MSVC 2015 warns C4554 on the shifts with casts in the fractional part
of zmij::to_shortest despite the parentheses; with /WX every target
fails to build. Name the shift amounts as constants instead.
The shortest-digits test built the candidates with one digit fewer via
a stream with std::setprecision(digits.size() - 2); the standard library
of MSVC 2017 ignores a precision of 0, so for two-digit values the
"shorter" candidates were the value itself. Truncate the digits
instead: the value lies within half a unit of the last digit, so the
truncated digits and their successor are the closest candidates.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Escape characters below 0x60 from a table straight into the write buffer
When the bulk scanner in dump_escaped_impl() stops at a control
character, a quote or a backslash, that byte used to go through the
UTF-8 decoder and a switch that wrote the escape byte by byte into
string_buffer, which was then copied into write_buffer at the next run.
A byte below 0x60 at a character boundary is a complete code point, so
it now skips the decoder: its escape sequence comes from a 96-entry
table of 8-byte rows (sequence padded with zeros, length in the last
byte), and all 8 bytes are copied straight into write_buffer with one
room check. The check sits inside the existing character-boundary
block, so it runs once per scanner stop and not once per byte; testing
every byte cost about 6% on twitter.json with ensure_ascii.
The switch in the decoder's accept branch lost its cases for
characters below 0x60, which can no longer reach it.
Output is byte-identical. The string-buffer boundary tests now use
\u-escaped non-ASCII characters, since simple escapes no longer reach
string_buffer, and a new test case checks every ASCII character at
every position against a reference escaper.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Avoid raw string literals with backslashes in CHECK macros
MSVC's traditional preprocessor mis-lexes raw string literals inside
macro arguments and reports illegal escape sequences (C2017, C3850).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Sweep invalid UTF-8 at every position through dump()
Compare dump() of strings with ill-formed UTF-8 (stray continuation
bytes, invalid lead bytes, overlong encodings, surrogates, code points
above U+10FFFF, and sequences cut short by ASCII, a quote, a backslash,
a lead byte, or the end of the string) with an independent reference
serializer, for every error handler and with ensure_ascii on and off.
The pattern is put at every position of short strings to cross the
strides of the bulk scanners, and of longer ones to cross the end of the
512-byte string buffer.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fuzz the streaming input adapters against the contiguous one
The fast paths for contiguous input (bulk string scanning, multi-byte
loads, lazy position tracking) are taken only when parsing from a
pointer range; streams and non-contiguous iterators still read byte by
byte. The fuzzer drivers parsed only contiguous input, so a difference
between the two would ship unnoticed.
check_input_adapters() in fuzzer_common.hpp reads the input through a
std::istringstream, a std::list iterator range, and (on POSIX) a FILE*
from fmemopen, and asserts the same value, or the same exception id,
message and byte position, as for the pointer range. All seven drivers
call it; the JSON driver checks parse and accept, each with and without
ignore_comments. It replaces the stream check of the BON8 driver.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Rename the reader parameter to avoid a flawfinder false positive
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Cppcheck 2.20 rejects invocations of unknown function-like macros in #if
conditions instead of evaluating them as 0. default_defines.hpp expands
the Hedley macros to the compiler's feature-test operators (__has_attribute
etc.), so define them as 0 in a generated include file.
It also reports a false bufferAccessOutOfBounds in the float overload of
float_from_bits, sizing the parameter as the double overload's; use the
size of the source instead (equal by the static_assert).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Allow the same type for number_integer_t and number_unsigned_t
basic_json did not compile with NumberUnsignedType == NumberIntegerType
because overloads on both types collided. Such a basic_json now stores
every integer as number_integer: the colliding overloads take a
placeholder type, unsigned C++ integers are converted to number_integer_t,
and value_t::number_unsigned maps to number_integer.
Also fixes two problems this configuration would hit: the unsigned
integer parser accepted values that wrap around to negative numbers, and
dump() overflowed when computing the absolute value of the minimum.
A repeated NumberFloatType gets a static_assert.
Refs #2573
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Use json::with_integers_t in the test and the docs
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* ✅ skip patch test on GCC 5 to avoid internal compiler error
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix CI: name the unsigned overload type through template parameters
basic_json used number_unsigned_overload_t<basic_json> while the class was
still incomplete. MSVC rejected it as a recursive alias declaration and GCC
4.9, 5 and 7 crashed with an internal compiler error. basic_json now uses
number_unsigned_overload_type<NumberIntegerType, NumberUnsignedType>, which
also makes the GCC 5 workaround in the patch test unnecessary.
Infer reported a null dereference for the number_unsigned case in
from_json, which cannot happen for a basic_json whose number_unsigned_t is
number_integer_t; assert it.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Address review: conjunction/negation trait, astyle-friendly alias, docs table order
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
patch_inplace() built its error messages inline. In particular get_value()
concatenated the context-sensitive "operation 'x'" string on every call, even
when nothing was wrong. The patch exceptions now come from NEVER_INLINE
helpers: the index and key errors reuse basic_json's existing
throw_array_index_out_of_range/throw_key_not_found, and the seven
patch-specific messages get one helper each in detail/exceptions.hpp. Ids
and messages are unchanged, and unit-throw_messages pins them (with and
without JSON_DIAGNOSTICS).
citm_catalog patch: -6.6 % (Apple clang), -9.5 % (GCC 16).
The other remaining throw sites (parser::syntax_error, binary_writer,
serializer, input_adapters, enum_key_duplicate, format_error) were prototyped
and measured: no gain beyond noise, so they stay as they are.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Speed up container destruction, iterator access and diff()
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Destroy the first levels of a value by recursion
The pointer-reversing walk in destroy_container() avoids the call stack but
moves a value around for every container it visits. Recurse for the first
recursion_depth_limit() levels instead - no allocation, no thread-local, a
bounded frame - and hand anything deeper to the walk. Destroy twitter
0.49 -> 0.36 ms, citm 0.80 -> 0.74 ms (3.12.0: 0.47 / 1.00).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Make the value lifecycle inline-friendly
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Compare standard containers without a thread-local depth count
operator== on objects and arrays recursed through the containers' own
operator== and counted the depth in a thread-local variable. Where reading a
thread-local costs a call (macOS), that added about 10% to comparing documents
with many small containers. For the standard containers, walk objects and
arrays directly, compare scalars and strings inline, and pass the depth as an
argument, falling back to compare_iteratively beyond the depth limit. Other
container types keep the previous path.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Address review: owns_memory() helper, leaf compare naming, astyle
- Add detail::owns_memory(value_t), a switch listing every enumerator so
-Wswitch-enum flags it when a type is added; destroy() and the copy
constructor use it instead of hand-written comparisons. Clang emits the
same bitmask test as before.
- Drop the parentheses around the scalar operands in
JSON_IMPLEMENT_SCALAR_CASES; they are only needed for the container
operands (#1530), which the comment now says.
- Build equal_bounded_containers from detail::conjunction so astyle no
longer pads the template brackets.
- Rename equal_scalar() to equal_leaf() and document why it also compares
strings, unlike JSON_IMPLEMENT_SCALAR_FAST.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix quadratic parent updates for ordered_json with JSON_DIAGNOSTICS
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Only trust containers known to keep their elements in place
Address review: decide whether an insertion may have moved the other
members by naming the containers that are known not to, instead of the
ones that are known to.
- detail::has_stable_elements: std::map, std::multimap,
std::unordered_map and std::unordered_multimap; object_capacity()
reports 0 for them.
- detail::is_nlohmann_ordered_map: only the real nlohmann::ordered_map
(which only appends) reports its capacity().
- Any other object type reports unknown_size(), so it gets a full
set_parents() on every insertion. Before, an object type without
capacity() was treated as never moving, and one with capacity() (a
sorted flat map, say) was trusted to move only on reallocation.
- set_parents_after_object_erase() re-parents unless the object type is
stable; the unused set_parent() is removed.
- Comments say "the object type" instead of "ordered_json".
- New test with an object type that moves its members without changing
its capacity; it fails without this change.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Move elements when converting rvalue containers to JSON
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Simplify the rvalue to_json traits
Use negation<disjunction<...>> instead of a conjunction of negations and
keep the template argument lists compact (astyle only keeps the spaced
'< ... >' form when it is written that way).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Speed up array-to-container from_json
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Hint the end when converting an object to a standard container
from_json for map-like containers called emplace() without a hint, which
searches the tree for every key. The keys of the source object are sorted, so
inserting with the end as the hint is constant time, as it was in 3.12.0 via
std::inserter. get<std::map<...>>() of a 25,000 entry object: 4.2 ms -> 3.25 ms
(3.12.0: 3.7 ms).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fall back to emplace() in object from_json for containers without emplace_hint()
ordered_map has no hinted insert, so the insert(end(), ...) call broke
from_json into types holding an ordered_map (regression test #5122).
Use emplace_hint() where available and plain emplace() otherwise.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Speed up JSON pointer parsing, formatting, comparison and lookup
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Restore unflatten() speed after the linear-time rewrite
The prefix tree introduced for linear unflatten() kept a std::map per prefix,
copied each token into it and looked every token up a second time while
creating the values. Replace the per-prefix maps by one open-addressing table
keyed by (parent, token) that refers to the parsed tokens, skip the lookups for
the prefix shared with the previous key (flattened objects are mostly sorted),
and derive the prefix numbers of the second pass from parent links instead of
lookups.
Time and memory stay linear in the pointer depth, and the result is still
independent of the iteration order of the flattened object.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Put [[noreturn]] before __attribute__ on the JSON pointer throw helpers
Clang 18 on Linux rejects a C++11 attribute list after a GNU
__attribute__ ("an attribute list cannot appear here").
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Make flatten() linear for ordered_json
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Use priority_tag instead of a C-style variadic overload in add_flattened
clang-tidy flags the ellipsis fallback (cert-dcl50-cpp,
modernize-avoid-variadic-functions).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
scan_after_whitespace() built the expected text of true, false, and null in
local std::array objects. With -fstack-protector-strong (the default of
Ubuntu's GCC, among others), a local array gives the function a stack canary
check, which then ran on every token. scan_literal() now takes the text
after the first character as a string literal, with its length as a template
parameter; for contiguous input it compares the rest at once and falls back
to the byte loop on a mismatch, so errors are reported at the same
character. It is out of line, which keeps the token dispatch small.
The parser kept the nesting of arrays and objects in a std::vector<bool>,
whose push_back() GCC calls out of line; it is a std::vector<std::uint8_t>
now.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
The string fast path found the end of a run of plain characters 8 bytes at
a time (SWAR) and validated non-ASCII text one sequence at a time, calling
validate_one_utf8() out of line for each.
The vector code of json_view's scanner, in the new shared header
detail/input/simd_scan.hpp, now does both 16 bytes at a time: SSE2 finds
the end of a run on x86-64, and the UTF-8 check of Keiser and Lemire runs
with NEON on AArch64 and with SSSE3 on x86-64. SSSE3 is not part of x86-64,
so the check is compiled for it with a function attribute and used when the
CPU has it (cached in a constant-initialized atomic), else the scalar check;
code compiled with -mssse3 skips the CPU check. The first word of a run is
still checked with SWAR (the first four with NEON, where a vector mask
costs more), as most keys and short values end there, and the scalar
helpers are always inlined. On anything else, the scalar code is unchanged.
The new macro JSON_NO_SIMD selects the portable code everywhere, including
the SSE2/NEON digit formatting of the floating-point writer. The lexer tests
now also run with JSON_NO_SIMD and, on x86-64, with -mssse3, and a new test
places every byte that ends a run at each offset across several 16-byte
blocks, after ASCII and after CJK text.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
After the digits are converted while scanning, copying the token text into
token_buffer (an out-of-line std::string::append) was about 10% of parsing
canada.json. The only reader is the string argument of
json_sax::number_float(), and the library's own handlers
(json_sax_dom_parser, json_sax_dom_callback_parser, json_sax_acceptor)
ignore it.
For contiguous input, the fast path now keeps a pointer to the token in the
input, and get_string() copies it on first use. The parser passes an empty
string to the handlers marked by the new internal trait
sax_ignores_number_text instead of calling get_string(), so json::parse()
and json::accept() never copy float text. User handlers, and
JSON_DIAGNOSTIC_POSITIONS (which reads get_string().size()), still see the
exact text. Keys and strings, which never have pending text, use the
unchecked get_string_token(). Input read through a window into a stream
(std::istream, FILE*) copies the text right away, because the next read may
refill the window.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
The contiguous number fast path validated a token byte by byte, copied it
into token_buffer, and then converted it in out-of-line functions that
scanned the digits a second time: parse_float_native() called
parse_float_significand(), which returned its result through memory and
called std::any_of (out of line with GCC) twice per number, and
parse_integer_signed() divided by ten on every digit for its overflow check.
scan_number_bulk_contiguous() now accumulates the significand while it
validates the digits, eight at a time where at least eight bytes remain
(is_eight_digits() and parse_eight_digits(), after fast_float). Floats with
at most 19 digits go straight to decimal_to_float(); integers with at most
18 digits are stored without an overflow check (a narrower number type
still rejects values that do not round-trip). Longer tokens take the
existing path. has_nonzero_digit() is a plain loop and
parse_float_significand() is always inlined, which helps that path.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Building a std::string from the iterators of a CBOR byte vector narrows
each byte, which -fsanitize=implicit-conversion reports for bytes above
0x7F. Construct the string from the raw bytes instead.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Resolve the macro_unscope.hpp conflict by keeping both sets of #undefs
and regenerate the amalgamation.
The \u fast path from #5738 (get_codepoint_bulk) assumed that bulk
scanning implies lazy position tracking and advanced the adapter with a
bare bulk_skip(4). Stream adapters now scan in bulk but track position
and token eagerly, so a lone-surrogate error was reported 4 columns too
early. Advance through bulk_consume() instead.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Follow-up to review comments on #5825: number_float_t uses double
precision only by default, and the stated limits apply to double.
Also say "default type for X_t" instead of "default value for X_t".
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
- Discarded values: the legacy macro only changes <= and >=; != is true by default
- Out-of-range numbers are rejected while parsing, not stored as NaN
- merge_diff round trip: point to the ordered_json caveat
- ensure_ascii example shows the escaped output
- out_of_range.405 is thrown for "remove"/"move" on the root, not "add"
- CBOR stores binary subtypes as tags
- Integer/float comparison uses exact values, not conversion
- diff, flatten, and the binary writers are depth-bounded
- FAQ: max_digits10 for serialization, trailing commas are supported
- out_of_range.407 is still thrown by to_bon8
- FetchContent_MakeAvailable needs CMake 3.14
- json_fwd.hpp contents, operator< link, RFC 7396 for merge patch
- README: dump() wording, merge_diff, MSYS2/build2, patch comment typos,
operator[] link, embedded CMake example aligned with the docs
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
The privacy plugin downloads img.shields.io badges at build time; when
shields.io times out, mkdocs build --strict aborts and the
ci_test_build_documentation job fails until it is restarted. Exclude
the host like repology.org; the version badges stay current this way.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Update Google Benchmark to 1.10.0
Bump the pinned release and its archive checksum. Reword the comment and
README: the version is fixed for reproducible builds and updated like any
other dependency; results stay comparable because both sides of a
comparison are built with the same version.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Do not name the Google Benchmark version in the README
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Write doubles with the shortest digits (Żmij), digits in registers
Write doubles with the conversion of Zmij by Victor Zverovich (MIT),
ported to C++11 (detail/conversions/zmij.hpp). It finds the
shortest decimal that reads back as the same double, and the
closest one if there are several. Grisu2, used until now, is fast
but not always shortest: it sometimes writes a 17th digit where 16
suffice, or a last digit that is not the closest. The layout is
unchanged (1.5, 100.0, 1e+100, -0.0); float keeps Grisu2.
Digits are converted eight at a time with the BCD conversion of
Xiang JunBo, as in Zmij, and written with one byte swap per eight
digits and fixed-size moves instead of per-digit loops. Leading and
trailing zeros are counted from those bytes. to_chars() uses a
local buffer when the caller's is shorter than the 41 bytes this
may write. The powers of ten come from the number-parsing table,
adjusted where it holds values rounded up, and extended with Zmij's
compressed tables beyond 10^308.
write_shortest() converts its 16 digits in one vector register
(SSE2 on x86-64, NEON on 64-bit Arm, both baseline) and inserts the
decimal point inside the register, avoiding a store-forwarding
stall that cost about 25% of the time to write a double. dump()
writes floats and integers straight into the serializer's write
buffer instead of copying them from a member buffer, and small
integers eight digits at a time. read_eight_bytes() and
parse_eight_digits() are marked always-inline, which GCC had been
calling out of line in the number-parsing loops.
Of one million random doubles, about 0.14% are now written with
different digits, always to a value that still reads back as the
same double.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix CI findings in the Zmij writer and its tests
- bit_ops.hpp: include macro_scope.hpp (JSON_HEDLEY_ALWAYS_INLINE); fixes IWYU
- to_chars.hpp: C4100 for the unused parameter in release builds, clang-tidy
sign comparison, cpplint runtime/int, GCC -Wstrict-overflow (unsigned abs)
- unit-to_chars.cpp: parse with the library instead of strtod (MinGW's strtod
rounds some 16/17 digit inputs wrongly); no floating-point std::to_chars
with icpc
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Select the Zmij conversion by a trait, not by the double overload
The non-template overloads for double asserted binary64 doubles wherever json.hpp was included, so the library no longer compiled where double is not IEEE 754 binary64 (AVR, -fshort-double). A trait now picks Zmij for any binary64 type, including a long double of that format (MSVC, Apple Arm), and Grisu2 for the others. Remove the unused write_short_decimal, powers_of_ten_16, zmij::decimal, zmij::to_decimal, and shortest_digits(double).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Credit Zmij's authors in to_chars.hpp, the README, and the license page
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Silence MSVC C4127 for a platform-constant check in the to_chars test
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Regenerate the amalgamated headers
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Speed up the lexer: own float parser, string scan, and \u table
Give the library its own correctly rounded float converter for
binary32 and binary64 (IEEE 754), and speed up the lexer's string
and escape scanning.
The converter splits a number token into sign, significand, and
decimal exponent, then tries Clinger's fast path, then a templated
Eisel-Lemire step, and falls back to an exact big-integer digit
comparison for tokens with more than 19 significant digits whose two
candidate values round differently. This replaces std::from_chars
and strtod/strtof for both formats, so parsed values no longer
depend on the C/C++ library or the current locale. The strtold
fallback kept for other long double formats (x87, binary128) now
also copies a multi-byte decimal point correctly, fixing #5660.
eisel_lemire() and decimal_to_float() are always inlined so callers
keep the whole conversion in their hot loop.
The string-scanning kernels in string_scan.hpp find a stop byte with
the trailing-zero count of the SWAR mask instead of a byte loop, and
scalar_string_bulk_run() validates a run of multi-byte UTF-8
sequences one after another instead of re-searching after each one.
get_codepoint() decodes a contiguous \uXXXX escape with one table
lookup per byte instead of four range-checked get() calls; the
streaming path and all error positions are unchanged.
Adds 508 generated hard float-parsing cases with expected binary32
and binary64 bits, and kernel-comparison tests for the string scans
and the escape table against byte-by-byte references.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix CI warnings in own float parser
- pow5_table.hpp: pow5_128_largest_power was unused in this branch's
own code (GCC -Werror=unused-const-variable); tie it to the table
size with a static_assert instead of removing it, since a later
branch in the stack (json-view/23-zmij) uses it.
- number_parse.hpp: rename the local variable `copy` to `buffer` to
satisfy cpplint's build/include_what_you_use check.
- unit-class_lexer.cpp: extend the NOLINT list on the seeded mt19937
with bugprone-random-generator-seed, and parenthesize
`8 * sizeof(Bits) - 1` for clang-tidy.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Load string scan words with memcpy and use MSVC bit-scan intrinsics
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Reword float conversion notes for the string type and number_float_t docs
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Fix clang-tidy 22 findings in unit-class_lexer.cpp
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Mark the unsigned long of the MSVC bit scans for cpplint
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Regenerate the amalgamated headers
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Drop the position updates in the bulk \u path, which #5809 made unused
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Address review: std::any_of in has_nonzero_digit, comment punctuation
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Use data() instead of c_str() in the multi-byte decimal point path
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
ParseFILE/* parses the nativejson-benchmark files from a FILE*, and
ParseStreamValues/* reads 10,000 small values from one std::istringstream
with operator>> (NDJSON objects and plain numbers), which measures the
per-value cost of stream input.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Parsing from a FILE* reads one character at a time with std::fgetc, which
locks and unlocks the FILE object for every character. file_input_adapter
now locks the file once when it is created and unlocks it when it is
destroyed - also when parsing throws - and reads with the unlocked stdio
functions in between: flockfile/getc_unlocked/funlockfile with glibc,
Apple, Android, and the BSDs, and _lock_file/_fgetc_nolock/_unlock_file
with the Microsoft C runtime. Other C libraries keep std::fgetc. The lock
is recursive, so fread() in the binary readers and stdio calls of a SAX
handler or parser callback on the same thread keep working.
On an Apple M1 Max, parsing from a FILE* is 2.9 to 4.0 times as fast
(twitter.json 14.3 -> 4.1 ms, citm_catalog.json 35.4 -> 8.8 ms); with
glibc, whose locking is cheaper, 1.1 to 1.3 times. The values and the
position of the file after parsing are unchanged.
Other threads that use the same FILE object now wait until parsing is
done instead of interleaving their reads with the parser's.
JSON_DISABLE_UNLOCKED_FILE_IO restores std::fgetc, for programs in which a
SAX handler waits for another thread that uses the same FILE. The new
tests check from a second thread that the lock is released after success
and after a parse error, and run again with the macro.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Parsing from a std::istream read one character at a time through the
streambuf, so none of the parser's fast paths applied: it was 1.8 to 2.5
times slower than reading the file into a string and parsing that.
input_stream_adapter now exposes the streambuf's get area (gptr() to
egptr()) to the lexer as a window. The protected get-area members are
reached through member pointers formed in a class derived from
std::streambuf, which is ordinary, standard-conforming member access. The
lexer scans the window with the same bulk paths as a string, and asks
the streambuf for the next character as before when the window is used
up. Nothing is read ahead, copied, or put back, so the stream is left
exactly where it was left before, with and without
JSON_PRECISE_STREAM_POSITION, and a streambuf without a get area is still
read one character at a time.
The adapter also keeps the current token for error messages as a pointer
into the get area, and copies it out only before the get area is
refilled, instead of the lexer copying every character.
On an Apple M1 Max (Apple Clang, -O3), parsing from an std::ifstream is
1.6 to 2.1 times as fast (twitter.json 4.08 -> 2.30 ms, citm_catalog.json
8.75 -> 4.08 ms), and reading 100,000 small values with operator>> 1.4
(objects) to 4.0 (integers) times as fast. With GCC 13 and libstdc++ on
x86-64 Linux, std::ifstream is 1.5 to 2.2 times as fast. Parsing from a
stream is now within 11 to 18 % of reading the file into a string first.
JSON_DISABLE_STREAM_BULK_SCAN restores reading one character at a time,
for streambufs that do not follow the usual get-area contract. The new
tests compare every observable result (values, exception messages, stream
state, remaining characters) for get areas of 1 to 4096 bytes against a
streambuf without a get area, and run again with the macro and with
JSON_PRECISE_STREAM_POSITION.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
The SWAR string scan, the bulk number scan, and the indentation skip were
only enabled for adapters that support seeking (contiguous random-access
input), because they rely on the lazy token and position tracking that
reconstructs both from the consumed input on demand.
This decouples the bulk paths from seeking, so that an adapter can expose
a contiguous block that is only a window into its input (bulk_window):
- bulk_consume() does the bookkeeping that get() would have done for the
skipped characters when token and position are tracked eagerly.
- A number that reaches the end of a window is finished character by
character, because it may continue after the window.
- After a bulk-scanned number, a window adapter reads the terminating
character and ungets it, as scan_number() does, so that the input
handed back to the user is consumed exactly as before.
- An adapter can keep the current token itself (supports_token_tracking,
begin_token(), mark_token_start(), copy_token()) so the lexer need not
copy every character for its error messages.
No adapter declares any of this yet, so the behavior is unchanged.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Drop the class wrapper from the C++20 comparison operator signatures in the docs
The signature boxes of operator<=> and the C++20 half of operator== wrapped
the member declarations in an incomplete "class basic_json { ... };". The
trailing const already marks them as members, and no other page uses such a
wrapper.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Drop the class wrapper from the json_pointer comparison signatures too
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Mark json_fwd.hpp as exported by json.hpp for IWYU / include-cleaner
json_fwd.hpp is the only header declaring the nlohmann::json alias, so
include-what-you-use and clang-tidy's misc-include-cleaner attributed
the name to it and told users of <nlohmann/json.hpp> to include
json_fwd.hpp as well (or instead, which breaks compilation).
Fixes#5812
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Export the remaining client-facing headers from json.hpp and check clients with IWYU
Besides json_fwd.hpp, include analyzers also told clients of
<nlohmann/json.hpp> to include detail/json_ref.hpp (brace
initialization), json_literals.hpp (_json, _json_pointer),
adl_serializer.hpp, detail/macro_scope.hpp (NLOHMANN_DEFINE_*) and
detail/abi_macros.hpp. Mark those includes as exported, too.
ci_single_binaries only checked each header on its own, so it could not
notice this. It now also runs IWYU and clang-tidy's misc-include-cleaner
on small client programs in tests/iwyu_consumer, which fail on develop.
With the exports, json_literals.hpp no longer gets the json_fwd.hpp
suggestion that kept it out of the IWYU gate, so it is gated again.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
Pretty-printed JSON indents every line with a run of spaces, which
skip_whitespace() read one character at a time. For contiguous input
(bulk_scan) the lexer now skips the spaces after a newline 8 bytes at a
time directly in the input buffer, including the leading spaces of the
last, partial word. This bypasses get(), which is safe because
bulk_scan implies the lazy position from the previous commit: nothing
is counted per character.
Compared to lazy positions alone, pretty-printed input (twitter,
citm_catalog with indent 2 and 4) parses up to 13% faster and accept()
gets up to 30% faster (GCC 13 and clang 18 on x86-64, Apple clang 21 on
arm64). Minified input is unchanged within noise.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Derive the lexer position lazily for seekable input
The lexer counted the read position on every character: two increments
and a newline check per get(), all on members kept in memory. For
seekable input (pointers, strings, other random-access byte ranges) the
position can be derived instead: chars_read_total is the adapter's read
count (including reads at EOF) minus a pending unget, and line and
column are recomputed from the consumed input only when get_position()
is called, i.e. on parse errors. JSON_DIAGNOSTIC_POSITIONS uses the new
O(1) get_chars_read_total(). Stream, file, wide-string and user-defined
adapters keep the eager counters.
Parsing pretty-printed input gets 8-12% faster and accept() 12-27%
(canada, citm_catalog, twitter; clang 21 arm64 and GCC 13 x86-64).
Error messages, byte offsets, SAX positions and diagnostic positions are
unchanged (differential comparison against develop on 548k inputs).
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Move the std::vector<std::byte> binary type tests out of unit-regression3
The #4804/#4209 tests instantiated a second basic_json specialization
(with_binary_t<std::vector<std::byte>>), including its parser and CBOR
reader. With clang 18, the C++20 object of unit-regression3.cpp had
32656 COFF sections on develop; the lazy position helpers pushed it to
32975, past what the MinGW linker of the Windows clang jobs handles
("relocation truncated to fit: IMAGE_REL_AMD64_REL32 against .rdata").
unit-custom-binary-type.cpp already instantiates the same type, so the
tests move there; unit-regression3.cpp drops to 27616 sections.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* Report the parse errors of the position test via SAX
The test caught json::parse_error, so it aborted in the
ci_test_noexceptions build, where JSON_NOEXCEPTION turns the throw into
std::abort(). sax_parse() hands the error position and the exception
object to parse_error() without throwing, so the test now records them
there and runs in every configuration.
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
---------
Signed-off-by: Niels Lohmann <mail@nlohmann.me>
* support JSON Merge Patch (RFC 7396) diff creation
The JSON merge patch document format describes the set of modifications
to a resource's content, that more closely mimics the syntax of the
resource being modified.
However, and in contrast to JSON Patch (RFC 6902), a JSON Merge Patch
cannot express certain modifications, e.g., changing an array element
at a specific index, or setting a specific object value to null.
The null value in a JSON Merge Patch is used to remove the key from the
object.
The diff algorithm is not part of the RFC 7396, but it was tested
against all examples provided, plus additional cases on how null values
are handled.
Signed-off-by: Luís Murta <luis@murta.dev>
* improve algorithm efficiency
Signed-off-by: Luís Murta <luis@murta.dev>
* fix inconsistent nulls and empty arrays
Also add CHECKs for the patch, documentation and run `make amalgamate`.
Signed-off-by: Luís Murta <luis@murta.dev>
* documentation cleanups
Signed-off-by: Luís Murta <luis@murta.dev>
* add an iterative fallback to merge_diff
Bound the recursion depth like diff() and merge_patch(), and test values
nested past the bound.
Signed-off-by: Luís Murta <luis@murta.dev>
* address the remaining review comments
Signed-off-by: Luís Murta <luis@murta.dev>
* move multi-line raw strings out of CHECK macros
GCC 4.8/5, ICPC and NVHPC mis-tokenize multi-line raw string literals
inside macro arguments, breaking the build or the parsed JSON.
Signed-off-by: Luís Murta <luis@murta.dev>
---------
Signed-off-by: Luís Murta <luis@murta.dev>
Niels Lohmann
dependabot[bot]
satelliteprogrammer