothing to remove.
+    nothing to remove. If the discarded slot instead holds a duplicate key's
+    stashed previous value pending restoration, that value is restored instead
+    of the slot being erased.
 
     @param[in,out] parent  the container to remove the rejected value from
     @param[in] key  the key the value was stored under; unused for arrays
     */
-    static void remove_discarded_value(BasicJsonType& parent, const string_t& key)
+    void remove_discarded_value(BasicJsonType& parent, const string_t& key)
     {
         if (parent.is_array())
         {
@@ -882,11 +942,14 @@ class json_sax_dom_callback_parser
             auto& object = *parent.m_data.m_value.object;
             const auto it = object.find(key);
             if (it != object.end() && it->second.is_discarded())
+            {
+                if (!resolve_duplicate_key_stash(&(it->second), true))
                 {
                     object.erase(it);
                 }
             }
         }
+    }
 
     /*!
     @param[in] v  value to add to the JSON value we build during parsing
@@ -985,6 +1048,16 @@ class json_sax_dom_callback_parser
 
         JSON_ASSERT(object_element);
         *object_element = std::move(value);
+        if (!skip_callback)
+        {
+            // this scalar value finally, definitively replaces whatever was
+            // at this slot; drop any pending duplicate-key stash entry for
+            // it since it can no longer be restored (a container value at
+            // this slot is resolved later, in end_object()/end_array(),
+            // since skip_callback is true for the placeholder handling that
+            // happens here for those)
+            resolve_duplicate_key_stash(object_element, false);
+        }
         return {true, object_element};
     }
 
@@ -1004,6 +1077,12 @@ class json_sax_dom_callback_parser
     std::vector<string_t> container_key_stack {}; // NOLINT(readability-redundant-member-init)
     /// helper to hold the reference for the next object element
     BasicJsonType* object_element = nullptr;
+    /// stash of (slot pointer, previous value) for object members that
+    /// already existed when key() was called again for the same key
+    /// (duplicate keys); used to restore the previous value if the new
+    /// value is later rejected by the callback, instead of erasing the
+    /// member entirely
+    std::vector<std::pair<BasicJsonType*, BasicJsonType>> duplicate_key_stash {};
     /// whether a syntax error occurred
     bool errored = false;
     /// callback function

include/nlohmann/detail/output/output_adapters.hpp

@@ -13,7 +13,6 @@
 #include <iterator> // back_inserter
 #include <memory> // shared_ptr, make_shared
 #include <string> // basic_string
-#include <utility> // move
 #include <vector> // vector
 
 #ifndef JSON_NO_IO
@@ -45,32 +44,22 @@ template<typename CharType> struct output_adapter_protocol
 template<typename CharType>
 using output_adapter_t = std::shared_ptr<output_adapter_protocol<CharType>>;
 
-/// @brief non-virtual output sink writing into a std::vector
-///
-/// This sink is not part of the virtual output_adapter_protocol hierarchy: it is
-/// passed to binary_writer by value as a template parameter, so
-/// write_character()/write_characters() are ordinary (inlinable) calls with no
-/// vtable lookup and no shared_ptr. It is used for the common
-/// `to_cbor`/`to_msgpack`/... into a std::vector. output_vector_adapter below
-/// wraps this same sink to provide the virtual interface.
+/// output adapter for byte vectors
 template<typename CharType, typename AllocatorType = std::allocator<CharType>>
-class output_vector_sink
+class output_vector_adapter : public output_adapter_protocol<CharType>
 {
   public:
-    explicit output_vector_sink(std::vector<CharType, AllocatorType>& vec) noexcept
+    explicit output_vector_adapter(std::vector<CharType, AllocatorType>& vec) noexcept
         : v(vec)
     {}
 
-    void write_character(CharType c)
+    void write_character(CharType c) override
     {
         v.push_back(c);
     }
 
-    // no JSON_HEDLEY_NON_NULL here: binary_writer legitimately passes a null
-    // pointer with length 0 for empty strings/binary values. Appending an empty
-    // range is a no-op; the type-erased path tolerates this via the (unattributed)
-    // virtual base, and the concrete sink must do the same.
-    void write_characters(const CharType* s, std::size_t length)
+    JSON_HEDLEY_NON_NULL(2)
+    void write_characters(const CharType* s, std::size_t length) override
     {
         v.insert(v.end(), s, s + length);
     }
@@ -79,34 +68,6 @@ class output_vector_sink
     std::vector<CharType, AllocatorType>& v;
 };
 
-/// output adapter for byte vectors
-///
-/// The appending itself lives in output_vector_sink; this class only adds the
-/// virtual output_adapter_protocol interface on top of it, so both the
-/// type-erased and the templated path share one implementation.
-template<typename CharType, typename AllocatorType = std::allocator<CharType>>
-class output_vector_adapter : public output_adapter_protocol<CharType>
-{
-  public:
-    explicit output_vector_adapter(std::vector<CharType, AllocatorType>& vec) noexcept
-        : sink(vec)
-    {}
-
-    void write_character(CharType c) override
-    {
-        sink.write_character(c);
-    }
-
-    JSON_HEDLEY_NON_NULL(2)
-    void write_characters(const CharType* s, std::size_t length) override
-    {
-        sink.write_characters(s, length);
-    }
-
-  private:
-    output_vector_sink<CharType, AllocatorType> sink;
-};
-
 #ifndef JSON_NO_IO
 /// output adapter for output streams
 template<typename CharType>
@@ -157,39 +118,6 @@ class output_string_adapter : public output_adapter_protocol<CharType>
     StringType& str;
 };
 
-/// @brief output sink forwarding to a type-erased output adapter
-///
-/// Wraps the polymorphic output_adapter_t so the same binary_writer template can
-/// also target arbitrary adapters (output streams, strings, user-provided
-/// adapters) via the `output_adapter`-based overloads. Each write still goes
-/// through one virtual call, exactly as before; only the concrete sinks above
-/// avoid it.
-template<typename CharType>
-class output_adapter_sink
-{
-  public:
-    explicit output_adapter_sink(output_adapter_t<CharType> adapter)
-        : oa(std::move(adapter))
-    {
-        JSON_ASSERT(oa);
-    }
-
-    void write_character(CharType c)
-    {
-        oa->write_character(c);
-    }
-
-    // no JSON_HEDLEY_NON_NULL: forwards (null, 0) for empty payloads, exactly as
-    // the type-erased path already did before this sink existed
-    void write_characters(const CharType* s, std::size_t length)
-    {
-        oa->write_characters(s, length);
-    }
-
-  private:
-    output_adapter_t<CharType> oa;
-};
-
 template<typename CharType, typename StringType = std::basic_string<CharType>>
 class output_adapter
 {

include/nlohmann/json.hpp

@@ -140,7 +140,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     friend ::nlohmann::detail::serializer<basic_json>;
     template<typename BasicJsonType>
     friend class ::nlohmann::detail::iter_impl;
-    template<typename BasicJsonType, typename CharType, typename OutputSinkType>
+    template<typename BasicJsonType, typename CharType>
     friend class ::nlohmann::detail::binary_writer;
     template<typename BasicJsonType, typename InputType, typename SAX>
     friend class ::nlohmann::detail::binary_reader;
@@ -188,14 +188,6 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     template<typename InputType>
     using binary_reader = ::nlohmann::detail::binary_reader<basic_json, InputType>;
     template<typename CharType> using binary_writer = ::nlohmann::detail::binary_writer<basic_json, CharType>;
-    // binary_writer over a concrete (non-virtual) sink appending into a std::vector,
-    // used by the vector-returning to_* overloads
-    template<typename CharType> using vector_binary_writer =
-        ::nlohmann::detail::binary_writer<basic_json, CharType, ::nlohmann::detail::output_vector_sink<CharType>>;
-    template<typename CharType> static vector_binary_writer<CharType> vector_writer(std::vector<CharType>& v)
-    {
-        return vector_binary_writer<CharType>(::nlohmann::detail::output_vector_sink<CharType>(v));
-    }
 
   JSON_PRIVATE_UNLESS_TESTED:
     using serializer = ::nlohmann::detail::serializer<basic_json>;
@@ -3610,6 +3602,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
         {
             using std::swap;
             swap(*(m_data.m_value.array), other);
+            set_parents();
         }
         else
         {
@@ -3626,6 +3619,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
         {
             using std::swap;
             swap(*(m_data.m_value.object), other);
+            set_parents();
         }
         else
         {
@@ -4382,8 +4376,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     static std::vector<std::uint8_t> to_cbor(const basic_json& j)
     {
         std::vector<std::uint8_t> result;
-        result.reserve(detail::binary_reserve_hint(j));
-        vector_writer(result).write_cbor(j);
+        to_cbor(j, result);
         return result;
     }
 
@@ -4406,8 +4399,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     static std::vector<std::uint8_t> to_msgpack(const basic_json& j)
     {
         std::vector<std::uint8_t> result;
-        result.reserve(detail::binary_reserve_hint(j));
-        vector_writer(result).write_msgpack(j);
+        to_msgpack(j, result);
         return result;
     }
 
@@ -4432,8 +4424,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
             const bool use_type = false)
     {
         std::vector<std::uint8_t> result;
-        result.reserve(detail::binary_reserve_hint(j));
-        vector_writer(result).write_ubjson(j, use_size, use_type);
+        to_ubjson(j, result, use_size, use_type);
         return result;
     }
 
@@ -4461,8 +4452,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
             const bjdata_version_t version = bjdata_version_t::draft2)
     {
         std::vector<std::uint8_t> result;
-        result.reserve(detail::binary_reserve_hint(j));
-        vector_writer(result).write_ubjson(j, use_size, use_type, true, true, version);
+        to_bjdata(j, result, use_size, use_type, version);
         return result;
     }
 
@@ -4489,8 +4479,7 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
     static std::vector<std::uint8_t> to_bson(const basic_json& j)
     {
         std::vector<std::uint8_t> result;
-        result.reserve(detail::binary_reserve_hint(j));
-        vector_writer(result).write_bson(j);
+        to_bson(j, result);
         return result;
     }
 
@@ -5281,21 +5270,96 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
 
             case value_t::object:
             {
-                // first pass: traverse this object's elements
+                // first pass: record, for every source key, whether it is
+                // common to both objects (in source's iteration order) or
+                // was deleted (i.e., in source but not in target) -- this is
+                // a by-product of the target.find() call already needed to
+                // tell the two cases apart, so it adds no extra lookups. The
+                // "remove" ops themselves are emitted later, interleaved
+                // with the recursive per-key diffs in the fast path below,
+                // to match source's original iteration order (as the
+                // original, pre-reordering-aware implementation did) instead
+                // of grouping all removes before all recursive diffs.
+                std::vector<typename object_t::key_type> common_keys_source_order;
                 for (auto it = source.cbegin(); it != source.cend(); ++it)
                 {
-                    // escape the key name to be used in a JSON patch
-                    const auto path_key = detail::concat<string_t>(path, '/', detail::escape(it.key()));
-
                     if (target.find(it.key()) != target.end())
                     {
-                        // recursive call to compare object values at key it
-                        auto temp_diff = diff(it.value(), target[it.key()], path_key);
-                        result.insert(result.end(), temp_diff.begin(), temp_diff.end());
+                        common_keys_source_order.push_back(it.key());
+                    }
+                }
+
+                // second pass: find keys that were added (i.e., in target but
+                // not in source), and record the keys common to both, in
+                // target's iteration order -- again a by-product of the
+                // source.find() call already needed to detect added keys. At
+                // the same time, determine whether every added key comes
+                // after every common key in target's order (a precondition
+                // for the fast path below, which only ever appends new keys
+                // at the very end): for an object_t whose iteration order is
+                // a pure function of the key set (e.g. the default std::map,
+                // which always iterates in sorted key order), the order
+                // check further below is always true and this whole
+                // mechanism is effectively a no-op; it only matters for a
+                // reorderable object_t such as the one backing `ordered_json`.
+                // patch ops for keys that were added (i.e., in target but not
+                // in source); built here so the fast path below can reuse
+                // them without a second source.find() per target key. Only
+                // used by the fast path -- the slow (reordering) path
+                // rebuilds "add" ops for every key itself.
+                std::vector<typename object_t::key_type> common_keys_target_order;
+                basic_json added_ops(value_t::array);
+                bool new_keys_form_suffix = true;
+                bool seen_new_key = false;
+                for (auto it = target.cbegin(); it != target.cend(); ++it)
+                {
+                    if (source.find(it.key()) == source.end())
+                    {
+                        seen_new_key = true;
+                        const auto path_key = detail::concat<string_t>(path, '/', detail::escape(it.key()));
+                        added_ops.push_back(
+                        {
+                            {"op", "add"}, {"path", path_key},
+                            {"value", it.value()}
+                        });
                     }
                     else
                     {
-                        // found a key that is not in o -> remove it
+                        common_keys_target_order.push_back(it.key());
+                        if (seen_new_key)
+                        {
+                            new_keys_form_suffix = false;
+                        }
+                    }
+                }
+
+                if (common_keys_source_order == common_keys_target_order && new_keys_form_suffix)
+                {
+                    // fast path: order of common keys already matches (or the
+                    // object_t's iteration order does not depend on
+                    // insertion history), so a plain per-key recursive diff
+                    // is correct and minimal, as before. common_keys_source_order
+                    // is, by construction, the subsequence of source's keys
+                    // that are common to both objects, in source's iteration
+                    // order -- so it can be walked in lockstep with `source`
+                    // using a cheap key comparison instead of another lookup.
+                    // Deleted keys (those source keys not in common_keys_source_order)
+                    // are interleaved here too, in source's original order, to
+                    // match the historical (pre-reordering-aware) output order.
+                    auto common_it = common_keys_source_order.cbegin();
+                    for (auto it = source.cbegin(); it != source.cend(); ++it)
+                    {
+                        if (common_it != common_keys_source_order.cend() && it.key() == *common_it)
+                        {
+                            const auto path_key = detail::concat<string_t>(path, '/', detail::escape(it.key()));
+                            auto temp_diff = diff(it.value(), target[it.key()], path_key);
+                            result.insert(result.end(), temp_diff.begin(), temp_diff.end());
+                            ++common_it;
+                        }
+                        else
+                        {
+                            // found a key that is not in target -> remove it
+                            const auto path_key = detail::concat<string_t>(path, '/', detail::escape(it.key()));
                             result.push_back(object(
                             {
                                 {"op", "remove"}, {"path", path_key}
@@ -5303,12 +5367,42 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
                         }
                     }
 
-                // second pass: traverse other object's elements
+                    // append the "add" ops for brand-new keys collected above
+                    // during the pass over target -- no second source.find()
+                    // per target key needed
+                    result.insert(result.end(), added_ops.begin(), added_ops.end());
+                }
+                else
+                {
+                    // slow path: the common keys are in a different relative
+                    // order in source and target (only possible for a
+                    // reorderable object_t like ordered_map). Building a
+                    // minimal reordering patch is a nontrivial (LCS-like)
+                    // problem; instead, remove every source key -- both
+                    // deleted keys (which must be removed regardless) and
+                    // common keys (removed so they can be re-added in
+                    // target's order) -- and re-add every key that should
+                    // remain, with its final target value, in target's
+                    // order. basic_json::patch()'s "add" operation on an
+                    // object uses operator[], which appends at the end for a
+                    // vector-backed insertion-ordered map when the key does
+                    // not already exist -- so removing a key and then adding
+                    // it moves it to the end, fixing its position.
+                    for (auto it = source.cbegin(); it != source.cend(); ++it)
+                    {
+                        const auto path_key = detail::concat<string_t>(path, '/', detail::escape(it.key()));
+                        result.push_back(object(
+                        {
+                            {"op", "remove"}, {"path", path_key}
+                        }));
+                    }
+
+                    // add every key that is either common (just removed
+                    // above) or brand new, in target's iteration order, so
+                    // that the final order after applying the patch matches
+                    // target exactly
                     for (auto it = target.cbegin(); it != target.cend(); ++it)
                     {
-                    if (source.find(it.key()) == source.end())
-                    {
-                        // found a key that is not in this -> add it
                         const auto path_key = detail::concat<string_t>(path, '/', detail::escape(it.key()));
                         result.push_back(
                         {

tests/src/unit-binary_writer_sinks.cpp

@@ -1,198 +0,0 @@
-//     __ _____ _____ _____
-//  __|  |   __|     |   | |  JSON for Modern C++ (supporting code)
-// |  |  |__   |  |  | | | |  version 3.12.0
-// |_____|_____|_____|_|___|  https://github.com/nlohmann/json
-//
-// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
-// SPDX-License-Identifier: MIT
-
-#include "doctest_compatibility.h"
-
-#include <nlohmann/json.hpp>
-using nlohmann::json;
-
-#include <cstdint>
-#include <string>
-#include <vector>
-
-namespace
-{
-
-// a spread of values exercising every writer path: scalars of each width, the
-// float paths, strings, binary, and containers big enough to reallocate
-std::vector<json> test_values()
-{
-    json big_array = json::array();
-    for (int i = 0; i < 5000; ++i)
-    {
-        big_array.push_back(i);
-    }
-
-    json big_object = json::object();
-    for (int i = 0; i < 1000; ++i)
-    {
-        big_object[std::to_string(i)] = i;
-    }
-
-    return
-    {
-        json(nullptr), json(true), json(false),
-        json(0), json(-1), json(255), json(-129), json(65535), json(-32769),
-        json(4294967295U), json(-2147483649LL), json(18446744073709551615ULL),
-        json(0.0), json(-0.5), json(3.1415926535897932),
-        json(""), json("hello"), json(std::string(1000, 'x')),
-        json::binary({0x00, 0x01, 0x02}, 42),
-        json::array(), json::object(),
-        json::array({1, 2, 3}), json({{"a", 1}, {"b", nullptr}}),
-        json({{"nested", {{"deep", json::array({1, "two", 3.0, nullptr})}}}}),
-        big_array, big_object
-    };
-}
-
-// values to_bson() accepts: the document must be an object
-std::vector<json> bson_values()
-{
-    json big_object = json::object();
-    for (int i = 0; i < 1000; ++i)
-    {
-        big_object[std::to_string(i)] = i;
-    }
-
-    return
-    {
-        json::object(),
-        json({{"a", 1}, {"b", nullptr}, {"c", true}, {"d", 2.5}, {"e", "text"}}),
-        json({{"arr", json::array({1, 2, 3})}, {"obj", {{"k", "v"}}}}),
-        big_object
-    };
-}
-
-} // namespace
-
-// The vector-returning to_*(j) overloads write through the non-virtual
-// output_vector_sink, while to_*(j, adapter) goes through output_adapter_sink.
-// The two are separate code paths that must stay byte-for-byte identical; these
-// checks fail if either overload is ever changed without the other.
-TEST_CASE("binary writer output sinks")
-{
-    SECTION("vector sink and adapter sink agree")
-    {
-        // note: no SUBCASE inside these loops - doctest keys subcases by
-        // name/file/line, so a subcase in a loop body would only ever run for
-        // the first iteration
-        for (const auto& j : test_values())
-        {
-            CAPTURE(j.dump(-1, ' ', false, json::error_handler_t::replace));
-
-            std::vector<std::uint8_t> cbor;
-            json::to_cbor(j, cbor);
-            CHECK(json::to_cbor(j) == cbor);
-
-            std::vector<std::uint8_t> msgpack;
-            json::to_msgpack(j, msgpack);
-            CHECK(json::to_msgpack(j) == msgpack);
-
-            for (const bool use_size :
-                    {
-                        false, true
-                    })
-            {
-                for (const bool use_type :
-                        {
-                            false, true
-                        })
-                {
-                    if (use_type && !use_size)
-                    {
-                        continue; // not a supported combination
-                    }
-                    CAPTURE(use_size);
-                    CAPTURE(use_type);
-                    std::vector<std::uint8_t> ubjson;
-                    json::to_ubjson(j, ubjson, use_size, use_type);
-                    CHECK(json::to_ubjson(j, use_size, use_type) == ubjson);
-                }
-            }
-
-            for (const auto version :
-                    {
-                        json::bjdata_version_t::draft2, json::bjdata_version_t::draft3
-                    })
-            {
-                std::vector<std::uint8_t> bjdata;
-                json::to_bjdata(j, bjdata, false, false, version);
-                CHECK(json::to_bjdata(j, false, false, version) == bjdata);
-            }
-        }
-
-        for (const auto& j : bson_values())
-        {
-            CAPTURE(j.dump());
-            std::vector<std::uint8_t> bson;
-            json::to_bson(j, bson);
-            CHECK(json::to_bson(j) == bson);
-        }
-    }
-
-    SECTION("the char adapter produces the same bytes")
-    {
-        for (const auto& j : test_values())
-        {
-            CAPTURE(j.dump(-1, ' ', false, json::error_handler_t::replace));
-
-            const std::vector<std::uint8_t> expected = json::to_cbor(j);
-            std::vector<char> as_char;
-            json::to_cbor(j, as_char);
-
-            REQUIRE(as_char.size() == expected.size());
-            std::vector<std::uint8_t> as_bytes;
-            as_bytes.reserve(as_char.size());
-            for (const char c : as_char)
-            {
-                as_bytes.push_back(static_cast<std::uint8_t>(c));
-            }
-            CHECK(as_bytes == expected);
-        }
-    }
-}
-
-// binary_reserve_hint() is documented as a *lower* bound on the serialized size,
-// so that reserving it up front can never leave the returned vector holding
-// capacity beyond what the value actually needs.
-TEST_CASE("binary_reserve_hint never over-reserves")
-{
-    for (const auto& j : test_values())
-    {
-        CAPTURE(j.dump(-1, ' ', false, json::error_handler_t::replace));
-
-        const std::size_t hint = nlohmann::detail::binary_reserve_hint(j);
-
-        CHECK(hint <= json::to_cbor(j).size());
-        CHECK(hint <= json::to_msgpack(j).size());
-        CHECK(hint <= json::to_ubjson(j).size());
-        CHECK(hint <= json::to_ubjson(j, true, true).size());
-        CHECK(hint <= json::to_bjdata(j).size());
-    }
-
-    for (const auto& j : bson_values())
-    {
-        CAPTURE(j.dump());
-        CHECK(nlohmann::detail::binary_reserve_hint(j) <= json::to_bson(j).size());
-    }
-
-    SECTION("scalars get no hint")
-    {
-        CHECK(nlohmann::detail::binary_reserve_hint(json(nullptr)) == 0);
-        CHECK(nlohmann::detail::binary_reserve_hint(json(42)) == 0);
-        CHECK(nlohmann::detail::binary_reserve_hint(json("a string")) == 0);
-        CHECK(nlohmann::detail::binary_reserve_hint(json::binary({0x01})) == 0);
-    }
-
-    SECTION("containers are hinted from their element count")
-    {
-        CHECK(nlohmann::detail::binary_reserve_hint(json::array()) == 1);
-        CHECK(nlohmann::detail::binary_reserve_hint(json::array({1, 2, 3})) == 4);
-        CHECK(nlohmann::detail::binary_reserve_hint(json::object()) == 1);
-        CHECK(nlohmann::detail::binary_reserve_hint(json({{"a", 1}, {"b", 2}})) == 5);
-    }
-}

tests/src/unit-diagnostics.cpp

@@ -273,5 +273,36 @@ TEST_CASE("Regression tests for extended diagnostics")
         CHECK(j1["numbers"]["two"] == 2);
         CHECK(j1["string"] == "t");
     }
+
+    SECTION("Regression test - swap(array_t&)/swap(object_t&) must update JSON_DIAGNOSTICS parent pointers")
+    {
+        // swap(array_t&)
+        {
+            json j = json::array();
+            json::array_t arr = {json::array({1})};
+            j.swap(arr);
+
+            // parent pointers of the moved-in elements must point into j, not
+            // into the now-defunct free-standing array_t
+            CHECK_THROWS_WITH_AS(j[0][0].get<std::string>(), "[json.exception.type_error.302] (/0/0) type must be string, but is number", json::type_error);
+
+            // must not trigger assert_invariant() in a debug/assert-enabled build
+            json const k = j;
+            CHECK(k == j);
+        }
+
+        // swap(object_t&)
+        {
+            json o = json::object();
+            json::object_t obj = {{"a", json::array({1})}};
+            o.swap(obj);
+
+            CHECK_THROWS_WITH_AS(o["a"][0].get<std::string>(), "[json.exception.type_error.302] (/a/0) type must be string, but is number", json::type_error);
+
+            // must not trigger assert_invariant() in a debug/assert-enabled build
+            json const p = o;
+            CHECK(p == o);
+        }
+    }
 }
 

tests/src/unit-ordered_json.cpp

@@ -81,3 +81,84 @@ TEST_CASE("regression test for issue #3732 - iteration_proxy_value<iter_impl<ord
     };
     static_cast<void>(fn);
 }
+
+TEST_CASE("regression test - diff() must account for ordered_json member order")
+{
+    SECTION("pure reorder, no value changes")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}};
+        ordered_json b = {{"b", 2}, {"a", 1}};
+        CHECK(a != b); // order-sensitive equality
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("new key must land at the front")
+    {
+        ordered_json c = {{"b", 2}};
+        ordered_json e = {{"a", 1}, {"b", 2}};
+        CHECK(c.patch(ordered_json::diff(c, e)) == e);
+    }
+
+    SECTION("reorder plus a value change on one of the reordered keys")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}};
+        ordered_json b = {{"b", 20}, {"a", 1}};
+        CHECK(a != b);
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("reorder plus a deleted key")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}, {"c", 3}};
+        ordered_json b = {{"b", 2}, {"a", 1}};
+        CHECK(a != b);
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("reorder plus a nested value that itself needs a recursive diff")
+    {
+        ordered_json a = {{"a", {{"x", 1}, {"y", 2}}}, {"b", 2}};
+        ordered_json b = {{"b", 2}, {"a", {{"x", 1}, {"y", 99}}}};
+        CHECK(a != b);
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("three or more keys shuffled into a different order")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}, {"c", 3}, {"d", 4}};
+        ordered_json b = {{"d", 4}, {"b", 2}, {"a", 1}, {"c", 3}};
+        CHECK(a != b);
+        CHECK(a.patch(ordered_json::diff(a, b)) == b);
+    }
+
+    SECTION("matching order still produces a minimal patch (fast path unaffected)")
+    {
+        ordered_json a = {{"a", 1}, {"b", 2}, {"c", 3}};
+        ordered_json b = {{"a", 1}, {"b", 20}, {"c", 3}};
+        auto p = ordered_json::diff(a, b);
+        // only the changed value should be touched, not a wholesale remove+add
+        CHECK(p.size() == 1);
+        CHECK(p[0]["op"] == "replace");
+        CHECK(p[0]["path"] == "/b");
+        CHECK(a.patch(p) == b);
+    }
+
+    SECTION("plain json (std::map-backed) is unaffected by same-key-different-insertion-order")
+    {
+        json a;
+        a["b"] = 2;
+        a["a"] = 1;
+
+        json b;
+        b["a"] = 1;
+        b["b"] = 2;
+
+        // std::map iteration is always sorted by key, so a == b regardless of
+        // insertion order, and diff() must still produce the same minimal
+        // (empty) result as before this fix
+        CHECK(a == b);
+        auto p = json::diff(a, b);
+        CHECK(p.empty());
+        CHECK(a.patch(p) == b);
+    }
+}

tests/src/unit-regression3.cpp

@@ -895,5 +895,76 @@ TEST_CASE("issue #5402 - update(merge_objects=true) overwrites a primitive with
     CHECK(mixed == json({{"keep", {{"a", 1}, {"b", 2}}}, {"replace", {{"x", 2}}}}));
 }
 
+TEST_CASE("regression test - parser callback must not lose a duplicate key's prior value")
+{
+    // a callback that rejects only the scalar value 2
+    const json::parser_callback_t drop_value_2 = [](int /*depth*/, json::parse_event_t ev, json & v) noexcept
+    {
+        return !(ev == json::parse_event_t::value && v == 2);
+    };
+
+    SECTION("duplicate key, second (scalar) value rejected - prior value is restored")
+    {
+        const json j = json::parse(R"({"a":1,"a":2})", drop_value_2);
+        CHECK(j.dump() == "{\"a\":1}");
+    }
+
+    SECTION("duplicate key, second value is an object rejected at object_end - prior value is restored")
+    {
+        const json j = json::parse(R"({"a":1,"a":{"x":2}})",
+                                   [](int depth, json::parse_event_t ev, json& /*parsed*/) noexcept
+        {
+            return !(ev == json::parse_event_t::object_end && depth == 1);
+        });
+        CHECK(j.dump() == "{\"a\":1}");
+    }
+
+    SECTION("duplicate key, second value is an array rejected at array_end - prior value is restored")
+    {
+        const json j = json::parse(R"({"a":1,"a":[9,9]})",
+                                   [](int depth, json::parse_event_t ev, json& /*parsed*/) noexcept
+        {
+            return !(ev == json::parse_event_t::array_end && depth == 1);
+        });
+        CHECK(j.dump() == "{\"a\":1}");
+    }
+
+    SECTION("duplicate key, second value accepted (scalar) - last value wins")
+    {
+        const json j = json::parse(R"({"a":1,"a":2})", [](int, json::parse_event_t, json&) noexcept
+        {
+            return true;
+        });
+        CHECK(j.dump() == "{\"a\":2}");
+    }
+
+    SECTION("duplicate key, second value accepted (object) - last value wins")
+    {
+        const json j = json::parse(R"({"a":1,"a":{"x":2}})", [](int, json::parse_event_t, json&) noexcept
+        {
+            return true;
+        });
+        CHECK(j.dump() == "{\"a\":{\"x\":2}}");
+    }
+
+    SECTION("brand new (non-duplicate) key, value rejected - member is fully absent")
+    {
+        const json j = json::parse(R"({"a":1,"b":2})", drop_value_2);
+        CHECK(j.dump() == "{\"a\":1}");
+    }
+
+    SECTION("duplicate key nested two levels deep")
+    {
+        const json j = json::parse(R"({"outer":{"a":1,"a":2}})", drop_value_2);
+        CHECK(j.dump() == "{\"outer\":{\"a\":1}}");
+    }
+
+    SECTION("three occurrences of the same key - middle rejected, last accepted")
+    {
+        const json j = json::parse(R"({"k":1,"k":2,"k":3})", drop_value_2);
+        CHECK(j.dump() == "{\"k\":3}");
+    }
+}
+
 
 DOCTEST_CLANG_SUPPRESS_WARNING_POP