* Keep the serializer conversion for objects whose keys cannot be converted
#5591 added a test converting nlohmann::json into a basic_json whose
string type cannot be constructed from std::string. That instantiates
convert_iteratively(), whose members.emplace_back(next.key(), ...) needs
exactly that key conversion, and broke the build of unit-alt-string.
Dispatch on the key's constructibility and leave such conversions to the
serializers, as the levels above the nesting bound already do (#3425).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix the remaining CI failures on develop
- unit-wstring: with a 16-bit wchar_t (Windows), a lone surrogate is
reported as the ill-formed byte 0xFF since #5704; the std::wstring
expectations still had the previous <U+0000>.
- ci_single_binaries: json_literals.hpp (#5610) and json.hpp include each
other on purpose, and IWYU, not following the cycle, asks to replace
json.hpp with json_fwd.hpp. Report its findings without failing the
build, as already done for json.hpp.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix the library warnings and noexcept specifications from the merged PRs
- binary_reader: rename the error_handler constructor parameter, which
shadowed the member (-Wshadow, -Wshadow-field-in-constructor; #5746)
- basic_json(copy_construct_tag, ...): declare it noexcept when copying
the base class is (GCC 16 -Wnoexcept; #5690)
- the scalar-on-left legacy comparison operators: noexcept only when
converting the scalar is, like their member counterparts (#5682, #5751)
- compare_leaves: use std::is_eq/is_lt/is_gt instead of comparing a
std::partial_ordering with 0 (-Wzero-as-null-pointer-constant; #5686)
- serializer: silence MSVC C4127 for the EnsureAscii template parameter
(#5741, #5746)
- clang-tidy: return the sanitized reference in binary_writer, take the
key of ordered_map::find_impl by const reference (#5727), and mark the
switches over parse_array_index (#5728)
- ordered_map: keep <memory> for std::allocator (IWYU)
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Split unit-conversions.cpp so MinGW can link it
clang 18 with the MinGW linker failed to link test-conversions_cpp17
("relocation truncated to fit: IMAGE_REL_AMD64_REL32"). As windows.yml
recommends, keep the objects small by splitting the test file.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix the tests added by the merged PRs for all CI configurations
- discard the results of dump() and from_*() in CHECK_THROWS with
utils::ignore_return_value (GCC -Werror=unused-result)
- give unit-bson's huge_string_t a default constructor (MSVC C2512,
GCC 5, clang 3.5)
- unit-disabled_exceptions: use the literals namespace when the global
UDLs are off (ci_test_noglobaludls; #5700)
- unit-binary_utf8_strict: expect the JSON pointer prefix with
JSON_DIAGNOSTICS (#5741)
- skip the tests that rely on exceptions under JSON_NOEXCEPTION
(#5678, #5732)
- clang-tidy and clang -Werror: static test data, CAPTURE(...);,
const-correctness, use-after-move alias, unused conversion operator,
a missing <iterator> include
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Title the macro examples and add JSON_STRICT_BINARY_UTF8 to the docset
The documentation style check requires "Example: ..." titles on pages with several examples (#5741, #5591) and a docset entry for every macro page.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Regenerate BUILD.bazel and nlohmann_json.natvis
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
#5746 added detail/output/error_handler.hpp and #5741 the json_abi_sbu8 ABI tag.
* Install libidn11 for the CMake 3.5.0 binary in ci_cmake_flags
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
#5733 moved ci_cmake_options from ubuntu:focal to ubuntu:24.04, which no longer ships libidn.so.11; the CMake 3.5.0 release binary links against it, so every ci_cmake_flags run has failed since. Install focal's libidn11 package for that matrix entry only.
* Suppress Infer's false STACK_VARIABLE_ADDRESS_ESCAPE in get_impl
get_impl() returns its local by value. A test added by the merged PRs instantiates it with a type Infer misreads, so ci_infer reported the 2021 code for the first time.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Document options to reduce compile times
Add an integration page that collects the ways to reduce compile times
with measurements: json_fwd.hpp in headers, JSON_NO_AUTOMATIC_UDLS,
explicit instantiation with extern template, modules, and precompiled
headers, and notes that JSON_NO_IO and JSON_USE_GLOBAL_UDLS have no
measurable effect.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Include only json_literals.hpp in the JSON_NO_AUTOMATIC_UDLS examples
json_literals.hpp includes json.hpp itself, so including both is redundant.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Add basic_json::as_base_class and document name conflicts with custom base classes
Members of basic_json hide members of a custom base class with the same
name, and future releases may add members that hide ones accessible
today. Document this in json_base_class_t and add as_base_class() to
reach hidden members without spelling out the cast.
Also make json_base_class_t a public member type. It was documented
since 3.12.0, but declared private, so users could not name it.
Supersedes #3899.
Co-authored-by: Raphael Grimm <1005058+barcode@users.noreply.github.com>
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Add as_base_class to the docset search index
New public members get an entry in docs/docset/docSet.sql (as done for
to_bon8/from_bon8 in #2998). Without it, the Dash/Zeal docset built from
the documentation cannot find basic_json::as_base_class.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Silence clang-tidy for the hidden type_name() in the base class test
ci_clang_tidy failed with readability-convert-member-functions-to-static
on base_class_with_hidden_members::type_name(). It must stay a
non-static member: the test shows that it is hidden by the non-static
basic_json::type_name() and reachable through as_base_class().
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Co-authored-by: Raphael Grimm <1005058+barcode@users.noreply.github.com>
* Document the macros that preview version 4.0 in the roadmap
List the macros that guard breaking changes planned to become the
default in version 4.0.0, and explain that 4.0.0 will be the sum of
these opt-in flags, which can be tried on the 3.x release train.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* List the deprecated functions removed in 4.0 in the roadmap
The roadmap lists what will be removed; the migration guide keeps the
examples for how to replace each item. Also mention the deprecated
(ptr, len) overloads of the from_* functions in the migration guide.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Wrap overlong line in cbor_tag_handler_t documentation
The line added in #5559 exceeds the 160-character limit enforced by the
documentation style check.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Add JSON_USE_LEGACY_DISCARDED_VALUE_COMPARISON to the 4.0 example
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Add JSON_STRICT_BINARY_UTF8 to the 4.0 roadmap
The macro comes from #5741: the CBOR, UBJSON, BJData, and BSON writers keep writing ill-formed UTF-8 unchanged in 3.x and are planned to check it by default in 4.0.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Remove unused private aliases from basic_json
The private aliases primitive_iterator_t, internal_iterator and
output_adapter_t are not used anywhere: iter_impl, binary_writer and
the tests refer to the detail:: names directly. As the aliases are
private, no user or derived class can depend on them. The
internal_iterator.hpp include stays because iter_impl needs it.
Part of #5724
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix meta()'s dead, syntactically invalid HP aCC branch
The HP aCC branch of basic_json::meta() was missing a semicolon and
has therefore never compiled; adding only a semicolon would also make
it throw type_error.305, since it assigned a plain string to
result["compiler"] and then indexed into it like the other branches
do into an object. Make the branch consistent with the others by
assigning an object with "family" and "version" keys, narrow the
condition to __HP_aCC (a C compiler cannot build this header-only
library), and fix meta.md, which documented the old (impossible)
plain-string behavior.
Part of #5724
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Stop the noexcept null constructor delegating to a throwing one
basic_json(std::nullptr_t) delegated to basic_json(value_t), whose
underlying json_value(value_t) constructor allocates for other types
and can therefore throw, which is why the noexcept had a
NOLINT(bugprone-exception-escape). The delegated-to constructor also
called assert_invariant() a second time. The default member
initializers of data already produce the same null state (a
value-initialized, i.e. zeroed, union with object == nullptr), so the
delegation and its NOLINT can simply be dropped.
Part of #5724
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Remove four cppcheck accessForwarded suppressions in the move constructor
basic_json(basic_json&&) built its base subobject with
std::forward<json_base_class_t>(other), so cppcheck saw the whole of
other as forwarded and flagged every subsequent access to it as
accessForwarded, three of them still marked "TODO check". Only the
base subobject is actually moved from; cast explicitly to the base
type instead, the way ordered_map already does, so cppcheck can tell
the two are unrelated. Behavior is unchanged: for a non-reference T,
std::forward<T>(x) is defined as static_cast<T&&>(x).
Part of #5724
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Remove stale cppcheck suppressions and name the local parser in parse()
Running the pinned cppcheck (ci_cppcheck's invocation) without
--inline-suppr across all configurations reports no syntaxError, no
ignoredReturnValue and no assertWithSideEffect, so the corresponding
suppressions in json_fwd.hpp, string_concat.hpp and
assert_invariant() no longer match anything (json_fwd.hpp's is kept,
since downstream users who run an older cppcheck against it could
still hit the warning it once silenced).
The three basic_json::parse() overloads still trigger a false-positive
accessMoved/accessForwarded because they build a temporary parser and
call .parse() on it in the same expression; giving that parser a name
makes the warning go away without changing behavior, and removes the
last of the inline suppressions on these functions.
Part of #5724
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Deduplicate the string/binary cleanup in the two erase() overloads
erase(pos) and erase(first, last) each carried a byte-identical
14-line block that destroys and deallocates a string or binary
value before resetting the type to null. That reimplements the
string/binary cases of json_value::destroy(), so any future change to
how those values are freed would have to be made in three places
instead of one. Both overloads now just call destroy() and reset the
union; for the other primitive types (boolean, numbers) destroy() is
a no-op, so behavior is unchanged.
Also fix erase(first, last)'s error-path branch hint, which used
JSON_HEDLEY_LIKELY where erase(pos), the iterator-range constructor,
and every other error path in the class use JSON_HEDLEY_UNLIKELY.
This only affects code layout, not semantics.
Part of #5724
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix stale and copy-pasted comments in basic_json
Several comments no longer match the code: the class invariant and
assert_invariant()'s doc still named the members m_value/m_type
(now m_data.m_value/m_data.m_type) and did not mention the binary
invariant that assert_invariant() already checks; the json_value note
and the get<PointerType>() @tparam list omitted binary_t even though
binary is a variable-length, pointer-stored type like the others; the
key-based value() overload's brief said "via JSON Pointer", which is
the other overload; and swap(binary_t&)/swap(binary_t::container_type&)
both carried "swap only works for strings", copied from swap(string_t&).
Comment-only change; behavior, the public API and the ABI are
unchanged. The private get_impl() doxygen and the emplace() comments
that border #5585's hunk are intentionally left alone.
Part of #5724
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Deduplicate the 16 copied from_cbor/msgpack/ubjson/bjdata/bon8/bson bodies
Each of the 16 binary deserialization overloads (from_cbor,
from_msgpack, from_ubjson, from_bjdata, from_bon8, from_bson, each in
an InputType&& and an iterator/sentinel version, plus the deprecated
span overloads of from_cbor/from_msgpack/from_ubjson/from_bson) had
the same body, differing only in the input_format_t value. Every copy
built a temporary binary_reader from std::move(ia) and called
sax_parse on it in the same expression, which also produced a
false-positive cppcheck accessMoved on all 16 lines and needed a
NOLINTNEXTLINE(hicpp-move-const-arg,performance-move-const-arg) on
the four span overloads.
Add a private from_binary_impl() helper that builds the reader as a
named local instead, and make each of the 16 overloads a one-line
forward to it. All public signatures, default arguments,
JSON_HEDLEY_WARN_UNUSED_RESULT and JSON_HEDLEY_DEPRECATED_FOR
attributes are unchanged, tag_handler keeps defaulting to
cbor_tag_handler_t::error for the non-CBOR formats (matching
binary_reader::sax_parse's own default), and the helper is placed in
the existing private section before the binary section banner rather
than between the from_* overloads, so from_binary_impl() itself does
not collide with #5688's insertion point. Collapsing the
from_bjdata/from_bon8 bodies into one-line forwards does rewrite the
"return result; }" context lines that #5688 inserts its two
deprecated overloads after, so that PR will need a small manual
rebase (reinserting its overloads after the new one-line bodies)
rather than applying cleanly.
Part of #5724
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Make insert(pos, basic_json&&) move its argument instead of copying it
insert(const_iterator pos, basic_json&& val) delegated to
insert(pos, val), but val is a named rvalue reference, so inside the
function it is an lvalue: the call always resolved to
insert(const_iterator, const basic_json&) and deep-copied the value.
This has been the case since the overload was introduced, in every
release. push_back(basic_json&&), by contrast, already moves.
Give the rvalue overload its own body with the same two checks
(type_error.309, invalid_iterator.202), then move the argument into a
local before inserting it. Moving into a local first, rather than
inserting std::move(val) directly, keeps this safe even when val
aliases an element of the same array (e.g.
arr.insert(arr.begin(), std::move(arr[1]))), since
std::vector::insert(pos, T&&) is not guaranteed to handle an argument
that aliases one of its own elements.
This is a deliberate, small behavior change: the moved-from argument
now ends up null afterwards, the same as after push_back(&&), instead
of keeping its old value unchanged. No signature changes, so the
public API and ABI are unaffected. Add unit-modifiers coverage for
the moved-from state and for self-aliasing insertion, both with and
without reallocation of the underlying array.
Part of #5724
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Deduplicate object key lookup and checked at() access
at()/find()/count()/contains()/const operator[]/erase_internal() each
repeated the raw object lookup (m_value.object->find(key)), and the
six at() overloads additionally repeated the type_error.304 check and
out_of_range.401/403 throw. Route them all through two new private
helpers, object_lookup()/object_at() (plus array_at() for the index
overloads of at()), templated on the constness of the receiver so one
body serves both the const and non-const overload. count() is left
untouched, since it already goes through object_t::count() rather
than a second find().
The at(KeyType&&) overloads used to forward the same key twice: once
into object->find() and again, on the not-found path, into the
string_t() conversion for the exception message. object_at() now
forwards it only into the lookup and reuses the (unmoved) key for the
message. clang-tidy 22 (Docker silkeh/clang:22) still flags that reuse
under bugprone-use-after-move/hicpp-invalid-access-moved even with the
single forward, since it cannot see that object_t::find() (a plain
std::map or ordered_map) never actually moves from its argument; add a
NOLINTNEXTLINE with that reasoning rather than avoid the pattern.
No signature, exception id/message, or set_parent() behavior changes.
Overlaps #5689, #5705, #5606, #5687 and #5585, which touch the same
hunks; whichever of this commit and those PRs lands second will need
a small rebase.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Deduplicate the lookup/default/throw body of value()
The six non-deprecated value() overloads each held a full copy of the
same body: the four key-based overloads looked up the key and either
returned the found element converted to the requested type or the
default value (throwing type_error.306 if this is not an object), and
the two json_pointer overloads did the same via
ptr.get_checked_or_null(), throwing type_error.306 unless
is_structured(). Replace the duplicated bodies with two private
helpers, value_member() and value_pointee(), that return a
const basic_json* (null when not found) and do the type check/throw
once each. Every value() overload now just picks between the found
pointer's get<T>() and the default.
Same signatures, template parameters, SFINAE conditions, exception id,
message and this context on every overload.
Overlaps #5689 (routes find() through lookup_key()) and #5705 (adds a
deleted integral-key value() next to these overloads); whichever of
this commit and those PRs lands second will need a small rebase.
#5724 item 4
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Deduplicate the null-to-container conversion into convert_null_to()
Nine sites wrote out the same "turn a null value into an empty array
or object" logic with two different idioms: operator[](size_type),
operator[](key_type), operator[](KeyType&&) and update() set m_type
then assigned m_value.array/object directly via create<T>(), while the
three push_back() overloads, emplace_back() and emplace() set m_type
then assigned m_value = value_t::array/object (going through
json_value's converting constructor and a temporary). Both idioms end
up calling create<T>() and produce the same state, just via a
different path; both also share a latent exception-safety bug, since
m_type is written before the (possibly throwing) allocation, so a
throwing allocator leaves m_type == array/object with a null pointer
behind it, violating the class invariant and crashing on the next
access to, or destruction of, the value.
Add a private convert_null_to(value_t) helper and call it from all
nine sites. Unlike the idioms it replaces, it allocates the container
first and only then writes m_type, so a throwing allocation leaves the
value as a valid null instead of a mistyped, half-constructed one;
verified with a throwing allocator (see unit-allocator.cpp's
bad_allocator) that j["x"] = ... on a null j now stays null, and no
longer trips assert_invariant()/crashes, when create<object_t>()
throws. Same allocator usage and assert_invariant() call as before,
otherwise.
Overlaps #5585, which reorders these same nine blocks for exception
safety; whichever of this commit and that PR lands second will need a
small rebase.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Deduplicate the linear key search in ordered_map
emplace, at, erase(key), count and find each repeated the same
"for (auto it = begin(); it != end(); ++it) if (m_compare(it->first,
key)) ..." loop (15 copies across their key_type and transparent
KeyType&& overloads), and both erase(key) overloads additionally
repeated the exception-sensitive in-place reconstruction (destroy,
placement-new, pop_back) used to remove an element while keeping the
const Key non-movable.
Add two private helpers: find_impl(Self&, KeyType&&), a static member
template that runs the search once for either constness of the
receiver, and erase_at(iterator), which keeps the existing
pop_back-based reconstruction instead of switching to erase()/resize()
(which would add a DefaultInsertable requirement). Route find, at,
count, emplace, insert(const value_type&) and both erase(key)
overloads through them.
Same signatures, is_usable_as_key_type constraints and exception
messages/types.
Overlaps #5609 and #5685, which both rewrite emplace (#5609 also
touches insert and adds private members at the end of the class);
whichever of this commit and those PRs lands second will need a
rebase.
#5724 item 6
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Re-enable bugprone-use-after-move/hicpp-invalid-access-moved
These two checks (and portability-template-virtual-member-function)
were disabled in #4489 (November 2024) "only removed to get the CI
going". portability-template-virtual-member-function is a separate,
still-open cleanup (#5725 item 3 on its own branch) and stays
disabled here; this commit only re-enables the move/forward checks
and cleans up what they flag on this branch.
The move constructor (json.hpp) already casts to the base type
instead of forwarding the whole object (#5724 item 9), so it no
longer trips either check. The at(KeyType&&) double-forward this
check used to flag was reduced to a single forward with the
now-unforwarded reuse annotated by a NOLINTNEXTLINE in #5724 item 3's
object_at() helper (clang-tidy 22 still flags that reuse even after a
single forward; see that commit's message). What is left here:
- from_json_inplace_array_impl(), from_json_tuple_impl_base() and the
std::pair overload of from_json_tuple_impl() forwarded j into every
j.at(...) call in a pack expansion or a pair of calls. at() has no
ref-qualified overloads, so the forward was a no-op; call j.at(...)
directly.
- container_input_adapter_factory::create() forwards container twice
on purpose, into begin() and end(), so both see the same value
category and produce matching iterator types. Annotate it with
NOLINTNEXTLINE and a comment instead of changing it.
- unit-class_parser.cpp's "move constructor resets the moved-from
value to npos" test still pointed at the pre-static_cast move
constructor by line number and mentioned the cppcheck-suppress
annotation that #5724 item 9 already removed; update the comment.
No behavior change anywhere in include/. Verified with clang-tidy 22.1.8
(Docker silkeh/clang:22, --platform linux/amd64) against a TU including
json.hpp with the repo's .clang-tidy: bugprone-use-after-move and
hicpp-invalid-access-moved report nothing unsuppressed.
Overlaps #5737 (open PR for the rest of #5725 item 3: the
from_json.hpp/input_adapters.hpp cleanup above, and
portability-template-virtual-member-function), which currently keeps
both checks disabled pending this move-constructor change; whichever
of this commit and that PR lands second will need a small rebase of
.clang-tidy.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Make convert_null_to() take the container type as a template argument
Passing array_t or object_t instead of a value_t makes an invalid target
a compile error instead of a runtime assertion, and removes the branch.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Make cross-string-type basic_json conversion explicit without implicit conversions
The converting constructor from another basic_json specialization was
always implicit, so a value with a different string_t (std::wstring, a
string with a custom allocator, ...) silently converted into a temporary,
e.g. when passed to a function taking const nlohmann::json&. Such
conversions do not produce correct values (#3425), and
JSON_USE_IMPLICIT_CONVERSIONS=0 did not catch them.
When JSON_USE_IMPLICIT_CONVERSIONS is 0, the constructor is now explicit
if the string types differ. Specializations sharing a string type (json
and ordered_json, different serializers or object maps) stay implicitly
convertible, so the NLOHMANN_DEFINE_TYPE_* macros keep working with
nested json members. get<BasicJsonType>() constructs explicitly and
works in both modes.
Fixes#2649.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Construct explicitly in get_to() and to_json(std::optional)
With JSON_USE_IMPLICIT_CONVERSIONS=0 the conversion from a basic_json
with a different string type is now explicit, but two library paths
still assigned such a value implicitly and failed to compile inside the
library:
- get_to() with a basic_json target (the #2175 overload) did
`v = *this`, so json(42).get_to(alt_json&) broke although
get<alt_json>() works.
- to_json(BasicJsonType&, const std::optional<T>&) is constrained on
std::is_constructible (which accepts the explicit constructor) but did
`j = *opt`, so converting a std::optional<alt_json> into a json broke.
Both now construct the value explicitly, as get_impl() already does.
Also replace static_cast<bool>(JSON_USE_IMPLICIT_CONVERSIONS) with a
comparison: clang-tidy's modernize-use-bool-literals rejected the cast
of the integer literal the macro expands to, failing ci_clang_tidy.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Since #4958, a key type implicitly convertible to std::string_view was
accepted by is_usable_as_basic_json_key_type without checking that the
object's comparator can actually compare object_t::key_type with that
key type. The key was then forwarded unchanged to the underlying map,
so const operator[], at, find, count, contains, erase and value failed
to compile (a hard error inside <map>) for a key convertible only to
std::string_view, and value() rejected such keys outright. For keys
convertible to both std::string and std::string_view, the KeyType&&
templates now won overload resolution over the object_t::key_type
overloads and then failed the same way, a regression from 3.12.0. Only
the non-const operator[] worked, because it uses emplace(), which
constructs a std::string from the key explicitly. ordered_json was not
affected, since ordered_map checks comparability itself.
Add a trait, is_string_view_convertible_key_type, that recognizes a key
type that is convertible to std::string_view but not directly
comparable with the object's key type, provided std::string_view itself
is comparable with it. at(), operator[], find(), count(), contains(),
erase() and value() now route such keys through a new lookup_key()
helper that converts them to std::string_view before they reach the
object, matching how the object's transparent comparator already
supports std::string_view lookups.
Fixes#5663.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Accept lvalues in ordered_map::emplace's value parameter
ordered_map::emplace(key, value) took the mapped value only by T&&, an
rvalue reference rather than a forwarding reference, so
ordered_json::emplace("a", value) failed to compile whenever value was
an lvalue or a const lvalue, even though the same call compiles for
json (whose object_t is std::map, with a variadic emplace). Turn the
value parameter into a separately-deduced forwarding reference,
constrained with std::is_constructible so the overloads still only
accept something convertible to the mapped type. std::map-compatible
semantics are unchanged: emplace still does nothing if the key already
exists.
Open PR #5609 also touches ordered_map.hpp (moving values on vector
growth); this change only touches the two emplace() overloads and
should not conflict.
Fixes#5673.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Avoid astyle's padding in ordered_map::emplace's template headers
Use detail::conjunction instead of && and drop the redundant V&& in detail::is_constructible, so astyle keeps the usual template formatting. Addresses review comment by @gregmarr.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
With JSON_USE_LEGACY_DISCARDED_VALUE_COMPARISON=1 and C++20, a scalar on
the left-hand side of <= or >= (e.g., `1 <= discarded`) yielded false
instead of the documented true. The C++20 legacy block only had member
operators, which are only candidates when the basic_json is the left
operand; for a scalar on the left, overload resolution picked the
candidate rewritten from operator<=>, which does not emulate the legacy
behavior. The C++17 branch already has scalar-on-the-left friend
overloads for <= and >=; add the equivalent pair to the C++20 legacy
block.
Added a regression test to tests/src/unit-comparison.cpp covering all
four operand orders for both operators.
Fixes#5665.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
The const operator[] overloads are unchecked by design, and a missing key
or index is undefined behavior. The key overload guards this with a
runtime assertion, but the index overload did not, although the element
access documentation says an assertion fires in both cases. The const
JSON pointer overload inherits both through json_pointer::get_unchecked(),
so a pointer to a missing array index read out of bounds even in debug
builds, and its documentation promised out_of_range.404 for any pointer
that cannot be resolved.
Add JSON_ASSERT(idx < size()) to const operator[](size_type), which also
covers the index leg of the const JSON pointer overload. Document the
undefined behavior for the const JSON pointer overload in operator[].md
and in the runtime assertions page. Release builds are unchanged.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Document GCC < 11 incomplete-type error with optional members (#3669)
With GCC 10 and older in C++11/14 mode, a free to_json() for a type
holding an optional<Dummy> (Dummy constructible from json) fails with
"invalid use of incomplete type detector<...to_json_function...>".
ADL for Dummy finds the unrelated to_json and closes an instantiation
cycle through optional's converting constructor. The same error
reproduces without the library, so it can't be fixed here.
Add a FAQ entry explaining the cause and the hidden-friend workaround,
recommend hidden friends in the arbitrary types docs, and add a
regression test that keeps the workaround compiling on GCC 7-10.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Use the #3669 fixture's optional member in to_json
Issue3669Holder::d is never read, so clang's -Weverything -Werror build
(ci_test_clang) fails with -Wunused-private-field. Reference the member
in the hidden-friend to_json; this does not affect the instantiation
cycle the fixture reproduces.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
The comparison operators taking a scalar (==, !=, <, <=, >, >=, and
C++20's <=>) convert the scalar to a basic_json and compare, but were
unconditionally noexcept. When that conversion throws, the program
called std::terminate instead of propagating the exception, e.g. when
comparing a json with a string literal under memory pressure
(std::bad_alloc) or with an enum value not mapped by
NLOHMANN_JSON_SERIALIZE_ENUM_STRICT (out_of_range.410). clang-tidy
22.1 reports the latter as bugprone-exception-escape.
Declare the 16 scalar overloads
noexcept(std::is_nothrow_constructible<basic_json, ScalarType>::value):
they stay noexcept for numbers, Booleans, nullptr, and plain enums,
and are noexcept(false) for strings and enums whose to_json may throw.
The comparisons of two basic_json values are unchanged.
Restore the strict-enum comparisons removed from unit-conversions.cpp
in the previous PR, check that comparing an unmapped strict enum now
throws, and pin the new exception specifications in unit-noexcept.cpp.
Document the exception safety of overload (2) on all seven operator
pages. Ran make amalgamate.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Conflicts in See also lists (docs/mkdocs/docs/api/basic_json/operator_ne.md), where develop (#5638) and this branch both edited: kept develop's entries and added this branch's basic_json_view links.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Conflicts in See also lists (docs/mkdocs/docs/api/basic_json/begin.md,docs/mkdocs/docs/api/basic_json/cbegin.md,docs/mkdocs/docs/api/basic_json/cend.md,docs/mkdocs/docs/api/basic_json/end.md,docs/mkdocs/docs/api/basic_json/type_name.md), where develop (#5638) and this branch both edited: kept develop's entries and added this branch's basic_json_view links.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Conflicts in the See also lists of nine basic_json pages, is_discarded.md, and features/index.md, where develop (#5638) and this branch both added entries: kept both. Ran make amalgamate.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Conflicts:
- number_parse.hpp: kept this branch's float parser, which replaces the
Eisel-Lemire code that develop's side changed (#5750 made its digit
counter unsigned; this parser has no such counter, and it compiles
cleanly with GCC's -Wstrict-overflow=5).
- number_handling.md, template_parameters.md: kept this branch's
description of the conversion and added develop's "Before version
3.13.0" sentence.
Ran make amalgamate.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Review and extend the documentation, and check it in CI
A review of all documentation pages found factual errors, dead links,
missing cross-references, and gaps in examples. This fixes them and adds
checks so the same problems are caught automatically.
Fixes:
- wrong signatures and version histories (operator!= C++20 member,
binary() subtype type, get<PointerType>(), JSON_NO_THREAD_LOCAL, ...)
- stale descriptions (number parsing since #5283, UBJSON table, SAX
example that no longer compiled, tsl::ordered_map advice)
- dead internal and external links; repology.org badges (the domain is
suspended) replaced by badges that query the registries directly
- deprecation notes link the migration guide; the guide itself fixed
Additions:
- "See also" sections, cross-references, 25 runnable examples, 12
Mermaid diagrams, new API pages for json_pointer::operator<=> and
byte_container_with_subtype::operator==/!=
- landing page, guides for untrusted input and performance
- "unreleased" badge after versions newer than the latest release
Checks:
- strict documentation build (broken links/anchors fail it); CI and
the publish workflow fetch the full history the build needs
- weekly external link check, Mermaid syntax check in CI
- check_structure.py: example titles, heading levels, alt texts,
header links, docset index coverage; its unused-example check works
again
- all examples produce the same output on every platform
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Keep the customer links that could not be fixed
A dead link on the customers page is still the evidence of where the
use of the library was documented. Keep the original URLs of the entries
without a working replacement (Marne, Cisco Webex Desk Camera, Philips
Hue, CyberArk) and exclude exactly these URLs from the link check.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Correct the duplicate-key recipe's claim about SAX positions
The SAX interface's key() receives no position either; only parse_error()
does. Also note that the recipe does not report the path to the repeated
key (see discussion #5085).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Say the library is available as a single header and mention json_fwd.hpp
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Correct documentation errors found while hunting for bugs
- patch/patch_inplace: list the JSON pointer errors parse_error.106-109
and out_of_range.402/404, and quote the actual parse_error.105 message.
- unflatten: list parse_error.106/107/108 and out_of_range.404.
- to_bson: list out_of_range.415 (binary subtype above 255) and note
that 412 and 415 are new in 3.13.0.
- to_string: state that string_t must be convertible to std::string, also
in the StringType requirements table.
- JSON Lines: a `while (input >> j)` loop also throws after the last value
for concatenated JSON values; show a loop that works for both.
- BON8: a string gets 0xFF only if nothing follows it in the message; a
string at the end of an array or object is ended by 0xFE.
- custom_string_type.hpp: add operator+=(char), which the "Always
required" list asks for (json_pointer::to_string, flatten, unflatten,
and diff did not compile), and an ADL int_to_string for diff and items.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Cache the release headers with functools.lru_cache
Codacy (Pylint) flagged the mutable default argument that header() used
as its cache. functools.lru_cache keeps the same memoization without it.
The script's output is unchanged.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Conflicted only in tests/src/unit-class_lexer.cpp, where develop's #5737
lint fix (CAPTURE(x); -> CAPTURE(x)) collided with this PR's rewrite of
the Eisel-Lemire float tests; kept the PR's new tests and applied the
lint-fixed CAPTURE style. single_include regenerated via make amalgamate.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix the ci_cmake_flags wiring so every option is checked
The CMake 3.31.6 flag list referred to itself before it was defined,
so only JSON_BuildTests was checked with that version. The targets for
the CMake running the build ("_2") were created but never added to
ci_cmake_flags, and the three versions shared one build directory.
JSON_StrictNulHandling was not in the list at all.
Use the 3.5.0 list for 3.31.6, add JSON_StrictNulHandling, and create
one ci_cmake_flag_<flag> target per option for the running CMake with
its own build directory. Also use the function parameter in the
COMMENT, refresh the stale version comment, and let ci_clean remove
the downloaded cmake-<version> directories instead of the long-gone
cmake-3.5.0-Darwin64.
Part of #5715
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix ci_test_clang_libcxx_cxx* jobs silently building without warnings
CMake only seeds CMAKE_CXX_FLAGS from the CXXFLAGS environment variable
when the cache entry is unset, so the explicit -DCMAKE_CXX_FLAGS="-stdlib=libc++"
argument made it ignore CXXFLAGS="${CLANG_CXXFLAGS}" entirely. The six
ci_test_standards_clang (..., libcxx) jobs therefore compiled without
-Weverything/-Werror while their libstdc++ siblings did use them.
Pass -stdlib=libc++ through the same CXXFLAGS value instead of a separate
-D argument, and give the target its own build directory
(build_clang_libcxx_cxx${CXX_STANDARD}) so it no longer shares a CMake
cache with the libstdc++ variant. Suppress the resulting
-Wthread-safety-negative finding from libc++'s std::mutex annotations,
which fires on doctest's reporters in this translation unit only.
Part of #5715
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Remove the no-op AppVeyor with_win_header job
The with_win_header matrix entry patched Windows.h into
single_include/nlohmann/json.hpp before building, but JSON_MultipleHeaders
has defaulted to ON since #3532 (2022-06), so CMakeLists.txt points the
tests at include/ and the patched single header is never compiled. The
job has been a no-op VS2015 build since then.
Windows.h coverage already exists through tests/src/unit-windows_h.cpp
(#3631), which runs in every MSVC job. Delete the dead matrix entry and
its before_build steps, and cite unit-windows_h.cpp from the QA page.
Part of #5715
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Remove unused ci_oclint and ci_pvs_studio targets
No workflow invokes ci_oclint, ci_pvs_studio, or their tool discovery.
ci_oclint also had a side effect on every JSON_CI configure: it copied
the single header into src_single/all.cpp and added an add_executable()
for it without EXCLUDE_FROM_ALL, so a plain build compiled a 1.2 MB
translation unit that only that unused target consumed. ci_pvs_studio
duplicates the Makefile's pvs_studio target, which is kept.
Also drop the duplicate --check-level=exhaustive flag passed twice to
the same ci_cppcheck invocation.
Part of #5715
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Stop Dependabot from proposing astyle bumps
astyle is deliberately pinned at 3.4.13 because newer versions reformat
unrelated lines and this version defines the formatting that
check_amalgamation.yml enforces. Without an ignore rule, Dependabot
keeps opening PRs for every new astyle release (most recently #4580,
#4942, #5445, #5448), each of which fails the amalgamation check and
gets closed unmerged.
Part of #5715
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Move Linux arm64 CI from dead Cirrus CI to ubuntu-24.04-arm
Cirrus CI stopped reporting check runs on develop sometime after
d10879bca (2026-05-26); every commit since has only github-actions
check runs, so .cirrus.yml silently lost its only consumer while
README.md, FILES.md and the QA page kept advertising the coverage.
Add a ci_test_arm64 job to ubuntu.yml using the same pinned
actions/checkout and lukka/get-cmake actions as the other jobs, on the
native ubuntu-24.04-arm runner, with a step that confirms uname -m
reports aarch64. Delete .cirrus.yml and its README badge and FILES.md
section, and update the QA page's arm64 row.
Part of #5715
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Make scan-build fail on findings and drop irrelevant checkers (#5715 item 4a)
ci_clang_analyze ran scan-build without --status-bugs, so the job
passed whenever the ninja build succeeded, no matter what the
analyzer found ("No bugs found" in a green run gave no signal either
way). It is also missing --use-analyzer=${CLANG_TOOL}, so scan-build
picks whichever clang happens to be first on PATH inside the
silkeh/clang:dev container instead of the one this file already
selected and versioned.
Add --status-bugs and --use-analyzer=${CLANG_TOOL} to the scan-build
invocation. While here, drop the osx.*, webkit.*, fuchsia.*, and
optin.mpi.* checkers from CLANG_ANALYZER_CHECKS: none of them apply
to this portable C++ library, and leaving them enabled only adds
noise once the job can actually fail on a finding.
The job is currently clean (0 bugs), so this alone does not surface
any new finding; it only makes the existing "no bugs found" result
authoritative. This is 4a of 3 independent steps in #5715 item 4;
4b (Infer) and 4c (IWYU) still need their existing findings triaged
before --fail-on-issue/-Xiwyu --error can be added, and are handled
in separate commits.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Deduplicate the amalgamation/format check's file set and add BUILD.bazel (#5715 item 5)
The amalgamation/format check existed three times with three different
file sets: the Makefile's pretty/check-amalgamation, the pull_request-only
check_amalgamation.yml workflow, and the ci_test_amalgamation CMake target
that also runs on direct pushes to develop/master/release/*. The CMake
target's glob was a strict subset of the workflow's (missing the
docs/mkdocs/docs/examples/*.hpp headers, tests/abi/, tests/cmake_*/project/,
tests/cuda_example/, tests/fmt_formatter/, and tests/module_cpp20/), and it
never checked BUILD.bazel at all, so a misformatted file in any of those
paths, or a stale BUILD.bazel, could reach develop through a direct push
even though the PR-only workflow would have caught it.
Make ci_test_amalgamation glob the same roots (docs/mkdocs/docs/examples,
include, tests) and extensions (*.hpp, *.cpp, *.cu) as check_amalgamation.yml,
excluding tests/thirdparty/ and tests/abi/include/nlohmann/ the same way, and
regenerate and diff BUILD.bazel next to json.hpp/json_fwd.hpp. Also add
docs/mkdocs/docs/examples/*.hpp to the Makefile's pretty/pretty_format
targets, which were missing the four custom_*_type.hpp example headers, and
drop the stale "called by Travis" comment on check-amalgamation (Travis is
gone; nothing currently calls that Makefile target from CI).
Leaves the workflow itself untouched: it deliberately runs amalgamate.py
from a fresh develop checkout so a PR cannot change the tool that checks it.
Overlaps #5610 and #5621, which each add a new amalgamated header and touch
the same INDENT_FILES/ci_test_amalgamation/check_amalgamation.yml hunks.
Verified with `make check-amalgamation` on this branch: clean, no diff.
#5715 item 5.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Download prebuilt CMake binaries on Linux x86_64 instead of building from source (#5715 item 6)
ci_get_cmake() downloaded the source tarball of CMake 3.5.0, 3.31.6, and
4.0.0 and compiled each one completely (including CMake's own test
helpers) with -DCMAKE_POLICY_VERSION_MINIMUM=3.5 as a workaround for
building old CMake with a newer one. On CI this made the
ci_cmake_options (ci_cmake_flags) job take about 11 minutes, most of it
spent building CMake itself, even though Kitware has published
ready-to-run Linux x86_64 archives for all three of these releases
since 3.20 (lowercase platform name).
On Linux x86_64, download and unpack the prebuilt
cmake-<version>-linux-x86_64.tar.gz archive instead and point the
existing ${var} output at its bin/cmake, skipping the configure/build
steps and CMAKE_POLICY_VERSION_MINIMUM entirely. Keep the previous
source build as a fallback for any other platform (macOS, Linux
aarch64), since Kitware does not publish binaries for every
CMake/platform combination this project might build on.
Verify the downloaded archive against Kitware's own published checksum
before unpacking it: download cmake-<version>-SHA-256.txt alongside the
archive and run `sha256sum -c` on the matching line. A CI job that wgets
and untars a binary from a release page with no integrity check is a
supply-chain gap; Kitware has published this file for every release
since 3.20, so checking it costs one extra download and one grep.
As a separate, mechanical change: the ci_cmake_options job's container
only needed to stay on ubuntu:focal for the source build's
libssl-dev dependency and its own aging toolchain; now that the
Linux/x86_64 path never compiles CMake, drop libssl-dev from its apt
install line and move the job to ubuntu:24.04 (Ubuntu 20.04 left
standard support in May 2025). ci_clean already removes the
cmake-3.5.0/cmake-3.31.6/cmake-4.0.0 directories from the #5715 item 3
fix, and the prebuilt path reuses those same directory names, so no
further cleanup changes are needed.
Overlaps #5598, which edits the same ci_cmake_options matrix line in
ubuntu.yml; a rebase may be needed once that lands.
Verified locally: `cmake -S . -B build -DJSON_CI=On` configures cleanly
on macOS/arm64 (source-build fallback branch) and on Linux/x86_64 in an
ubuntu:24.04 Docker container (47 `ci_cmake_flag_*` targets generated,
one built and run successfully); `.github/workflows/ubuntu.yml` still
parses as valid YAML; downloaded the real v3.31.6 Linux x86_64 archive
and SHA-256 file from Kitware and confirmed the `grep | sha256sum -c`
pipeline both accepts the genuine file and is anchored to the exact
filename (not a prefix match).
CI must confirm: the prebuilt-binary path actually runs on the
ubuntu-latest/ubuntu:24.04 x86_64 runner, all `ci_cmake_options`
entries still pass with the new container's GCC, and the job's
runtime drops from roughly 11 minutes.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Make Infer fail on findings, with a type-level baseline for the ~174 pre-existing ones (#5715 item 4b)
ci_infer ran `infer run` without --fail-on-issue, so the job passed
regardless of what Pulse found; the last recorded run (35829411620,
commit 1054b2097) logged "Found 174 issues" and still went green.
report.txt was also never uploaded, so the full finding list was only
ever visible in the truncated 5-issue console excerpt.
Add a repository-root .inferconfig (auto-discovered by Infer; passing
--project-root on the `infer run` invocation makes sure it is found
even though the analysis runs from build/build_infer) that sets
fail-on-issue and disables the six PULSE issue types that made up all
174 findings in that run: PULSE_UNNECESSARY_COPY_ASSIGNMENT (129),
PULSE_UNNECESSARY_COPY (22), PULSE_UNNECESSARY_COPY_INTERMEDIATE (15),
PULSE_RESOURCE_LEAK (5), PULSE_CONST_REFABLE (2), and
PULSE_UNNECESSARY_COPY_OPTIONAL (1).
This is a deliberate, narrower fix than "triage and fix everything in
this PR": the visible sample is entirely doctest-macro copies in test
code (for example tests/src/unit-algorithms.cpp:141 and
tests/src/unit-bjdata.cpp:3706), but 169 of the 174 findings were never
uploaded anywhere and this PR cannot respectably claim to have fixed
issues it never saw, including the resource-leak and const-refable
ones that are the most likely to be genuine bugs. Disabling by issue
type is a coarser baseline than a per-finding one (Infer has no
built-in per-finding baseline short of the two-run `infer reportdiff`
workflow, which this repository does not have the CI infrastructure
for), but it has the same effect today: the job goes from always green
to green-only-when-clean-of-everything-else, so CI now fails the
moment a *new* issue type appears, and report.txt is uploaded as a
workflow artifact on every run (including failures) so the six
disabled types can be triaged and re-enabled incrementally in follow-up
PRs.
#5715 item 4b. 4a (scan-build) and 4c (IWYU) are handled in separate
commits.
Verified: .inferconfig parses as JSON, ubuntu.yml still parses as
YAML. Infer itself is not available in this environment (v1.3.0 tar.xz
requires a Linux x86_64 runner), so CI must confirm that `infer run
--project-root ... -- make` picks up .inferconfig, that fail-on-issue
takes effect, and that the six disabled types actually suppress the
existing findings without also hiding an unrelated new one.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix IWYU findings for json.hpp/json_fwd.hpp/ordered_map.hpp and make CI fail on new ones (#5715 item 4c)
ci_single_binaries ran IWYU via CMake's CXX_INCLUDE_WHAT_YOU_USE launcher
property, which only printed "Warning: include-what-you-use reported
diagnostics" without failing the build: CMake's own __run_co_compile
wrapper does not propagate the launched tool's exit code, so even
`-Xiwyu --error` could never fail `cmake --build` this way. Verified
this empirically by injecting a deliberately-unused #include and
confirming the build still exited 0.
Fix the findings from the last recorded run (issue #5715 item 4, log
35829411620):
- ordered_map.hpp: add <new> (placement new) and
nlohmann/detail/abi_macros.hpp; drop <memory> (std::allocator is
still visible transitively via <vector>, confirmed by full local and
containerized test suite runs).
- json_fwd.hpp: drop <memory> (same reasoning). Keep every forward
declaration IWYU wanted removed (adl_serializer, basic_json,
json_pointer, ordered_map): this file's only job is to forward-declare
them for downstream users, so "nothing in this TU uses them" is
expected, not a real finding. Mark each with `// IWYU pragma: keep`.
- json.hpp: add <cmath>, <cstdint>, <set>, <type_traits>,
<unordered_map>, and the detail/abi_macros.hpp, detail/input/json_sax.hpp,
detail/meta/detected.hpp, thirdparty/hedley/hedley.hpp includes IWYU
says it needs. Do NOT remove adl_serializer.hpp,
detail/conversions/from_json.hpp, detail/conversions/to_json.hpp,
detail/macro_unscope.hpp, or ordered_map.hpp as IWYU suggests: nothing
else in include/nlohmann includes adl_serializer.hpp or
ordered_map.hpp, so basic_json<>'s own default template arguments
(JSONSerializer = adl_serializer, and ordered_json = basic_json<ordered_map>)
would lose their complete type; detail/macro_unscope.hpp is what
undoes the JSON_* macros detail/macro_scope.hpp defines earlier in
this same file, and removing it leaks those macros into every
translation unit that includes <nlohmann/json.hpp>. Verified by
actually removing them in a scratch test: the header still "compiles"
stand-alone but ordered_json and every macro-using translation unit
break. Marked each `// IWYU pragma: keep`.
Enforce it with `iwyu_tool` (ships with IWYU, e.g. as /usr/bin/iwyu_tool
on Debian/Ubuntu) instead of relying on the launcher property: it reads
compile_commands.json (now exported project-wide under JSON_CI) and
does return a real exit code for its own analysis, independent of
CMake's wrapper. ci_single_binaries now runs it over every
src_single/*.cpp with `-Xiwyu --error`, so a *new* finding fails CI.
json.hpp itself is excluded from that hard gate: even after every fix
above, IWYU's suggestion for one remaining symbol (a container
`swap, operator!=` used somewhere via a templated comparator) is not
deterministic — repeated, otherwise-identical containerized runs
reported <set>, then <unordered_map>, then <map> as "the" header to
add/remove for the exact same source. Gating a whole CI job on a
nondeterministic suggestion would make ci_single_binaries flaky rather
than informative, so json.hpp keeps the existing informational warning
(still shown during its normal compile) without failing the build on
it. Every other one of the ~50 single-header checks is included in the
hard gate.
#5715 item 4c. 4a (scan-build) and 4b (Infer) are separate commits.
Verified: full local ctest suite (129/129) and the ci_single_binaries
target itself both green in a containerized silkeh/clang:dev run
(matching the actual CI job) after this fix; a deliberately-reintroduced
unused #include in ordered_map.hpp was confirmed to fail
`cmake --build ... --target ci_single_binaries` (exit 2) with this
change, and to pass without it, on the same container/IWYU version CI
uses. `make check-amalgamation` is clean. Compiled with Clang and GCC
at -std=c++11/14/17/20 locally with no new warnings.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix math formulas rendering as raw TeX in the published docs
The privacy plugin self-hosts MathJax 2.7.0 but drops its
?config=TeX-MML-AM_CHTML query string, so the rehosted script loads no
input jax and the 10 formulas across 6 pages render as raw TeX to
readers. Remove pymdownx.arithmatex and the MathJax extra_javascript
entry, and rewrite the formulas in plain HTML (<sup>, <i>) instead.
This also drops a nine-year-old third-party script from every page.
Part of #5718
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix publish_documentation triggers and persist unneeded git credentials
publish_documentation.yml only triggered on docs/mkdocs/** pushes, but
the site also embeds .github/CODE_OF_CONDUCT.md, CONTRIBUTING.md,
SECURITY.md, cmake/{clang,gcc}_flags.cmake, .clang-tidy,
tools/astyle/.astylerc and tests/fmt_formatter/project/main.cpp via
pymdownx.snippets, so changes to those files never republished the
site. Extend the path filter to cover them, and switch runs-on from
the long-pinned ubuntu-22.04 to ubuntu-latest to match
ci_test_documentation.
Also add persist-credentials: false to the checkouts in ci_icpx and
ci_nvhpc (ubuntu.yml) and msvc-vs2026/msvc-arm64 (windows.yml), none
of which pushes with git, matching every other checkout in these
workflows.
Part of #5718
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix example build: broken debug echo, deprecations hidden for everything
docs/Makefile's debug echo used a space instead of a comma in
$(call cxx_standard ...), so it always printed an empty standard.
Every example was also compiled with -Wno-deprecated-declarations,
which would silently hide an accidental deprecated-API call in any of
them. Factor the duplicated compile flags into EXAMPLE_CPPFLAGS/
EXAMPLE_WARNFLAGS, build with -Werror=deprecated-declarations by
default, and only allow the three examples that intentionally
document deprecated API (the DEPRECATED_EXAMPLES list) to suppress it.
Part of #5718
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix check_structure.py NOLINT parsing and an off-by-one report line
`line.strip("<!-- NOLINT")` followed by `.strip(" -->")` strips any of
the characters in those sets from both ends, not a literal prefix; it
only happened to work for "Examples". A NOLINT'd section name
starting with N, O, L, I or T (e.g. "Notes", "Template parameters",
"Iterator invalidation", "Literals") was silently mangled, so the
suppression did not apply and the checker could report a spurious
missing/misordered section. Parse the comment with a regex instead.
The same fragile strip() pattern was used for heading text; replace it
with a plain prefix slice. Also fix the admonition_title report, which
used the 0-based line index while every other report in the file uses
lineno+1.
Part of #5718
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix docset README's non-existent make target and stale fallback URL
The README told readers to run `make nlohmann_json.docset`, but the
Makefile's targets are `all`, `JSON_for_Modern_C++.docset` and
`install_docset_zeal`; the documented command has failed with "No
rule to make target" since #2967 (2021). Point the README at the real
target and folder name. Info.plist's DashDocSetFallbackURL also still
pointed at the old nlohmann.github.io/json/ URL instead of the
canonical https://json.nlohmann.me/ from mkdocs.yml's site_url.
Leave list_missing_pages/list_removed_paths alone: they may become
redundant once #5638's check_docset() lands, which is a follow-up.
Part of #5718
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Remove two leftover Doxygen-era .link files from the examples directory
parse__iterator_pair.link and parse__pointers.link each held only a
Wandbox "online" permalink from the old Doxygen docs. #3071 deleted
every other .link file in 2021; these two came in through a parallel
PR (#3100) and were never referenced by any page, script or config.
Part of #5718
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix MacPorts CMake example to include its own snippet files
The MacPorts "Example: CMake" block included
integration/homebrew/example.cpp and
integration/homebrew/CMakeLists.txt instead of the MacPorts files
right next to it, a copy-paste slip from the Homebrew section. Nothing
referenced integration/macports/CMakeLists.txt as a result. The page
rendered correctly only because the homebrew, macports and
vcpkg/CMakeLists.txt snippets are byte-identical, so a future edit to
the MacPorts files would not have shown up on the page.
Part of #5718 item 6
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Do not persist git credentials in publish_documentation's checkout
The checkout step in publish_documentation.yml left the default
persist-credentials: true, so GITHUB_TOKEN stayed writable in
.git/config for the rest of the job (zizmor's artipacked finding). The
Deploy documentation step authenticates through its own github_token
input to peaceiris/actions-gh-pages and does not push with the
checked-out credentials, so persist-credentials: false is safe here,
matching every other checkout in the workflow set.
Overlaps #5638, which edits this same checkout step (adds
fetch-depth: 0); expect a rebase conflict there.
Part of #5718 item 2
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Replace list_missing_pages/list_removed_paths with a comm(1)-based diff
The docset Makefile's list_missing_pages ran one sqlite3 query per
mkdocs page, and list_removed_paths nested a loop over all mkdocs pages
inside a loop over all docset index paths (O(n*m) shell iteration).
Issue #5718 item 5 suggested removing or reducing these targets once
#5638's check_docset() lands, but that PR is still open and covers only
API pages and macros, not the full page set these targets check.
Replace the loops with two sorted path lists (DOCSET_PAGE_PATHS from
mkdocs' markdown sources, DOCSET_INDEX_PATHS from the built docset
index) compared with a single comm(1) call each, verified to produce
output identical to the old loops against the current docSet.dsidx.
The sed expression used '#' as its delimiter, which GNU Make reads as
a comment character even inside a variable assignment, truncating the
line and orphaning the closing paren of $(shell ...) ("unterminated
call to function 'shell': missing ')'"). Use '@' as the delimiter
instead.
Part of #5718 item 5
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Lookups in objects are linear, as for ordered_json. Objects with 128
members or more now get a hash table after parsing (open addressing; the
first of duplicate keys is kept, as for the linear search), so that
operator[], at(), find(), contains(), count(), value(), and JSON pointers
take constant time on average in them; the idea of switching to a hash
table for large objects is Boost.JSON's. The parser notes such objects when
it closes them (out of line, so that the parse loop only has a call for
it), and the object node keeps the number of its table.
Looking up each key of an object with 10,000 members: 59.8 ms -> 0.16 ms.
Parsing (json_document::parse, best of 7, separate processes): most files
within 1%; canada +5%, mesh.pretty +3%, citm +3%.
Tests: objects with 127, 128, 129, and 10,000 members (escaped, empty,
and duplicate keys, missing keys, comparisons), nested large objects, and
documents reused with read().
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Long runs of string bytes are scanned 16 at a time with NEON (AArch64, with
GCC and Clang) and SSE2 (x86-64): both belong to the baseline instruction
sets. A signed compare with 0x20 finds control characters and non-ASCII
bytes at once. Keys keep 16 table checks before the vector loop (their
lengths repeat from record to record, so the branches predict well);
string values have 8, as their lengths vary more.
Non-ASCII text is validated 16 bytes at a time with the "lookup4" check of
simdjson (J. Keiser and D. Lemire, "Validating UTF-8 In Less Than One
Instruction Per Byte", 2021): with NEON, and on x86-64 with SSSE3 if
JSON_VIEW_USE_SSSE3 is defined (SSSE3 is not part of x86-64, and the code
must not depend on the flags of a translation unit). JSON_VIEW_NO_SIMD
selects the portable code. The vector code sits in
detail/view/simd.hpp; the same input is accepted either way.
json_document::parse, best of 7 runs in separate processes (M1 Max):
poet.json (CJK text) -72%, random.json -25%, twitter.json -22%,
gsoc-2018.json -20%, semanticscholar -19%, github_events -11%,
apache_builds -9.5%, canada/citm -5/-6%; lottie +4%, tree-pretty +2.5%.
Tests: every two-byte sequence and three- and four-byte sequences with
continuation bytes at the edges of their ranges, at every offset around
the vector blocks of keys and values, cut short, and long runs of text
with a damaged byte, against json::accept and json::parse. CMake builds
the parser tests again with JSON_VIEW_NO_SIMD, and on x86-64 with
JSON_VIEW_USE_SSSE3 and -mssse3; the macros are documented.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
- API pages for operator== and operator!= of basic_json_view, linked
both ways with the basic_json pages
- the feature page and the class overview list the comparisons
- the examples show when the view helps: detecting a changed document
without building json values
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
- API pages for dump, number_format, and operator<< of basic_json_view,
linked both ways with the basic_json pages
- the feature page describes document order and number_format::source
- the examples show when the view helps: forwarding part of a message
and writing numbers exactly as they were read
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Niels Lohmann
