Merge branch 'develop' into bench/json-view-comparison-fairness

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

docs/mkdocs/docs/api/basic_json/at.md

@@ -90,7 +90,8 @@ Strong exception safety: if an exception occurs, the original value stays intact
 1. Constant.
 2. Logarithmic in the size of the container.
 3. Logarithmic in the size of the container.
-4. Logarithmic in the size of the container.
+4. Linear in the number of reference tokens of `ptr`; for each of them, logarithmic in the size of the object, or
+   constant for an array.
 
 ## Notes
 

docs/mkdocs/docs/api/basic_json/diff.md

@@ -19,9 +19,10 @@ The equality holds up to [`operator==`](operator_eq.md), so it has the following
 - A value that is not equal to itself cannot round-trip. This applies to `NaN` (also nested in an array or object) and
   to [discarded](is_discarded.md) values: `diff` emits a `replace` operation for them. A `NaN` in the patch is
   serialized as `#!json null`, so a patch that was dumped and parsed again would set `#!json null`.
-- Numbers that compare equal produce no operation, even if their types or signs differ: `diff(1, 1.0)`,
+- Numbers that compare equal produce no operation, even if their types differ or if one is zero and the other is
-  `diff(1u, 1)` (unsigned and signed integer) and `diff(0.0, -0.0)` all return `#!json []`. The patched value is then
+  negative zero: `diff(1, 1.0)` (integer and floating-point), `diff(1u, 1)` (unsigned and signed integer) and
-  the `source` and keeps its number type.
+  `diff(0.0, -0.0)` (positive and negative zero) all return `#!json []`. The patched value is then the `source` and
+  keeps its number type and sign.
 
 ## Parameters
 

docs/mkdocs/docs/api/basic_json/operator[].md

@@ -92,12 +92,20 @@ Strong exception safety: if an exception occurs, the original value stays intact
     For the **const** version, an object key or array index in `ptr` that does not exist is not reported by an
     exception, but is undefined behavior (see the notes below). Use [`at`](at.md) for checked access.
 
+    If the **non-const** version throws, the value is left unchanged: values it created for earlier reference tokens
+    of `ptr` are removed again. This also holds for exceptions such as `#!cpp std::bad_alloc` or
+    `#!cpp std::length_error` thrown while padding an array up to a very large index. With custom
+    [template parameters](../../features/types/template_parameters.md), this requires that `erase()` of the object and
+    array types and the move operations of the custom base class do not throw.
+
 ## Complexity
 
 1. Constant if `idx` is in the range of the array. Otherwise, linear in `idx - size()`.
 2. Logarithmic in the size of the container.
 3. Logarithmic in the size of the container.
-4. Logarithmic in the size of the container.
+4. Linear in the number of reference tokens of `ptr`; for each of them, logarithmic in the size of the object, or
+   constant for an array. When the non-const version pads an array with `#!json null` values, it is additionally
+   linear in `idx - size()` for that array.
 
 ## Notes
 
@@ -145,6 +153,14 @@ Strong exception safety: if an exception occurs, the original value stays intact
         specified by the JSON Pointer RFC; it is this library's own, intentional disambiguation rule. See also
         [JSON Pointer](../../features/json_pointer.md).
 
+    !!! warning "Large array indices"
+
+        As the array is filled up with `#!json null` values to the given index, the memory used grows with the index
+        and not with the size of the pointer: on a 64-bit platform, the index `10000000` allocates about 160 MB.
+        The library does not limit the index, so check the array indices of a JSON pointer from untrusted input
+        before using it with the non-const version. See
+        [Parsing untrusted input](../../features/parsing/untrusted_input.md#json-pointers-and-flattened-keys).
+
 !!! warning "Deprecation"
 
     Overload (4) also accepts a [`json_pointer`](../json_pointer/index.md) whose template argument is a `basic_json`
@@ -296,4 +312,6 @@ Strong exception safety: if an exception occurs, the original value stays intact
 4. Added in version 2.0.0. A missing array index in the const version is guarded by a runtime assertion since
    version 3.13.0. Throws [`parse_error.109`](../../home/exceptions.md#jsonexceptionparse_error109) instead of
 [`out_of_range.404`](../../home/exceptions.md#jsonexceptionout_of_range404) for a one-character array index that is not
-a digit (e.g., `/x`) in version 3.13.0, as it already did for longer ones.
+a digit (e.g., `/x`) in version 3.13.0, as it already did for longer ones. Fixed in version 3.13.0 to leave the value
+unchanged if the non-const version throws (strong exception safety, as documented); before, values created for earlier
+reference tokens of `ptr` were kept.

docs/mkdocs/docs/features/json_patch.md

@@ -78,8 +78,9 @@ flowchart LR
     - `NaN` and [discarded](../api/basic_json/is_discarded.md) values are not equal to themselves, so the invariant
       cannot hold for them. `diff` emits a `replace` operation for such a value, and a `NaN` in a patch is serialized
       as `#!json null`.
-    - Numbers that compare equal produce no operation even if their types differ: `diff(1, 1.0)`, `diff(1u, 1)`, and
+    - Numbers that compare equal produce no operation even if their types differ or if one is zero and the other is
-      `diff(0.0, -0.0)` return `#!json []`, so the patched value keeps the number type of *source*.
+      negative zero: `diff(1, 1.0)`, `diff(1u, 1)`, and `diff(0.0, -0.0)` return `#!json []`, so the patched value
+      keeps the number type and sign of *source*.
     - A [binary value](binary_values.md) in a patch is serialized as an object
       (`#!json {"bytes":[...],"subtype":null}`), so a patch that was dumped and parsed again no longer holds a binary
       value.

docs/mkdocs/docs/features/parsing/untrusted_input.md

@@ -130,6 +130,17 @@ parameters of [`parse()`](../../api/basic_json/parse.md) or [`accept()`](../../a
 enable either for input whose conformance you cannot otherwise control, since interoperability with strictly
 conforming JSON consumers is exactly what the default rejects.
 
+## JSON Pointers and flattened keys
+
+A [JSON Pointer](../json_pointer.md) can come from untrusted input as well, for instance as a path in a request or as a
+key of a flattened object. The non-const [`operator[]`](../../api/basic_json/operator%5B%5D.md) with a JSON Pointer and
+[`unflatten`](../../api/basic_json/unflatten.md) create missing values, and when an array index lies beyond the end of
+an array, they fill the array with `#!json null` values up to that index. The memory used therefore grows with the
+index, not with the size of the pointer: on a 64-bit platform, the index `10000000` allocates about 160 MB, the index
+`100000000` about 1.6 GB. The library does not limit the index. For pointers from untrusted input, check the array
+indices (or the pointer as a whole) before using them, or use [`at`](../../api/basic_json/at.md) or
+[`contains`](../../api/basic_json/contains.md), which never add values.
+
 ## Checklist
 
 - Wrap parsing in a `#!cpp try`/`#!cpp catch` block, or use `allow_exceptions=false`/`accept()` if your environment
@@ -145,6 +156,8 @@ conforming JSON consumers is exactly what the default rejects.
   `JSON_STRICT_NUL_HANDLING`.
 - Decide whether duplicate object keys should be an error for your application, and add a callback if so.
 - Do not enable `ignore_comments` or `ignore_trailing_commas` for input that must be strictly conforming JSON.
+- Check the array indices of untrusted JSON Pointers and flattened keys before passing them to the non-const
+  `operator[]` or `unflatten()`, which fill arrays with `#!json null` values up to the index.
 
 For the broader design rationale and how it is tested (fuzzing, sanitizers, static analysis), see the
 [assurance case](../../community/assurance_case.md) and [quality assurance](../../community/quality_assurance.md). To

include/nlohmann/detail/json_pointer.hpp

@@ -8,7 +8,7 @@
 
 #pragma once
 
-#include <algorithm> // all_of, copy
+#include <algorithm> // copy
 #include <array> // array
 #include <cctype> // isdigit
 #include <cerrno> // errno, ERANGE
@@ -16,9 +16,11 @@
 #include <cstdint> // uint64_t
 #include <cstdlib> // strtoull
 #include <cstring> // memchr, memcmp, memcpy
+#include <functional> // less
 #ifndef JSON_NO_IO
     #include <iosfwd> // ostream
 #endif  // JSON_NO_IO
+#include <iterator> // next
 #include <limits> // max
 #include <string> // string
 #include <utility> // move
@@ -28,6 +30,7 @@
 #include <nlohmann/detail/macro_scope.hpp>
 #include <nlohmann/detail/meta/cpp_future.hpp>
 #include <nlohmann/detail/meta/type_traits.hpp>
+#include <nlohmann/detail/meta/void_t.hpp>
 #include <nlohmann/detail/string_concat.hpp>
 #include <nlohmann/detail/string_escape.hpp>
 #include <nlohmann/detail/string_utils.hpp>
@@ -134,6 +137,26 @@ JSON_HEDLEY_ALWAYS_INLINE array_index_status parse_array_index(const StringType&
     return array_index_status::ok;
 }
 
+/// whether two keys of type @a StringType are equivalent under @a Compare
+/// exactly if they are equal: std::less<StringType> and std::less<>
+template<typename Compare, typename StringType>
+struct is_less_comparator : std::false_type {};
+
+template<typename StringType>
+struct is_less_comparator<std::less<StringType>, StringType> : std::true_type {};
+
+template<typename StringType>
+struct is_less_comparator<std::less<void>, StringType> : std::true_type {};
+
+/// whether the objects of type @a ObjectType find a key of type @a StringType
+/// exactly if it equals a stored key
+template<typename ObjectType, typename StringType, typename = void>
+struct finds_equal_keys : std::false_type {};
+
+template<typename ObjectType, typename StringType>
+struct finds_equal_keys<ObjectType, StringType, void_t<typename ObjectType::key_compare, typename ObjectType::key_type>>
+    : conjunction<is_less_comparator<typename ObjectType::key_compare, StringType>, std::is_same<typename ObjectType::key_type, StringType>> {};
+
 }  // namespace detail
 
 /// @brief JSON Pointer defines a string syntax for identifying a specific value within a JSON document
@@ -589,6 +612,225 @@ class json_pointer
         return object_member_impl(object, key, detail::priority_tag<1> {});
     }
 
+    /// @brief append a null value to @a array, like operator[] with the size of @a array
+    template<typename BasicJsonType>
+    static BasicJsonType& array_append(BasicJsonType& array)
+    {
+        auto& elements = *array.m_data.m_value.array;
+        const auto old_capacity = array.container_capacity();
+        elements.emplace_back();
+        return array.set_parent_after_insert(elements.back(), old_capacity);
+    }
+
+    /// @brief add @a key, which is missing, to @a object
+    template<typename BasicJsonType>
+    static auto object_add_impl(BasicJsonType& object, const string_t& key, detail::priority_tag<1> /*unused*/)
+    -> decltype(object.m_data.m_value.object->emplace(key, nullptr), std::declval<BasicJsonType&>())
+    {
+        const auto old_capacity = object.container_capacity();
+        const auto it = object.m_data.m_value.object->emplace(key, nullptr).first;
+        return object.set_parent_after_insert(it->second, old_capacity);
+    }
+
+    template<typename BasicJsonType>
+    static BasicJsonType& object_add_impl(BasicJsonType& object, const string_t& key, detail::priority_tag<0> /*unused*/)
+    {
+        return object.operator[](key);
+    }
+
+    template<typename BasicJsonType>
+    static BasicJsonType& object_add(BasicJsonType& object, const string_t& key)
+    {
+        return object_add_impl(object, key, detail::priority_tag<1> {});
+    }
+
+    /// @brief search @a key in @a members: the result is where the key is or,
+    /// for objects that can insert at a position (std::map), where it belongs
+    template<typename ObjectType>
+    static auto object_search_impl(ObjectType& members, const string_t& key, detail::priority_tag<1> /*unused*/)
+    -> decltype(members.key_comp()(key, members.lower_bound(key)->first), members.lower_bound(key))
+    {
+        return members.lower_bound(key);
+    }
+
+    template<typename ObjectType>
+    static auto object_search_impl(ObjectType& members, const string_t& key, detail::priority_tag<0> /*unused*/)
+    -> decltype(members.find(key))
+    {
+        return members.find(key);
+    }
+
+    template<typename ObjectType>
+    static auto object_search(ObjectType& members, const string_t& key)
+    -> decltype(object_search_impl(members, key, detail::priority_tag<1> {}))
+    {
+        return object_search_impl(members, key, detail::priority_tag<1> {});
+    }
+
+    /// @brief whether @a it, the result of @ref object_search for @a key, points to @a key
+    ///
+    /// With std::less, the key at the lower bound is equivalent to @a key
+    /// exactly if it is equal to it, and comparing for equality is cheaper.
+    template<typename ObjectType, typename Iterator,
+             detail::enable_if_t<detail::finds_equal_keys<ObjectType, string_t>::value, int> = 0>
+    static auto object_search_found_impl(ObjectType& members, const Iterator& it, const string_t& key, detail::priority_tag<2> /*unused*/)
+    -> decltype(members.lower_bound(key), it->first == key)
+    {
+        return it != members.end() && it->first == key;
+    }
+
+    template<typename ObjectType, typename Iterator>
+    static auto object_search_found_impl(ObjectType& members, const Iterator& it, const string_t& key, detail::priority_tag<1> /*unused*/)
+    -> decltype(members.key_comp()(key, it->first), members.lower_bound(key), bool())
+    {
+        return it != members.end() && !members.key_comp()(key, it->first);
+    }
+
+    template<typename ObjectType, typename Iterator>
+    static bool object_search_found_impl(ObjectType& members, const Iterator& it, const string_t& /*key*/, detail::priority_tag<0> /*unused*/)
+    {
+        return it != members.end();
+    }
+
+    template<typename ObjectType, typename Iterator>
+    static bool object_search_found(ObjectType& members, const Iterator& it, const string_t& key)
+    {
+        return object_search_found_impl(members, it, key, detail::priority_tag<2> {});
+    }
+
+    /// @brief insert @a key, which is missing, into @a object at @a it, the
+    /// result of @ref object_search for @a key
+    template<typename BasicJsonType, typename Iterator>
+    static auto object_insert_impl(BasicJsonType& object, const Iterator& it, const string_t& key, detail::priority_tag<1> /*unused*/)
+    -> decltype(object.m_data.m_value.object->emplace_hint(it, key, nullptr), std::declval<BasicJsonType&>())
+    {
+        const auto old_capacity = object.container_capacity();
+        const auto pos = object.m_data.m_value.object->emplace_hint(it, key, nullptr);
+        return object.set_parent_after_insert(pos->second, old_capacity);
+    }
+
+    template<typename BasicJsonType, typename Iterator>
+    static BasicJsonType& object_insert_impl(BasicJsonType& object, const Iterator& /*it*/, const string_t& key, detail::priority_tag<0> /*unused*/)
+    {
+        return object_add(object, key);
+    }
+
+    template<typename BasicJsonType, typename Iterator>
+    static BasicJsonType& object_insert(BasicJsonType& object, const Iterator& it, const string_t& key)
+    {
+        return object_insert_impl(object, it, key, detail::priority_tag<1> {});
+    }
+
+    /*!
+    @brief undoes the change @ref get_unchecked_create made to an existing value
+
+    @ref get_unchecked walks existing values without changing them. At the
+    first reference token that does not select an existing value, @ref
+    get_unchecked_create makes exactly one change to an existing value: it
+    converts a null value, inserts a key into an object, or grows an array.
+    Everything it creates afterwards lies below that change. If a later step
+    throws, the destructor of this guard reverts that one change, which
+    removes everything created after it, so the value is left as it was
+    (strong exception safety). The guard needs no try/catch and therefore also
+    compiles with JSON_NOEXCEPTION.
+
+    Reverting the change must not throw, as it happens in a destructor. It
+    erases the inserted member by its iterator (no key is converted or
+    compared, nothing is allocated), erases the added elements at the end of
+    the array, or swaps the null value back. This holds for std::map,
+    std::vector, ordered_map, and every object_t, array_t, and
+    json_base_class_t whose erase() and move operations do not throw.
+    Between recording and reverting, nothing is inserted into or erased from
+    the object or array, so the iterator stays valid.
+    */
+    template<typename BasicJsonType>
+    class get_unchecked_rollback
+    {
+      public:
+        using object_iterator = typename BasicJsonType::object_t::iterator;
+
+        /// record that @a member was inserted into the object @a value
+        get_unchecked_rollback(BasicJsonType* value, const object_iterator& member)
+            : m_value(value)
+            , m_member(member)
+            , m_change(change::key_inserted)
+        {}
+
+        /// record the change that the next reference token makes to the null value or array @a value
+        explicit get_unchecked_rollback(BasicJsonType* value)
+            : m_value(value)
+        {
+            if (value->is_null())
+            {
+                // keep the null value (with its base class and positions) and
+                // continue with a fresh null value in its place, which is what
+                // assigning an array or object to it would have left there
+                m_saved_null.swap(*value);
+                m_change = change::null_converted;
+            }
+            else
+            {
+                JSON_ASSERT(value->is_array());
+                m_old_size = value->m_data.m_value.array->size();
+                m_change = change::array_grown;
+            }
+        }
+
+        get_unchecked_rollback(const get_unchecked_rollback&) = delete;
+        get_unchecked_rollback(get_unchecked_rollback&&) = delete;
+        get_unchecked_rollback& operator=(const get_unchecked_rollback&) = delete;
+        get_unchecked_rollback& operator=(get_unchecked_rollback&&) = delete;
+
+        ~get_unchecked_rollback()
+        {
+            undo();
+        }
+
+        /// keep the change
+        void commit() noexcept
+        {
+            m_change = change::none;
+        }
+
+      private:
+        void undo() noexcept
+        {
+            switch (m_change)
+            {
+                case change::null_converted:
+                    // the values created since go away with m_saved_null
+                    m_value->swap(m_saved_null);
+                    break;
+                case change::key_inserted:
+                    m_value->m_data.m_value.object->erase(m_member);
+                    break;
+                case change::array_grown:
+                {
+                    auto& elements = *m_value->m_data.m_value.array;
+                    elements.resize(m_old_size);
+                    break;
+                }
+                case change::none:
+                default:
+                    break;
+            }
+        }
+
+        enum class change
+        {
+            none,
+            null_converted,
+            key_inserted,
+            array_grown
+        };
+
+        BasicJsonType* m_value;
+        object_iterator m_member {};
+        typename BasicJsonType::size_type m_old_size = 0;
+        change m_change = change::none;
+        BasicJsonType m_saved_null {};
+    };
+
   JSON_PRIVATE_UNLESS_TESTED:
     json_pointer top() const
     {
@@ -972,47 +1214,245 @@ class json_pointer
 
     Complexity: Linear in the length of the JSON pointer.
 
-    @throw parse_error.106   if an array index begins with '0'
+    Exception safety: Strong guarantee. This function walks existing values
-    @throw parse_error.109   if an array index was not a number
+    until the first reference token that does not select one. If that is the
-    @throw out_of_range.404  if the JSON pointer can not be resolved
+    last reference token, inserting the key into the object or growing the
+    array is strongly exception safe by itself; any other change is made by
+    @ref get_unchecked_create, which undoes it if a later step throws.
+
+    @throw parse_error.106   if an array index of an existing array begins with '0'
+    @throw parse_error.109   if an array index of an existing array was not a number
+    @throw out_of_range.404  if the JSON pointer can not be resolved, or if an
+                             array index is too large for strtoull
+    @throw out_of_range.410  if an array index exceeds size_type
+    @throw std::length_error if padding an array exceeds its maximal size
+    @throw std::bad_alloc    if an allocation fails
     */
     template<typename BasicJsonType>
     BasicJsonType& get_unchecked(BasicJsonType* ptr) const
     {
-        for (const auto& reference_token : reference_tokens)
+        const auto* token = reference_tokens.data();
+        const auto* const end = token + reference_tokens.size();
+        if (token == end)
         {
+            return *ptr;
+        }
+        if (ptr->is_null())
+        {
+            // the null value is converted to an array or an object
+            return get_unchecked_create(ptr, token, end);
+        }
+
+        // walk existing values; the walk itself inserts nothing, which keeps
+        // the loop fast (insertions are done below or in get_unchecked_create)
+        typename BasicJsonType::object_t::iterator it;
+        do
+        {
+            if (ptr->is_object())
+            {
+                auto& members = *ptr->m_data.m_value.object;
+                it = object_search(members, *token);
+                if (!object_search_found(members, it, *token))
+                {
+                    // the key is inserted below or in get_unchecked_create
+                    break;
+                }
+                ptr = &it->second;
+            }
+            else if (ptr->is_array())
+            {
+                if (is_dash(*token))
+                {
+                    if (token + 1 == end)
+                    {
+                        // explicitly treat "-" as index beyond the end
+                        return array_append(*ptr);
+                    }
+                    // an element will be appended in get_unchecked_create
+                    break;
+                }
+                if (token + 1 == end)
+                {
+                    // convert array index to number; unchecked access
+                    return ptr->operator[](array_index<BasicJsonType>(*token));
+                }
+                const auto idx = array_index<BasicJsonType>(*token);
+                if (idx >= ptr->m_data.m_value.array->size())
+                {
+                    // the array will be padded in get_unchecked_create
+                    break;
+                }
+                ptr = &(*ptr->m_data.m_value.array)[idx];
+            }
+            else if (ptr->is_null())
+            {
+                // the null value will be converted in get_unchecked_create
+                break;
+            }
+            else
+            {
+                // primitive value with a reference token left
+                throw_unresolved(*token, ptr);
+            }
+        }
+        while (++token != end);
+
+        if (token == end)
+        {
+            return *ptr;
+        }
+        if (token + 1 == end && ptr->is_object())
+        {
+            // insert the missing key where the walk found it belongs
+            return object_insert(*ptr, it, *token);
+        }
+        return get_unchecked_create(ptr, token, end);
+    }
+
+    /*!
+    @brief create the values for the reference tokens from @a first on
+
+    Continues @ref get_unchecked at the first reference token that does not
+    select an existing value. That token changes @a ptr, an existing value (see
+    @ref get_unchecked_rollback); all later tokens only change values created
+    here.
+
+    If the emplace() of the object type finds a member that the lookup in @ref
+    get_unchecked did not find (the two disagree, e.g. for fifo_map), nothing
+    is inserted, and the walk continues with that member as in @ref
+    get_unchecked, which is what basic_json::operator[] does.
+
+    @param[in] ptr    the value the reference token @a first is applied to
+    @param[in] first  the first reference token to resolve (an element of
+                      @ref reference_tokens)
+    @param[in] end    the end of @ref reference_tokens
+
+    @throw parse_error.106   if an array index of an existing array begins with
+                             '0' (only after emplace() found a member, see above)
+    @throw parse_error.109   if an array index of an existing array was not a
+                             number (only after emplace() found a member)
+    @throw out_of_range.404  if an array index is too large for strtoull, or (only
+                             after emplace() found a member) if the JSON pointer
+                             can not be resolved
+    @throw out_of_range.410  if an array index exceeds size_type
+    @throw std::length_error if padding an array exceeds its maximal size
+    @throw std::bad_alloc    if an allocation fails
+    */
+    template<typename BasicJsonType>
+    JSON_HEDLEY_NEVER_INLINE static BasicJsonType& get_unchecked_create(BasicJsonType* ptr, const string_t* first, const string_t* end)
+    {
+        const auto* token = first;
+        while (ptr->is_object())
+        {
+            const auto old_capacity = ptr->container_capacity();
+            const auto inserted = ptr->m_data.m_value.object->emplace(*token, nullptr);
+            if (JSON_HEDLEY_LIKELY(inserted.second))
+            {
+                get_unchecked_rollback<BasicJsonType> rollback(ptr, inserted.first);
+                BasicJsonType& result = create_values(&ptr->set_parent_after_insert(inserted.first->second, old_capacity), token + 1, end);
+                rollback.commit();
+                return result;
+            }
+
+            // nothing was inserted: walk existing values from the member
+            // emplace() found on, like get_unchecked
+            ptr = &inserted.first->second;
+            while (++token != end)
+            {
+                if (ptr->is_object())
+                {
+                    auto& members = *ptr->m_data.m_value.object;
+                    const auto it = object_search(members, *token);
+                    if (!object_search_found(members, it, *token))
+                    {
+                        break;
+                    }
+                    ptr = &it->second;
+                }
+                else if (ptr->is_array())
+                {
+                    if (is_dash(*token))
+                    {
+                        break;
+                    }
+                    const auto idx = array_index<BasicJsonType>(*token);
+                    if (idx >= ptr->m_data.m_value.array->size())
+                    {
+                        break;
+                    }
+                    ptr = &(*ptr->m_data.m_value.array)[idx];
+                }
+                else if (ptr->is_null())
+                {
+                    break;
+                }
+                else
+                {
+                    // primitive value with a reference token left
+                    throw_unresolved(*token, ptr);
+                }
+            }
+            if (token == end)
+            {
+                return *ptr;
+            }
+        }
+
+        get_unchecked_rollback<BasicJsonType> rollback(ptr);
+        BasicJsonType& result = create_values(ptr, token, end);
+        rollback.commit();
+        return result;
+    }
+
+    /*!
+    @brief the loop of @ref get_unchecked_create
+
+    @a ptr is a null value or an array. Every value created here is a fresh
+    one, so every object a key is added to is empty.
+
+    @throw out_of_range.404  if an array index is too large for strtoull
+    @throw out_of_range.410  if an array index exceeds size_type
+    @throw std::length_error if padding an array exceeds its maximal size
+    @throw std::bad_alloc    if an allocation fails
+    */
+    template<typename BasicJsonType>
+    static BasicJsonType& create_values(BasicJsonType* ptr, const string_t* first, const string_t* end)
+    {
+        for (const auto* token = first; token != end; ++token)
+        {
+            const auto& reference_token = *token;
+
             // convert null values to arrays or objects before continuing
             if (ptr->is_null())
             {
-                // check if the reference token is a valid array index, that is
+                // change value to an array for array indices or "-" or to an
-                // a nonempty sequence of digits without a leading '0'
+                // object otherwise; every null value here is a fresh one (see
-                // (cf. RFC 6901, Sect. 4); tokens that could never be a valid
+                // get_unchecked_rollback), so converting it in place leaves
-                // array index (such as "01" or "") are treated as object keys
+                // the same value as assigning an array or object to it
-                const bool nums = !reference_token.empty()
+                if (is_array_index_token(reference_token) || is_dash(reference_token))
-                                  && (reference_token.size() == 1 || reference_token[0] != '0')
-                                  && std::all_of(reference_token.begin(), reference_token.end(),
-                                                 [](const unsigned char x)
                 {
-                    return std::isdigit(x);
+                    ptr->template convert_null_to<typename BasicJsonType::array_t>();
-                });
+                }
-
+                else
-                // change value to an array for array indices or "-" or to object otherwise
+                {
-                *ptr = (nums || reference_token == "-")
+                    ptr->template convert_null_to<typename BasicJsonType::object_t>();
-                       ? detail::value_t::array
+                }
-                       : detail::value_t::object;
             }
 
             if (ptr->is_object())
             {
-                // use unchecked object access
+                // the object is new and empty
-                ptr = &object_member(*ptr, reference_token);
+                ptr = &object_add(*ptr, reference_token);
             }
-            else if (ptr->is_array())
+            else
             {
-                if (reference_token == "-")
+                // only the first value is not created here, and it is a null
+                // value or an array (see get_unchecked_create)
+                JSON_ASSERT(ptr->is_array());
+                if (is_dash(reference_token))
                 {
                     // explicitly treat "-" as index beyond the end
-                    ptr = &ptr->operator[](ptr->m_data.m_value.array->size());
+                    ptr = &array_append(*ptr);
                 }
                 else
                 {
@@ -1020,16 +1460,31 @@ class json_pointer
                     ptr = &ptr->operator[](array_index<BasicJsonType>(reference_token));
                 }
             }
-            else
-            {
-                // primitive value with a reference token left
-                throw_unresolved(reference_token, ptr);
-            }
         }
 
         return *ptr;
     }
 
+    /// whether @a s is a valid array index, that is a nonempty sequence of
+    /// digits without a leading '0' (cf. RFC 6901, Sect. 4); tokens that could
+    /// never be a valid array index (such as "01" or "") are object keys
+    static bool is_array_index_token(const string_t& s)
+    {
+        const std::size_t size = s.size();
+        if (size == 0 || (size > 1 && s[0] == '0'))
+        {
+            return false;
+        }
+        for (std::size_t i = 0; i < size; ++i)
+        {
+            if (std::isdigit(static_cast<unsigned char>(s[i])) == 0)
+            {
+                return false;
+            }
+        }
+        return true;
+    }
+
     /*!
     @throw parse_error.106   if an array index begins with '0'
     @throw parse_error.109   if an array index was not a number

single_include/nlohmann/json.hpp

@@ -24064,7 +24064,7 @@ NLOHMANN_JSON_NAMESPACE_END
 
 
 
-#include <algorithm> // all_of, copy
+#include <algorithm> // copy
 #include <array> // array
 #include <cctype> // isdigit
 #include <cerrno> // errno, ERANGE
@@ -24072,9 +24072,11 @@ NLOHMANN_JSON_NAMESPACE_END
 #include <cstdint> // uint64_t
 #include <cstdlib> // strtoull
 #include <cstring> // memchr, memcmp, memcpy
+#include <functional> // less
 #ifndef JSON_NO_IO
     #include <iosfwd> // ostream
 #endif  // JSON_NO_IO
+#include <iterator> // next
 #include <limits> // max
 #include <string> // string
 #include <utility> // move
@@ -24088,6 +24090,8 @@ NLOHMANN_JSON_NAMESPACE_END
 
 // #include <nlohmann/detail/meta/type_traits.hpp>
 
+// #include <nlohmann/detail/meta/void_t.hpp>
+
 // #include <nlohmann/detail/string_concat.hpp>
 
 // #include <nlohmann/detail/string_escape.hpp>
@@ -24198,6 +24202,26 @@ JSON_HEDLEY_ALWAYS_INLINE array_index_status parse_array_index(const StringType&
     return array_index_status::ok;
 }
 
+/// whether two keys of type @a StringType are equivalent under @a Compare
+/// exactly if they are equal: std::less<StringType> and std::less<>
+template<typename Compare, typename StringType>
+struct is_less_comparator : std::false_type {};
+
+template<typename StringType>
+struct is_less_comparator<std::less<StringType>, StringType> : std::true_type {};
+
+template<typename StringType>
+struct is_less_comparator<std::less<void>, StringType> : std::true_type {};
+
+/// whether the objects of type @a ObjectType find a key of type @a StringType
+/// exactly if it equals a stored key
+template<typename ObjectType, typename StringType, typename = void>
+struct finds_equal_keys : std::false_type {};
+
+template<typename ObjectType, typename StringType>
+struct finds_equal_keys<ObjectType, StringType, void_t<typename ObjectType::key_compare, typename ObjectType::key_type>>
+    : conjunction<is_less_comparator<typename ObjectType::key_compare, StringType>, std::is_same<typename ObjectType::key_type, StringType>> {};
+
 }  // namespace detail
 
 /// @brief JSON Pointer defines a string syntax for identifying a specific value within a JSON document
@@ -24653,6 +24677,225 @@ class json_pointer
         return object_member_impl(object, key, detail::priority_tag<1> {});
     }
 
+    /// @brief append a null value to @a array, like operator[] with the size of @a array
+    template<typename BasicJsonType>
+    static BasicJsonType& array_append(BasicJsonType& array)
+    {
+        auto& elements = *array.m_data.m_value.array;
+        const auto old_capacity = array.container_capacity();
+        elements.emplace_back();
+        return array.set_parent_after_insert(elements.back(), old_capacity);
+    }
+
+    /// @brief add @a key, which is missing, to @a object
+    template<typename BasicJsonType>
+    static auto object_add_impl(BasicJsonType& object, const string_t& key, detail::priority_tag<1> /*unused*/)
+    -> decltype(object.m_data.m_value.object->emplace(key, nullptr), std::declval<BasicJsonType&>())
+    {
+        const auto old_capacity = object.container_capacity();
+        const auto it = object.m_data.m_value.object->emplace(key, nullptr).first;
+        return object.set_parent_after_insert(it->second, old_capacity);
+    }
+
+    template<typename BasicJsonType>
+    static BasicJsonType& object_add_impl(BasicJsonType& object, const string_t& key, detail::priority_tag<0> /*unused*/)
+    {
+        return object.operator[](key);
+    }
+
+    template<typename BasicJsonType>
+    static BasicJsonType& object_add(BasicJsonType& object, const string_t& key)
+    {
+        return object_add_impl(object, key, detail::priority_tag<1> {});
+    }
+
+    /// @brief search @a key in @a members: the result is where the key is or,
+    /// for objects that can insert at a position (std::map), where it belongs
+    template<typename ObjectType>
+    static auto object_search_impl(ObjectType& members, const string_t& key, detail::priority_tag<1> /*unused*/)
+    -> decltype(members.key_comp()(key, members.lower_bound(key)->first), members.lower_bound(key))
+    {
+        return members.lower_bound(key);
+    }
+
+    template<typename ObjectType>
+    static auto object_search_impl(ObjectType& members, const string_t& key, detail::priority_tag<0> /*unused*/)
+    -> decltype(members.find(key))
+    {
+        return members.find(key);
+    }
+
+    template<typename ObjectType>
+    static auto object_search(ObjectType& members, const string_t& key)
+    -> decltype(object_search_impl(members, key, detail::priority_tag<1> {}))
+    {
+        return object_search_impl(members, key, detail::priority_tag<1> {});
+    }
+
+    /// @brief whether @a it, the result of @ref object_search for @a key, points to @a key
+    ///
+    /// With std::less, the key at the lower bound is equivalent to @a key
+    /// exactly if it is equal to it, and comparing for equality is cheaper.
+    template<typename ObjectType, typename Iterator,
+             detail::enable_if_t<detail::finds_equal_keys<ObjectType, string_t>::value, int> = 0>
+    static auto object_search_found_impl(ObjectType& members, const Iterator& it, const string_t& key, detail::priority_tag<2> /*unused*/)
+    -> decltype(members.lower_bound(key), it->first == key)
+    {
+        return it != members.end() && it->first == key;
+    }
+
+    template<typename ObjectType, typename Iterator>
+    static auto object_search_found_impl(ObjectType& members, const Iterator& it, const string_t& key, detail::priority_tag<1> /*unused*/)
+    -> decltype(members.key_comp()(key, it->first), members.lower_bound(key), bool())
+    {
+        return it != members.end() && !members.key_comp()(key, it->first);
+    }
+
+    template<typename ObjectType, typename Iterator>
+    static bool object_search_found_impl(ObjectType& members, const Iterator& it, const string_t& /*key*/, detail::priority_tag<0> /*unused*/)
+    {
+        return it != members.end();
+    }
+
+    template<typename ObjectType, typename Iterator>
+    static bool object_search_found(ObjectType& members, const Iterator& it, const string_t& key)
+    {
+        return object_search_found_impl(members, it, key, detail::priority_tag<2> {});
+    }
+
+    /// @brief insert @a key, which is missing, into @a object at @a it, the
+    /// result of @ref object_search for @a key
+    template<typename BasicJsonType, typename Iterator>
+    static auto object_insert_impl(BasicJsonType& object, const Iterator& it, const string_t& key, detail::priority_tag<1> /*unused*/)
+    -> decltype(object.m_data.m_value.object->emplace_hint(it, key, nullptr), std::declval<BasicJsonType&>())
+    {
+        const auto old_capacity = object.container_capacity();
+        const auto pos = object.m_data.m_value.object->emplace_hint(it, key, nullptr);
+        return object.set_parent_after_insert(pos->second, old_capacity);
+    }
+
+    template<typename BasicJsonType, typename Iterator>
+    static BasicJsonType& object_insert_impl(BasicJsonType& object, const Iterator& /*it*/, const string_t& key, detail::priority_tag<0> /*unused*/)
+    {
+        return object_add(object, key);
+    }
+
+    template<typename BasicJsonType, typename Iterator>
+    static BasicJsonType& object_insert(BasicJsonType& object, const Iterator& it, const string_t& key)
+    {
+        return object_insert_impl(object, it, key, detail::priority_tag<1> {});
+    }
+
+    /*!
+    @brief undoes the change @ref get_unchecked_create made to an existing value
+
+    @ref get_unchecked walks existing values without changing them. At the
+    first reference token that does not select an existing value, @ref
+    get_unchecked_create makes exactly one change to an existing value: it
+    converts a null value, inserts a key into an object, or grows an array.
+    Everything it creates afterwards lies below that change. If a later step
+    throws, the destructor of this guard reverts that one change, which
+    removes everything created after it, so the value is left as it was
+    (strong exception safety). The guard needs no try/catch and therefore also
+    compiles with JSON_NOEXCEPTION.
+
+    Reverting the change must not throw, as it happens in a destructor. It
+    erases the inserted member by its iterator (no key is converted or
+    compared, nothing is allocated), erases the added elements at the end of
+    the array, or swaps the null value back. This holds for std::map,
+    std::vector, ordered_map, and every object_t, array_t, and
+    json_base_class_t whose erase() and move operations do not throw.
+    Between recording and reverting, nothing is inserted into or erased from
+    the object or array, so the iterator stays valid.
+    */
+    template<typename BasicJsonType>
+    class get_unchecked_rollback
+    {
+      public:
+        using object_iterator = typename BasicJsonType::object_t::iterator;
+
+        /// record that @a member was inserted into the object @a value
+        get_unchecked_rollback(BasicJsonType* value, const object_iterator& member)
+            : m_value(value)
+            , m_member(member)
+            , m_change(change::key_inserted)
+        {}
+
+        /// record the change that the next reference token makes to the null value or array @a value
+        explicit get_unchecked_rollback(BasicJsonType* value)
+            : m_value(value)
+        {
+            if (value->is_null())
+            {
+                // keep the null value (with its base class and positions) and
+                // continue with a fresh null value in its place, which is what
+                // assigning an array or object to it would have left there
+                m_saved_null.swap(*value);
+                m_change = change::null_converted;
+            }
+            else
+            {
+                JSON_ASSERT(value->is_array());
+                m_old_size = value->m_data.m_value.array->size();
+                m_change = change::array_grown;
+            }
+        }
+
+        get_unchecked_rollback(const get_unchecked_rollback&) = delete;
+        get_unchecked_rollback(get_unchecked_rollback&&) = delete;
+        get_unchecked_rollback& operator=(const get_unchecked_rollback&) = delete;
+        get_unchecked_rollback& operator=(get_unchecked_rollback&&) = delete;
+
+        ~get_unchecked_rollback()
+        {
+            undo();
+        }
+
+        /// keep the change
+        void commit() noexcept
+        {
+            m_change = change::none;
+        }
+
+      private:
+        void undo() noexcept
+        {
+            switch (m_change)
+            {
+                case change::null_converted:
+                    // the values created since go away with m_saved_null
+                    m_value->swap(m_saved_null);
+                    break;
+                case change::key_inserted:
+                    m_value->m_data.m_value.object->erase(m_member);
+                    break;
+                case change::array_grown:
+                {
+                    auto& elements = *m_value->m_data.m_value.array;
+                    elements.resize(m_old_size);
+                    break;
+                }
+                case change::none:
+                default:
+                    break;
+            }
+        }
+
+        enum class change
+        {
+            none,
+            null_converted,
+            key_inserted,
+            array_grown
+        };
+
+        BasicJsonType* m_value;
+        object_iterator m_member {};
+        typename BasicJsonType::size_type m_old_size = 0;
+        change m_change = change::none;
+        BasicJsonType m_saved_null {};
+    };
+
   JSON_PRIVATE_UNLESS_TESTED:
     json_pointer top() const
     {
@@ -25036,47 +25279,245 @@ class json_pointer
 
     Complexity: Linear in the length of the JSON pointer.
 
-    @throw parse_error.106   if an array index begins with '0'
+    Exception safety: Strong guarantee. This function walks existing values
-    @throw parse_error.109   if an array index was not a number
+    until the first reference token that does not select one. If that is the
-    @throw out_of_range.404  if the JSON pointer can not be resolved
+    last reference token, inserting the key into the object or growing the
+    array is strongly exception safe by itself; any other change is made by
+    @ref get_unchecked_create, which undoes it if a later step throws.
+
+    @throw parse_error.106   if an array index of an existing array begins with '0'
+    @throw parse_error.109   if an array index of an existing array was not a number
+    @throw out_of_range.404  if the JSON pointer can not be resolved, or if an
+                             array index is too large for strtoull
+    @throw out_of_range.410  if an array index exceeds size_type
+    @throw std::length_error if padding an array exceeds its maximal size
+    @throw std::bad_alloc    if an allocation fails
     */
     template<typename BasicJsonType>
     BasicJsonType& get_unchecked(BasicJsonType* ptr) const
     {
-        for (const auto& reference_token : reference_tokens)
+        const auto* token = reference_tokens.data();
+        const auto* const end = token + reference_tokens.size();
+        if (token == end)
         {
+            return *ptr;
+        }
+        if (ptr->is_null())
+        {
+            // the null value is converted to an array or an object
+            return get_unchecked_create(ptr, token, end);
+        }
+
+        // walk existing values; the walk itself inserts nothing, which keeps
+        // the loop fast (insertions are done below or in get_unchecked_create)
+        typename BasicJsonType::object_t::iterator it;
+        do
+        {
+            if (ptr->is_object())
+            {
+                auto& members = *ptr->m_data.m_value.object;
+                it = object_search(members, *token);
+                if (!object_search_found(members, it, *token))
+                {
+                    // the key is inserted below or in get_unchecked_create
+                    break;
+                }
+                ptr = &it->second;
+            }
+            else if (ptr->is_array())
+            {
+                if (is_dash(*token))
+                {
+                    if (token + 1 == end)
+                    {
+                        // explicitly treat "-" as index beyond the end
+                        return array_append(*ptr);
+                    }
+                    // an element will be appended in get_unchecked_create
+                    break;
+                }
+                if (token + 1 == end)
+                {
+                    // convert array index to number; unchecked access
+                    return ptr->operator[](array_index<BasicJsonType>(*token));
+                }
+                const auto idx = array_index<BasicJsonType>(*token);
+                if (idx >= ptr->m_data.m_value.array->size())
+                {
+                    // the array will be padded in get_unchecked_create
+                    break;
+                }
+                ptr = &(*ptr->m_data.m_value.array)[idx];
+            }
+            else if (ptr->is_null())
+            {
+                // the null value will be converted in get_unchecked_create
+                break;
+            }
+            else
+            {
+                // primitive value with a reference token left
+                throw_unresolved(*token, ptr);
+            }
+        }
+        while (++token != end);
+
+        if (token == end)
+        {
+            return *ptr;
+        }
+        if (token + 1 == end && ptr->is_object())
+        {
+            // insert the missing key where the walk found it belongs
+            return object_insert(*ptr, it, *token);
+        }
+        return get_unchecked_create(ptr, token, end);
+    }
+
+    /*!
+    @brief create the values for the reference tokens from @a first on
+
+    Continues @ref get_unchecked at the first reference token that does not
+    select an existing value. That token changes @a ptr, an existing value (see
+    @ref get_unchecked_rollback); all later tokens only change values created
+    here.
+
+    If the emplace() of the object type finds a member that the lookup in @ref
+    get_unchecked did not find (the two disagree, e.g. for fifo_map), nothing
+    is inserted, and the walk continues with that member as in @ref
+    get_unchecked, which is what basic_json::operator[] does.
+
+    @param[in] ptr    the value the reference token @a first is applied to
+    @param[in] first  the first reference token to resolve (an element of
+                      @ref reference_tokens)
+    @param[in] end    the end of @ref reference_tokens
+
+    @throw parse_error.106   if an array index of an existing array begins with
+                             '0' (only after emplace() found a member, see above)
+    @throw parse_error.109   if an array index of an existing array was not a
+                             number (only after emplace() found a member)
+    @throw out_of_range.404  if an array index is too large for strtoull, or (only
+                             after emplace() found a member) if the JSON pointer
+                             can not be resolved
+    @throw out_of_range.410  if an array index exceeds size_type
+    @throw std::length_error if padding an array exceeds its maximal size
+    @throw std::bad_alloc    if an allocation fails
+    */
+    template<typename BasicJsonType>
+    JSON_HEDLEY_NEVER_INLINE static BasicJsonType& get_unchecked_create(BasicJsonType* ptr, const string_t* first, const string_t* end)
+    {
+        const auto* token = first;
+        while (ptr->is_object())
+        {
+            const auto old_capacity = ptr->container_capacity();
+            const auto inserted = ptr->m_data.m_value.object->emplace(*token, nullptr);
+            if (JSON_HEDLEY_LIKELY(inserted.second))
+            {
+                get_unchecked_rollback<BasicJsonType> rollback(ptr, inserted.first);
+                BasicJsonType& result = create_values(&ptr->set_parent_after_insert(inserted.first->second, old_capacity), token + 1, end);
+                rollback.commit();
+                return result;
+            }
+
+            // nothing was inserted: walk existing values from the member
+            // emplace() found on, like get_unchecked
+            ptr = &inserted.first->second;
+            while (++token != end)
+            {
+                if (ptr->is_object())
+                {
+                    auto& members = *ptr->m_data.m_value.object;
+                    const auto it = object_search(members, *token);
+                    if (!object_search_found(members, it, *token))
+                    {
+                        break;
+                    }
+                    ptr = &it->second;
+                }
+                else if (ptr->is_array())
+                {
+                    if (is_dash(*token))
+                    {
+                        break;
+                    }
+                    const auto idx = array_index<BasicJsonType>(*token);
+                    if (idx >= ptr->m_data.m_value.array->size())
+                    {
+                        break;
+                    }
+                    ptr = &(*ptr->m_data.m_value.array)[idx];
+                }
+                else if (ptr->is_null())
+                {
+                    break;
+                }
+                else
+                {
+                    // primitive value with a reference token left
+                    throw_unresolved(*token, ptr);
+                }
+            }
+            if (token == end)
+            {
+                return *ptr;
+            }
+        }
+
+        get_unchecked_rollback<BasicJsonType> rollback(ptr);
+        BasicJsonType& result = create_values(ptr, token, end);
+        rollback.commit();
+        return result;
+    }
+
+    /*!
+    @brief the loop of @ref get_unchecked_create
+
+    @a ptr is a null value or an array. Every value created here is a fresh
+    one, so every object a key is added to is empty.
+
+    @throw out_of_range.404  if an array index is too large for strtoull
+    @throw out_of_range.410  if an array index exceeds size_type
+    @throw std::length_error if padding an array exceeds its maximal size
+    @throw std::bad_alloc    if an allocation fails
+    */
+    template<typename BasicJsonType>
+    static BasicJsonType& create_values(BasicJsonType* ptr, const string_t* first, const string_t* end)
+    {
+        for (const auto* token = first; token != end; ++token)
+        {
+            const auto& reference_token = *token;
+
             // convert null values to arrays or objects before continuing
             if (ptr->is_null())
             {
-                // check if the reference token is a valid array index, that is
+                // change value to an array for array indices or "-" or to an
-                // a nonempty sequence of digits without a leading '0'
+                // object otherwise; every null value here is a fresh one (see
-                // (cf. RFC 6901, Sect. 4); tokens that could never be a valid
+                // get_unchecked_rollback), so converting it in place leaves
-                // array index (such as "01" or "") are treated as object keys
+                // the same value as assigning an array or object to it
-                const bool nums = !reference_token.empty()
+                if (is_array_index_token(reference_token) || is_dash(reference_token))
-                                  && (reference_token.size() == 1 || reference_token[0] != '0')
-                                  && std::all_of(reference_token.begin(), reference_token.end(),
-                                                 [](const unsigned char x)
                 {
-                    return std::isdigit(x);
+                    ptr->template convert_null_to<typename BasicJsonType::array_t>();
-                });
+                }
-
+                else
-                // change value to an array for array indices or "-" or to object otherwise
+                {
-                *ptr = (nums || reference_token == "-")
+                    ptr->template convert_null_to<typename BasicJsonType::object_t>();
-                       ? detail::value_t::array
+                }
-                       : detail::value_t::object;
             }
 
             if (ptr->is_object())
             {
-                // use unchecked object access
+                // the object is new and empty
-                ptr = &object_member(*ptr, reference_token);
+                ptr = &object_add(*ptr, reference_token);
             }
-            else if (ptr->is_array())
+            else
             {
-                if (reference_token == "-")
+                // only the first value is not created here, and it is a null
+                // value or an array (see get_unchecked_create)
+                JSON_ASSERT(ptr->is_array());
+                if (is_dash(reference_token))
                 {
                     // explicitly treat "-" as index beyond the end
-                    ptr = &ptr->operator[](ptr->m_data.m_value.array->size());
+                    ptr = &array_append(*ptr);
                 }
                 else
                 {
@@ -25084,16 +25525,31 @@ class json_pointer
                     ptr = &ptr->operator[](array_index<BasicJsonType>(reference_token));
                 }
             }
-            else
-            {
-                // primitive value with a reference token left
-                throw_unresolved(reference_token, ptr);
-            }
         }
 
         return *ptr;
     }
 
+    /// whether @a s is a valid array index, that is a nonempty sequence of
+    /// digits without a leading '0' (cf. RFC 6901, Sect. 4); tokens that could
+    /// never be a valid array index (such as "01" or "") are object keys
+    static bool is_array_index_token(const string_t& s)
+    {
+        const std::size_t size = s.size();
+        if (size == 0 || (size > 1 && s[0] == '0'))
+        {
+            return false;
+        }
+        for (std::size_t i = 0; i < size; ++i)
+        {
+            if (std::isdigit(static_cast<unsigned char>(s[i])) == 0)
+            {
+                return false;
+            }
+        }
+        return true;
+    }
+
     /*!
     @throw parse_error.106   if an array index begins with '0'
     @throw parse_error.109   if an array index was not a number

tests/benchmarks/json_view/README.md

@@ -21,6 +21,10 @@ python3 tests/benchmarks/json_view/compare.py --data <json_test_data directory>
 - The other libraries come from the system: pkg-config, or Homebrew (`brew install yyjson simdjson boost`). With
   `--download`, pinned releases are downloaded instead and checked against their SHA-256. Without Boost headers (or
   with `--no-boost`), the Boost.JSON columns are skipped, and the results say so.
+- `--aligned` additionally builds and runs everything with all functions aligned to 64 bytes (`-mllvm
+  -align-all-functions=6` for Clang, `-falign-functions=64 -falign-loops=64` for GCC), and reports both result
+  tables as "default build" and "aligned build". Code layout alone can move a result by a few percent; a difference
+  that appears in only one of the two tables is a layout artifact. Off by default, which doubles the run time.
 - `--corpus file...` adds files to the corpus benchmark, e.g. those of
   [simdjson-data](https://github.com/simdjson/simdjson-data) or the
   [yyjson benchmark](https://github.com/ibireme/yyjson_benchmark).
@@ -39,7 +43,7 @@ where `json_view` stands on another architecture; they are not meant for publica
 
 ## What is measured
 
-`bench_view.cpp` runs four workloads on twitter, citm_catalog, canada, jeopardy, a single tweet (`status`), and a
+`bench_view.cpp` runs these workloads on twitter, citm_catalog, canada, jeopardy, a single tweet (`status`), and a
 JSON-RPC request (`rpc`):
 
 | workload | what it does |
@@ -48,6 +52,7 @@ JSON-RPC request (`rpc`):
 | traverse | parse, then visit every value, convert every number, touch every string and key |
 | select | parse, then read a few fields per record (e.g. id, user name, and retweet count of each tweet) |
 | dump | serialize a parsed document (compact) |
+| parse+dump | build a fresh document, then serialize it (compact), in one timed call |
 
 `bench_corpus.cpp` runs parse, traverse, and dump on any list of files, so that no library is tuned to a handful of
 documents. Its dump also writes the numbers as they are in the input: `json_view` with `number_format::source`, and
@@ -72,11 +77,32 @@ call follows an untimed call of the same engine: otherwise the engine after `jso
 cleaning up the tens of thousands of nodes `json::parse` just freed (with glibc, this made `json_view` look 1.7 times
 slower on citm_catalog traverse).
 
+`json_view` is an index into the source text, not a tree of owning nodes. Its like-for-like peers are yyjson and
+simdjson DOM (immutable, indexed, no per-node allocation). Boost.JSON and `json::parse` build mutable, owning DOMs and
+are reference points for what that costs, not equals; simdjson On-Demand is lazy and forward-only, so it is the other
+way round: it has no random access to pay for. All engines validate UTF-8 and syntax by default.
+
+Number conversion differs: `json_view` converts integers while parsing but converts floating-point numbers lazily,
+when they are read or dumped; yyjson (default flags) and simdjson DOM convert every number while parsing. The cost of
+float conversion therefore appears in the traverse and dump rows for `json_view`, and in the parse row for the others.
+For float-heavy documents (canada), the fair reading is a sum of parse and use: the `traverse` workload already
+includes the parse, and `parse+dump` measures a fresh document and its dump in one timed call; the `dump` row alone
+leaves the conversion in `json_view` and out of the others. Eager conversion at parse time was prototyped as an opt-in
+and measured: canada traverse (with its parse) was 4-7 % faster, but parse+dump 11-30 % slower and parse 36-90 % slower. It was
+dropped, so lazy conversion is the only mode.
+
+`json_view` can also be edited: `json_editable_document` has the same read API, and the `(editable)` rows in
+`bench_view` measure it on the same workloads. They are the fair counterpart of Boost.JSON and `json::parse`, which are
+mutable too; the plain `json_view` rows are the counterpart of yyjson and simdjson DOM, which are not. `bench_edit` adds
+the read-modify-write workloads. Read the tables with that in mind: an immutable-only library is not compared with an
+editable `json_view` unless it says so.
+
 The engines do not all offer the same features, which the numbers should be read with:
 
 | engine | document | random access | editable | notes |
 |---|---|---|---|---|
 | `json_view` | immutable index into the text | yes | no | a fresh document per parse; "reused" parses into the same document |
+| `json_view` (editable) | index into the text plus edits | yes | yes, in place | `json_editable_document`; keeps member order and number spellings |
 | yyjson | immutable (`yyjson_read`) | yes | via a mutable copy | |
 | simdjson DOM | immutable, parser reused | yes | no | "fresh" uses a new parser per parse |
 | simdjson On-Demand | none: forward-only, lazy | no | no | only traverse and select |

tests/benchmarks/json_view/bench_view.cpp

@@ -8,8 +8,9 @@
 
 // Same-feature-set benchmark: read-only JSON documents with random access.
 //
-//   json_view      nlohmann/json_view.hpp (fresh document per parse / reused)
+//   json_view      nlohmann/json_view.hpp (fresh document per parse / reused;
-//   yyjson         yyjson_read(): immutable document, random access
+//                  "(editable)": json_editable_document, same read API)
+//   yyjson        yyjson_read(): immutable document, random access
 //   simdjson DOM   dom::parser (reused, as recommended; "fresh": a new parser
 //                  per parse): immutable, random access
 // references (different feature sets):
@@ -19,7 +20,11 @@
 //
 // Workloads: parse (build + free), traverse (visit everything, convert every
 // number, touch every string and key), select (a few fields per document),
-// dump (compact serialization of the parsed document).
+// dump (compact serialization of the parsed document), and parse+dump (a fresh
+// document, then its dump, in one timed call). json_view converts floating-point
+// numbers lazily (when read or dumped), yyjson and simdjson DOM while parsing:
+// for float-heavy documents, traverse (which includes the parse) and parse+dump
+// are the fair reading of the parse and dump rows.
 // All engines run interleaved in every round, each timed call after an untimed
 // one of the same engine; the best round is reported.
 #include <nlohmann/json_view.hpp>
@@ -43,6 +48,8 @@
 
 using nlohmann::json;
 using nlohmann::json_document;
+using nlohmann::json_editable_document;
+using nlohmann::json_editable_view;
 using nlohmann::json_view;
 
 static volatile double g_sink;
@@ -55,7 +62,8 @@ struct stats
 
 // ---------------- traversal ----------------
 
-static void walk(json_view v, stats& st)
+template<class V>
+static void walk_view(V v, stats& st)
 {
     ++st.nodes;
     switch (v.type())
@@ -64,32 +72,42 @@ static void walk(json_view v, stats& st)
             for (auto it = v.begin(); it != v.end(); ++it)
             {
                 st.str += it.key().size();
-                walk(*it, st);
+                walk_view<V>(*it, st);
             }
             break;
         case json::value_t::array:
-            for (const json_view e : v)
+            for (const V e : v)
             {
-                walk(e, st);
+                walk_view<V>(e, st);
             }
             break;
         case json::value_t::string:
             st.str += v.get_string().size();
             break;
         case json::value_t::number_integer:
-            st.num += static_cast<double>(v.get<std::int64_t>());
+            st.num += static_cast<double>(v.template get<std::int64_t>());
             break;
         case json::value_t::number_unsigned:
-            st.num += static_cast<double>(v.get<std::uint64_t>());
+            st.num += static_cast<double>(v.template get<std::uint64_t>());
             break;
         case json::value_t::number_float:
-            st.num += v.get<double>();
+            st.num += v.template get<double>();
             break;
         default:
             break;
     }
 }
 
+static void walk(json_view v, stats& st)
+{
+    walk_view<json_view>(v, st);
+}
+
+static void walk(json_editable_view v, stats& st)
+{
+    walk_view<json_editable_view>(v, st);
+}
+
 static void walk(const json& j, stats& st)
 {
     ++st.nodes;
@@ -300,20 +318,21 @@ static void walk(const boost::json::value& v, stats& st)
 // status (one tweet): id, user.screen_name, retweet_count
 // rpc: method, params.minuend, id
 
-static double pick(const std::string& name, json_view r)
+template<class V>
+static double pick_view(const std::string& name, V r)
 {
     double acc = 0;
     if (name == "twitter" || name == "status")
     {
-        auto one = [&](json_view s)
+        auto one = [&](V s)
         {
-            acc += static_cast<double>(s["id"].get<std::uint64_t>());
+            acc += static_cast<double>(s["id"].template get<std::uint64_t>());
             acc += static_cast<double>(s["user"]["screen_name"].get_string().size());
-            acc += static_cast<double>(s["retweet_count"].get<std::int64_t>());
+            acc += static_cast<double>(s["retweet_count"].template get<std::int64_t>());
         };
         if (name == "twitter")
         {
-            for (const json_view s : r["statuses"])
+            for (const V s : r["statuses"])
             {
                 one(s);
             }
@@ -325,20 +344,20 @@ static double pick(const std::string& name, json_view r)
     }
     else if (name == "citm_catalog")
     {
-        for (const json_view p : r["performances"])
+        for (const V p : r["performances"])
         {
-            acc += static_cast<double>(p["id"].get<std::uint64_t>() + p["eventId"].get<std::uint64_t>() + p["seatCategories"].size());
+            acc += static_cast<double>(p["id"].template get<std::uint64_t>() + p["eventId"].template get<std::uint64_t>() + p["seatCategories"].size());
         }
         acc += static_cast<double>(r["events"].size());
     }
     else if (name == "canada")
     {
-        const json_view g = r["features"][0]["geometry"];
+        const V g = r["features"][0]["geometry"];
         acc += static_cast<double>(r["type"].get_string().size() + g["type"].get_string().size() + g["coordinates"].size());
     }
     else if (name == "jeopardy")
     {
-        for (const json_view q : r)
+        for (const V q : r)
         {
             acc += q["round"].get_string() == "Final Jeopardy!" ? 1 : 0;
             acc += static_cast<double>(q["category"].get_string().size());
@@ -347,11 +366,21 @@ static double pick(const std::string& name, json_view r)
     else if (name == "rpc")
     {
         acc += static_cast<double>(r["method"].get_string().size());
-        acc += static_cast<double>(r["params"]["minuend"].get<std::int64_t>() + r["id"].get<std::int64_t>());
+        acc += static_cast<double>(r["params"]["minuend"].template get<std::int64_t>() + r["id"].template get<std::int64_t>());
     }
     return acc;
 }
 
+static double pick(const std::string& name, json_view r)
+{
+    return pick_view<json_view>(name, r);
+}
+
+static double pick(const std::string& name, json_editable_view r)
+{
+    return pick_view<json_editable_view>(name, r);
+}
+
 static double pick(const std::string& name, const json& r)
 {
     double acc = 0;
@@ -634,15 +663,35 @@ int main(int argc, char** argv)
         walk(json::parse(dc.text), a);
         auto d = json_document::parse(dc.text);
         walk(d.root(), b);
+        auto ed = json_editable_document::parse(dc.text);
+        stats e;
+        walk(ed.root(), e);
         yyjson_doc* y = yyjson_read(dc.text.data(), dc.text.size(), 0);
         walk(yyjson_doc_get_root(y), c);
         simdjson::dom::parser p;
         walk(p.parse(dc.text).value(), dd);
-        const bool ok = a.nodes == b.nodes && a.nodes == c.nodes && a.nodes == dd.nodes && a.str == b.str && a.str == c.str && a.str == dd.str
+        const bool ok = a.nodes == b.nodes && a.nodes == e.nodes && a.str == e.str && std::fabs(a.num - e.num) <= 1e-9 * std::fabs(a.num) && a.nodes == c.nodes && a.nodes == dd.nodes && a.str == b.str && a.str == c.str && a.str == dd.str
                         && std::fabs(a.num - b.num) <= 1e-9 * std::fabs(a.num) && std::fabs(a.num - c.num) <= 1e-9 * std::fabs(a.num);
-        const double pa = pick(dc.name, json::parse(dc.text)), pb = pick(dc.name, d.root()), pc = pick(dc.name, yyjson_doc_get_root(y)), pd = pick(dc.name, p.parse(dc.text).value());
+        const double pa = pick(dc.name, json::parse(dc.text)), pb = pick(dc.name, d.root()), pe = pick(dc.name, ed.root()), pc = pick(dc.name, yyjson_doc_get_root(y)), pd = pick(dc.name, p.parse(dc.text).value());
-        std::printf("check %-13s traverse %s select %s\n", dc.name.c_str(), ok ? "OK" : "MISMATCH", (pa == pb && pa == pc && pa == pd) ? "OK" : "MISMATCH");
+        std::printf("check %-13s traverse %s select %s\n", dc.name.c_str(), ok ? "OK" : "MISMATCH", (pa == pb && pa == pe && pa == pc && pa == pd) ? "OK" : "MISMATCH");
         yyjson_doc_free(y);
+
+        // the dumps of the engines may format floats differently, so only: every dump is non-empty, and the
+        // dump of json_view (also editable) and of json::parse parse back to the same number of nodes
+        const std::string dj = json::parse(dc.text).dump();
+        const std::string dv = d.root().dump();
+        const std::string de = ed.root().dump();
+        std::size_t yn = 0;
+        yyjson_doc* y2 = yyjson_read(dc.text.data(), dc.text.size(), 0);
+        char* yo = yyjson_write(y2, 0, &yn);
+        const bool dumps_ok = !dj.empty() && !dv.empty() && !de.empty() && yo != nullptr && yn > 0 && !simdjson::to_string(p.parse(dc.text).value()).empty();
+        std::free(yo);
+        yyjson_doc_free(y2);
+        stats sj_nodes, sv_nodes, se_nodes;
+        walk(json::parse(dj), sj_nodes);
+        walk(json::parse(dv), sv_nodes);
+        walk(json::parse(de), se_nodes);
+        std::printf("check %-13s dump %s\n", dc.name.c_str(), (dumps_ok && sj_nodes.nodes == a.nodes && sv_nodes.nodes == a.nodes && se_nodes.nodes == a.nodes) ? "OK" : "MISMATCH");
     }
 
     std::FILE* csv = std::fopen("bench_view.csv", "w");
@@ -669,6 +718,7 @@ int main(int argc, char** argv)
         workloads.push_back({"parse", {
                 {"json_view", [&] { auto d = json_document::parse(s); g_sink = static_cast<double>(d.node_count()); }},
                 {"json_view (reused)", [&] { reused.read(s); g_sink = static_cast<double>(reused.node_count()); }},
+                {"json_view (editable)", [&] { auto d = json_editable_document::parse(s); g_sink = static_cast<double>(d.root().size()); }},
                 {"yyjson", [&] { yyjson_doc* d = yyjson_read(s.data(), s.size(), 0); g_sink = static_cast<double>(yyjson_doc_get_val_count(d)); yyjson_doc_free(d); }},
                 {"simdjson DOM", [&] { auto e = sj.parse(ps).value_unsafe(); g_sink = e.is_object(); }},
                 {"simdjson DOM (fresh)", [&] { simdjson::dom::parser p; auto e = p.parse(ps).value_unsafe(); g_sink = e.is_object(); }},
@@ -680,6 +730,7 @@ int main(int argc, char** argv)
         workloads.push_back({"traverse", {
                 {"json_view", [&] { auto d = json_document::parse(s); stats st; walk(d.root(), st); g_sink = st.num; }},
                 {"json_view (reused)", [&] { reused.read(s); stats st; walk(reused.root(), st); g_sink = st.num; }},
+                {"json_view (editable)", [&] { auto d = json_editable_document::parse(s); stats st; walk(d.root(), st); g_sink = st.num; }},
                 {"yyjson", [&] { yyjson_doc* d = yyjson_read(s.data(), s.size(), 0); stats st; walk(yyjson_doc_get_root(d), st); g_sink = st.num; yyjson_doc_free(d); }},
                 {"simdjson DOM", [&] { stats st; walk(sj.parse(ps).value_unsafe(), st); g_sink = st.num; }},
                 {"simdjson OD", [&] { auto d = od.iterate(ps).value_unsafe(); stats st; walk_od(d.get_value().value_unsafe(), st); g_sink = st.num; }},
@@ -691,15 +742,31 @@ int main(int argc, char** argv)
         workloads.push_back({"select", {
                 {"json_view", [&] { auto d = json_document::parse(s); g_sink = pick(name, d.root()); }},
                 {"json_view (reused)", [&] { reused.read(s); g_sink = pick(name, reused.root()); }},
+                {"json_view (editable)", [&] { auto d = json_editable_document::parse(s); g_sink = pick(name, d.root()); }},
                 {"yyjson", [&] { yyjson_doc* d = yyjson_read(s.data(), s.size(), 0); g_sink = pick(name, yyjson_doc_get_root(d)); yyjson_doc_free(d); }},
                 {"simdjson DOM", [&] { g_sink = pick(name, sj.parse(ps).value_unsafe()); }},
                 {"simdjson OD", [&] { auto d = od.iterate(ps).value_unsafe(); g_sink = pick_od(name, d); }},
                 {"json::parse", [&] { json j = json::parse(s); g_sink = pick(name, j); }},
             }});
+        // pipeline: a fresh document, then a dump of it, in one timed call
+        // (simdjson DOM reuses its parser, as in the parse row); the traverse
+        // workload above is already parse+traverse
+        workloads.push_back({"parse+dump", {
+                {"json_view", [&] { auto d = json_document::parse(s); std::string o = d.root().dump(); g_sink = static_cast<double>(o.size()); }},
+                {"json_view (editable)", [&] { auto d = json_editable_document::parse(s); std::string o = d.root().dump(); g_sink = static_cast<double>(o.size()); }},
+                {"yyjson", [&] { yyjson_doc* d = yyjson_read(s.data(), s.size(), 0); std::size_t n = 0; char* o = yyjson_write(d, 0, &n); g_sink = static_cast<double>(n); std::free(o); yyjson_doc_free(d); }},
+                {"simdjson DOM", [&] { std::string o = simdjson::to_string(sj.parse(ps).value_unsafe()); g_sink = static_cast<double>(o.size()); }},
+#if JSON_VIEW_BENCH_BOOST
+                {"Boost.JSON", [&] { boost::json::monotonic_resource mr; auto v = boost::json::parse(s, &mr); std::string o = boost::json::serialize(v); g_sink = static_cast<double>(o.size()); }},
+#endif
+                {"json::parse", [&] { json j = json::parse(s); std::string o = j.dump(); g_sink = static_cast<double>(o.size()); }},
+            }});
         {
             // serialization of an already parsed document
             static json_document vd;
             vd.read(s);
+            static json_editable_document ed = json_editable_document::parse(s);
+            ed = json_editable_document::parse(s);
             static yyjson_doc* yd = nullptr;
             if (yd)
             {
@@ -713,6 +780,7 @@ int main(int argc, char** argv)
             jd = json::parse(s);
             workloads.push_back({"dump", {
                     {"json_view", [&] { std::string o = vd.root().dump(); g_sink = static_cast<double>(o.size()); }},
+                    {"json_view (editable)", [&] { std::string o = ed.root().dump(); g_sink = static_cast<double>(o.size()); }},
                     {"yyjson", [&] { std::size_t n = 0; char* o = yyjson_write(yd, 0, &n); g_sink = static_cast<double>(n); std::free(o); }},
                     {"simdjson DOM", [&] { std::string o = simdjson::to_string(se); g_sink = static_cast<double>(o.size()); }},
                     {"json::parse", [&] { std::string o = jd.dump(); g_sink = static_cast<double>(o.size()); }},
@@ -740,7 +808,7 @@ int main(int argc, char** argv)
                 }
             }
             const double ref = best[0];
-            std::printf("%-13s %-9s", dc.name.c_str(), wl.first.c_str());
+            std::printf("%-13s %-14s", dc.name.c_str(), wl.first.c_str());
             for (std::size_t k = 0; k < wl.second.size(); ++k)
             {
                 const double us = best[k] / 1e3;

tests/benchmarks/json_view/compare.py

@@ -239,6 +239,8 @@ def main():
     ap.add_argument('--download', action='store_true', help='use pinned downloads instead of system libraries')
     ap.add_argument('--no-boost', action='store_true', help='skip Boost.JSON')
     ap.add_argument('--native', action='store_true', help='compile for this CPU (-march=native / -mcpu=native)')
+    ap.add_argument('--aligned', action='store_true',
+                    help='also build and run with all functions (and loops) aligned, to make code layout artifacts visible')
     ap.add_argument('--rounds', type=int, default=30, help='rounds of bench_view (default: 30)')
     ap.add_argument('--corpus', nargs='*', default=[], help='more files for bench_corpus')
     ap.add_argument('--build-dir', default=os.path.join(HERE, 'build'), help='where to build (default: build/ next to this script)')
@@ -282,32 +284,46 @@ def main():
                 + ['-I' + d for d in lib.include])
             objects.append(obj)
 
-    binaries = {}
+    cxx_version = output(cxx + ['--version'])
-    for bench in ['bench_view', 'bench_corpus', 'bench_edit', 'bench_decode']:
+    is_clang = 'clang' in cxx_version.lower()
-        exe = os.path.join(args.build_dir, bench)
+    # alignment flags: functions and (gcc) loops on 64-byte boundaries
-        run(cxx + flags + include + [os.path.join(HERE, bench + '.cpp')] + objects + link + ['-o', exe])
+    align_flags = ['-mllvm', '-align-all-functions=6'] if is_clang else ['-falign-functions=64', '-falign-loops=64']
-        binaries[bench] = exe
+
+    # the builds: the default one and, with --aligned, one with aligned code
+    variants = [('default build', args.build_dir, [])]
+    if args.aligned:
+        variants.append(('aligned build', os.path.join(args.build_dir, 'aligned'), align_flags))
 
-    # run: bench_view on its documents, bench_corpus on those and the given files
     corpus = [os.path.join(args.data, f) for f in DEFAULT_CORPUS] + args.corpus
-    outputs = {}
+    # label -> {benchmark: output}
-    outputs['bench_view'] = run([binaries['bench_view'], args.data, str(args.rounds)], cwd=args.build_dir,
+    all_outputs = {}
-                                capture_output=True, text=True).stdout
+    for label, build_dir, extra in variants:
-    outputs['bench_corpus'] = run([binaries['bench_corpus']] + corpus, cwd=args.build_dir,
+        os.makedirs(build_dir, exist_ok=True)
-                                  capture_output=True, text=True).stdout
+        binaries = {}
-    outputs['bench_edit'] = run([binaries['bench_edit'], args.data, str(max(1, args.rounds // 2))], cwd=args.build_dir,
+        for bench in ['bench_view', 'bench_corpus', 'bench_edit', 'bench_decode']:
-                                capture_output=True, text=True).stdout
+            exe = os.path.join(build_dir, bench)
-    outputs['bench_decode'] = run([binaries['bench_decode'], args.data, str(max(1, args.rounds // 2))], cwd=args.build_dir,
+            run(cxx + flags + extra + include + [os.path.join(HERE, bench + '.cpp')] + objects + link + ['-o', exe])
-                                  capture_output=True, text=True).stdout
+            binaries[bench] = exe
-    for name, text in outputs.items():
+
-        print(text)
+        # run: bench_view on its documents, bench_corpus on those and the given files
+        outputs = {}
+        outputs['bench_view'] = run([binaries['bench_view'], args.data, str(args.rounds)], cwd=build_dir,
+                                    capture_output=True, text=True).stdout
+        outputs['bench_corpus'] = run([binaries['bench_corpus']] + corpus, cwd=build_dir,
+                                      capture_output=True, text=True).stdout
+        outputs['bench_edit'] = run([binaries['bench_edit'], args.data, str(max(1, args.rounds // 2))], cwd=build_dir,
+                                    capture_output=True, text=True).stdout
+        outputs['bench_decode'] = run([binaries['bench_decode'], args.data, str(max(1, args.rounds // 2))],
+                                      cwd=build_dir, capture_output=True, text=True).stdout
+        for name, text in outputs.items():
+            print(text)
+        all_outputs[label] = outputs
 
     # results with their metadata
     now = datetime.datetime.now()
     host = re.sub(r'[^A-Za-z0-9-]+', '-', platform.node().split('.')[0]) or 'host'
     stem = os.path.join(HERE, 'results', f'{now:%Y-%m-%d}-{host}')
     os.makedirs(os.path.dirname(stem), exist_ok=True)
-    cxx_version = output(cxx + ['--version'])
     compiler_version = cxx_version.splitlines()[0] if cxx_version else ' '.join(cxx)
     meta = [
         ('date', f'{now:%Y-%m-%d %H:%M}'),
@@ -316,6 +332,7 @@ def main():
         ('OS', f'{platform.system()} {platform.release()} ({platform.machine()})'),
         ('compiler', compiler_version),
         ('flags', ' '.join(flags)),
+        ('aligned results', ('included (flags: ' + ' '.join(align_flags) + ')') if args.aligned else 'not included'),
         ('yyjson', libs['yyjson'].version),
         ('simdjson', libs['simdjson'].version),
         ('Boost.JSON', libs['boost'].version if with_boost else 'skipped (not found)'),
@@ -328,15 +345,20 @@ def main():
         f.write('| | |\n|---|---|\n')
         for key, value in meta:
             f.write(f'| {key} | {value} |\n')
-        for name, text in outputs.items():
+        for label, outputs in all_outputs.items():
-            f.write(f'\n## {name}\n\n```\n{text.rstrip()}\n```\n')
+            # (without --aligned there is only the default build, and no extra heading level)
+            prefix = f'{label}: ' if args.aligned else ''
+            for name, text in outputs.items():
+                f.write(f'\n## {prefix}{name}\n\n```\n{text.rstrip()}\n```\n')
     with open(stem + '.csv', 'w', encoding='utf-8') as out:
         out.write(''.join(f'# {key}: {value}\n' for key, value in meta))
-        for name in ['bench_view', 'bench_corpus', 'bench_edit', 'bench_decode']:
+        for label, build_dir, _ in variants:
-            path = os.path.join(args.build_dir, name + '.csv')
+            for name in ['bench_view', 'bench_corpus', 'bench_edit', 'bench_decode']:
-            if os.path.isfile(path):
+                path = os.path.join(build_dir, name + '.csv')
-                with open(path, encoding='utf-8') as f:
+                if os.path.isfile(path):
-                    out.write(f'# {name}\n' + f.read())
+                    with open(path, encoding='utf-8') as f:
+                        # (the default build keeps its plain section names)
+                        out.write(f'# {name}' + (f' ({label})' if args.aligned else '') + '\n' + f.read())
     print(f'results: {stem}.md, {stem}.csv')
 
 

tests/src/unit-alt-string.cpp

@@ -373,6 +373,23 @@ TEST_CASE("alternative string type")
         CHECK_FALSE(j.contains(alt_json::json_pointer("/foo/bar")));
     }
 
+    SECTION("non-const operator[](json_pointer)")
+    {
+        // creating values must compile without string_t::begin()
+        alt_json j = alt_json::parse(R"({"a":{"b":[1]},"c":null})");
+        const alt_json before = j;
+        j[alt_json::json_pointer("/a/b/3/x/-/y")] = 1; // pad, append, create
+        j[alt_json::json_pointer("/a/z")] = 2;         // insert a key
+        j[alt_json::json_pointer("/c/0/q")] = 3;       // convert null
+        j[alt_json::json_pointer("/a/b/-")] = 4;       // append
+        CHECK(j.dump() == R"({"a":{"b":[1,null,null,{"x":[{"y":1}]},4],"z":2},"c":[{"q":3}]})");
+
+        // a failing access leaves the value unchanged
+        j = before;
+        CHECK_THROWS_AS(j[alt_json::json_pointer("/a/n/99999999999999999999999")], alt_json::out_of_range&);
+        CHECK(j == before);
+    }
+
     SECTION("operator/(std::size_t)")
     {
         // json_pointer::operator/=(std::size_t) must compile without string_t

tests/src/unit-bson.cpp

@@ -1311,7 +1311,7 @@ TEST_CASE("BSON declared document size bounds the content")
     // records the events that a SAX parser receives
     struct recorder : nlohmann::json_sax<json>
     {
-        std::string events;
+        std::string events {};
 
         bool null() override
         {

tests/src/unit-json_patch.cpp

@@ -2178,8 +2178,13 @@ TEST_CASE("JSON patch - behavior documented in the API reference")
 
     SECTION("patch() has the strong guarantee")
     {
-        const json doc = R"({"a": [1, 2], "o": {"k": "v"}})"_json;
+        // independent values, so that the comparison below checks the untouched document
-        const json before = doc;
+        const auto make_doc = []
+        {
+            return R"({"a": [1, 2], "o": {"k": "v"}})"_json;
+        };
+        const json doc = make_doc();
+        const json before = make_doc();
 
         const std::vector<json> failing =
         {
@@ -2202,7 +2207,7 @@ TEST_CASE("JSON patch - behavior documented in the API reference")
             CHECK_THROWS_AS(doc.patch(patch), json::exception&);
             CHECK(doc == before);
 
-            json copy = doc;
+            json copy = make_doc();
             CHECK_THROWS_AS(copy.patch(patch), json::exception&);
             CHECK(copy == before);
         }

tests/src/unit-json_pointer.cpp

@@ -29,9 +29,64 @@ using nlohmann::json;
 #include <map>
 #include <random>
 #include <sstream>
+#include <stdexcept>
 #include <string>
 #include <vector>
 
+namespace
+{
+// the non-const operator[] must throw for pointer and leave the value parsed
+// from init unchanged (strong exception safety)
+template<typename Json>
+void check_strong_guarantee(const char* init, const std::string& pointer, const char* message)
+{
+    CAPTURE(init)
+    CAPTURE(pointer)
+    Json j = Json::parse(init);
+    const Json before = Json::parse(init);
+    const typename Json::json_pointer ptr(pointer);
+    CHECK_THROWS_WITH_AS(j[ptr], message, typename Json::exception&);
+    CHECK(j == before);
+}
+
+// as above, for an index that fits into size_type, but not into an array
+template<typename Json>
+void check_strong_guarantee_length_error(const char* init, const std::string& pointer)
+{
+    CAPTURE(init)
+    CAPTURE(pointer)
+    Json j = Json::parse(init);
+    const Json before = Json::parse(init);
+    const typename Json::json_pointer ptr(pointer);
+    CHECK_THROWS_AS(j[ptr], std::length_error&);
+    CHECK(j == before);
+}
+
+template<typename Json>
+void check_strong_guarantee_cases()
+{
+    const char* const unresolved = "[json.exception.out_of_range.404] unresolved reference token '99999999999999999999999'";
+
+    // the first change converts a null value
+    check_strong_guarantee<Json>("null", "/99999999999999999999999", unresolved);
+    check_strong_guarantee<Json>("null", "/a/99999999999999999999999", unresolved);
+    check_strong_guarantee<Json>(R"({"a":null})", "/a/b/-/99999999999999999999999", unresolved);
+    // the first change inserts a key into an object
+    check_strong_guarantee<Json>("{}", "/a/99999999999999999999999", unresolved);
+    check_strong_guarantee<Json>(R"({"a":{"b":1}})", "/a/c/d/99999999999999999999999", unresolved);
+    // the first change pads or appends to an array
+    check_strong_guarantee<Json>(R"({"a":[1]})", "/a/5/99999999999999999999999", unresolved);
+    check_strong_guarantee<Json>(R"({"a":[1]})", "/a/-/x/1/99999999999999999999999", unresolved);
+    // an index that does not fit into size_type
+    check_strong_guarantee<Json>("null", "/x/18446744073709551615",
+                                 "[json.exception.out_of_range.410] array index 18446744073709551615 exceeds size_type");
+    // an index that fits into size_type, but not into an array
+    const auto too_large = std::to_string((std::numeric_limits<typename Json::size_type>::max)() - 1);
+    check_strong_guarantee_length_error<Json>("null", "/" + too_large);
+    check_strong_guarantee_length_error<Json>(R"({"a":[[1]]})", "/a/0/3/" + too_large);
+}
+} // namespace
+
 TEST_CASE("JSON pointers")
 {
     SECTION("errors")
@@ -557,6 +612,23 @@ TEST_CASE("JSON pointers")
             CHECK_THROWS_WITH_AS(j.at("/01"_json_pointer),
                                  "[json.exception.parse_error.106] parse error: array index '01' must not begin with '0'", json::parse_error&);
         }
+
+        SECTION("a failing operator[] leaves the value unchanged")
+        {
+            check_strong_guarantee_cases<json>();
+            check_strong_guarantee_cases<nlohmann::ordered_json>();
+        }
+
+        SECTION("a successful operator[] still creates the same values")
+        {
+            json j = {{"a", {1}}};
+            j["/a/3/x/-/y"_json_pointer] = 1;
+            CHECK(j == json::parse(R"({"a":[1,null,null,{"x":[{"y":1}]}]})"));
+            j["/a/-"_json_pointer] = 2;
+            j["/a/0"_json_pointer] = 3;
+            j["/b"_json_pointer] = 4;
+            CHECK(j == json::parse(R"({"a":[3,null,null,{"x":[{"y":1}]},2],"b":4})"));
+        }
     }
 
     SECTION("flatten")

tests/src/unit-json_view_builder.cpp

@@ -592,7 +592,7 @@ TEST_CASE("json_view node integer bits")
                 })
         {
             CAPTURE(v)
-            n.kind = static_cast<nlohmann::detail::view::node_kind>(0x5A);
+            n.kind = nlohmann::detail::view::node_kind::link;
             n.flags = 0xA5;
             n.extra = 0x1234;
             n.off = 0x89ABCDEFu;
@@ -601,7 +601,7 @@ TEST_CASE("json_view node integer bits")
             CHECK(n.len == static_cast<std::uint32_t>(v));
             CHECK(n.next == static_cast<std::uint32_t>(v >> 32));
             // the other fields are untouched
-            CHECK(static_cast<std::uint8_t>(n.kind) == 0x5A);
+            CHECK(n.kind == nlohmann::detail::view::node_kind::link);
             CHECK(n.flags == 0xA5);
             CHECK(n.extra == 0x1234);
             CHECK(n.off == 0x89ABCDEFu);

tests/src/unit-regression1.cpp

@@ -1499,6 +1499,23 @@ TEST_CASE("regression tests 1")
         CHECK(order.dump() == R"({"a":2,"m":{"y":1,"b":2},"new_key":3})");
     }
 
+    SECTION("operator[](json_pointer) when emplace() finds a key the lookup does not")
+    {
+        // a copied fifo_map restarts its timestamps, so a new key "k" gets
+        // the timestamp of "x": the lookup does not find "k", but emplace()
+        // finds "x" and inserts nothing; the pointer then continues with the
+        // member of "x", as basic_json::operator[] does
+        const my_json orig = my_json::parse(R"({"x":1})");
+        const std::array<const char*, 3> pointers = {{"/k/0", "/k/a/b", "/k/-"}};
+        for (const char* pointer : pointers)
+        {
+            CAPTURE(pointer)
+            my_json j = orig;
+            CHECK_THROWS_AS(j[my_json::json_pointer(pointer)], my_json::out_of_range&);
+            CHECK(j.dump() == R"({"x":1})");
+        }
+    }
+
     SECTION("issue #977 - Assigning between different json types")
     {
         foo_json lj = ns::foo{3};