Compare commits

..

672 Commits

Author SHA1 Message Date
Francisco Geiman Thiesen d326f2ce9f Working! 2025-10-10 21:32:10 -07:00
Francisco Geiman Thiesen ca42a49fba Compile-time support for parsing json objects! 2025-10-10 18:40:09 -07:00
0xflotus a7f8fb71c5 chore: fix small error in docs (#2497) 2025-10-03 11:03:17 -04:00
Howard Guo a553db4c67 Update workflow name to Ubuntu aarch64 (GCC 13) (#2484) 2025-10-03 09:56:07 -04:00
Jaël Champagne Gareau 1fa1af8c15 Fix yyjson leaks when running ./bench_ondemand (#2485) 2025-10-03 09:55:32 -04:00
Daniel Lemire 5ed1044056 4.0.7 2025-09-30 11:26:03 -04:00
Francisco Geiman Thiesen 62913867ff Merge pull request #2475 from simdjson/francisco/extract_from
Adding extract_from functionality + unit tests
2025-09-29 20:34:35 -07:00
Francisco Geiman Thiesen 6700d48b57 Merge pull request #2480 from simdjson/francisco/extract_from3
minor tweaks... ;-)
2025-09-29 13:54:05 -07:00
Francisco Geiman Thiesen b5577d5e85 Merge branch 'master' into francisco/extract_from 2025-09-29 12:46:48 -07:00
wszqkzqk b84a4ec2b9 Fix: Correct narrowing conversion in lsx string parsing (#2481)
Resolves a build failure on the loong64 architecture caused by a narrowing conversion error.

The compiler, with the -Werror=narrowing flag, was flagging the implicit conversion from 'int' (the return
type of to_bitmask()) to 'uint64_t'.

This is fixed by adding an explicit static_cast to uint64_t in include/simdjson/lsx/stringparsing_defs.h.

Signed-off-by: Zhou Qiankang <wszqkzqk@qq.com>
2025-09-29 11:57:06 -04:00
Daniel Lemire 1638a185f7 minor tweaks... ;-) 2025-09-29 11:12:37 -04:00
Francisco Geiman Thiesen 6fe450f5ce Updateing single_header 2025-09-29 03:44:29 -07:00
Francisco Geiman Thiesen c7b70de070 Merge branch 'master' into francisco/extract_from 2025-09-29 03:23:02 -07:00
Francisco Geiman Thiesen 3279fbd55b Merge pull request #2474 from simdjson/complete_extract_into
this completes the extract_into work.
2025-09-29 03:12:20 -07:00
Francisco Geiman Thiesen 66e64e0e5f Merge branch 'master' into francisco/extract_from 2025-09-29 03:02:12 -07:00
Daniel Lemire 03f81e66af updating single header 2025-09-27 12:26:04 -04:00
Daniel Lemire e3b7eddb37 fixing off-by-one mistake in the documentation (#2477) 2025-09-27 12:19:44 -04:00
Daniel Lemire 99c4ba6e8f tweak 2025-09-26 23:44:42 -04:00
Francisco Geiman Thiesen 6aa7eea334 Adding extract_from functionality + unit tests 2025-09-26 19:48:16 -07:00
Daniel Lemire 6a47cda07f guarding 2025-09-26 22:34:36 -04:00
Daniel Lemire 617c69e104 completing doc 2025-09-26 21:33:10 -04:00
Daniel Lemire 625adceb24 updating single-header 2025-09-26 21:29:28 -04:00
Daniel Lemire bd0e9c1336 tweak 2025-09-26 21:04:53 -04:00
Daniel Lemire 7bf82b02d5 this completes the extra_into work. 2025-09-26 21:00:46 -04:00
Francisco Geiman Thiesen 88a1b3e83b Merge pull request #2471 from simdjson/francisco/extract_into
Adding extract_into functionality + test (targets simdjson >= 4.0 as it relies on reflection)
2025-09-26 01:29:33 -07:00
Francisco Geiman Thiesen c72954eade Addressing reviews. 2025-09-25 21:08:38 -07:00
Francisco Geiman Thiesen 4456a10469 Francisco/using iterators for containers (#2470)
* Using iterators instead of subscript operators and size. This helps us work with a broader range of containers.

* Adding list test

* Using std::ranges::input_range<T> as suggested by moisrex
2025-09-25 17:30:42 -04:00
Francisco Geiman Thiesen d8ed2417ad Adding coverage for extract_into with types that have custom serialization 2025-09-25 03:35:46 -07:00
Francisco Geiman Thiesen 5517df7aee Adding extract_into functionality + test 2025-09-24 22:58:23 -07:00
evbse 6e618b0805 Improve DOM implementation (#2434) 2025-09-21 11:26:33 -06:00
Pavel Novikov bde288a623 Fixed string_builder::operator std::string() (#2465)
* clang format

* fixed `string_builder::operator std::string()`

* fixed variable shadowing error false positive
2025-09-21 11:25:42 -06:00
Daniel Lemire b2932d1b8f release candidate 4.0.6 (#2464) 2025-09-21 08:13:48 -06:00
Daniel Lemire a7811090ef fixing issue 2458 (#2461) 2025-09-20 22:23:09 -06:00
Daniel Lemire 3320885fac Fixing issue 2462 (#2463)
* fun

* progress

* completing the documentation
2025-09-20 22:22:57 -06:00
Daniel Lemire 786c68b158 release 4.0.5 2025-09-18 15:17:32 -06:00
Daniel Lemire 703ef54bd9 allow string reuse (#2454)
* allow string reuse

* portability fix

* using data and not begin
2025-09-18 15:16:38 -06:00
Daniel Lemire 2526068e2f Add mamba link to README 2025-09-18 09:22:03 -06:00
Daniel Lemire ddc7b8c7dd update 2025-09-17 18:59:06 -06:00
Daniel Lemire d8f90bdd14 modifying simdjson::from to avoid exceptions when needed. (#2452)
* modifying simdjson::from to avoid exceptions when needed.

* moved the function

* moving the strings.

* more moving around

* updating cmake version in ci
2025-09-17 18:58:25 -06:00
Daniel Lemire e38a4923e5 adding keys as templates in builder (#2453)
* adding keys as templates in builder

* more guarding

* cmake update in ci

* guarding.

* guarding
2025-09-17 18:58:05 -06:00
Dirk Stolle e6dfa2e0ed remove trailing whitespace + fix typos (#2451) 2025-09-16 22:41:14 -06:00
Daniel Lemire 1f369ef210 minor patch which allows us to pass mutable strings to simdjson::from… (#2448)
* minor patch which allows us to pass mutable strings to simdjson::from and fix
an issue with ambiguous integrals

* compatibility patch.
2025-09-15 22:22:25 -06:00
Daniel Lemire 22dcdc9f1e Update README.md 2025-09-15 18:52:15 -06:00
Dirk Stolle dda2dafa30 fix some typos (#2446) 2025-09-15 18:23:13 -06:00
Daniel Lemire 72e9d44e10 fixing indent 2025-09-15 17:08:12 -06:00
Daniel Lemire 8aae14931d release candidate 4.0.2 (#2441)
* release candidate 4.0.2

* more fixes

* fixing typos

* adding macro check
2025-09-15 09:17:43 -06:00
Daniel Lemire ef3d1ac25f fixing 2440 2025-09-14 08:09:27 -06:00
Daniel Lemire 9292480a9b removing leftover <experimental/...> (#2438) 2025-09-14 08:00:26 -06:00
Daniel Lemire 611ea97dbd Add C++26 reflection example link to README 2025-09-12 21:22:45 -04:00
Daniel Lemire f249e7e128 patch release 2025-09-12 19:26:58 -04:00
Marian Klymov 44fddaa807 Replace POSITION_INDEPENDENT_CODE with INTERFACE_POSITION_INDEPENDENT_CODE for older CMake (#2437) 2025-09-12 19:26:22 -04:00
Daniel Lemire be16e1ae04 removing space 2025-09-11 20:58:31 -04:00
Daniel Lemire 124a33c160 Update JSON example and add section on special cases
Removed code block formatting for JSON example and added a section header for special cases.
2025-09-11 19:32:38 -04:00
Daniel Lemire a0b870e9c2 Add example for deserializing 'Car' type
Updated the documentation to include an example of deserializing a 'Car' type using simdjson.
2025-09-11 19:31:15 -04:00
Daniel Lemire 68699eb73c release 4.0.0 2025-09-11 19:25:10 -04:00
Daniel Lemire 06453f1637 Remove Fuzzing Status badge from README
Removed Fuzzing Status badge from README.
2025-09-09 18:44:21 -04:00
Daniel Lemire eb828d97ba fixing a few typos 2025-09-09 18:41:00 -04:00
Daniel Lemire 00f84fb448 simplify 2025-09-09 14:21:14 -04:00
Daniel Lemire 5bf0954c6e hop 2025-09-08 19:26:34 -04:00
Daniel Lemire 3056d54fa3 adding instructions 2025-09-08 17:47:14 -04:00
Daniel Lemire 241b9f6ea7 adding bad/good player test 2025-09-08 17:04:52 -04:00
Daniel Lemire 015daad6a9 Merge branch 'master' of github.com:simdjson/simdjson 2025-09-05 19:50:05 -04:00
Daniel Lemire fc3b7766c0 typo fix 2025-09-05 19:49:57 -04:00
Daniel Lemire b5c78288e5 can we run the address sanitizer under VS in CI ? (#2435)
* can we run the address sanitizer under VS in CI ?

* fixed typo
2025-09-05 18:32:15 -04:00
Daniel Lemire bd86bb0cfb adding UTF-8 test 2025-09-05 14:56:02 -04:00
Daniel Lemire 4840a0347c extra documentation. 2025-09-05 14:44:53 -04:00
Daniel Lemire 7f68baec1e minor tweak 2025-09-05 08:01:04 -04:00
Daniel Lemire e2ea5fb8de This PR adds -> and * operators to our error types and it (#2433)
improves slightly the documentation.
2025-09-04 22:35:01 -04:00
Daniel Lemire 7dfd165ecd Merge branch 'master' of github.com:simdjson/simdjson 2025-09-04 09:46:07 -04:00
Daniel Lemire 4dfa0a2407 guarding 2025-09-03 16:45:09 -04:00
Daniel Lemire e252a21f65 removing expand workaround (#2431)
* removing expland workaround

* minor fixes
2025-09-03 13:30:35 -04:00
Daniel Lemire 89f2b634df guarding from benchmark + rvv 2025-09-02 13:06:25 -04:00
Daniel Lemire addca203d5 adding a 'car builder' benchmark (#2428)
Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2025-08-26 09:16:18 -04:00
Felipe Monteiro 1825d7e88c Add "node:" prefix on fs import (#2426) 2025-08-25 10:21:13 -04:00
Daniel Lemire f87d6af3e9 fix issue 2424 (#2425) 2025-08-24 16:04:56 -04:00
Daniel Lemire 8fff57578c let us default on developer mode when building with VScode 2025-08-23 20:59:58 -04:00
Daniel Lemire 056d66926a Renames a few macros and extends slightly our basic builder (#2422)
* This PR renames a few macros and extends slightly our basic builder

* minor tuning
2025-08-20 09:01:58 -04:00
Daniel Lemire b434a50681 doing more to discourage value_unsafe(); (#2421) 2025-08-19 16:10:38 -04:00
Daniel Lemire e8ff2fa692 This is a small reorg of the new convert code so that we only expose 'simdjson::from' as experimental (#2418)
* This is a small reorg of the new convert code so that we only expose 'simdjson::from'.
This can be changed in a future release, but we don't want our users to start depending
on code that we might need to change.

* marking simdjson::from as experimental

* updating tests to match recent changes
2025-08-15 15:13:36 -04:00
Daniel Lemire 1ebf115b2b improving the documentation of raw json access (#2416)
* improving the documentation of raw json access

* minor fix

* documentation update

* guarding for exceptions

* fixing exception issue

* fix test

* update.

* saving comments

* more technical fixes

* correcting path in ci test

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2025-08-15 10:26:20 -04:00
Joshua Gawley 172fbe785d Use std::string_view in parser.load function (#2417) 2025-08-15 10:26:03 -04:00
Daniel Lemire a10f096af2 let us just say that it should not work 2025-08-14 11:49:42 -04:00
Daniel Lemire a85e474360 documentation update 2025-08-14 09:33:11 -04:00
Daniel Lemire 05752d6c6f Update ubuntu22-cxx20.yml 2025-08-13 21:17:41 -04:00
Joseph Olabisi fb9a689dff Adding DOM support for json path with wildcard (#2346)
* wip brute-force wildcard for json_path

* wip - bruteforce surface wildcard with result

* partially handle keys with wildcard

* wip - nested paths/pointers on wildcard results

* wip

* wip - handling child properties of wildcard result

* done - handling child properties of wildcard result

* fix key

* add support for wildcard for arrays

* handle array INCORRECT_TYPE

* rename at_path_new to at_path_with_wildcard

* add benchmark

* fix benchmark

* use memcmp

* minor improvements

* refactor to tail recursion

* nit

* approximately 30% improvement in runtime

* nit

* corrected logic

* nit

* cleanup

* add some initial tests

* cleanup 2

* modified:   CMakeLists.txt

* cleanup

* cleanup

* cleanup

* cleanup

* restore examples/quickstart/CMakeLists.txt

* cleanup

* restore quickstart.cpp

* revert array-inl.h

* cleanup array-inl.h

* revert object-inl.h

* cleanup object-inl.h

* cleanup

* nit

* nit

* add test

* address some feedbacks

* address additional feedbacks (copilot)

* final changes based on feedback - reduce string allocations

* fix bug in object-inl.h

* fix logic for wildcards inside arrays

* add test for wildcard in nested array

* refactor and create util for getting key and json path

* minor error handling

* refactor process_json_path_of_child_element from recursion to loop in order to prevent stack overflow

* fix bug with array, add boundary check to get_next_key_and_json_path function

* add more tests

* fix boundary check and unnecessary string allocation

* minor changes

* fix

* fix jsonpathutil

* remove unneccessary string allocation

* some minor fixes

* documentation

* removing printout

* various fixes

* reorg of the CI test file

---------

Co-authored-by: Daniel Lemire <daniel@lemire.me>
2025-08-13 21:16:32 -04:00
Daniel Lemire 472f818fe5 trimming repeated code 2025-08-13 20:40:04 -04:00
Daniel Lemire 5c9c133632 improve how we pad strings. (#2415) 2025-08-13 18:41:14 -04:00
Daniel Lemire faf921bc7e introducing a thread-local parser and removing ranges (#2412)
* introducing a thread-local parser

* adding functionality to release the memory

* some more documentation.

* fixing build

* adding benchmarks for 'from'

* generalizing the code somewhat.

* adding tests, fixing the benchmark (now with arrays and streams), and a
minor update to document_stream

* adding missing files (I forgot to check them).

* We cannot use [[nodiscard]] without guarding it, it is C++17

* fixing the cmake

* marking it as experimental

* removing ranges support (it is too experimental)

* putting back documentation.

* guarding SIMDJSON_CONSTEVAL more carefully.

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2025-08-13 18:08:25 -04:00
Daniel Lemire 626eedc3d8 documentation update 2025-08-13 16:24:55 -04:00
Daniel Lemire 294aa64f94 Update bug_report.md 2025-08-13 15:01:53 -04:00
Daniel Lemire 397b40e574 more documentation tuning 2025-08-11 15:39:10 -04:00
Daniel Lemire fdc93528d1 updating documentation. 2025-08-11 13:48:01 -04:00
Francisco Geiman Thiesen 3e00a431a0 Merge pull request #2406 from simdjson/fix-convert-ci-failures
Introducing simplified api with from/to adapters (thanks to @the-moisrex for driving this) and fixing all the CI errors.
2025-08-08 10:42:32 -07:00
Pavel Novikov d3626c6a37 removed now unnecessary <utility> header, (#2411)
disabled copy ctor and assignment for `vector_with_small_buffer`
2025-08-08 08:39:28 -04:00
Daniel Lemire 686a1869c8 updating commit 2025-08-08 08:38:41 -04:00
Daniel Lemire 4c0f86db44 minor fixes 2025-08-07 23:59:30 -04:00
Daniel Lemire 6c4c934457 added a new test. 2025-08-07 18:11:44 -04:00
Daniel Lemire fcae795042 minor fix 2025-08-07 16:05:24 -04:00
Daniel Lemire b7c51156b9 Sped up serialization fix (#2409)
* clang format

* added `chars()` method

* implemented vector with small buffer instead of `std::vector`

* added missing <utility> header

* minor fixes

---------

Co-authored-by: Pavel Novikov <dev-ape@yandex.ru>
Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2025-08-07 16:04:20 -04:00
Pavel Novikov 662e3d96c7 Sped up serialization by 10..70%-ish (#2408)
* clang format

* added `chars()` method

* implemented vector with small buffer instead of `std::vector`

* added missing <utility> header
2025-08-07 16:03:39 -04:00
Daniel Lemire 64009f7063 more code simplification. 2025-08-07 13:25:18 -04:00
Daniel Lemire 74fb3ecac7 removing another pragma and some code simplification. 2025-08-07 13:23:21 -04:00
Daniel Lemire 26abf1d180 removing macros 2025-08-07 11:39:51 -04:00
Francisco Geiman Thiesen e878dc6be3 Addressing @the-moisrex review. 2025-08-06 03:05:32 +00:00
Francisco Geiman Thiesen efa03f5733 Fix to_bad_array test to handle both exception and non-exception error cases
The test was failing in CI with g++-13 because it only handled the
exception case. However, the array() method is marked noexcept and
returns a simdjson_result that may contain an error code instead of
throwing an exception.

This fix checks for both cases:
1. If array_result.error() is not SUCCESS, verify it's INCORRECT_TYPE
2. If no error is returned initially, the exception may be thrown when
   iterating over the result

This ensures the test passes regardless of whether the error is
reported via error code or exception.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-05 20:51:17 +00:00
Francisco Geiman Thiesen 8519623257 Fix unused parameter warning in json_iterator::assert_valid_position for SIMDJSON_CLANG_VISUAL_STUDIO
Added (void)position; to suppress unused parameter warning when compiling with SIMDJSON_CLANG_VISUAL_STUDIO defined, where the position parameter isn't used in the SIMDJSON_ASSUME statements.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-05 19:27:01 +00:00
Francisco Geiman Thiesen 484a092c31 Adding a few tests for the simdjson::to adapter. 2025-08-05 16:29:59 +00:00
Francisco Geiman Thiesen 8bb520adbf Merge master and regenerate amalgamated files
Resolved conflicts by regenerating the amalgamated single-header
files (simdjson.h, simdjson.cpp, and singleheader.zip) using the
amalgamate.py script after merging latest changes from master.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-03 20:16:47 +00:00
Francisco Geiman Thiesen ec8e6a6758 Fix -Werror=effc++ warnings and test issues in convert.h
This commit addresses multiple issues:

1. Fixed -Werror=effc++ warnings by using #pragma to disable the
   warning for constructors that cannot initialize all members in
   the member initialization list due to error handling requirements.

2. Added proper error tracking (m_error member) to handle cases where
   document initialization fails, preventing segfaults when using
   invalid documents.

3. Fixed lifetime issues in tests where temporary auto_parser objects
   were being used, causing dangling references. Tests now properly
   store the parser object before using it.

4. Simplified range adaptor tests that were expecting features not
   yet implemented in simdjson's ondemand API.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-03 19:26:58 +00:00
Francisco Geiman Thiesen 769528b6c0 Fix segmentation fault in auto_parser constructor for C++20
The issue was that we were trying to initialize ondemand::document
directly from simdjson_result<ondemand::document> in the member
initializer list. This caused a segfault in C++20 builds.

The fix explicitly handles the simdjson_result in the constructor
body, checking for errors and using value_unsafe() to extract the
document. This avoids potential issues with implicit conversions
and ensures proper error handling.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-03 19:01:42 +00:00
Francisco Geiman Thiesen 9fbc577be7 Fix member initialization order warning in auto_parser
The compiler was warning about member initialization order mismatch.
C++ initializes members in the order they are declared in the class,
not the order they appear in the initializer list.

Fixed by reordering member declarations to match the initialization
order needed: m_doc must be initialized before m_parser since we
need to call parser.iterate() before moving the parser.

This fixes the -Werror=reorder compilation error in CI.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-03 18:48:07 +00:00
Francisco Geiman Thiesen a7c95e9cc8 Fix initialization order in auto_parser constructor
The issue was that we were calling m_parser.iterate() after moving
the parser, which could leave it in an invalid state. In C++20,
this might behave differently than C++17.

Fixed by reordering the member initializer list to call
parser.iterate() BEFORE moving the parser into m_parser.

This ensures the document is created while the parser is still valid.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-03 18:36:14 +00:00
Francisco Geiman Thiesen 248d4eb2aa Try using parentheses instead of braces for document initialization
The issue might be related to how brace initialization vs parentheses
initialization handles implicit conversion from simdjson_result<document>
to document. This could be compiler-specific behavior.

Using parentheses initialization to ensure the conversion operator
is called properly.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-03 18:12:50 +00:00
Francisco Geiman Thiesen 559493e2bf Fix auto_parser constructor to handle document initialization correctly
The issue was that the auto_parser constructor was using implicit
conversion from simdjson_result<document> to document, which could
cause issues with certain implementations (particularly fallback).

Changed to use value_unsafe() to explicitly extract the document
after the parser is fully initialized. This ensures the document
is in a valid state for subsequent operations.

This fixes the ondemand_convert_tests failure in CI with clang++-16.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-03 15:28:40 +00:00
Francisco Geiman Thiesen 75acae5c46 Fix C++20 compatibility issues in convert.h
- Remove constexpr from functions that call non-constexpr methods
- The no_errors and to<T> adaptors were marked constexpr but call
  simdjson_result methods that are not constexpr in C++20
- This was causing compilation failures in CI for C++20 builds
- Tests now compile and pass with both C++17 and C++20

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-03 06:01:50 +00:00
Francisco Geiman Thiesen bd761ef573 Fix ranges support for C++20 compatibility
Instead of disabling the feature, provide C++20-compatible implementation
of the pipe operators for ranges support. The range_adaptor_closure is
C++23-only, so we implement our own pipe operators for C++20.

This preserves the core functionality of the PR while ensuring
compatibility across different compiler versions.
2025-08-03 05:39:10 +00:00
Francisco Geiman Thiesen 76ed73f07f Disable ranges-dependent tests to fix compilation
The test_no_errors() and to_clean_array() tests depend on the C++23
ranges features that we disabled. This commit conditionally compiles
these tests out when ranges support is disabled.
2025-08-03 05:34:34 +00:00
Francisco Geiman Thiesen 1b59b38de8 Disable C++23 ranges features to fix CI compatibility
The ranges features were causing compatibility issues across different
compilers and platforms. Disabling them for now until C++23 support
is more widespread.

This should fix the remaining Ubuntu and Windows CI failures.
2025-08-03 05:11:31 +00:00
Francisco Geiman Thiesen 4794b5d936 Update amalgamated files with convert.h fixes
Regenerate singleheader/simdjson.cpp and singleheader/simdjson.h
to include all the fixes for CI compatibility issues.
2025-08-02 17:06:45 +00:00
Francisco Geiman Thiesen 114f14924c Simplify ranges feature detection for C++23
Only enable range_adaptor_closure features when compiling with C++23
or later, as this feature is not available in C++20 implementations.
2025-08-02 16:04:58 +00:00
Francisco Geiman Thiesen e94825c9a8 Fix preprocessor check for __cpp_lib_ranges_zip
Add defined() check before comparing the value to avoid preprocessor
errors in compilers where this macro doesn't exist (like g++-13
with certain configurations).
2025-08-02 15:43:31 +00:00
Francisco Geiman Thiesen 333fd72f98 Fix CI failures in convert.h implementation
- Fix deprecated reflect_value warning by using reflect_constant
- Fix std::const_iterator C++23 requirement by using auto_iterator
- Fix C++23 std::ranges::range_adaptor_closure availability check
- Add convert.h to main simdjson.h includes

These changes ensure compatibility across different C++ standards
and compiler versions, fixing the Ubuntu CI failures.

Co-Authored-By: Claude <noreply@anthropic.com>
2025-08-02 15:04:25 +00:00
Daniel Lemire dd4d02617f removing unnecessary hack 2025-07-31 18:40:54 -04:00
Daniel Lemire e28d63334e fix 2025-07-31 13:07:29 -04:00
Daniel Lemire 3d1ab87ecb Merge branch 'master' into release_candidate_4_0_0 2025-07-31 10:24:54 -04:00
Borislav Stanimirov d365cfb4a1 remove cmake_policy (#2404)
* properly gitignore Visual Studio artifacts

* remove cmake_policy
2025-07-31 10:18:28 -04:00
Daniel Lemire b9e308a727 marking them as 'really inline' 2025-07-28 11:47:09 -04:00
Brad Bramble aa6817d5aa Fix linker errors for downstream users caused by non-inline symbols in header (#2403) 2025-07-28 11:44:32 -04:00
Francisco Geiman Thiesen 90e4a66c93 Bringing a bit more use-cases for reflection based serializations (optional). Also adding string-based enum handling as requested on X. (#2395)
* Adding type validation, enhancing optional type support and adding test a few more tests.

* Adding support for string-based enum serlalization and deserialization.

* Removing unintentional endline.

* Removing trailing whitespace.

* Adding simpler api as suggested by moisrex.

* Removing explicit optiona<int> and optional<std::string> references and using concepts instead! Credit goes to Lemire for pointing this out and suggesting a concepts based approach here.

* Removing tests that are not relevant for this branch.

* Removing api related changes. That will be done by moisrex.

* Removing unnecessary new endlines.

* Removing tests related to api changes and cleaning-up irrelevant tests.

* removing broken reference

* Removing trailing whitespace
2025-07-23 09:24:05 +02:00
evbse e6240f18c4 Add version to amalgamated files (#2400) 2025-07-23 09:20:06 +02:00
M. Bahoosh d43fb6ff84 test for no_errors 2025-07-21 05:24:12 -10:00
M. Bahoosh 5eec29a6db Moving iterator's storage to auto_parser 2025-07-21 05:18:13 -10:00
M. Bahoosh 228501f786 From/To adaptors 2025-07-21 03:22:20 -10:00
M. Bahoosh 9d9f2427c5 Make auto_parser and auto_iterator comply with ranges. 2025-07-20 03:22:07 -10:00
M. Bahoosh 8d53840253 Removing unneeded code 2025-07-19 08:37:04 -10:00
M. Bahoosh 0a82fb110f Auto Iterator 2025-07-19 08:35:49 -10:00
M. Bahoosh 11f273c580 Moving ondemand::document into auto_parser 2025-07-19 04:37:13 -10:00
M. Bahoosh 31662531ac Basic Auto Parser 2025-07-19 03:58:21 -10:00
Daniel Lemire 97eb557388 Merge branch 'release_candidate_4_0_0' of github.com:simdjson/simdjson into release_candidate_4_0_0 2025-07-16 12:08:03 -04:00
Daniel Lemire 41ced28821 adding a space 2025-07-16 12:07:49 -04:00
Daniel Lemire 1078eb4034 minor update to the release candidate (#2394) 2025-07-16 12:05:09 -04:00
evbse 6029af75b3 Improve fallback implementation (#2393) 2025-07-15 14:39:09 -04:00
Daniel Lemire 0cd774097f [skip ci] doc fixes 2025-07-14 23:46:30 -04:00
Daniel Lemire b5e27af4da release candidate 4.0.0 2025-07-14 15:55:00 -04:00
Daniel Lemire c806e955c4 C++26 static reflection (#2282)
* Initial work on JSON builder

* moving the files back to ondemand for now.

* tweak

* more later

* update

* minor edits

* dropping vs arm (missing support)

* adding tests. we still specialized write_string_escaped

* tweaking

* fix typo

* tweaking the approach

* minor fix

* missing store

* another missing store

* Attempt at fixing failing serialization tests. (#2292)

* Fixing appeand_float typo (#2294)

* applying a couple of fixes

* updating single header

* fix for pre C++17 if constexpr

* Fixing unused argument problem and updating the singleheader file

* various pedantic fixes

* Sketch of builder

* reordering.

* simplify

* Adding draft of static reflection based deserialization

* Updating simdjson singleheader

* patching the automated deserialization.

* automated

* Adding support for smart pointers of user defined types.

* Adding specialization for smart pointers for basic types. I think it is highly likely that this can be done in a more generic way.

* Referncing a later version of rapidjson that fixed the issue related with assignment attempt of a const variable for GenericStringRef class.

* guarding the tests

* adding documentation for string_builder

* saving

* rename to 'append'

* saving

* non-functional benchmarks (#2342)

* non-functional benchmarks

* Fix typo

* various fixes

* tweaking

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
Co-authored-by: Francisco Geiman Thiesen <franciscogthiesen@gmail.com>

* tuning

* various minor fixes

* minor tweak

* minor simplification

* updating amal

* adding a cast

* update

* fancy casting

* removing dead code

* Pushing latest changes. CITM benchmark is still not working.

* Still not working, but now I am getting only 10 errors.

* add static reflection benchmark to 'large random' benchmark and allows (#2349)

deserialization (with static reflection) from objects and arrays.

Co-authored-by: Daniel Lemire <dlemire@lemire.me>

* Removing std::map from CitmCatalog definition, since that is not currently supported.

* Added free to rust bench, segfault is still happening..

* The syntax changed: ^E became ^^E. (#2350)

* The syntax changed: ^E became ^^E.

* guarding

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>

* Adding support for string_view_keyed_map types.

* Adding concepts as a conditional include.

* updating single-header

* Adding concepts to ondemand deps

* rust benchmark is finally working

* Fixing small typo in docs.

* adding docker config and instructions so that our users can test the static reflection (#2358)

* adding docker config and instructions so that our users can test the
static reflection

* completing the instructions

* pruning white spaces

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>

* minor optimizations on the JSON builder branch

* avoiding undef behaviour

* saving

* somewhat nicer builder

* make it possible to run just one benchmark

* adding linux perf

* fixing minor issue

* updating swar

* Adding real world compilation benchmark (#2379)

* Adding compilation benchmark for json parsing with and without reflection

* Moving it to the benchmark folder, also reducing a bit the number of iterations.

* Removing script from root folder.

* Reducing number of iterations

* Update benchmark/benchmark_reflection_usage_compilation.sh

Co-authored-by: Daniel Lemire <daniel@lemire.me>

* Update benchmark/benchmark_reflection_usage_compilation.sh

Co-authored-by: Daniel Lemire <daniel@lemire.me>

* Update benchmark/benchmark_reflection_usage_compilation.sh

Co-authored-by: Daniel Lemire <daniel@lemire.me>

* Making the script more customizable and also test whether the compiler being used supports reflection before actually running the benchmark

---------

Co-authored-by: Daniel Lemire <daniel@lemire.me>

* Using define_static_string from  https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2025/p3491r2.html (#2389)

* Applying changes needed after latest reflection paper updates.

* Working, but no template for yet.

* Updating single-header to incldue the use of define_static_string.

* copying over master

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
Co-authored-by: Francisco Geiman Thiesen <franciscogthiesen@gmail.com>
2025-07-14 15:43:52 -04:00
Daniel Lemire 4acaf2ea1c improving DOM ranges test and silencing a warning (#2385)
* improving DOM ranges test and silencing a warning

* moving test_main to macros (even though it is not a macro)

* moving test_main
2025-07-04 15:46:50 -04:00
Daniel Lemire 86adda06d0 clarifying NDEBUG usage (#2388)
* clarifying NDEBUG usage

* init
2025-07-04 15:46:37 -04:00
Daniel Lemire 8df5f96eba guarding the POSITION_INDEPENDENT_CODE (#2386) 2025-06-30 16:41:49 -04:00
xkszltl 937667796c Replace -fPIC with POSITION_INDEPENDENT_CODE. (#2383) 2025-06-30 11:26:27 -04:00
Cuda Chen 7d86b04247 Add document and tests of std::ranges support of DOM API (#2381)
* Add document and tests of std::ranges support of DOM API

* Alter lambda

* Fix require of range
2025-06-24 15:13:44 -04:00
Daniel Lemire 265022ad8f Fixing MSYS2 issue (#2380)
* Google Benchmarks does not support MSYS2.

* narrowing it down
2025-06-21 21:26:52 -04:00
Daniel Lemire 0c0ce1bd48 3.13.0 2025-06-04 00:46:03 -04:00
Daniel Lemire f64c004cb7 making it easier to convert to std::string (#2378)
* making it easier to convert to std::string

* typo

* guarding C++20

* bad semi-colon
2025-06-04 00:22:10 -04:00
Daniel Lemire d7d19db6ae Emscripten (#2377)
* emscripten

* cmake_policy

* space
2025-06-03 09:10:59 -04:00
Daniel Lemire a260c967ed adding test for issue 2375 (#2376)
* adding test for issue 2375

* update

* removing space
2025-06-02 09:24:40 -04:00
Daniel Lemire 16e390d81a fix for issue 2373 (#2374)
* fix for issue 2373

* bumping google bench

* removing json11 since dropbox retired it

* removing json11 as a dependency
2025-05-31 00:56:19 -04:00
Daniel Lemire 8d34e14000 Update dom.md 2025-05-31 00:27:28 -04:00
shuiyisong b717136fd9 chore: add greptimedb in real world usage (#2372) 2025-05-07 16:33:24 -04:00
Daniel Lemire f3ac74caf8 slightly better documemntation of the C++20 features (#2371)
* slightly better documemntation of the C++20 features

* adding another remark
2025-04-23 18:32:51 -04:00
Daniel Lemire 2369927661 Moving from Ubuntu 20 to 24 (#2366) 2025-04-18 17:56:46 -04:00
Daniel Lemire 818c0491a1 Treat -0 as -0.0 when SIMDJSON_MINUS_ZERO_AS_FLOAT is set (#2364)
* In the DOM API, treat -0 as -0.0

* documenting...

* adding it to On-Demand
2025-04-18 13:56:02 -04:00
Daniel Lemire 9f14f90a64 [documentation] We run on practically all systems. (#2365)
* We run on practically all systems.

* Update doc/basics.md

Co-authored-by: Antoine Pitrou <pitrou@free.fr>

---------

Co-authored-by: Antoine Pitrou <pitrou@free.fr>
2025-04-18 13:48:13 -04:00
Daniel Lemire 00843d2711 better documentation 2025-04-16 11:45:52 -04:00
Daniel Lemire 2887a17bab bumping up nlohmann/json 2025-04-13 14:43:30 -04:00
huangqinjin d84c934768 do not try to include <string_view> without C++17 for MSVC (#2360)
Doing that will emit warning at https://github.com/microsoft/STL/blob/vs-2022-17.13/stl/inc/string_view#L12.
2025-03-28 19:23:15 -04:00
huangqinjin 7cec7c7ae4 fix VS2017 exception detection (#2359)
MSVC option /EHsc defines __cpp_exceptions only in VS2019 and above,
_CPPUNWIND should be used before VS2019.

https://learn.microsoft.com/en-us/cpp/preprocessor/predefined-macros#microsoft-specific-predefined-macros
2025-03-28 19:22:56 -04:00
Daniel Lemire c52b010a57 missing string_view include (#2361) 2025-03-28 19:22:38 -04:00
Daniel Lemire ee301599b1 Update basics.md 2025-03-28 11:53:07 -04:00
Daniel Lemire 7382dc2be8 preparing patch release 3.12.3 (#2357) 2025-03-28 11:26:19 -04:00
Daniel Lemire 8c14e0c56f fixing issue 2354: get_bool() on document with trailing spaces (#2356) 2025-03-27 15:01:10 -04:00
Daniel Lemire a9a62feb75 Map proto (#2353)
* std::map support

* completed

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2025-03-23 16:32:48 -04:00
Ezequiel Ramis Folberg b9228b4d3c Update README.md (#2352) 2025-03-22 15:49:02 -04:00
Dirk Stolle 726c3eb611 fix several typos (#2348) 2025-03-17 09:57:08 -04:00
Daniel Lemire 4cdc4f18ef documenting fatal errors 2025-03-13 13:26:50 -04:00
Dirk Eddelbuettel f3b034ac38 Address 'whitespace in literal' decreation warning from clang++-20 2025-03-06 14:28:01 -06:00
Daniel Lemire 9c2e8a8f39 upgrading uraimo/run-on-arch-action 2025-03-01 14:32:13 -05:00
yun dfa43f6cdd doc: replace ndjson[dot]org with ndjson spec (#2340)
the `ndjson[dot]org` expired, and point to incorrect website with malware.
more in https://github.com/ndjson/ndjson.github.io/issues/24

follow up #2234
2025-02-21 09:07:39 +00:00
Daniel Lemire 797e61742c release bump 2025-02-14 16:12:28 -05:00
Daniel Lemire f289412e0a fixing performance issue under Zen 4 processors (#2335)
* fixing performance issue under Zen 4 processors

* trying something else

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2025-02-14 16:11:36 -05:00
Daniel Lemire 7bd79b4445 Update README.md 2025-02-14 16:03:01 -05:00
Daniel Lemire 078e2c9073 patch release 3.12.1 2025-02-11 13:46:03 -05:00
Daniel Lemire d7b6b20511 creating zip files. 2025-02-11 13:43:34 -05:00
Daniel Lemire dbea3bbd62 Update README.md 2025-01-29 11:48:36 -05:00
Daniel Lemire e422933414 release 3.12.0 2025-01-27 20:35:57 -05:00
Daniel Lemire de4d69b367 Introducing dual licensing (#2328)
* Introducing dual licensing

* adding missing file
2025-01-27 20:34:35 -05:00
Paul Dreik b8675a7f7b fix fuzzing when running on icelake (#2327) 2025-01-26 05:36:33 +01:00
Paul Dreik 5642bb93a4 fix fuzzer CI job (#2324) 2025-01-16 13:33:53 -05:00
Daniel Lemire 1b23a77e03 3.11.6 2025-01-14 21:33:29 -05:00
Daniel Lemire 57699bfed8 release candidate (#2323) 2025-01-09 11:01:48 -05:00
Daniel Lemire 648303b26a typo 2025-01-08 21:29:11 -05:00
Daniel Lemire 9008960e36 fixed comment 2025-01-08 20:35:30 -05:00
Daniel Lemire 8a9e8a1792 fixing issue 2322 2025-01-08 20:28:25 -05:00
Daniel Lemire ba33e9e78f release 2025-01-06 19:33:53 -05:00
Daniel Lemire d98b351eef let us be more careful with concepts, we require __cpp_concepts >= 201907L and check for old versions of Apple clang (#2321)
* let us be more careful with concepts, we require __cpp_concepts >= 202002L

* checking Apple version

* let us be more specific

* removing extra endif
2025-01-06 19:23:09 -05:00
Daniel Lemire 5488dca126 Accidentally, it appears that PPC64 was always disabled by default. (#2320) 2025-01-03 11:01:04 -05:00
Daniel Lemire 7712ecf164 adding char bit check (pedantic) (#2316) 2024-12-31 14:34:13 -05:00
Paul Dreik 2803ca3093 use clang 19 in fuzzer job (#2315)
* use clang 19 in fuzzer job

* fix syntax error in msys2-clang.yml
2024-12-22 15:00:58 -05:00
Daniel Lemire e7f2463920 some tweaks 2024-12-17 16:27:31 -05:00
Daniel Lemire 5bfa0b098c patch release 3.11.3 (#2313)
* preparing patch release 3.11.3
2024-12-12 13:38:22 -05:00
Daniel Lemire f7ba9cb11b Update basics.md 2024-12-10 11:47:21 -05:00
Daniel Lemire d4bf0cc7ec sync 2024-12-09 21:36:43 -05:00
Daniel Lemire 2fbbea0b15 Merge branch 'master' of github.com:simdjson/simdjson 2024-12-09 21:36:12 -05:00
Daniel Lemire c16486f702 release candidate 3.11.2 2024-12-09 21:35:21 -05:00
zhanglistar f615112093 Disabling memory sanitizer with parse_string (#2310) 2024-12-09 10:20:33 -05:00
Daniel Lemire 6bbcbfbb95 version 3.11.1 2024-12-07 11:13:00 -05:00
Valery Mironov e6578fea39 Fix missing override (#2305) 2024-12-07 11:12:25 -05:00
Joyee Cheung 6fb050d04e fix: fix deprecation from -Wdeprecated-literal-operator (#2307)
Otherwise simdjson doesn't build with V8's Node.js fork, which
uses -Werror,-Wdeprecated-literal-operator and latest version
of clang.
2024-12-07 11:12:14 -05:00
Daniel Lemire b4242d3b4f release candidate (#2304) 2024-12-05 22:05:39 -05:00
Daniel Lemire cbfe2d4a21 Update vs17-arm-ci.yml 2024-12-02 20:24:24 -05:00
halx99 79126ca323 fix: processor checks inconsistently problem (#2299) 2024-12-02 20:23:28 -05:00
Daniel Lemire 3f87f958d7 documentation 2024-12-02 00:06:20 -05:00
Daniel Lemire cab383e1de Update README.md 2024-11-30 20:43:20 -05:00
Daniel Lemire 9b314922a1 additional documentation 2024-11-08 11:53:34 -05:00
Daniel Lemire 587beae307 prevent int to pointer implicit conversion for operator[](const char *) (#2286) 2024-11-01 18:11:43 -04:00
Daniel Lemire f53981d945 Update dom.md 2024-11-01 12:23:35 -04:00
Daniel Lemire 41b4ffd6aa Update basics.md 2024-11-01 12:15:40 -04:00
Daniel Lemire 6cd8a6d776 adding pad function (#2283) 2024-10-29 21:26:01 -04:00
Daniel Lemire 3c0d032ded Candidate for C++20 deserialization features (#2276)
* tag_invoke based custom types (#2219)

* tag_invoke based custom types

Now you can use tag_invoke to add a custom type or a group of custom types.

* Fixing macro usage + Fixing noexcept

* Fixing the usage of #include

We don't need <concepts> at all seems like it

* Fixing tag_invoke impl for MSVC

* Making `tag_invoke` to support `ondemand::document` as well + docs (#2228)

* Making `tag_invoke` to support `ondemand::document` as well + docs

* Fix typos and doc update by @lemire

Co-authored-by: Daniel Lemire <daniel@lemire.me>

* Better docs by @lemire

Co-authored-by: Daniel Lemire <daniel@lemire.me>

* Preserving the old, disallowing in the new

I'm disabling `document::get() &&` if the user has provided a `tag_invoke`d version; otherwise, we retain the compatibility.

---------

Co-authored-by: Daniel Lemire <daniel@lemire.me>

* fix: correct small issues with deserialize (#2232)

* Extending the deserialization code with more defaults + docs (#2233)

* Make custom types easier with some predefined cases + docs

* missing include

* adding Ubuntu 24 CXX 20

* using concepts all the way

* minor tweak

* tiny tweak

* tweaks

* more tweaking

* saving

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>

* Making `tag_invoke` a "put" as opposed to a "get" (#2256)

* fix: add tests related to issue 2227 (#2229)

* fix: add tests related to issue 2227

* avoiding name clash

* pedantic fix

* deprecate rvalue get on document

* selectively deprecating

* Fix ndjson spec link (#2234)

* fix ndjson spec link

The link in the readme of parse_many links to a casino spam site

* fix link

* [no-ci] Update README.md

* Make simdjson compile again

* Enable SIMDJSON_SINGLEHEADER=OFF in VS Code

With singleheader on, clangd can't find the right
include files.

* Add missing include directives to static build targets of simdjson. (#2240)

* adding a warning

* adding warning regarding SIMDJSON_BUILD_STATIC_LIB

* release candidate

* pedantic viable size

* Making tag_invoke a feeder instead of a producer

* adding missing undef silencer (#2253)

* Ignore pragma once when amalgamating source files (#2248)

With gcc it causes an error in `simdjson.cpp`:
```
simdjson.cpp:548:9: warning: #pragma once in main file
  548 | #pragma once
      |         ^~~~
```

It had previously been commented out in:
https://github.com/simdjson/simdjson/commit/6ef555e6fb79363fae057a9a46b52cd208d9e305

However, this was lost in an upgrade:
https://github.com/simdjson/simdjson/commit/2a4ff7346813b120f2b5b40e95d69352b593cc9c

* Update CI (#2254)

* adding missing undef silencer

* Updating CI

* more fixes

* fix

* big endian fix

* Moving to the new tag_invoke signature

* Fix nlohmann ambiguity on C++23-enabled clang

* Revert "Merge branch 'master' of https://github.com/simdjson/simdjson into builder_development_branch_extra"

This reverts commit 3eeecbab34, reversing
changes made to 6858b208b4.

---------

Co-authored-by: Daniel Lemire <daniel@lemire.me>
Co-authored-by: Sasha Lopoukhine <superlopuh@gmail.com>
Co-authored-by: John Keiser <john@johnkeiser.com>
Co-authored-by: Tan Li Boon <undisputed-seraphim@users.noreply.github.com>
Co-authored-by: tobil4sk <tobil4sk@outlook.com>

* update CI on the builder_development_branch (no code change) (#2262)

* typo

* General madness simpler, no simpler!!! (#2267)

* Minimal tag_invokes for STL types

* simpler madness

* adding a comment

* missing file

* minor tweaks to style

* fixing incorrect max/min usage

* updating single

* simplify

* validating the idea

* putting back the concept

* moving the include

* guarding

* Cheap General Madness (#2268)

* Some General Concepts and their deserializations

* Resolving ambiguity

* Add missing #include

* C++20 custom deserializer: better documentation (#2269)

* mostly a documentation update.

* missing cpp

* [no-ci] fix comment

* various minor fixes

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>

---------

Co-authored-by: M. Bahoosh <12122474+the-moisrex@users.noreply.github.com>
Co-authored-by: Daniel Lemire <dlemire@lemire.me>
Co-authored-by: M. Bahoosh <moisrex@gmail.com>

* minor update

* More documentation regarding builder (#2270)

* minor update

* more improvment to our documentation (builder branch)

* putting back missing functions

* merge candidate

---------

Co-authored-by: M. Bahoosh <moisrex@gmail.com>
Co-authored-by: Daniel Lemire <dlemire@lemire.me>
Co-authored-by: Sasha Lopoukhine <superlopuh@gmail.com>
Co-authored-by: John Keiser <john@johnkeiser.com>
Co-authored-by: Tan Li Boon <undisputed-seraphim@users.noreply.github.com>
Co-authored-by: tobil4sk <tobil4sk@outlook.com>
Co-authored-by: M. Bahoosh <12122474+the-moisrex@users.noreply.github.com>
2024-10-18 20:37:38 -04:00
Daniel Lemire 36f5dbcb75 documentation for threading... (#2277)
* documentation for threading...

* remove const

* doc
2024-10-11 12:34:11 -04:00
Daniel Lemire 49901fb254 doc 2024-10-10 21:32:51 -04:00
Daniel Lemire c066b5421b Jsonpath (#2272)
* Add support for JSONPath with '$' prefix and integrate into dom::at_path (#2266)

- Updated `json_path_to_pointer_conversion` to support JSONPath starting with the '$' prefix, while maintaining compatibility with the existing implementation.
- Moved `json_path_to_pointer_conversion` to a separate header file for better modularity and to support JSONPath queries in `dom` mode.
- Implemented `at_path` functionality in `dom` mode to enable querying JSON using JSONPath.
- Added unit tests to validate the new JSONPath support in `dom` mode and ensure compatibility with both standard and existing JSONPath formats.

* two minor fixes

* more tests and documentation

* damn compiler warnings

* more documentation fixes

---------

Co-authored-by: Zhengguo Yang <yangzhgg@gmail.com>
2024-10-08 10:37:59 -04:00
Daniel Lemire ec7550a70e fixes issue 2271 (#2273) 2024-10-08 10:36:12 -04:00
Daniel Lemire 3ef3078e51 documentation for issue 2259 (clarifying) (#2260) 2024-09-21 09:10:35 -04:00
tobil4sk fd06782c97 Make document::is_null behavior consistent with documentation (#2258)
* Make null-like value test consistent with docs

According to the docs: `INCORRECT_TYPE If the JSON value begins with 'n'
and is not 'null'.`

* Match document::is_null behavior to documentation

* Test other token beginning with n with is_null
2024-09-20 20:58:41 -04:00
Daniel Lemire 9303efbd0c [no-ci] Update basics.md 2024-09-20 20:56:11 -04:00
Daniel Lemire 6c979f15cc deps: switch to CPM (#2257)
* deps: switch to CPM

* missing file

* setting the dependencies on URL download

* using fixed version of cxxopts
2024-09-20 13:29:47 -04:00
Daniel Lemire 09ccabbe6c enable BMI, but only when LLVM is the compiler (#2255) 2024-09-20 10:59:54 -04:00
Daniel Lemire e00cc8c6dc Update CI (#2254)
* adding missing undef silencer

* Updating CI

* more fixes

* fix

* big endian fix
2024-09-18 09:09:11 -04:00
tobil4sk c10b32d463 Ignore pragma once when amalgamating source files (#2248)
With gcc it causes an error in `simdjson.cpp`:
```
simdjson.cpp:548:9: warning: #pragma once in main file
  548 | #pragma once
      |         ^~~~
```

It had previously been commented out in:
https://github.com/simdjson/simdjson/commit/6ef555e6fb79363fae057a9a46b52cd208d9e305

However, this was lost in an upgrade:
https://github.com/simdjson/simdjson/commit/2a4ff7346813b120f2b5b40e95d69352b593cc9c
2024-09-17 15:41:12 -04:00
Daniel Lemire 025a44348a adding missing undef silencer (#2253) 2024-09-16 21:04:09 -04:00
Daniel Lemire 70a68da941 pedantic viable size 2024-09-06 01:19:47 -04:00
Daniel Lemire e341c8b438 release candidate 2024-08-26 09:37:33 -04:00
Daniel Lemire 0ac0a80e28 Merge branch 'master' of github.com:simdjson/simdjson 2024-08-26 09:37:03 -04:00
Daniel Lemire 6b9117c029 adding warning regarding SIMDJSON_BUILD_STATIC_LIB 2024-08-26 09:36:44 -04:00
Daniel Lemire dd92151971 adding a warning 2024-08-26 09:34:39 -04:00
Tan Li Boon 0679c247f4 Add missing include directives to static build targets of simdjson. (#2240) 2024-08-26 09:24:01 -04:00
John Keiser 615218a3ad Merge pull request #2237 from simdjson/jkeiser/simdjson-vscode
Reduce errors in vscode
2024-08-18 22:04:20 -07:00
John Keiser 4c1b0a41d8 Enable SIMDJSON_SINGLEHEADER=OFF in VS Code
With singleheader on, clangd can't find the right
include files.
2024-08-18 14:21:00 -07:00
John Keiser ef563a4b09 Make simdjson compile again 2024-08-18 11:21:26 -07:00
Daniel Lemire 7a9ff93388 [no-ci] Update README.md 2024-08-15 12:52:47 -04:00
Sasha Lopoukhine fc61d7c7ba Fix ndjson spec link (#2234)
* fix ndjson spec link

The link in the readme of parse_many links to a casino spam site

* fix link
2024-08-10 10:11:05 -04:00
Daniel Lemire d506af0a79 fix: add tests related to issue 2227 (#2229)
* fix: add tests related to issue 2227

* avoiding name clash

* pedantic fix

* deprecate rvalue get on document

* selectively deprecating
2024-08-07 20:15:06 -04:00
Daniel Lemire ccf8694510 v3.10.0 2024-08-01 09:32:54 -04:00
Daniel Lemire 9b67497ed0 Allows field::unescape_key to take in a string parameter + additional dev. checks for string overflow (#2224)
* This PR does the following:

1. Upgrade cxxopts.
2. Allows field::unescape_key to take in a string parameter (syntaxic sugar).
3. Adds a dev. check to detect a string buffer overflow (indicating broken code). Note that this is unrecoverable and indicates bad code.

* tweak
2024-08-01 09:31:50 -04:00
Daniel Lemire 0336684df7 [no-ci] Update basics.md 2024-07-31 11:10:01 -04:00
Daniel Lemire c19320dd6e making it more precise 2024-07-31 10:08:04 -04:00
Daniel Lemire 412a5680e8 update 2024-07-31 10:06:47 -04:00
Daniel Lemire a05a56856d fix: use On-Demand throughout. (#2222) 2024-07-29 15:54:21 -04:00
didarpin 58173a6a1f Added the functionality to convert dom::object and dom::array to dom::element. (#2221)
Co-authored-by: didarpin <didarpin@163.com>
2024-07-25 22:26:20 -04:00
Francisco Geiman Thiesen 1721032cfd Merge pull request #2220 from simdjson/adding_macros_for_cpp20_cpp23
fix: add macros to detected C++20 and C++23
2024-07-25 17:44:36 -07:00
Daniel Lemire b73877f95e Update compiler_check.h 2024-07-23 14:25:27 -04:00
Daniel Lemire 09723897e9 fix: add macros to detected C++20 and C++23 2024-07-23 14:23:00 -04:00
Daniel Lemire 49e231b634 [no-ci] Update README.md 2024-07-16 16:05:32 -04:00
Daniel Lemire acdbbab916 adding an example of value capture with std::string_view (#2216)
* adding an example of value capture with std::string_view

* minor fix

* minor fix
2024-07-16 15:51:43 -04:00
Daniel Lemire 5090247c34 adding test (#2214) 2024-07-13 11:40:13 -04:00
Daniel Lemire 4180e05730 [no-ci] fix 'null_ptr' written as 'null_nullptrptr' in the comments 2024-07-11 08:25:27 -04:00
Daniel Lemire feea2bce2c Create config.yml 2024-07-04 16:48:09 -04:00
Daniel Lemire 692f43cd84 Update standard-issue-template.md 2024-07-04 16:46:34 -04:00
Daniel Lemire 3240d55bcc chore: add RelWithDebInfo to our CI tests (#2209)
* chore: add RelWithDebInfo to our tests

* fix

* chore: add various build types to ubuntu ci
2024-07-04 16:26:22 -04:00
Daniel Lemire 66fd28fc00 Marking a few functions as pure (no side-effect) (#2210)
* marking a few trivial functions as pure

* adding other marks

* additional marks

* vs will issue warnings, so don't use [[gnu::pure]] when __clang__ or __GNUC__ is not defined
2024-07-04 16:26:11 -04:00
Daniel Lemire 5f638951c6 Update basics.md 2024-06-27 14:49:57 -04:00
Daniel Lemire 3e94eea939 adding another example (#2206)
Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2024-06-27 13:22:44 -04:00
Tunghohin 0e8311f812 reset moved object's viable_size to 0 in the move ctor of padded_string (#2204) 2024-06-25 18:06:32 -04:00
Daniel Lemire 3620e9d151 version bump 2024-06-11 15:37:22 -04:00
Daniel Lemire d017cd7ca4 Merge branch 'master' of github.com:simdjson/simdjson 2024-06-11 14:08:20 -04:00
Daniel Lemire d9d1ff5856 chore: remove unneeded gcc13 ci tests 2024-06-11 14:07:56 -04:00
Daniel Lemire eb8f2bce14 fix: add test for issue 2199 (#2200)
* fix: add test for issue 2199

* added ubuntu 24 workflow + silencing a warning
2024-06-11 13:55:55 -04:00
Dirk Stolle 2a4ff73468 update string_view lite to version 1.8.0 (#2197)
This is the header as seen for the tag v1.8.0,
commit a47222b9855dd6e6d1eac38acaa495822e2caa69, on
<https://github.com/martinmoene/string-view-lite>.
2024-06-10 10:29:25 -04:00
Daniel Lemire 77fc2b8447 doc: explaining the page trick (#2196)
* doc: explaining the page trick

* simplify

* did as john said

* trying something else

* flipping order

* trying some other order

* hmmm
2024-06-07 22:12:08 -04:00
Daniel Lemire ba8b66a633 Update basics.md 2024-06-07 12:49:32 -04:00
Daniel Lemire 66eec5feaf Update basics.md 2024-06-05 08:55:12 -04:00
Janeczko Jakub 3964f3e5d2 pull size_t from the std namespace (#2191) 2024-06-03 14:08:23 -04:00
Daniel Lemire ee8515122d version bump 2024-05-30 10:53:40 -04:00
halx99 5d35e7ca1f Fix compile error on llvm-19 (#2187) 2024-05-30 10:52:38 -04:00
spershin deefc88b9c Adding path for parsing incomplete json. (#2189)
1. Allows processing inclomplete, damaged, corrupted json to some extent.
2. Pariity with the Presto Java functionality.
3. Protected with SIMDJSON_EXPERIMENTAL_ALLOW_INCOMPLETE_JSON define.
4. Does not interfere with the normal path (can co-exist).
5. Tested in production forkflow.
2024-05-30 10:52:09 -04:00
Yuriy Chernyshov c80dda7c58 Fix building simdjson against libc++ with _LIBCPP_REMOVE_TRANSITIVE_INCLUDES defined (#2184)
```
src/implementation.cpp:193:20: error: no template named 'is_trivially_destructible' in namespace 'std'; did you mean 'is_trivially_move_constructible'?
static_assert(std::is_trivially_destructible<detect_best_supported_implementation_on_first_use>::value, "detect_best_supported_implementation_on_first_use should be trivially destructible");
              ~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~
                   is_trivially_move_constructible
```
2024-05-23 16:54:23 -04:00
Daniel Lemire d2954ef68b Update ubuntu22-gcc13.yml 2024-05-23 16:53:48 -04:00
Daniel Lemire ac719827ff fix: solve issue 2181 (#2182) 2024-05-11 20:44:38 -04:00
Daniel Lemire 6ea77392a7 Update basics.md 2024-05-10 12:21:26 -04:00
pnck e2f879751c fix: issue #2154 (#2178) 2024-05-10 00:33:09 -04:00
Daniel Lemire bf7834179c version bump 2024-05-07 19:15:43 -04:00
Daniel Lemire 69ab8848bf fix: satisfy a C++20 compiler warning (#2177) 2024-05-07 18:04:59 -04:00
Daniel Lemire 7288323e36 Update basics.md 2024-05-06 21:09:56 -04:00
Daniel Lemire 3f381cf3ee Update README.md 2024-05-06 11:11:34 -04:00
Daniel Lemire 52406402ed fix: replace vm*vq_u8 by vm*vq_u32 for better performance under some systems. 2024-05-01 14:53:48 -04:00
Daniel Lemire 799d8e3fc9 fix ref 2024-04-26 14:09:42 -04:00
Daniel Lemire 6c647d9f3b Update README.md 2024-04-24 16:23:45 -04:00
Daniel Lemire 8519e24f12 issue 2170 (#2172) 2024-04-24 01:41:40 -04:00
Daniel Lemire fffb62743c removing power fuzzer. 2024-04-23 23:50:49 -04:00
Daniel Lemire c888075f8d makes implementations trivially destructible (#2171) 2024-04-23 23:49:20 -04:00
Daniel Lemire 9b8aac5f22 [skip ci] updating documentation 2024-04-22 09:18:49 -04:00
Javier Blazquez 40b414d184 add Windows ARM64EC build and CI support (#2168)
* add Windows ARM64EC support

* add ARM64EC to vs17-arm-ci workflow
2024-04-09 14:56:58 -04:00
Daniel Lemire c85e8a7db1 version bump 2024-04-05 15:19:02 -04:00
Daniel Lemire 83f390f8e7 fix: adopt recommendation from issue 2163 (#2164) 2024-04-05 15:17:57 -04:00
Daniel Lemire 82563dcf70 correcting version bump 2024-04-04 14:08:25 -04:00
Daniel Lemire 646bd3f2c3 version bump 2024-04-04 12:24:07 -04:00
Daniel Lemire 4b0a908bb2 [skip ci] adding documentation regarding new hardware support 2024-04-04 12:22:03 -04:00
Jinyang He 4c98e51c53 Add LoongArch LSX and LASX support (#2159)
* Add LoongArch SX support

* Add LoongArch ASX support
2024-04-04 12:09:36 -04:00
Daniel Lemire 58e3d5d9cc fix: fix path to cross-compiler 2024-04-03 22:37:46 -04:00
Daniel Lemire a22a92dc72 trying to add loong to our ci tests (#2160)
* trying to add loong to our ci tests

* tuning loongarch

* updating the loong toolset
2024-04-03 15:18:45 -04:00
Daniel Lemire c4654553dd [skip ci] windows and unicode 2024-03-26 09:31:50 -04:00
Daniel Lemire 5997f82c37 Update basics.md 2024-03-20 14:26:11 -04:00
Daniel Lemire bee410b315 comment 2024-03-20 01:52:42 -04:00
Daniel Lemire b115e3ef61 some updates to our CI (#2151) 2024-03-17 11:59:50 -04:00
Daniel Lemire 669229df5d more documentation. 2024-03-17 10:29:41 -04:00
Daniel Lemire b6a2286e62 adding some documentation for escaped_key(). 2024-03-17 10:27:13 -04:00
Twice 3a97c57773 Add escaped_key method to ondemand::field (#2150) 2024-03-17 10:10:31 -04:00
Daniel Lemire e1c6a778f8 3.8.0 2024-03-10 13:25:53 -04:00
Daniel Lemire d413dd5b0e v3.7.1 2024-03-10 13:24:07 -04:00
Daniel Lemire a33fdc2770 Unconfusing clang (#2147)
* In some cases, clang might try to cast an ondemand::document to an
ondemand::document, instead of calling the move constructor. So we
can disable the template cast.

* In some cases, clang might get confused when constructing an ondemand
document in a constructor: instead of calling the move constructor, it
somehow ends up trying to cast a document to a document. We can easily
disallow this behavior with std::enable_if.

* making compatible with C++11

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2024-03-10 13:20:39 -04:00
Daniel Lemire 87f5f7a250 added some documentation in the code related to custom types (#2146) 2024-03-09 18:45:05 -05:00
Daniel Lemire 9923d27994 upper case 2024-03-06 16:01:21 -05:00
Piotr Rżysko 876be23545 Add link to simdjson-java (#2145) 2024-03-06 15:55:16 -05:00
Daniel Lemire 2028bdaeac This documents the big-int feature, and adds a few tests. (#2142)
* This documents the big-int feature, and adds a few tests.

* moving check_if_integer

* trimming the example.

* More trimming.

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2024-02-26 20:39:20 -05:00
Daniel Lemire 1ac3a00ca9 verify issue 2137 (#2138)
* fix issue 2137

* updating the single header

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2024-02-26 20:39:08 -05:00
Daniel Lemire 6d7cc2c7c1 fixing build 2024-02-26 20:00:17 -05:00
Daniel Lemire 61a52287fc Fixing apple builds 2024-02-26 19:50:22 -05:00
Daniel Lemire 37a9aaad48 [no ci] tweaking the wording 2024-02-26 14:05:12 -05:00
Serge Aleynikov 78325888ea Add functionality to detect big integers (#2139) 2024-02-26 14:00:33 -05:00
Shi Pujin a90cb5d900 add loongarch64 (#2141) 2024-02-26 13:54:49 -05:00
Daniel Lemire 34a01ec3af removing unnecessary spaces 2024-02-23 19:21:31 -05:00
Daniel Lemire e510ed2774 Release candidate (3.7.0) (#2136)
* Release candidate (3.7.0)

* some extra comments.
2024-02-23 16:50:55 -05:00
Dirk Stolle dabac13dde Add missing word in documentation (#2133) 2024-02-18 11:50:12 -05:00
Daniel Lemire aef7111c05 Documenting Unicode support in JSONPath and JSON Pointer (#2132)
* Documenting our support of JSON Path and JSON Pointer vs. Unicode characters

* using more standard terminology (nitpicking)

* specifying UTF-8 encoding

* allow char8_t when compiling as C++20

* casting

* minor doc corrections
2024-02-15 20:28:38 -05:00
Daniel Lemire 0c08894a54 adopting string literals (#2128)
* adopting string literals

* String suffixes are C++14
2024-02-15 10:40:11 -05:00
Daniel Lemire 2fe3a1467d This PR adds a convenience method (is_string) as well as a new method (#2131)
key_raw_json_token() to get the raw key token.
2024-02-15 10:39:30 -05:00
Daniel Lemire 167a46a0e7 adding remark 2024-02-15 09:56:05 -05:00
Francisco Geiman Thiesen 2b532489cc Adding on-demand support for a subset of json path. (#2127)
* First implementation of the conversion logic.

- Not yet tested
- Not yet compiled
- Not yet used in ondemand array/object/etc...

* Adding at_path to ondemand arrays

* Adding at_path to array, document and value.

* Adding at_path to ondemand object as well.

Pending:
- Building
- Adding tests for each of the on-demand classes usage of at_path
- Properly documenting the subset of json path that is currently supported.

* Adding a simple json_path test to on_demand

* Still trying to compile the ondemand_readme_examples test with the at_path() call

* Fixing linking issues with array::at_path

* Fixing issues with the path -> pointer conversion and removing comments

* Adding a clone of json_pointer to test json_path extensively

* Adding ondemand_json_path_tests (all tests passing)

* Adding documentation for at_path()

* Removing some newlines

* Removing simdjson_result from the return type of the string conversion function. Now json_path_to_pointer_conversion returns a std::string.

* Fixing typo

* Making string concatenation explicit to avoid ubuntu gcc12 issue with -O3 flag.

* Addressing latest reviews

* Addressing latest reviews
2024-02-15 09:47:25 -05:00
Daniel Lemire 85163df88c Update ondemand_design.md 2024-02-14 23:36:25 -05:00
Daniel Lemire 6a5be0f0db Create SECURITY.md 2024-02-09 17:16:54 -05:00
Daniel Lemire 710a3c7241 Update implementation-selection.md 2024-02-09 11:46:22 -05:00
Daniel Lemire c93e02e172 Update implementation-selection.md 2024-02-09 11:45:15 -05:00
Daniel Lemire 517143aec4 Update implementation-selection.md 2024-02-09 11:36:38 -05:00
Francisco Geiman Thiesen 9feda5e703 Update HACKING.md (#2122)
Updating windows command
2024-02-09 10:29:07 -05:00
Daniel Lemire cda198b7b9 Update ondemand_design.md 2024-02-09 10:22:44 -05:00
Daniel Lemire aad17a6a8d Update basics.md 2024-02-09 10:20:42 -05:00
Daniel Lemire 32d072a208 fixing issue2120 (#2121) 2024-02-08 11:34:36 -05:00
Daniel Lemire 5eab555053 documenting the support for custom types (#2098)
* documenting the support for custom types

* fixes

* getting around -Weffc++

* calling get_raw_json_string()

* documenting the support for custom types

* fixes

* getting around -Weffc++

* calling get_raw_json_string()
2024-02-06 13:20:35 -05:00
Daniel Lemire 6430fbd39a Better documentation and support for the case where you have a single implementation. (#2118) 2024-02-06 13:20:23 -05:00
Daniel Lemire 4d2fc31b27 removes simdjson.cpp include from the amalgated demo (#2119)
Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2024-02-05 19:34:50 -05:00
Bo Anderson 33610b59eb Allow some iterators to be used in standard C++ iterator functions (#2106) 2024-02-05 17:52:01 -05:00
Daniel Lemire 23290ae48d retiring deprecated DOM APIs (#2109) 2024-02-04 21:58:48 -05:00
Daniel Lemire eb3a085e0f Bash check is unneeded. (#2112) 2024-02-04 21:58:36 -05:00
Daniel Lemire 4c8ec35a48 fixing the order of the apple counters 2024-01-31 14:30:44 -05:00
Daniel Lemire 88bf5cb6d9 version bump 2024-01-29 10:43:12 -05:00
Niles Salter 9b0435d864 [utf8-validator] eliminate unnecessary comparison from must_be_2_3_continuation (#2113)
* [utf8-validator] eliminate unnecessary comparison from must_be_2_3_continuation

* Fix comment in
2024-01-28 12:42:59 -05:00
Daniel Lemire ddecadde86 Update README.md 2024-01-23 09:35:12 -05:00
Bo Anderson 02bee7d1b9 Add load_into_document to simdjson::dom::parser (#2103) 2024-01-08 20:30:26 -05:00
Daniel Lemire 6ef555e6fb This PR updates the backup string_view implementation. It also adds a few comments to our README.md file. (#2105)
* This PR updates the backup string_view implementation. It also adds a few comments to our README.md file.

* removing trailing space
2024-01-08 14:52:00 -05:00
Daniel Lemire 6952f8dfce adding gcc 13 (#2101) 2024-01-04 18:57:44 -05:00
Daniel Lemire 03d0a4731a Update README.md 2024-01-03 17:49:43 -05:00
Daniel Lemire 185ff4e8fe Update basics.md 2023-12-29 09:35:35 -05:00
Daniel Lemire 5bf6d9f3f0 Update README.md 2023-12-29 09:30:01 -05:00
Daniel Lemire 2fa729922e fix for issue 2099 (#2100)
* fix for issue 2099

* avoid exceptions.

* fix for exception-less code
2023-12-19 12:15:59 -05:00
Daniel Lemire ebd09cb2a3 Update basics.md 2023-12-15 08:29:16 -05:00
Dirk Stolle 3978b9f159 update actions/checkout in GitHub Actions to v4 (#2095) 2023-12-09 23:21:28 -05:00
Dirk Stolle edc36a87ab fix a small typo (#2096) 2023-12-09 23:20:51 -05:00
Daniel Lemire 6060be2fdf version bump 2023-12-08 12:09:45 -05:00
Daniel Lemire 3251f61c9d fix for issue 2093 (#2094) 2023-12-07 12:42:28 -05:00
Ikraduya Edian 561ffcd519 Add benchmarks for boost json (#2092) 2023-12-04 10:33:53 -05:00
Daniel Lemire 5541e78402 version bump 2023-12-01 13:56:12 -05:00
Daniel Lemire 9a9ed0e2da Fix Visual Studio 64-bit ARM regression (#2091)
* fix for ARM64 kernel under Visual Studio (bug introduced in release 3.6.1)

* revert comment

* updating amal.
2023-12-01 13:55:28 -05:00
Daniel Lemire bca8d3e277 version bump. 2023-11-30 18:13:02 -05:00
Daniel Lemire 97196fa969 Fixing warnings issued by clang (Node issue 50930) (#2090)
* Fixing warnings issued by clang (Node issue 50930)

* removing space

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2023-11-30 18:06:50 -05:00
John Keiser 23fab307ca Merge pull request #2088 from luhenry/upstream-riscv64
Make simdjson riscv64 aware
2023-11-30 11:06:43 -08:00
Daniel Lemire 0a14e05ec7 Update README.md 2023-11-30 13:42:00 -05:00
Daniel Lemire b88f831dbd Update basics.md 2023-11-30 11:52:24 -05:00
Daniel Lemire 71a73c2c47 Update basics.md 2023-11-30 11:51:35 -05:00
Daniel Lemire 07603aa7f7 fix space 2023-11-29 18:57:37 -05:00
Ludovic Henry b8a6b889c7 review: Revert changes in include/simdjson/implementation_detection.h 2023-11-29 22:09:25 +00:00
Ludovic Henry 7898234bb3 Make simdjson riscv64 aware 2023-11-29 19:28:14 +00:00
Daniel Lemire f52533538a Update basics.md 2023-11-28 23:49:40 -05:00
Daniel Lemire 3bb2f0a186 Return a better error when a user tries to cast a document that has already been accessed to a value (#2085)
* Fix.

* adding a test, cleaning some comments, adding documentation

* Correcting the JSON input.

* adding another test
2023-11-26 14:13:05 -05:00
Daniel Lemire d6551bf278 Update README.md 2023-11-22 11:36:33 -05:00
Daniel Lemire 0173eaaad7 Macro-based workaround for Visual Studio ARM (#2082) 2023-11-20 09:58:50 -05:00
Daniel Lemire 5459dbd6c6 typos. 2023-11-13 21:48:21 -05:00
Saint 9c45f1f292 Added link to Nim bindings (#2081) 2023-11-11 04:06:32 -05:00
int mian 665cbba159 fix typo (#2080) 2023-11-06 07:34:12 -05:00
Daniel Lemire a26f664907 removing space 2023-10-30 12:56:26 -04:00
Daniel Lemire 17cb457ffd version bump 2023-10-30 12:52:49 -04:00
Daniel Lemire 8b2af0473a skip UTF-8 BOM if present. (#2079) 2023-10-30 12:51:02 -04:00
Daniel Lemire f1bfa43385 Update ubuntu20-sani.yml 2023-10-27 20:58:54 -04:00
Daniel Lemire 29fbbee852 version bump 2023-10-27 20:57:30 -04:00
Daniel Lemire baa7d961ce Allowing users to write directly to std::optional<std::string> (#2078)
* Allowing users to write directly to std::optional<std::string>

* better fallback

* do not force the cast to std::string

* missing header

* removing abort
2023-10-27 20:56:12 -04:00
Yagiz Nizipli 6412b27c7e Merge pull request #2077 from simdjson/raw_json_on_values 2023-10-27 19:11:01 -04:00
Daniel Lemire 503da5ec83 raw_json() on values 2023-10-27 16:22:22 -04:00
Daniel Lemire 9dd6e9c959 version bump 2023-10-25 19:35:51 -04:00
Daniel Lemire 6db330c1a2 Automating string reallocation when it is possible (#2076)
* Automating string reallocation

* Typo
2023-10-25 19:34:52 -04:00
Daniel Lemire 993ac4b87c Adding the ability to write directly to an std::string. It is mostly syntaxic sugar (#2075) 2023-10-25 19:32:12 -04:00
Daniel Lemire 35e87896f0 minor tweak 2023-10-25 10:50:27 -04:00
Daniel Lemire f7e281cadc removing noisy warning 2023-10-20 15:45:40 -04:00
Daniel Lemire 13405afd4b This provides verbose static_assert messages (#2074)
* This provides verbose static_assert messages

* minor fix
2023-10-20 15:42:18 -04:00
Daniel Lemire 24b44309fb Removing commented code 2023-10-15 14:56:32 -04:00
Momtchil Momtchev e11ad58aad add a total order for elements (#2072) 2023-10-09 15:41:57 -04:00
Daniel Lemire 1875ed6550 Version bump. 2023-09-20 10:13:47 -04:00
Daniel Lemire 80d26298a0 Redesigning visit_primitive so that it is optimized for strings and (#2060)
numbers.

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2023-09-20 10:09:56 -04:00
Antoine Prouvost 25b5015c09 Add simdjson_static lib (#2068)
* Add simdjson_static lib

* Make simdjson_static a separate optional export file
2023-09-20 10:09:27 -04:00
Antoine Prouvost 26f8c566c7 Fix issue template bold text (#2067) 2023-09-19 11:15:54 -04:00
Daniel Lemire ff77ac801e Additional documentation based on issue 2063 (#2065)
* Additional documentation based on issue 2063

* Update basics.md
2023-09-11 10:01:35 -04:00
Daniel Lemire e0699994ef Testing macOS shared library support (#2062) 2023-09-09 12:48:42 -04:00
Daniel Lemire 68ba9a1b2a Adding a few spaces (minor formatting) (#2059) 2023-08-30 20:23:58 -04:00
Daniel Lemire 6fed6bd29b Update README.md 2023-08-30 19:26:34 -04:00
Piotr Rżysko cc24bb4114 Removed number parsing fallback (#2056) 2023-08-29 18:28:21 -04:00
John Keiser 14ec0ca8f4 Merge pull request #2052 from simdjson/jkeiser/structural_indexer_step
Write out 4 structural indexes at a time instead of 8
2023-08-29 15:24:54 -07:00
John Keiser 8dabd02c3a Default to step = 4 2023-08-29 15:48:52 -04:00
Daniel Lemire c9692005fd This allows us to pass SIMDJSON_STRUCTURAL_INDEXER_STEP as a command-line option. (#2055) 2023-08-25 13:47:20 -04:00
Daniel Lemire 2e0a9397c9 Merge branch 'master' into jkeiser/structural_indexer_step 2023-08-25 11:34:13 -04:00
Daniel Lemire f7922075c3 Adding support for Apple event counters (#2054) 2023-08-25 11:33:24 -04:00
Daniel Lemire b2e20e04c9 Standard compatibility fixes (#2053)
* Standard compatibility fixes

* missing commit

* Should work.

* Fix.

* Fix.

* Should work now.

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2023-08-25 10:18:02 -04:00
John Keiser adc9d18efd Change step for structural_indexes to 2 2023-08-23 19:02:05 -07:00
Daniel Lemire bf849e3619 Preparing release 2023-08-22 14:05:14 -04:00
Daniel Lemire c09e8dc225 Icelake processes 64 bytes at a time when processing strings. (#2050)
Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2023-08-18 14:37:10 -04:00
Daniel Lemire 5bf585ef02 Fixing failing thread sanitizer (#2051) 2023-08-18 14:35:12 -04:00
Daniel Lemire caff09cafc Update iterate_many.md 2023-08-14 12:18:09 -04:00
Daniel Lemire f043db4dca Preparing release. 2023-08-02 16:01:39 -04:00
Daniel Lemire e3bd0828d1 Fix for issue 2045. (#2046)
* Fix for issue 2045.

* Retauring single-header code.
2023-08-02 16:00:45 -04:00
John Keiser 9408298e0e Merge pull request #2042 from simdjson/jkeiser/escape3
New escape algorithm with significantly less data dependency
2023-08-02 10:38:09 -07:00
John Keiser 573bbac2d2 Reamalgamate again 2023-08-02 08:28:35 -07:00
John Keiser ac78c625df simdjson_inline -> simdjson_really_inline 2023-08-02 08:28:05 -07:00
John Keiser 390a66c6e8 Make next_structurals inline 2023-08-02 08:28:05 -07:00
John Keiser 210e0a56a1 Reamalgamate 2023-08-02 08:28:05 -07:00
John Keiser 084f662f32 Don't use C++20 designated initializers 2023-08-02 08:28:05 -07:00
John Keiser c878ff2500 Fix compiler warnings 2023-08-02 08:28:05 -07:00
John Keiser 06afe9ecb3 New escape algorithm with significantly less data dependency 2023-08-02 08:28:05 -07:00
Daniel Lemire e658502adb We are having issues with llvm.sh (#2044) 2023-08-02 11:24:28 -04:00
Daniel Lemire 6e7d415040 These lines should not be necessary. (#2043) 2023-08-01 19:33:27 -04:00
John Keiser 2ccd5b9c97 Merge pull request #2039 from simdjson/jkeiser/more-headers
Disallow including headers from implementation files
2023-07-21 12:09:57 -07:00
Daniel Lemire 1cdcbf79b3 We do actually support legacy 32-bit systems, although not with fast kernels. (#2041) 2023-07-21 13:53:10 -04:00
Calum McConnell 7013ed98cc Peek(0) is an error on 32-bit systems, where pointers are indistinguishable from int32_t (#2040) 2023-07-21 13:52:53 -04:00
John Keiser b383c717d7 Have amalgamate.py verify it's actually amalgamating all the files 2023-07-20 15:11:52 -07:00
John Keiser b8743bced0 Fix amalgamate.py under python 3.7 2023-07-20 14:14:44 -07:00
John Keiser ae9ba2bed2 Rename SIMDJSON_AMALGAMATED to SIMDJSON_CONDITIONAL_INCLUDE 2023-07-20 11:43:19 -07:00
Daniel Lemire 939844d79f Adding tests for issue 2037. (#2038)
* Adding tests for issue 2037.

* Fixed typo

* Simplifying.
2023-07-20 14:11:33 -04:00
John Keiser 4f4e81668e Change all include paths to <> instead of "" to avoid relative path includes 2023-07-20 10:48:25 -07:00
John Keiser bb54946b78 Disallow including headers from implementation files 2023-07-20 10:01:22 -07:00
John Keiser 4dec6ed5a7 Merge pull request #2036 from simdjson/jkeiser/more-header-fixes
A few more header fixes
2023-07-18 10:26:40 -07:00
John Keiser 1dfd48195d Merge pull request #2035 from simdjson/jkeiser/sanitizer-ci
Run sanitizers in parallel
2023-07-17 23:37:34 -07:00
John Keiser 23afc6b695 Merge branch 'master' into jkeiser/sanitizer-ci 2023-07-17 16:09:20 -07:00
John Keiser e3cac71afe Enforce that dependencies.h and generic/*.h include the same dependencies. 2023-07-17 16:07:42 -07:00
John Keiser cbe562c1d5 Fix header define names for inline files 2023-07-17 13:55:07 -07:00
John Keiser 30437d0ff6 A few header fixes for dom and ondemand 2023-07-17 13:53:50 -07:00
John Keiser 1e97af7dfb Merge pull request #2031 from simdjson/jkeiser/vscode-green
Get headers compiling in editor
2023-07-17 13:07:20 -07:00
John Keiser 80bc9e5051 Run sanitizers in parallel 2023-07-17 12:57:50 -07:00
John Keiser a74e87df17 Merge pull request #2026 from Cuda-Chen/pretty-print-for-dom
Add pretty print for DOM
2023-07-17 12:38:20 -07:00
John Keiser 3274299647 Merge pull request #2030 from yongxiangng/lazy-log-error
Add info and error logging, defer log message evaluation
2023-07-17 12:38:05 -07:00
John Keiser 22ca25faa6 Reamalgamate; make amalgamation normalize paths 2023-07-17 11:27:43 -07:00
John Keiser 37399fd41b Merge pull request #2034 from simdjson/add_add_emul_for_32bits
Adding missing umul
2023-07-17 10:02:55 -07:00
Daniel Lemire a170145d5b Missing header. 2023-07-17 11:48:07 -04:00
Daniel Lemire 7d1d635883 Adding missing umul 2023-07-17 10:04:57 -04:00
John Keiser ddfa8ae928 Reamalgamate 2023-07-16 22:53:42 -07:00
John Keiser 25dfb078c7 A few random header include ordering fixes 2023-07-16 22:50:54 -07:00
John Keiser 9f0710e3ad Move full_multiplication to <implementation>/numberparsing_defs.h 2023-07-16 22:34:21 -07:00
John Keiser 19b9cd9b05 Explain new directory structure 2023-07-16 21:42:50 -07:00
John Keiser 7540248396 Include simdprune_tables.h in generic implementations 2023-07-16 19:19:10 -07:00
John Keiser ab09e96de9 Only use amalgamated.h in generic/ 2023-07-16 19:18:58 -07:00
John Keiser 66252e4925 Honor --output-on-failure again 2023-07-15 17:44:44 -07:00
John Keiser 944778c186 Reamalgamate 2023-07-15 17:30:42 -07:00
John Keiser 62868a2f53 Catch more headers without which it won't compile 2023-07-15 17:16:18 -07:00
John Keiser 3c446c551b Move AMALGAMATED up to include implementation.cpp 2023-07-15 17:04:14 -07:00
John Keiser efe3d4b7d7 Few more missing headers, make it even stricter 2023-07-15 16:55:27 -07:00
John Keiser 28db262b47 Make SIMDJSON_AMALGAMATED checking much stricter 2023-07-15 16:46:15 -07:00
John Keiser bf076bfb5f Begin amalgamation early in simdjson.cpp 2023-07-15 15:55:38 -07:00
John Keiser 7bf1ecff9e Mark amalgamation early in 2023-07-15 15:53:09 -07:00
John Keiser e25a40facf Check if amalgamation rules are not being followed while compiling 2023-07-15 15:47:14 -07:00
John Keiser a442842f0d Use SIMDJSON_AMALGAMATED (which works in-editor and out) instead of editor-specific tweaks 2023-07-15 14:53:02 -07:00
John Keiser 421ed4f390 Try to make output on failure actually work 2023-07-14 20:57:24 -07:00
John Keiser 7dff5fa37c Include intrin.h from jsoncharutils.h since it's needed 2023-07-14 20:53:47 -07:00
John Keiser 9a0527c380 Only make editor stuff work in editor 2023-07-14 20:05:05 -07:00
John Keiser b160cf0610 Make simdjson.cpp compile on VSCode 2023-07-08 11:53:58 -07:00
John Keiser 553ee69b7c Make implementation .h/.cpp compile 2023-07-08 11:42:57 -07:00
John Keiser bf70cd0ebf Get headers compiling in editor 2023-07-08 10:47:15 -07:00
Cuda-Chen cf0551edf5 Add pretty print for DOM
Add pretty print for DOM with documentation.
Currently, the indentation is fixed to four spaces.

Close #1329.
2023-07-08 21:09:57 +08:00
yongxiangng ddb094d550 Move should log into separate if clause 2023-07-08 13:07:25 +08:00
Yong Xiang Ng 975fdda03c Add fine grain logging 2023-07-07 12:11:27 +08:00
Daniel Lemire c5c43e9c7f Release 2023-07-06 21:48:59 -04:00
Daniel Lemire 031ee32969 Undoing PR 2004 due to performance concerns. (#2029)
* Undoing PR 2004 due to performance concerns.

Fixes https://github.com/simdjson/simdjson/issues/2028
Fixes https://github.com/simdjson/simdjson/issues/2021

* Adding final lines.

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2023-07-06 21:34:14 -04:00
John Keiser b613326d69 Merge pull request #2027 from simdjson/jkeiser/number_type-namespace
Move numberparsing functions to architecture top level
2023-07-06 12:52:16 -07:00
John Keiser 53dad74d8e Don't put number_type into ondemand 2023-07-05 14:26:23 -07:00
John Keiser e4da19f5cf Make more prominent hacking quickstart 2023-07-05 13:36:47 -07:00
John Keiser 342b58986a Add clangd file for VS Code 2023-07-05 13:13:32 -07:00
Daniel Lemire 7edc475886 Update README.md 2023-07-03 18:46:21 -04:00
Dirk Stolle d2f8e84fc1 update various actions in GitHub Actions to v3 (#2024)
The updated actions are actions/cache, actions/checkout and
actions/upload-artifact.
2023-06-26 09:09:09 -04:00
Dirk Stolle 08000518cb Fix some typos (#2025) 2023-06-26 08:37:52 -04:00
Daniel Lemire 90040d96a5 Update basics.md 2023-06-22 21:38:45 -04:00
Daniel Lemire 2934892c0b Improving documentation. (#2020)
* Improving documentation.

* Fix typo

* Moving the definition.

* [skip ci] typo
2023-06-20 13:51:53 -04:00
Daniel Lemire 92174c927d New release 2023-06-15 09:21:07 -04:00
Ng Yong Xiang b399c0165e Add comma separated value parsing as an option in iterate_many (#2016)
* Add comma separated value parsing

* Fix failing tests

* Make tests work for exceptions

* Fix test

* Fix try catch making test fail

---------

Co-authored-by: Yong Xiang Ng <yxng@drwholdings.com>
2023-06-15 09:11:14 -04:00
Daniel Lemire 445ae024ba fix: bug identified by Yagiz Nizipli (#2019) 2023-06-15 09:10:13 -04:00
Daniel Lemire 65f9822542 Adding tests for issue 2017 (#2018) 2023-06-07 09:57:28 -04:00
Daniel Lemire 74bb7b2533 Adding popcnt 2023-06-05 09:04:36 -04:00
Daniel Lemire fd97288561 Adding popcnt. 2023-06-02 20:33:15 -04:00
Daniel Lemire 1da0d8ac65 Making sure popcnt is included 2023-06-02 20:32:54 -04:00
Daniel Lemire 85661a2901 Some systems release without NDEBUG set, let us not assert in these cases. (#2014) 2023-06-01 16:18:36 -04:00
Daniel Lemire 3fcdb65d88 Update README.md 2023-05-26 09:26:17 -04:00
Daniel Lemire 8b00d91a2c More documentation. (#2011) 2023-05-26 09:23:21 -04:00
Daniel Lemire 7db3f97ba8 Adding CXX 20 to CI (#2005)
* Adding CXX 20 to CI

* side-stepping new CXX 20 guard.

* Going another way

* Saving.

* Explicit.

* Saving...

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2023-05-26 09:23:02 -04:00
Daniel Lemire 412a8f7c4d Removing iso header. (#2009)
* Removing iso header.

* Fix.

* Another one.

* Caught another one.

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2023-05-23 16:19:21 -04:00
Daniel Lemire d4b31c9feb Update README.md 2023-05-23 16:16:26 -04:00
Daniel Lemire ecdcf276b2 Update README.md 2023-05-23 16:15:13 -04:00
Daniel Lemire 6d92563280 Fixing issue 1943 (#2006)
* Fixing issue 1943

* Tweaking.
2023-05-23 11:15:22 -04:00
Daniel Lemire 0998233d58 More documentation. (#2007) 2023-05-23 11:15:07 -04:00
Daniel Lemire cb735818ec Document key queries. (#2008) 2023-05-23 11:14:55 -04:00
Daniel Lemire 41ee548e3f Adding godbolt examples. (#2010)
* Adding godbolt examples.

* Updating the examples.
2023-05-23 11:14:41 -04:00
Ng Yong Xiang e9d5d5088a Introduce logging levels for conditional logging and log key for NO_SUCH_FIELD error (#2004)
* Add info and error logging

* Add tests for error logging

* Add logging for missing field

* Update docs for logging usage

* Fix style and pass by ref for string format args

* Make log_level explicit and simplify get log level from env

* Make log level int32_t

* Format enum class

* Fix ci

* Move enum to header

* Fix compilation for noexception build in test case

* Disable warnings and putenv

---------

Co-authored-by: Yong Xiang Ng <yxng@drwholdings.com>
2023-05-22 12:58:55 -04:00
Daniel Lemire 78151e80e5 Update README.md 2023-05-16 12:49:49 -04:00
Daniel Lemire 083a5d7351 Update README.md 2023-05-15 15:03:11 -04:00
Daniel Lemire 25fcc8b432 Version bump. 2023-05-14 17:18:01 -04:00
Daniel Lemire a9c120f779 Directly check altivec macro. (#2002) 2023-05-14 17:17:10 -04:00
Daniel Lemire b1487e34c3 Adding more PPC64 testing. (#2001) 2023-05-14 16:11:49 -04:00
Daniel Lemire 2a68a32012 Checking OS support for AVX-512 (#1998)
* Adding support for AVX-512 on macOS.

* Fix.

* Fix.

* Minor fix

* Setting the variable to zero.

* Fixing include

* Checking if the OS supports AVX-512

* Tweaking.

---------

Co-authored-by: Daniel Lemire <dlemire@lemire.me>
2023-05-12 19:25:02 -04:00
Daniel Lemire 21a2ccaeb8 Another fun one 2023-05-05 22:12:59 -04:00
Daniel Lemire 6cbb6ac89b On ARM processors, rapidjson may struggle. Let us help it out. (#1996) 2023-05-05 21:42:37 -04:00
Dirk Stolle e9f0f1bb9b Mention IceLake in documentation about implementation selection. (#1995)
Since the creation of the IceLake implementation there are now
four available implementations on Intel/AMD x64 processors.
2023-04-28 21:29:01 -04:00
Serge Aleynikov 7ef1ebd4fb Add Erlang bindings (#1994) 2023-04-28 09:54:49 -04:00
Daniel Lemire 8de540e267 Update README.md 2023-04-27 14:27:05 -04:00
Daniel Lemire ebc51fbd1a Update bug_report.md 2023-04-24 08:05:26 -04:00
Daniel Lemire 84a31c1615 Update bug_report.md 2023-04-23 12:06:47 -04:00
Daniel Lemire f77ac963f4 Trying the undefined behaviour sanitizer (#1988)
* Trying the undefined behaviour sanitizer

* Trimming quiet
2023-04-23 11:50:05 -04:00
pjuhasz ae3d912f8d Add link to Perl binding (JSON::SIMD) (#1991)
Co-authored-by: Peter Juhasz <juhasz.peter@uhusystems.com>
2023-04-23 10:37:56 -04:00
Daniel Lemire abf3d7d440 Slightly safer linux counters 2023-04-21 13:26:49 -04:00
Daniel Lemire 3c313b8530 Removing old file. 2023-04-13 11:52:02 -04:00
Dirk Stolle f435fddda1 update string_view lite to version 1.7.0 (#1984)
This is the header as seen for the tag v1.7.0,
commit 7b1867b7c08f9ac010fb8b9dae79ff8c8359bafb, on
<https://github.com/martinmoene/string-view-lite>.
2023-04-11 13:58:38 -04:00
Daniel Lemire 14d927128b New release 2023-04-08 16:13:48 -04:00
Daniel Lemire 48152a0326 Fix for issue 1981 (#1982)
* Fix

* Rewording.

* Doing it differently.
2023-04-08 16:11:40 -04:00
Daniel Lemire b5cf53232c More documentation. 2023-04-06 10:52:18 -04:00
Daniel Lemire b3d0b732e6 Update README.md 2023-04-04 14:22:12 -04:00
Daniel Lemire 3c29f40f2f Trying to verify issue 1979 (#1980) 2023-04-04 11:03:37 -04:00
Daniel Lemire 4a89d79984 Guarding the benchmarks. 2023-04-01 18:39:11 -04:00
Daniel Lemire 9cb639dcd7 Update basics.md 2023-03-30 12:35:38 -04:00
Daniel Lemire 4699dd0ce2 Update basics.md 2023-03-30 10:49:19 -04:00
Daniel Lemire cfd337bb50 This PR adds an 'at_end()' method. (#1978)
* This PR adds an 'at_end()' method.

* Adding 1111 }

* Tweaking test.
2023-03-30 10:47:13 -04:00
Daniel Lemire 9bef628982 It is no longer necessary to check separately for trailing content. 2023-03-24 20:04:04 -04:00
Daniel Lemire 59025bc8b1 Avoiding the issue where doc.get_value().get_object() differs from doc.get_object() in how errors are reported. (#1975)
* Avoiding the issue where doc.get_value().get_object() differs from doc.get_object() in how errors are reported.

* Minor tweaks
2023-03-24 12:44:18 -04:00
Daniel Lemire ffe96dde19 Update HACKING.md 2023-03-22 09:02:24 -04:00
Daniel Lemire f9664cf184 Update HACKING.md 2023-03-22 08:56:12 -04:00
Daniel Lemire 830323f383 Update HACKING.md 2023-03-22 08:55:26 -04:00
Daniel Lemire eb8e15d9c0 Update HACKING.md 2023-03-22 08:53:44 -04:00
Daniel Lemire 3571f3a8d0 Additional documentation. 2023-03-18 16:38:16 -04:00
Daniel Lemire 4b13020d65 Patch release 2023-03-13 21:44:34 -04:00
Daniel Lemire 75c6c950e5 Ok. 2023-03-13 21:26:32 -04:00
Daniel Lemire 5ad7698af1 ... 2023-03-13 17:07:25 -04:00
Daniel Lemire f6f359d3eb Very silly. 2023-03-13 17:05:53 -04:00
Daniel Lemire 185e157e0c Silencing the stupid compiler warnings. 2023-03-13 16:06:10 -04:00
Daniel Lemire 3f3f6cde7c Bad compiler 2023-03-13 13:33:20 -04:00
Daniel Lemire 005054916d Bad compiler 2023-03-13 13:02:01 -04:00
Daniel Lemire 338224849d Disabling memory sanitizer with one function. Might help with issue 1965 (#1966)
* Might help with issue 1965

* Fix macro

* Adding test.

* Stupid compiler

* Silly compiler

* Stupid compilers

* Unnecessary fixes

* Update developer-options.cmake
2023-03-13 12:34:47 -04:00
Ashot Vardanian ada52641b4 Fix: Defining maximal DOM capacity (#1970) 2023-03-13 12:33:37 -04:00
Daniel Lemire ede9d57f22 Adding tests with _GLIBCXX_ASSERTIONS (#1969)
* Allowing users to specify -D_GLIBCXX_ASSERTIONS

* Adding CI tests for GLIBCXX_ASSERTIONS

* Adding curly brackets.
2023-03-13 12:32:43 -04:00
Daniel Lemire 7bf33f6c8a Bad compiler 2023-03-13 12:29:01 -04:00
Daniel Lemire d62b789cb3 More curly. 2023-03-13 10:49:15 -04:00
Daniel Lemire 33dbd44098 Better way to put -Og in debug mode (#1967)
* Better way to put -Og in debug mode

* Silly compiler

* Stupid compiler.

* Stupid compiler.

* Silly compiler
2023-03-13 10:11:33 -04:00
Daniel Lemire e0dcf8adc9 Silly compiler 2023-03-12 20:57:45 -04:00
Daniel Lemire 34dcd33a88 Patch. 2023-03-09 11:24:06 -05:00
Daniel Lemire 17a1a8e187 Adding -Og to debug builds. (#1964)
* Adding -Og to debug builds.

* Stupid compiler.

* bad, bad, bad compiler
2023-03-09 11:14:42 -05:00
Daniel Lemire 77ad00b63c Producing a minor patch release 2023-03-09 11:13:01 -05:00
Daniel Lemire 25e7f1765a Various tweaks. 2023-03-09 10:42:23 -05:00
Daniel Lemire 8d01255ceb Update basics.md 2023-03-09 08:27:32 -05:00
Daniel Lemire 8ec016b093 Update README.md 2023-03-08 15:30:52 -05:00
Daniel Lemire 9e726b9856 Removing obsolete badge 2023-03-08 14:50:37 -05:00
Daniel Lemire 312f1f26d3 Now obsolete. 2023-03-08 14:50:06 -05:00
Daniel Lemire ba50571d66 Merge branch 'master' of github.com:simdjson/simdjson 2023-03-06 16:05:10 -05:00
Daniel Lemire 4a587cb898 Adding guards. 2023-03-06 16:04:42 -05:00
Daniel Lemire bdf9123c92 Update README.md 2023-03-06 11:08:56 -05:00
Dominik Lohmann 6e4c29b7c5 Fix export component for simdjsonTargets.cmake (#1962)
This fixes a bug that caused simdjsonTargets.cmake not to be included in
CPack-generated packages, which—unlike `cmake --install`—does not
pick up this mislabeled install component.

I git-grepped through the code base, after this change all components
are either `simdjson_Development` or `simdjson_Runtime`.
2023-03-04 17:34:40 -05:00
Daniel Lemire ec0b48b772 Version bump. 2023-03-02 14:27:49 -05:00
Daniel Lemire d65acbd47b Tuning documentation. 2023-03-02 09:39:35 -05:00
Daniel Lemire 75240ad8e1 Update basics.md 2023-03-01 19:05:45 -05:00
Daniel Lemire 3efb44dc0f Adding link 2023-03-01 19:03:03 -05:00
Daniel Lemire 5c5cd5b42d [skip ci] rephrasing. 2023-03-01 10:08:00 -05:00
Daniel Lemire 12a1b6db53 This adds some examples to the documentation. (#1959)
* This adds some examples to the documentation.

* Fix
2023-02-28 22:59:42 -05:00
Daniel Lemire 37e87f6f99 Adding support for wobbly strings to On Demand and lossy trancoding from escaped (with replacement) (#1947)
* Adding support for wobbly strings.

* Minor tweak.

* Adopt substitution code point.

* Lossy

* Patching
2023-02-28 22:56:44 -05:00
wanweiqiangintel cd13b02bf0 add UTF-8 validation flag (#1957) 2023-02-28 09:47:08 -05:00
Daniel Lemire 624174d916 Minor doc. tweaks. 2023-02-22 21:14:15 -05:00
Daniel Lemire 8a247f1321 Patch release. 2023-02-22 20:39:31 -05:00
Daniel Lemire 7aa35bffd3 Fixing support for windows ARM64 (#1955)
* Fixing Windows ARM64 support

* Updating the singleheader files.
2023-02-22 20:34:51 -05:00
Rawley d6c910686b Update README.md (#1953) 2023-02-13 21:11:05 -05:00
Daniel Lemire 22cc523fd4 Patching release script. 2023-02-06 19:11:52 -05:00
Daniel Lemire 3177cd1b5d Fix. 2023-02-06 19:08:11 -05:00
Daniel Lemire 63d9ce3c58 Patch release. 2023-02-06 18:31:02 -05:00
Daniel Lemire 8e8180560d Fix for issue 1950 (#1951) 2023-02-06 18:28:54 -05:00
Daniel Lemire 1594af5d5d Tweak. 2023-02-06 08:51:11 -05:00
Dirk Stolle b207338365 Remove trailing whitespace in README.md (#1949)
This should make the whitespace test pass again.
2023-01-31 17:53:35 -05:00
Daniel Lemire 68be9c2170 Update README.md 2023-01-30 16:17:42 -05:00
Daniel Lemire 5347771091 Update documentation.yml 2023-01-30 15:14:27 -05:00
Daniel Lemire 8bf8315651 Update README.md 2023-01-30 14:50:06 -05:00
Daniel Lemire 269ae7ad6b [skip ci] stupid mistake 2023-01-30 14:37:40 -05:00
Daniel Lemire 2aa91f741f [skip ci] new doc runner 2023-01-30 14:29:48 -05:00
Daniel Lemire 5430544bbb [skip ci] improving documentation. (#1948) 2023-01-30 14:27:35 -05:00
Daniel Lemire 7b55f16682 Trying s390x in CI (#1945)
* Trying s390x

* Providing big endian testing.

* adding token.

* Fixing branch
2023-01-27 11:34:00 -05:00
Daniel Lemire d8fec59e85 Safer add-dependency script. (#1946) 2023-01-26 22:40:26 -05:00
Daniel Lemire 57b6034a0c Let verbose 2023-01-26 21:12:23 -05:00
Daniel Lemire 7500d7bb4f Patching. 2023-01-25 22:19:40 -05:00
Daniel Lemire 82c9876aec Preparing release. 2023-01-21 18:07:06 -05:00
Daniel Lemire b9234be80f complete number_in_string for documents, stringify the version macro and improving trailing content errors (#1934)
* This completes the '*_in_string' support within documents.

* Minor updates.

* Additional fixes.
2023-01-21 15:29:08 -05:00
Daniel Lemire 0a2563e13e Testing and fixing log issues. (#1939)
* Testing and fixing log issues.

* Cleaning it a bit.
2023-01-21 15:28:56 -05:00
Daniel Lemire 727ec73c33 The fuzzer needs to be updated. (#1941) 2023-01-21 15:28:25 -05:00
Daniel Lemire 9a55a830ec Adding clang14 ci (#1940) 2023-01-21 15:27:21 -05:00
Daniel Lemire 8c89ae0cd9 Removing extra ';'. 2023-01-13 09:17:05 -05:00
Daniel Lemire 8c71ca1b1d Update implementation-selection.md 2023-01-07 13:43:26 -05:00
Daniel Lemire c72b4bc0f4 Update basics.md 2023-01-07 13:42:01 -05:00
Daniel Lemire 8a26cb3c8e Adding a couple of string tests. (#1935)
* Adding a couple of string tests.
2023-01-05 21:18:54 -05:00
Daniel Lemire bf2ba5c675 Update basics.md 2023-01-05 08:38:16 -05:00
Daniel Lemire f151300a76 Disabling fallback kernel on systems where it is not needed (#1930)
* Fixing issue 1772

* More cleaning.

* Allow disabling other kernels.

* Minor tweaks.
2023-01-03 09:52:41 -05:00
Daniel Lemire 9aa3563d99 Updating Google Benchmark. (#1929) 2023-01-03 09:52:26 -05:00
Daniel Lemire 9444c1a04f Merge branch 'master' of github.com:simdjson/simdjson 2022-12-19 21:07:17 -05:00
Daniel Lemire 883901a81b Updating users. 2022-12-19 21:07:07 -05:00
Daniel Lemire a2c96c9cca Update basics.md 2022-12-16 22:12:25 -05:00
Daniel Lemire c63da0ef2c Added a remark regarding assemblers. 2022-12-16 21:22:48 -05:00
Daniel Lemire c61cefda85 Merge branch 'master' of github.com:simdjson/simdjson 2022-12-05 12:25:52 -05:00
Daniel Lemire bffcbb444d Clarifying comment. 2022-12-05 12:25:29 -05:00
Daniel Lemire 0b2891aaab Trying fuzzer with llvm 14. (#1922) 2022-12-01 14:26:17 -05:00
wanweiqiangintel 983fb41802 Update README.md (#1925)
Add a real-word usage
2022-11-30 17:26:58 -05:00
Daniel Lemire 0d2455029d Preparing patch release. 2022-11-23 10:59:48 -05:00
Daniel Lemire e94f027310 Serialize 0 as 0.0 (#1921)
* Serialize 0 as 0.0

* Extending the '.0'.
2022-11-23 10:31:42 -05:00
sean 2b766e4c09 Fix: Add padded_string_view overload for parser::parse (#1916) 2022-11-23 09:22:39 -05:00
Daniel Lemire a5db75e1a6 Adding more development checks to the DOM front-end (#1915)
* This adds SIMDJSON_DEVELOPMENT_CHECKS to the DOM API to help users
in the scenario of issue 1914.

* More documentation and warnings.

* Updating following comments by Tyson
2022-10-17 09:11:09 -04:00
Daniel Lemire 4bd66cb891 We branch on GCC for performance. (#1913) 2022-10-16 13:15:35 -04:00
Daniel Lemire 960a7ebba1 Switching to iterate_many 2022-10-06 11:59:12 -04:00
Daniel Lemire d520062f8e Bumping API 2022-10-06 11:50:33 -04:00
Daniel Lemire db08d78ed0 Documents better the type method and makes is_null return an error condition in some instances (#1909) 2022-10-06 11:47:40 -04:00
Daniel Lemire 46292b8d4b Update basics.md 2022-10-05 08:50:52 -04:00
Tyson Andre 46241287c6 Check for trailing tokens in json2msgpack ondemand benchmark (#1908)
Related to #1904

Users of the simdjson library will see json2msgpack as an available
example of how to recursively process json with the ondemand parser,
and checking for trailing tokens in a document is one part of json validation.

These checks shouldn't affect benchmark results performance.
The benchmark is run on the 631KB twitter.json file.
2022-10-05 08:49:09 -04:00
Daniel Lemire 137cb14bcc Documenting how one can check for the end of the document. (#1907) 2022-10-04 20:23:52 -04:00
Daniel Lemire 5e6be3ed7a Minor fix (documentation and safety) regarding max. depth in ondemand. (#1906) 2022-10-04 12:24:20 -04:00
Tyson Andre c6ab52eebb [skip ci] Add an .editorconfig for .cpp/.h/.md for whitespace settings (#1901)
Make it less likely to accidentally introduce tabs, trailing whitespace,
carriage returns, non-utf8 in files, or files without trailing newlines.

https://editorconfig.org/ has plugins for various editors/IDEs and is
enabled by default in some IDEs.
2022-10-03 11:14:00 -04:00
Daniel Lemire a1aab1fafe Preparing release 2.2.3 2022-10-02 16:32:37 -04:00
Daniel Lemire b92cbbe280 Fixing issue 1898 (#1899)
* Fixing issue 1898 Preserve sign for number with underflowing exponent (#1900)


Before this commit, simdjson parsed "-1e-999" and "-0e-999" and "-1e-342"
as 0.0.
After this commit, those JSON strings get parsed as -0.0.
(https://en.wikipedia.org/wiki/Signed_zero)

The old behavior was inconsistent with the way simdjson parsed "-0.0" as -0.0.

Co-authored-by: Daniel Lemire <daniel@lemire.me>

Co-authored-by: Tyson Andre <tysonandre775@hotmail.com>
2022-10-02 16:25:39 -04:00
Tyson Andre 6d2a09f8e5 [skip ci] Fix typo, formatting nit in HACKING.md (#1902)
vim syntax highlighting doesn't work properly without space between a
bulleted list and the start of a code snippet.
2022-10-02 16:19:34 -04:00
Tyson Andre 00564bf7d5 Micro-optimization for parsing surrogate pairs (#1897)
Load 2 bytes and compare the 2 bytes against `"\u"`
Compilers with optimizations turned on will turn this into a 16-bit load
then 16-bit compare on supported platforms
(with smaller compiled code size).

Make it obvious to the compiler that it's reading two
consecutive bytes of the same pointer

Add parse_surrogate_pairs to show the difference exists.
See discussion in #1896
2022-10-02 12:10:53 -04:00
Tyson Andre 5809e51ae4 fix: Reject surrogate pairs with invalid low surrogate (#1896)
Closes #1894

Reject low surrogates outside of the range U+DC00—U+DFFF

Related to https://unicodebook.readthedocs.io/unicode_encodings.html#utf-16-surrogate-pairs

A surrogate pair should consist of a high surrogate and low surrogate.
They're used to represent 0x010000-0x10FFFF in the JSON spec because
the JavaScript specification originally only supported `\uXXXX`.

Previously, simdjson would accept some combinations of valid high
surrogates and invalid low surrogates due to a bug in the check.
(e.g. `\uD888\u1234` was accepted)

U+D800—U+DBFF (1,024 code points): high surrogates
U+DC00—U+DFFF (1,024 code points): low surrogates
2022-09-30 12:13:16 -04:00
Tyson Andre d27e7cce71 Fix typos in doc/basics.md (#1893) 2022-09-30 08:34:37 -04:00
David Korenchuk f7dc03f93d Fix documentation of description() method in implementation (#1895) 2022-09-30 08:34:06 -04:00
sean d4ac1b51d0 Fix various warnings & if constexpr (#1888) 2022-09-27 23:22:47 -04:00
Daniel Lemire 6a4222da71 Adding a remark to the documentation. 2022-09-19 15:28:33 -04:00
Daniel Lemire bba88eb5e5 Fixing typo in the documentation. 2022-09-19 10:21:11 -04:00
Daniel Lemire 16b3816455 This fixes an error caused by overeager gcc static analyser (#1891) 2022-09-19 10:20:48 -04:00
Andrea Pappacoda e5a408386b build: add pkg-config support (#1767)
* build: add pkg-config support

The CMake build script now generates a simple pkg-config files that can
be easily used by non-CMake users.

The file is generated from a template file that gets filled in at
configure time.

As CMake doesn't have anything similar to Meson's pkg-config generator
the file is quite static, i.e. new simdjson public defines/dependencies
won't be picked up automatically.

This approach also suffers from one minor issue, mentioned in
[jtojnar/cmake-snips][]; in short, it doesn't work well when users
specify CMAKE_INSTALL_INCLUDEDIR and similar as absolute paths. It's not
a big deal, and it will easily fixable once you'll require CMake >=3.20.

Fixes #1763

[jtojnar/cmake-snips]: https://github.com/jtojnar/cmake-snips#concatenating-paths-when-building-pkg-config-files

* build: handle absolute paths in .pc generation

As mentioned in the previous commit message, correct concatenation of
paths is only available in CMake >=3.20, so handling absolute paths in
pkg-config file generation requires using jtojnar's JoinPaths module.

* ci: add debian job

This new jobs compiles simdjson on Debian Testing, a semi-rolling
release, so that new compilers are always tested.

This job also tests the pkg-config file introduced in commit
1096c3b299
2022-08-26 16:20:51 -04:00
Herman Semenov e65f28e61a Fixed if condition, Win64 _fseeki64, trivial constructors C++11 (#1883) 2022-08-18 14:16:46 -04:00
Daniel Lemire fbb46b99e2 Fixes and verifies issue 1878. (#1880)
* Fixes and verifies issue 1878.

* Changing how NULL is handled.

* Different design.
2022-08-08 22:00:38 -04:00
Daniel Lemire 933c2ebeac Preparing release 2022-07-28 21:46:45 -04:00
Daniel Lemire db3e813aa6 Verifying and fixing issue 1876 (#1877)
* Verifying and fixing issue 1876

* Typo
2022-07-28 21:45:54 -04:00
Daniel Lemire 9c95a48fe6 cleaning on-demand benchmarks (#1875)
* Setting RapidJSON and yyjson to their latest version.

* Let us stop dumping all of the benchmarks (it is confusing) on screen.
2022-07-28 20:28:29 -04:00
Daniel Lemire cb20f7e7df Update CONTRIBUTING.md 2022-07-20 09:46:52 -04:00
Dirk Stolle 18b9168eec remove empty if block (#1873)
I guess it will be thrown away by the compiler's optimizer
anyway, but there is no need to keep this in the code.
2022-07-19 20:36:51 -04:00
Daniel Lemire 241ce7304c New version. 2022-07-19 16:47:06 -04:00
Daniel Lemire 5dbe96da96 We change slightly on development checks are enabled. (#1869)
* We change slightly on development checks are enabled.

* Removing garbagy code.
2022-07-19 16:40:02 -04:00
Daniel Lemire 40b397a3d4 Fixing issue 1870 (#1871) 2022-07-19 15:17:12 -04:00
strager 5510089d45 Improve build times for debug builds (#1859)
* Rename simdjson_really_inline -> simdjson_inline

I want to change the simdjson_really_inline macro to sometimes not force
inlining. After that upcoming change, the name simdjson_really_inline
will no longer makes sense.

Rename simdjson_really_inline to simdjson_inline. This patch should not
change semantics; simdjson_inline still forces inlining as before.

Some functions still need to be really inlined for ABI reasons.
(GCC's -Wpsabi complains otherwise.) Leave those functions marked as
simdjson_really_inline.

* Improve build times for debug builds

simdjson_inline is used for most simdjson functions. It forces inlining.
In unoptimized/debug builds, this can lead to a lot of machine code
being generated (especially with Address Sanitizer), causing slow
compilation.

Change simdjson_inline to force inlining only for optimized builds.

Sometimes, the programmer might want a slightly-optimized build and want
fast compilation (e.g. GCC's -Og mode). Allow simdjson users to define
the simdjson_inline macro themselves (e.g. on the command line:
-Dsimdjson_inline=inline) in cases where the default behavior is
undesired.

This patch reduced build times by over 75% for ondemand_object_tests.cpp
with GCC 9.4.0 and CMAKE_BUILD_TYPE=Debug on my AMD 5950X:

Before: 6.885 6.683 6.971 6.957 6.949 seconds (5 samples)
After:  1.492 1.551 1.494 1.490 1.531 seconds (5 samples)
2022-07-19 15:14:33 -04:00
Daniel Lemire 62a57907a7 Adding DOM benchmark to msgpack (#1866)
* We need simdjson dom for the json2msgpack benchmark

* Minor tweaking.
2022-07-14 13:47:10 -04:00
Daniel Lemire becbe99e81 Update README.md 2022-07-14 09:45:04 -04:00
Daniel Lemire bd3d67e889 Documenting a specific use case where you need a value if and only if it is another key is not present (#1865)
* Documenting a specific use case.

* Adding more comments and documentation.
2022-07-09 10:54:28 -04:00
Dirk Stolle a89d57d0d4 update dependency nlohmann/json for benchmarks to current version 3.10.5 (#1862) 2022-07-08 12:32:05 -04:00
Dirk Stolle e04d400c64 update jsoncpp to version 1.9.5 (#1863) 2022-07-08 12:31:43 -04:00
Daniel Lemire 1075e8609c Updating single header, prior to release. 2022-07-04 21:55:09 -04:00
Daniel Lemire 3fde8a4eac New release candidate (#1856)
* Patch for possible AVX-512 overflow.

* Updating the test for new padding.

* Preparing new version.

* replace binary integer literals with hex literals for C++11 compatibility (#1855)

Binary integer literals are a C++14 feature, so those are not supported
in C++11 and should be replaced by hexadecimal literals instead.

Fixes #1854.

Co-authored-by: Dirk Stolle <striezel-dev@web.de>
2022-07-04 21:54:19 -04:00
Daniel Lemire 62bdb9a2f7 [noci] update users 2022-07-02 21:44:48 -04:00
Daniel Lemire 2fbacb0058 New version 2022-06-30 11:49:26 -04:00
Daniel Lemire 509066f06a adding msgpack benchmarks (#1853) 2022-06-30 10:29:50 -04:00
Daniel Lemire 070f0b26a3 Removing dead code. (#1852) 2022-06-25 00:11:31 -04:00
Daniel Lemire 1a195623a5 Improve string performance in ondemand by making the string processing runtime dispatched. (#1849)
* This should improve string performance in ondemand by making the string processing runtime dispatched.
2022-06-24 09:57:16 -04:00
Dirk Stolle 5f7a56e7f1 add SIMDJSON_IMPLEMENTATION_ICELAKE to implementation-selection.md (#1848) 2022-06-22 13:09:19 -04:00
Daniel Lemire 5e60f0482d Update implementation-selection.md 2022-06-21 21:29:12 -04:00
Daniel Lemire e6c90b8efb Preparing release. 2022-06-15 15:23:03 -04:00
Daniel Lemire 7c450fbb70 Adding clang 13 tests. (#1844) 2022-06-15 15:21:33 -04:00
Daniel Lemire 4e1e002cb0 This verifies and fixes issue 1834. (#1843) 2022-06-15 13:42:04 -04:00
Benson Muite de196dd7a3 [skip ci] Grammar and typo fixes (#1842) 2022-06-14 16:16:39 -04:00
Daniel Lemire 08cb8dd81c Simpler counters. (#1841) 2022-06-07 15:19:51 -04:00
Dirk Stolle 57d54792ba update actions/cache + actions/checkout in GitHub Actions to v3 (#1839) 2022-06-06 08:51:50 -04:00
Dirk Stolle 79879802f9 run tests with GCC 8, too (#1837) 2022-06-03 20:52:07 -04:00
Daniel Lemire fbe955e9a4 gcc12 without warnings. (#1836) 2022-06-02 22:53:39 -04:00
Daniel Lemire a49ac04046 Version 2.0.3 2022-06-02 13:57:46 -04:00
Daniel Lemire 49c7654a70 We will be enabling AVX-512 under Visual Studio 2019 by default. (#1833) 2022-06-02 13:56:20 -04:00
541 changed files with 266769 additions and 34368 deletions
+47
View File
@@ -0,0 +1,47 @@
CompileFlags:
CompilationDatabase: build
Add:
- -Werror -Wall -Wextra -Weffc++ -Wsign-compare -Wshadow -Wwrite-strings -Wpointer-arith -Winit-self -Wconversion -Wno-sign-conversion
- -Wundefined-inline
Diagnostics:
Suppress:
- misc-unused-alias-decls
- misc-unused-using-decls
- misc-definitions-in-headers # TODO fix and remove these violations
---
If:
PathMatch:
- include/.*
- src/.*
PathExclude:
- include/simdjson.h
- src/simdjson.cpp
CompileFlags:
Add:
- -Wno-unneeded-internal-declaration
- -Wno-undefined-internal # TODO fix and remove these violations
- -Wno-unused-function
- -Wno-unused-const-variable
Diagnostics:
Suppress:
- pp_including_mainfile_in_preamble
- unused-includes
---
# Amalgamated files that require or partly define an implementation
If:
PathMatch:
- .*/(arm64|fallback|haswell|icelake|ppc64|westmere)/begin.h
- .*/generic/.*
Diagnostics:
Suppress:
- pragma_attribute_no_pop_eof
---
# clang has a bad time detecting the push/pop together in src/ for some reason
If:
PathMatch:
- include/simdjson/.*/end.h
- src/(arm64|fallback|haswell|icelake|ppc64|westmere).cpp
Diagnostics:
Suppress:
- pragma_attribute_no_pop_eof
- pragma_attribute_stack_mismatch
+12
View File
@@ -0,0 +1,12 @@
# https://editorconfig.org/
root = true
# Conservatively avoid changing defaults for other file types, e.g. raw json files for test cases,
# Makefiles, etc.
[*.{cpp,h,md}]
charset = utf-8
end_of_line = lf
indent_size = 2
indent_style = space
insert_final_newline = true
tab_width = 2
trim_trailing_whitespace = true
+1
View File
@@ -78,6 +78,7 @@
.gitattributes export-ignore
.gitignore export-ignore
.editorconfig export-ignore
# Sources
*.c text eol=lf diff=c
+18 -9
View File
@@ -9,27 +9,30 @@ assignees: ''
Before submitting an issue, please ensure that you have read the documentation:
* Basics is an overview of how to use simdjson and its APIs: https://github.com/simdjson/simdjson/blob/master/doc/basics.md
* Basics is an overview of how to use simdjson and its APIs to parse JSON: https://github.com/simdjson/simdjson/blob/master/doc/basics.md
* Builder is an overview of how to use simdjson to generate JSON: https://github.com/simdjson/simdjson/blob/master/doc/builder.md
* Performance shows some more advanced scenarios and how to tune for them: https://github.com/simdjson/simdjson/blob/master/doc/performance.md
* Contributing: https://github.com/simdjson/simdjson/blob/master/CONTRIBUTING.md
* We follow the [JSON specification as described by RFC 8259](https://www.rfc-editor.org/rfc/rfc8259.txt) (T. Bray, 2017).
* We follow the [JSON specification as described by RFC 8259](https://www.rfc-editor.org/rfc/rfc8259.txt) (T. Bray, 2017). If you wish to support features that are not part of RFC 8259, then you should not refer to your issue as a bug.
**Describe the bug**
A clear and concise description of what the bug is.
A clear and concise description of what the bug is. A bug is a failure to build with normal compiler settings or a misbehaviour: when running the code, you get a result that differs from the expected result from our documentation.
A compiler or static-analyzer warning is not a bug.
A compiler or static-analyzer warning is not a bug. It is possible with tools such as Visual Studio to require that rarely enabled warnings are considered errors. Do not report such cases as bugs. We do accept pull requests if you want to silence warnings issued by code analyzers, however.
We are committed to providing good documentation. We accept the lack of documentation or a misleading documentation as a bug (a 'documentation bug').
An unexpected poor software performance can be accepted as a bug (a 'performance bug').
We accept the identification of an issue by a sanitizer or some checker tool (e.g., valgrind) as a bug, but you must first ensure that it is not a false positive.
We recommend that you run your tests using different optimization levels.
We recommend that you run your tests using different optimization levels. In particular, we recommend your run tests with the simdjson library and you code compiled in debug mode. The simdjson then sets the SIMDJSON_DEVELOPMENT_CHECKS macro to 1, and this triggers additional checks on your code and on the internals of the library. If possible, we recommend that you run tests with sanitizers (e.g., see [No more leaks with sanitize flags in gcc and clang](https://lemire.me/blog/2016/04/20/no-more-leaks-with-sanitize-flags-in-gcc-and-clang/)). You can compile the library with sanitizers for debugging purposes (e.g., set SIMDJSON_SANITIZE to ON using CMake), but you should also turn on sanitizers on your own code. You may also use tools like valgrind or the commercial equivalent.
Before reporting a bug, please ensure that you have read our documentation.
**To Reproduce**
Steps to reproduce the behaviour: provide a code sample if possible.
Steps to reproduce the behaviour: provide a code sample if possible. Please provide a complete test with data. Remember that a bug is either a failure to build or an unexpected result when running the code.
If we cannot reproduce the issue, then we cannot address it. Note that a stack trace from your own program is not enough. A sample of your source code is insufficient: please provide a complete test for us to reproduce the issue. Please reduce the issue: use as small and as simple an example of the bug as possible.
@@ -43,14 +46,20 @@ It is fine to report bugs against our main branch, but if that is what you are d
**Configuration (please complete the following information if relevant)**
- OS: [e.g. Ubuntu 16.04.6 LTS]
- Compiler [e.g. Apple clang version 11.0.3 (clang-1103.0.32.59) x86_64-apple-darwin19.4.0]
- Compiler* [e.g. Apple clang version 11.0.3 (clang-1103.0.32.59) x86_64-apple-darwin19.4.0]
- Version [e.g. 22]
- Optimization setting (e.g., -O3)
We support up-to-date 64-bit ARM and x64 FreeBSD, macOS, Windows and Linux systems. Please ensure that your configuration is supported before labelling the issue as a bug. In particular, we do not support legacy 32-bit systems.
We support up-to-date 64-bit ARM and x64 FreeBSD, macOS, Windows and Linux systems. Please ensure that your configuration is supported before labelling the issue as a bug.
* We do not support unreleased or experimental compilers. If you encounter an issue with a
pre-release version of a compiler, do not report it as a bug to simdjson. However, we always
invite contributions either in the form an analysis or of a code contribution.
Under Windows, we support Visual Studio (both with LLVM and without). We do not support MinGW and other alternate compiler systems. Windows users should be aware that there [is a long-running bug with GCC under Windows](https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54412).
**Indicate whether you are willing or able to provide a bug fix as a pull request**
If you plan to contribute to simdjson, please read our
If you plan to contribute to simdjson, please read our guide:
* CONTRIBUTING guide: https://github.com/simdjson/simdjson/blob/master/CONTRIBUTING.md and our
* HACKING guide: https://github.com/simdjson/simdjson/blob/master/HACKING.md
+1
View File
@@ -0,0 +1 @@
blank_issues_enabled: false
+3 -2
View File
@@ -9,7 +9,8 @@ assignees: ''
Before submitting an issue, please ensure that you have read the documentation:
* Basics is an overview of how to use simdjson and its APIs: https://github.com/simdjson/simdjson/blob/master/doc/basics.md
* Basics is an overview of how to use simdjson and its APIs to parse JSON: https://github.com/simdjson/simdjson/blob/master/doc/basics.md
* Builder is an overview of how to use simdjson to generate JSON: https://github.com/simdjson/simdjson/blob/master/doc/builder.md
* Performance shows some more advanced scenarios and how to tune for them: https://github.com/simdjson/simdjson/blob/master/doc/performance.md
* Contributing: https://github.com/simdjson/simdjson/blob/master/CONTRIBUTING.md
* We follow the [JSON specification as described by RFC 8259](https://www.rfc-editor.org/rfc/rfc8259.txt) (T. Bray, 2017).
@@ -31,7 +32,7 @@ A clear and concise description of any alternative solutions or features you've
**Additional context**
Add any other context or screenshots about the feature request here.
** Are you willing to contribute code or documentation toward this new feature? **
**Are you willing to contribute code or documentation toward this new feature?**
If you plan to contribute to simdjson, please read our
* CONTRIBUTING guide: https://github.com/simdjson/simdjson/blob/master/CONTRIBUTING.md and our
* HACKING guide: https://github.com/simdjson/simdjson/blob/master/HACKING.md
@@ -9,7 +9,8 @@ assignees: ''
Before submitting an issue, please ensure that you have read the documentation:
* Basics is an overview of how to use simdjson and its APIs: https://github.com/simdjson/simdjson/blob/master/doc/basics.md
* Basics is an overview of how to use simdjson and its APIs to parse JSON: https://github.com/simdjson/simdjson/blob/master/doc/basics.md
* Builder is an overview of how to use simdjson to generate JSON: https://github.com/simdjson/simdjson/blob/master/doc/builder.md
* Performance shows some more advanced scenarios and how to tune for them: https://github.com/simdjson/simdjson/blob/master/doc/performance.md
* Contributing: https://github.com/simdjson/simdjson/blob/master/CONTRIBUTING.md
* We follow the [JSON specification as described by RFC 8259](https://www.rfc-editor.org/rfc/rfc8259.txt) (T. Bray, 2017).
@@ -18,7 +19,7 @@ We do not make changes to simdjson without clearly identifiable benefits, which
Is your issue:
1. A bug report? If so, please point at a reproducible test. Indicate whether you are willing or able to provide a bug fix as a pull request.
1. A bug report? If so, please point at a reproducible test. Indicate whether you are willing or able to provide a bug fix as a pull request. As a matter of policy, we do not consider a compiler warning to be a bug.
2. A build issue? If so, provide all possible details regarding your system configuration. If we cannot reproduce your issue, we cannot fix it.
+49 -4
View File
@@ -1,8 +1,53 @@
Short title (summary):
Description
- What did you change and why? (1-3 sentences)
- Issue reproduced / related issue: link the issue if relevant (e.g. #123)
Type of change
- [ ] Bug fix
- [ ] New feature
- [ ] Refactor / cleanup
- [ ] Documentation / tests
- [ ] Other (please describe):
How to verify / test
- Add additional tests to verify bugs or new features.
- If you claim performance gains, you should provide benchmark numbers using high quality benchmarking code.
Our tests check whether you have introduced trailing white space. If such a test fails, please check the "artifacts button" above, which if you click it gives a link to a downloadable file to help you identify the issue. You can also run scripts/remove_trailing_whitespace.sh locally if you have a bash shell and the sed command available on your system.
Please read before contributing:
- CONTRIBUTING: https://github.com/simdjson/simdjson/blob/master/CONTRIBUTING.md
- HACKING: https://github.com/simdjson/simdjson/blob/master/HACKING.md
If you plan to contribute to simdjson, please read our
CONTRIBUTING guide: https://github.com/simdjson/simdjson/blob/master/CONTRIBUTING.md and our
HACKING guide: https://github.com/simdjson/simdjson/blob/master/HACKING.md
If you can, we recommend running our tests with the sanitizers turned on.
For non-Visual Studio users, it is as easy as doing:
```bash
cmake -B build -D SIMDJSON_SANITIZE=ON -D SIMDJSON_DEVELOPER_MODE=ON
cmake --build build
ctest --test-dir build
```
Our CI checks, among other things, for trailing whitespace. If a test fails for that reason,
use the "artifacts" button to download the artifact and inspect the problematic lines,
or run `scripts/remove_trailing_whitespace.sh` locally if you have a bash shell and `sed`.
Checklist before submitting
- [ ] I added/updated tests covering my change (if applicable)
- [ ] Code builds locally and passes my check
- [ ] Documentation / README updated if needed
- [ ] Commits are atomic and messages are clear
- [ ] I linked the related issue (if applicable)
Final notes
- For large PRs, prefer smaller incremental PRs or request staged review.
Thanks for the contribution!
+29
View File
@@ -0,0 +1,29 @@
name: Ubuntu aarch64 (GCC 13)
on:
push:
branches:
- master
pull_request:
branches:
- master
jobs:
build:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: uraimo/run-on-arch-action@v3
name: Test
id: runcmd
with:
arch: aarch64
distro: ubuntu_latest
githubToken: ${{ github.token }}
install: |
apt-get update -q -y
apt-get install -y cmake make g++
run: |
cmake -DSIMDJSON_SANITIZE_UNDEFINED=ON -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_COMPETITION=OFF -B build
cmake --build build -j=2
ctest --output-on-failure --test-dir build
+3 -3
View File
@@ -9,8 +9,8 @@ jobs:
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
@@ -31,4 +31,4 @@ jobs:
./alpine.sh cmake --build build_for_alpine
- name: test
run: |
./alpine.sh bash -c "cd build_for_alpine && ctest -LE explicitonly"
./alpine.sh bash -c "cd build_for_alpine && ctest -LE explicitonly --output-on-failure"
+1 -1
View File
@@ -17,7 +17,7 @@ jobs:
fuzz-seconds: 600
dry-run: false
- name: Upload Crash
uses: actions/upload-artifact@v1
uses: actions/upload-artifact@v4
if: failure() && steps.build.outcome == 'success'
with:
name: artifacts
+33
View File
@@ -0,0 +1,33 @@
name: Debian
on: [push, pull_request]
defaults:
run:
shell: sh
permissions:
contents: read
jobs:
pkg-config:
runs-on: ubuntu-latest
container:
image: debian:testing
steps:
- uses: actions/checkout@v4
- name: Install dependencies
run: |
apt -y update
apt -y --no-install-recommends install g++ cmake make pkg-config
- name: Build and install
run: |
cmake -B build
cmake --build build
cmake --install build
- name: Test pkg-config
run: g++ examples/quickstart/quickstart.cpp $(pkg-config --cflags --libs simdjson)
+33
View File
@@ -0,0 +1,33 @@
name: Doxygen GitHub Pages
on:
release:
types: [created]
# Allows you to run this workflow manually from the Actions tab
workflow_dispatch:
permissions:
contents: write
pages: write
id-token: write
jobs:
deploy:
environment:
name: github-pages
url: ${{ steps.deployment.outputs.page_url }}
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Install Doxygen
run: sudo apt-get install doxygen graphviz -y
- run: mkdir docs
- name: Install theme
run: ./tools/prepare_doxygen.sh
- name: Generate Doxygen Documentation
run: doxygen
- name: Deploy to GitHub Pages
uses: peaceiris/actions-gh-pages@v3
with:
github_token: ${{ secrets.GITHUB_TOKEN }}
publish_dir: doc/api/html
+17
View File
@@ -0,0 +1,17 @@
on: [push, pull_request]
jobs:
build:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
- uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
- uses: mymindstorm/setup-emsdk@6ab9eb1bda2574c4ddb79809fc9247783eaf9021 # v14
- name: Verify
run: emcc -v
- name: Checkout
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v3.6.0
- name: Configure
run: emcmake cmake -B build
- name: Build # We build but do not test
run: cmake --build build
@@ -4,9 +4,9 @@ on: [push, pull_request]
jobs:
whitespace:
runs-on: ubuntu-20.04
runs-on: ubuntu-24.04
steps:
- uses: actions/checkout@v2
- uses: actions/checkout@v4
- name: Remove whitespace and check the diff
run: |
set -eu
@@ -24,7 +24,7 @@ jobs:
echo "no trailing whitespace found, good!"
fi
- name: Archive whitespace patch
uses: actions/upload-artifact@v2
uses: actions/upload-artifact@v4
if: always()
with:
name: whitespace-patch
+14 -13
View File
@@ -24,7 +24,7 @@ jobs:
implementations: haswell westmere fallback
UBSAN_OPTIONS: halt_on_error=1
MAXLEN: -max_len=4000
CLANGVERSION: 11
CLANGVERSION: 19
# which optimization level to use for the sanitizer build (see build_fuzzer.variants.sh)
OPTLEVEL: -O3
@@ -32,30 +32,31 @@ jobs:
- name: Install packages necessary for building
run: |
sudo apt update
sudo apt-get install --quiet ninja-build valgrind zip unzip
sudo apt-get install --quiet ninja-build valgrind zip unzip lsb-release wget software-properties-common gnupg
wget https://apt.llvm.org/llvm.sh
sudo apt-get purge --auto-remove llvm python3-lldb-15 llvm-15
chmod +x llvm.sh
sudo ./llvm.sh $CLANGVERSION
- uses: actions/checkout@v1
- uses: actions/checkout@v4
- uses: actions/cache@v2
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- uses: actions/cache@v2
- uses: actions/cache@v4
id: cache-corpus
with:
path: out/
key: corpus-${{ github.run_id }}
restore-keys: corpus-
- name: show statistics for the cached corpus
run: |
echo number of files in github action corpus cache:
find out -type f |wc -l
- name: Create and prepare the initial seed corpus
run: |
fuzz/build_corpus.sh
@@ -92,7 +93,7 @@ jobs:
export SIMDJSON_FORCE_IMPLEMENTATION=$implementation
build-sanitizers$OPTLEVEL/fuzz/fuzz_$fuzzer out/$fuzzer $others seedcorpus -max_total_time=20 $MAXLEN
done
echo now have $(ls out/$fuzzer |wc -l) files in corpus
echo now have $(ls out/$fuzzer |wc -l) files in corpus
done
- name: Fuzz differential impl. fuzzers with sanitizer+asserts (good at detecting errors)
@@ -119,12 +120,12 @@ jobs:
- name: Package the corpus into an artifact
run: |
for fuzzer in $defaultimplfuzzers $implfuzzers; do
for fuzzer in $defaultimplfuzzers $implfuzzers; do
tar rf corpus.tar out/$fuzzer
done
- name: Save the corpus as a github artifact
uses: actions/upload-artifact@v2
uses: actions/upload-artifact@v4
with:
name: corpus
path: corpus.tar
@@ -141,13 +142,13 @@ jobs:
run: |
for fuzzer in $defaultimplfuzzers $implfuzzers; do
find out/$fuzzer -type f |sort|head -n200|xargs -n40 valgrind build-replay/fuzz/fuzz_$fuzzer 2>&1|tee valgrind-$fuzzer.txt
done
done
- name: Compress the valgrind output
run: tar cf valgrind.tar valgrind-*.txt
- name: Save valgrind output as a github artifact
uses: actions/upload-artifact@v2
uses: actions/upload-artifact@v4
if: always()
with:
name: valgrindresults
@@ -155,7 +156,7 @@ jobs:
if-no-files-found: ignore
- name: Archive any crashes as an artifact
uses: actions/upload-artifact@v2
uses: actions/upload-artifact@v4
if: always()
with:
name: crashes
+50
View File
@@ -0,0 +1,50 @@
name: LoongArch64-CI
on: [push, pull_request]
jobs:
loongarch64:
runs-on: ubuntu-latest
strategy:
fail-fast: false
matrix:
platform:
- { toolchain-version: 2023.08.08 }
steps:
- uses: actions/checkout@v4
- name: Install build requirements
run: |
sudo apt-get update -y
sudo apt-get install -y --no-install-recommends cmake
- uses: actions/cache/restore@v4
id: restore-cache
with:
path: /opt/cross-tools
key: loongarch64-${{ matrix.platform.toolchain-version }}
- name: Download LoongArch64 gcc+glibc toolchain
if: ${{ !steps.restore-cache.outputs.cache-hit }}
run: |
url="https://github.com/loongson/build-tools/releases/download/${{ matrix.platform.toolchain-version }}/x86_64-cross-tools-loongarch64-gcc-libc.tar.xz"
wget "$url" -O /tmp/toolchain.tar.xz
mkdir -p /opt
tar -C /opt -x -f /tmp/toolchain.tar.xz
- uses: actions/cache/save@v3
if: ${{ !steps.restore-cache.outputs.cache-hit }}
with:
path: /opt/cross-tools
key: loongarch64-${{ matrix.platform.toolchain-version }}
- name: setup Loongarch64 build environment
run: |
echo "/opt/cross-tools/bin" >> $GITHUB_PATH
echo "CC=loongarch64-unknown-linux-gnu-gcc" >> $GITHUB_ENV
echo "CXX=loongarch64-unknown-linux-gnu-g++" >> $GITHUB_ENV
- name: configure
run: cmake -B build -DCMAKE_SYSTEM_PROCESSOR=loongarch64 -DARCH=lonngarch64 -DCMAKE_SYSTEM_NAME=Linux -DCMAKE_C_COMPILER=loongarch64-unknown-linux-gnu-gcc -DCMAKE_CXX_COMPILER=loongarch64-unknown-linux-gnu-g++
- name: build
run: cmake --build build
-34
View File
@@ -1,34 +0,0 @@
name: Macos (Xcode 11)
on: [push, pull_request]
jobs:
macos-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: macos-latest
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake
run: |
xcversion select 11.7
mkdir builddebug &&
cd builddebug &&
cmake -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
cd .. &&
mkdir build &&
cd build &&
cmake -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
cmake --install . &&
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp && c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson && ./linkandrun jsonexamples/twitter.json &&
cd ../tests/installation_tests/find &&
mkdir build && cd build && cmake -DCMAKE_INSTALL_PREFIX:PATH=../../../build/destination .. && cmake --build .
+15 -4
View File
@@ -9,8 +9,8 @@ jobs:
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: macos-latest
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
@@ -20,14 +20,25 @@ jobs:
cd builddebug &&
cmake -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
ctest --output-on-failure -LE explicitonly -j &&
cd .. &&
mkdir build &&
cd build &&
cmake -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
ctest --output-on-failure -LE explicitonly -j &&
cmake --install . &&
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp && c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson && ./linkandrun jsonexamples/twitter.json &&
cd ../tests/installation_tests/find &&
mkdir build && cd build && cmake -DCMAKE_INSTALL_PREFIX:PATH=../../../build/destination .. && cmake --build .
- name: Use cmake (shared)
run: |
mkdir buildshared &&
cd buildshared &&
cmake -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=ON -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
ctest --output-on-failure -LE explicitonly -j &&
cmake --install . &&
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp && c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson && ./linkandrun jsonexamples/twitter.json &&
cd ../tests/installation_tests/find &&
mkdir buildshared && cd buildshared && cmake -DCMAKE_INSTALL_PREFIX:PATH=../../../buildshared/destination .. && cmake --build .
+5 -2
View File
@@ -20,12 +20,15 @@ jobs:
- msystem: "MINGW64"
install: mingw-w64-x86_64-libxml2 mingw-w64-x86_64-cmake mingw-w64-x86_64-ninja mingw-w64-x86_64-clang
type: Debug
- msystem: "MINGW64"
install: mingw-w64-x86_64-libxml2 mingw-w64-x86_64-cmake mingw-w64-x86_64-ninja mingw-w64-x86_64-clang
type: RelWithDebInfo
env:
CMAKE_GENERATOR: Ninja
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
+5 -2
View File
@@ -22,12 +22,15 @@ jobs:
- msystem: "MINGW64"
install: mingw-w64-x86_64-cmake mingw-w64-x86_64-ninja mingw-w64-x86_64-gcc
type: Debug
- msystem: "MINGW64"
install: mingw-w64-x86_64-cmake mingw-w64-x86_64-ninja mingw-w64-x86_64-gcc
type: RelWithDebInfo
env:
CMAKE_GENERATOR: Ninja
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
-71
View File
@@ -1,71 +0,0 @@
name: short fuzz on the power arch
on:
push:
branches: [ master ]
pull_request:
branches: [ master ]
jobs:
armv7_job:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
# The host should always be Linux
runs-on: ubuntu-20.04
name: Build on ubuntu-20.04 ppc64le
steps:
- uses: actions/checkout@v2.1.0
- uses: uraimo/run-on-arch-action@v2.0.5
name: Run commands
id: runcmd
env:
DEBIAN_FRONTEND: noninteractive
with:
arch: ppc64le
distro: buster
# Not required, but speeds up builds by storing container images in
# a GitHub package registry.
githubToken: ${{ github.token }}
run: |
export CLANGSUFFIX="-7"
apt-get -qq update
apt-get install -q -y clang-7 libfuzzer-7-dev git wget zip ninja-build gnupg software-properties-common
wget -q -O - "https://raw.githubusercontent.com/simdjson/debian-ppa/master/key.gpg" | apt-key add -
apt-add-repository "deb https://raw.githubusercontent.com/simdjson/debian-ppa/master simdjson main"
apt-get -qq update
apt-get purge cmake cmake-data
apt-get -t simdjson -y install cmake
mkdir -p build ; cd build
cmake .. -GNinja \
-DCMAKE_CXX_COMPILER=clang++$CLANGSUFFIX \
-DCMAKE_C_COMPILER=clang$CLANGSUFFIX \
-DBUILD_SHARED_LIBS=OFF \
-DSIMDJSON_DEVELOPER_MODE=ON \
-DSIMDJSON_ENABLE_FUZZING=On \
-DSIMDJSON_COMPETITION=OFF \
-DSIMDJSON_GOOGLE_BENCHMARKS=OFF \
-DSIMDJSON_DISABLE_DEPRECATED_API=On \
-DSIMDJSON_FUZZ_LDFLAGS=-lFuzzer \
-DCMAKE_CXX_FLAGS="-fsanitize=fuzzer-no-link -DFUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION=" \
-DCMAKE_C_FLAGS="-fsanitize=fuzzer-no-link" \
-DCMAKE_BUILD_TYPE=Release \
-DSIMDJSON_FUZZ_LINKMAIN=Off
cd ..
builddir=build
cmake --build $builddir
wget -O corpus.tar.gz https://readonly:readonly@www.pauldreik.se/fuzzdata/index.php?project=simdjson
tar xf corpus.tar.gz
fuzzernames=$(cmake --build $builddir --target print_all_fuzzernames |tail -n1)
for fuzzer in $fuzzernames ; do
exe=$builddir/fuzz/$fuzzer
shortname=$(echo $fuzzer |cut -f2- -d_)
echo found fuzzer $shortname with executable $exe
mkdir -p out/$shortname
others=$(find out -type d -not -name $shortname -not -name out -not -name cmin)
$exe -max_total_time=20 -max_len=4000 out/$shortname $others
echo "*************************************************************************"
done
echo "all is good, no errors found in any of these fuzzers: $fuzzernames"
+29
View File
@@ -0,0 +1,29 @@
name: Ubuntu ppc64le (GCC 11)
on:
push:
branches:
- master
pull_request:
branches:
- master
jobs:
build:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: uraimo/run-on-arch-action@v3
name: Test
id: runcmd
with:
arch: ppc64le
distro: ubuntu_latest
githubToken: ${{ github.token }}
install: |
apt-get update -q -y
apt-get install -y cmake make g++
run: |
cmake -DCMAKE_BUILD_TYPE=Release -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_COMPETITION=OFF -B build
cmake --build build -j=2
ctest --output-on-failure --test-dir build
+29
View File
@@ -0,0 +1,29 @@
name: Ubuntu riscv64 (GCC 11)
on:
push:
branches:
- master
pull_request:
branches:
- master
jobs:
build:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: uraimo/run-on-arch-action@v3
name: Test
id: runcmd
with:
arch: riscv64
distro: ubuntu_latest
githubToken: ${{ github.token }}
install: |
apt-get update -q -y
apt-get install -y cmake make g++
run: |
cmake -DCMAKE_BUILD_TYPE=Release -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_COMPETITION=OFF -B build
cmake --build build -j=2
ctest --output-on-failure --test-dir build
+29
View File
@@ -0,0 +1,29 @@
name: Ubuntu rvv VLEN=1024 (clang 18)
on:
push:
branches:
- master
pull_request:
branches:
- master
jobs:
build:
runs-on: ubuntu-24.04
steps:
- uses: actions/checkout@v4
- name: Install packages
run: |
sudo apt-get update -q -y
sudo apt-get install -y cmake make g++-riscv64-linux-gnu qemu-user-static clang-18
- name: Build
run: |
CXX=clang++-18 CXXFLAGS="--target=riscv64-linux-gnu -march=rv64gcv_zvbb" \
cmake --toolchain=cmake/toolchains-ci/riscv64-linux-gnu.cmake -DCMAKE_BUILD_TYPE=Release -B build
cmake --build build/ -j$(nproc)
- name: Test VLEN=1024
run: |
export QEMU_LD_PREFIX="/usr/riscv64-linux-gnu"
export QEMU_CPU="rv64,v=on,zvbb=on,vlen=1024,rvv_ta_all_1s=on,rvv_ma_all_1s=on"
ctest --timeout 1800 --output-on-failure --test-dir build -j $(nproc)
+29
View File
@@ -0,0 +1,29 @@
name: Ubuntu rvv VLEN=128 (clang 17)
on:
push:
branches:
- master
pull_request:
branches:
- master
jobs:
build:
runs-on: ubuntu-24.04
steps:
- uses: actions/checkout@v4
- name: Install packages
run: |
sudo apt-get update -q -y
sudo apt-get install -y cmake make g++-riscv64-linux-gnu qemu-user-static clang-17
- name: Build
run: |
CXX=clang++-17 CXXFLAGS="--target=riscv64-linux-gnu -march=rv64gcv" \
cmake --toolchain=cmake/toolchains-ci/riscv64-linux-gnu.cmake -DCMAKE_BUILD_TYPE=Release -B build
cmake --build build/ -j$(nproc)
- name: Test VLEN=128
run: |
export QEMU_LD_PREFIX="/usr/riscv64-linux-gnu"
export QEMU_CPU="rv64,v=on,vlen=128,rvv_ta_all_1s=on,rvv_ma_all_1s=on"
ctest --timeout 1800 --output-on-failure --test-dir build -j $(nproc)
+29
View File
@@ -0,0 +1,29 @@
name: Ubuntu rvv VLEN=256 (gcc 14)
on:
push:
branches:
- master
pull_request:
branches:
- master
jobs:
build:
runs-on: ubuntu-24.04
steps:
- uses: actions/checkout@v4
- name: Install packages
run: |
sudo apt-get update -q -y
sudo apt-get install -y cmake make g++-14-riscv64-linux-gnu qemu-user-static
- name: Build
run: |
CXX=riscv64-linux-gnu-g++-14 CXXFLAGS=-march=rv64gcv \
cmake --toolchain=cmake/toolchains-ci/riscv64-linux-gnu.cmake -DCMAKE_BUILD_TYPE=Release -B build
cmake --build build/ -j$(nproc)
- name: Test VLEN=256
run: |
export QEMU_LD_PREFIX="/usr/riscv64-linux-gnu"
export QEMU_CPU="rv64,v=on,zvbb=on,vlen=256,rvv_ta_all_1s=on,rvv_ma_all_1s=on"
ctest --timeout 1800 --output-on-failure --test-dir build -j $(nproc)
+29
View File
@@ -0,0 +1,29 @@
name: Ubuntu s390x (GCC 11)
on:
push:
branches:
- master
pull_request:
branches:
- master
jobs:
build:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: uraimo/run-on-arch-action@v3
name: Test
id: runcmd
with:
arch: s390x
distro: ubuntu_latest
githubToken: ${{ github.token }}
install: |
apt-get update -q -y
apt-get install -y cmake make g++
run: |
cmake -DCMAKE_BUILD_TYPE=Release -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_COMPETITION=OFF -B build
cmake --build build -j=2
ctest --output-on-failure --test-dir build
-29
View File
@@ -1,29 +0,0 @@
name: Performance check on Ubuntu 18.04 CI (GCC 7)
on:
push:
branches:
- master
pull_request:
branches:
- master
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-18.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake
run: |
mkdir build &&
cd build &&
cmake -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_ENABLE_DOM_CHECKPERF=ON -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . --target checkperf &&
ctest --output-on-failure -R checkperf ubuntu18-checkperf.yml
-31
View File
@@ -1,31 +0,0 @@
name: Ubuntu 18.04 CI (LLVM 7)
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-18.04
env:
CC: clang-7
CXX: clang++-7
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: install clang 7
run: |
sudo apt update
sudo apt install clang-7
- name: Use cmake
run: |
mkdir build &&
cd build &&
cmake -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly
-25
View File
@@ -1,25 +0,0 @@
name: Ubuntu 18.04 CI (GCC 7) with Thread Sanitizer
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-18.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake
run: |
mkdir build &&
cd build &&
cmake -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_SANITIZE_THREADS=ON .. &&
cmake --build . --target document_stream_tests --target ondemand_document_stream_tests --target parse_many_test &&
ctest --output-on-failure -R parse_many_test &&
ctest --output-on-failure -R document_stream_tests
-25
View File
@@ -1,25 +0,0 @@
name: Ubuntu 18.04 CI (GCC 7)
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-18.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake
run: |
mkdir build &&
cd build &&
cmake -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
make install &&
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp && c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson && ./linkandrun jsonexamples/twitter.json
-23
View File
@@ -1,23 +0,0 @@
name: Ubuntu 20.04 CI (GCC 9) With Memory Sanitizer
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-20.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake
run: |
mkdir builddebug &&
cd builddebug &&
cmake -DSIMDJSON_SANITIZE=ON -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly
-33
View File
@@ -1,33 +0,0 @@
name: Ubuntu 20.04 CI (GCC 9)
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-20.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake
run: |
mkdir builddebug &&
cd builddebug &&
cmake -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
cd .. &&
mkdir build &&
cd build &&
cmake -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
cmake --install . &&
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp && c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson && ./linkandrun jsonexamples/twitter.json &&
cd ../tests/installation_tests/find &&
mkdir build && cd build && cmake -DCMAKE_INSTALL_PREFIX:PATH=../../../build/destination .. && cmake --build .
+23
View File
@@ -0,0 +1,23 @@
name: Ubuntu 22.04 CI (CLANG 13)
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-22.04
steps:
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake
run: |
mkdir build &&
cd build &&
CXX=clang++-13 cmake -DSIMDJSON_DEVELOPER_MODE=ON .. &&
cmake --build . &&
ctest --output-on-failure -LE explicitonly -j
+23
View File
@@ -0,0 +1,23 @@
name: Ubuntu 22.04 CI (CLANG 14)
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-22.04
steps:
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake
run: |
mkdir build &&
cd build &&
CXX=clang++-14 cmake -DSIMDJSON_DEVELOPER_MODE=ON .. &&
cmake --build . &&
ctest --output-on-failure -LE explicitonly -j
+65
View File
@@ -0,0 +1,65 @@
name: Ubuntu 22.04 CI (GCC 12, CXX 20)
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-22.04
steps:
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Configure Debug Build
run: |
CXX=g++-12 cmake -DSIMDJSON_CXX_STANDARD=20 -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF -B builddebug
- name: Compile Debug Build
run: |
cmake --build builddebug
- name: Test Debug Build
run: |
ctest --output-on-failure -LE explicitonly -j --test-dir builddebug
- name: Configure Release Build
run: |
CXX=g++-12 cmake -DSIMDJSON_CXX_STANDARD=20 -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination -B build
- name: Compile Release Build
run: |
cmake --build build
- name: Test Release Build
run: |
ctest --output-on-failure -LE explicitonly -j --test-dir build
- name: Install Release Build
run: |
cmake --install build
- name: Generate Example Code
run: |
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp
- name: Compile Example Code
run: |
c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson
- name: Run Example Code
run: |
./linkandrun jsonexamples/twitter.json
- name: Configure Find Tests
run: |
cd tests/installation_tests/find && \
cmake -DCMAKE_INSTALL_PREFIX:PATH=../../../destination -B build
- name: Compile Find Tests
run: |
cd tests/installation_tests/find && cmake --build build
+23
View File
@@ -0,0 +1,23 @@
name: Ubuntu 22.04 CI (GCC 12)
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-22.04
steps:
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake
run: |
mkdir build &&
cd build &&
CXX=g++-12 cmake -DSIMDJSON_DEVELOPER_MODE=ON .. &&
cmake --build . &&
ctest --output-on-failure -LE explicitonly -j
@@ -0,0 +1,24 @@
name: Ubuntu 22.04 CI GCC 12 with GLIBCXX_ASSERTIONS
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-22.04
steps:
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Install gcc12
run: sudo apt-get install -y g++-12
- name: Use cmake
run: |
mkdir build &&
cd build &&
CXX=g++-12 cmake -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GLIBCXX_ASSERTIONS=ON -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON .. &&
cmake --build . &&
ctest . -E avoid_
@@ -1,4 +1,4 @@
name: Ubuntu 20.04 CI (GCC 9) with Thread Sanitizer
name: Ubuntu 22.04 CI (GCC 12) with Thread Sanitizer
on: [push, pull_request]
@@ -7,10 +7,10 @@ jobs:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-20.04
runs-on: ubuntu-22.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
@@ -18,7 +18,7 @@ jobs:
run: |
mkdir build &&
cd build &&
cmake -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_SANITIZE_THREADS=ON .. &&
CXX=g++-12 cmake -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_SANITIZE_THREADS=ON .. &&
cmake --build . --target document_stream_tests --target ondemand_document_stream_tests --target parse_many_test &&
ctest --output-on-failure -R parse_many_test &&
ctest --output-on-failure -R document_stream_tests
+18 -4
View File
@@ -9,24 +9,38 @@ jobs:
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-22.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake to build just the library
run: |
mkdir buildjustlib &&
cd buildjustlib &&
cmake -DCMAKE_BUILD_TYPE=Release -DBUILD_SHARED_LIBS=OFF -DSIMDJSON_DEVELOPER_MODE=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
cmake --install . &&
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp &&
c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson &&
cd ../tests/installation_tests/find &&
mkdir buildjustlib &&
cd buildjustlib &&
cmake -DCMAKE_INSTALL_PREFIX:PATH=../../../buildjustlib/destination .. &&
cmake --build .
- name: Use cmake
run: |
mkdir builddebug &&
cd builddebug &&
cmake -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
ctest --output-on-failure -LE explicitonly -j &&
cd .. &&
mkdir build &&
cd build &&
cmake -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
ctest --output-on-failure -LE explicitonly -j &&
cmake --install . &&
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp && c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson && ./linkandrun jsonexamples/twitter.json &&
cd ../tests/installation_tests/find &&
@@ -13,10 +13,10 @@ jobs:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-20.04
runs-on: ubuntu-24.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
@@ -0,0 +1,22 @@
name: Ubuntu 24.04 CI (CXX 20, noexcept)
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-24.04
strategy:
matrix:
cxx: [g++-13, clang++-16]
steps:
- uses: actions/checkout@a5ac7e51b41094c92402da3b24376905380afc29 # v4.1.6
- name: Prepare
run: cmake -DSIMDJSON_CXX_STANDARD=20 -DSIMDJSON_EXCEPTIONS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -B build
env:
CXX: ${{matrix.cxx}}
- name: Build
run: cmake --build build -j=2
- name: Test
run: ctest --output-on-failure --test-dir build
+22
View File
@@ -0,0 +1,22 @@
name: Ubuntu 24.04 CI (CXX 20)
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-24.04
strategy:
matrix:
cxx: [g++-13, clang++-16]
steps:
- uses: actions/checkout@a5ac7e51b41094c92402da3b24376905380afc29 # v4.1.6
- name: Prepare
run: cmake -DSIMDJSON_CXX_STANDARD=20 -DSIMDJSON_DEVELOPER_MODE=ON -B build
env:
CXX: ${{matrix.cxx}}
- name: Build
run: cmake --build build -j=2
- name: Test
run: ctest --output-on-failure --test-dir build
@@ -7,10 +7,10 @@ jobs:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-20.04
runs-on: ubuntu-24.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
@@ -20,15 +20,15 @@ jobs:
cd builddebug &&
cmake -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_EXCEPTIONS=OFF -DBUILD_SHARED_LIBS=OFF .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
ctest --output-on-failure -LE explicitonly -j &&
cd .. &&
mkdir build &&
cd build &&
cmake -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DSIMDJSON_EXCEPTIONS=OFF -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
ctest --output-on-failure -LE explicitonly -j &&
make install &&
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp && c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson && ./linkandrun jsonexamples/twitter.json &&
mkdir testfindpackage &&
cd testfindpackage &&
echo -e 'cmake_minimum_required(VERSION 3.1)\nproject(simdjsontester)\nset(CMAKE_CXX_STANDARD 17)\nfind_package(simdjson REQUIRED)'> CMakeLists.txt && mkdir build && cd build && cmake -DCMAKE_INSTALL_PREFIX:PATH=../destination .. && cmake --build .
echo -e 'cmake_minimum_required(VERSION 3.14)\nproject(simdjsontester)\nset(CMAKE_CXX_STANDARD 17)\nfind_package(simdjson REQUIRED)'> CMakeLists.txt && mkdir build && cd build && cmake -DCMAKE_INSTALL_PREFIX:PATH=../destination .. && cmake --build .
@@ -7,10 +7,10 @@ jobs:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-20.04
runs-on: ubuntu-24.04
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
@@ -20,15 +20,15 @@ jobs:
cd builddebug &&
cmake -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_ENABLE_THREADS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
ctest --output-on-failure -LE explicitonly -j &&
cd .. &&
mkdir build &&
cd build &&
cmake -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_GOOGLE_BENCHMARKS=ON -DSIMDJSON_ENABLE_THREADS=OFF -DBUILD_SHARED_LIBS=OFF -DCMAKE_INSTALL_PREFIX:PATH=destination .. &&
cmake --build . &&
ctest -j --output-on-failure -LE explicitonly &&
ctest --output-on-failure -LE explicitonly -j &&
make install &&
echo -e '#include <simdjson.h>\nint main(int argc,char**argv) {simdjson::dom::parser parser;simdjson::dom::element tweets = parser.load(argv[1]); }' > tmp.cpp && c++ -Idestination/include -Ldestination/lib -std=c++17 -Wl,-rpath,destination/lib -o linkandrun tmp.cpp -lsimdjson && ./linkandrun jsonexamples/twitter.json &&
mkdir testfindpackage &&
cd testfindpackage &&
echo -e 'cmake_minimum_required(VERSION 3.1)\nproject(simdjsontester)\nset(CMAKE_CXX_STANDARD 17)\nfind_package(simdjson REQUIRED)'> CMakeLists.txt && mkdir build && cd build && cmake -DCMAKE_INSTALL_PREFIX:PATH=../destination .. && cmake --build .
echo -e 'cmake_minimum_required(VERSION 3.14)\nproject(simdjsontester)\nset(CMAKE_CXX_STANDARD 17)\nfind_package(simdjson REQUIRED)'> CMakeLists.txt && mkdir build && cd build && cmake -DCMAKE_INSTALL_PREFIX:PATH=../destination .. && cmake --build .
+41
View File
@@ -0,0 +1,41 @@
name: Ubuntu 20.04 CI (GCC 9) With Memory Sanitizer
on: [push, pull_request]
jobs:
ubuntu-build-address-sanitizier:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-24.04
steps:
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake with address sanitizer
run: |
mkdir builddebug &&
cd builddebug &&
cmake -DSIMDJSON_SANITIZE=ON -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF .. &&
cmake --build . &&
ctest --output-on-failure -LE explicitonly -j
ubuntu-build-undefined-sanitizer:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-24.04
steps:
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
- name: Use cmake with undefined sanitizer
run: |
mkdir builddebugundefsani &&
cd builddebugundefsani &&
cmake -DSIMDJSON_SANITIZE_UNDEFINED=ON -DCMAKE_BUILD_TYPE=Debug -DSIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_DEVELOPER_MODE=ON -DBUILD_SHARED_LIBS=OFF .. &&
cmake --build . &&
ctest --output-on-failure -LE explicitonly -j
+25
View File
@@ -0,0 +1,25 @@
name: Ubuntu 24.04 CI
on: [push, pull_request]
jobs:
ubuntu-build:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
runs-on: ubuntu-24.04
strategy:
matrix:
shared: [ON, OFF]
cxx: [g++-13, clang++-16]
sanitizer: [ON, OFF]
build_type: [RelWithDebInfo, Debug, Release]
steps:
- uses: actions/checkout@a5ac7e51b41094c92402da3b24376905380afc29 # v4.1.6
- name: Prepare
run: cmake -DCMAKE_BUILD_TYPE=${{matrix.build_type}} -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_SANITIZE=${{matrix.sanitizer}} -DBUILD_SHARED_LIBS=${{matrix.shared}} -B build
env:
CXX: ${{matrix.cxx}}
- name: Build
run: cmake --build build -j=2
- name: Test
run: ctest --output-on-failure --test-dir build
+4 -4
View File
@@ -10,12 +10,12 @@ jobs:
fail-fast: false
matrix:
include:
- {arch: ARM}
- {arch: ARM64}
- {arch: ARM64EC}
steps:
- name: checkout
uses: actions/checkout@v2
uses: actions/checkout@v4
- name: Use cmake
run: |
cmake -A ${{ matrix.arch }} -DCMAKE_CROSSCOMPILING=1 -DSIMDJSON_DEVELOPER_MODE=ON -D SIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_EXCEPTIONS=OFF -B build &&
cmake --build build --verbose
cmake -A ${{ matrix.arch }} -DCMAKE_SYSTEM_VERSION="10.0.22621.0" -DCMAKE_CROSSCOMPILING=1 -DSIMDJSON_DEVELOPER_MODE=ON -D SIMDJSON_GOOGLE_BENCHMARKS=OFF -DSIMDJSON_EXCEPTIONS=OFF -B build &&
cmake --build build --verbose
+44
View File
@@ -0,0 +1,44 @@
name: VS17-CI CXX20
on: [push, pull_request]
jobs:
ci:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
name: windows-vs17
runs-on: windows-latest
strategy:
fail-fast: false
matrix:
include:
- {gen: Visual Studio 17 2022, arch: Win32, shared: ON}
- {gen: Visual Studio 17 2022, arch: Win32, shared: OFF}
- {gen: Visual Studio 17 2022, arch: x64, shared: ON}
- {gen: Visual Studio 17 2022, arch: x64, shared: OFF}
steps:
- name: checkout
uses: actions/checkout@v4
- name: Configure
run: |
cmake -DSIMDJSON_CXX_STANDARD=20 -G "${{matrix.gen}}" -A ${{matrix.arch}} -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_COMPETITION=OFF -DBUILD_SHARED_LIBS=${{matrix.shared}} -B build
- name: Build Debug
run: cmake --build build --config Debug --verbose
- name: Build Release
run: cmake --build build --config Release --verbose
- name: Run Release tests
run: |
cd build
ctest -C Release -LE explicitonly --output-on-failure
- name: Run Debug tests
run: |
cd build
ctest -C Debug -LE explicitonly --output-on-failure
- name: Install
run: |
cmake --install build --config Release
- name: Test Installation
run: |
cmake -G "${{matrix.gen}}" -A ${{matrix.arch}} -B build_install_test tests/installation_tests/find
cmake --build build_install_test --config Release
+30
View File
@@ -0,0 +1,30 @@
name: VS17-CI-SANITIZE
on: [push, pull_request]
jobs:
ci:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
name: windows-vs17
runs-on: windows-latest
strategy:
fail-fast: false
matrix:
include:
- {gen: Visual Studio 17 2022, arch: x64, shared: OFF, build_type: Debug}
- {gen: Visual Studio 17 2022, arch: x64, shared: OFF, build_type: Release}
- {gen: Visual Studio 17 2022, arch: x64, shared: OFF, build_type: RelWithDebInfo}
steps:
- name: checkout
uses: actions/checkout@v4
- name: Configure
run: |
cmake -G "${{matrix.gen}}" -A ${{matrix.arch}} -DSANITIZE=ON -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_COMPETITION=OFF -DBUILD_SHARED_LIBS=${{matrix.shared}} -B build
- name: Build
run: cmake --build build --config ${{matrix.build_type}} --verbose
- name: Run tests
run: |
cd build
ctest -C ${{matrix.build_type}} -LE explicitonly --output-on-failure
+12 -16
View File
@@ -13,32 +13,28 @@ jobs:
fail-fast: false
matrix:
include:
- {gen: Visual Studio 17 2022, arch: Win32, shared: ON}
- {gen: Visual Studio 17 2022, arch: Win32, shared: OFF}
- {gen: Visual Studio 17 2022, arch: x64, shared: ON}
- {gen: Visual Studio 17 2022, arch: x64, shared: OFF}
- {gen: Visual Studio 17 2022, arch: Win32, shared: ON, build_type: Release}
- {gen: Visual Studio 17 2022, arch: Win32, shared: OFF, build_type: Release}
- {gen: Visual Studio 17 2022, arch: x64, shared: ON, build_type: Release}
- {gen: Visual Studio 17 2022, arch: x64, shared: OFF, build_type: Debug}
- {gen: Visual Studio 17 2022, arch: x64, shared: OFF, build_type: Release}
- {gen: Visual Studio 17 2022, arch: x64, shared: OFF, build_type: RelWithDebInfo}
steps:
- name: checkout
uses: actions/checkout@v2
uses: actions/checkout@v4
- name: Configure
run: |
cmake -G "${{matrix.gen}}" -A ${{matrix.arch}} -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_COMPETITION=OFF -DBUILD_SHARED_LIBS=${{matrix.shared}} -B build
- name: Build Debug
run: cmake --build build --config Debug --verbose
- name: Build Release
run: cmake --build build --config Release --verbose
- name: Run Release tests
run: cmake --build build --config ${{matrix.build_type}} --verbose
- name: Run tests
run: |
cd build
ctest -C Release -LE explicitonly --output-on-failure
- name: Run Debug tests
run: |
cd build
ctest -C Debug -LE explicitonly --output-on-failure
ctest -C ${{matrix.build_type}} -LE explicitonly --output-on-failure
- name: Install
run: |
cmake --install build --config Release
cmake --install build --config ${{matrix.build_type}}
- name: Test Installation
run: |
cmake -G "${{matrix.gen}}" -A ${{matrix.arch}} -B build_install_test tests/installation_tests/find
cmake --build build_install_test --config Release
cmake --build build_install_test --config ${{matrix.build_type}}
+37
View File
@@ -0,0 +1,37 @@
name: VS17-CLANG-CI
on: [push, pull_request]
jobs:
ci:
if: >-
! contains(toJSON(github.event.commits.*.message), '[skip ci]') &&
! contains(toJSON(github.event.commits.*.message), '[skip github]')
name: windows-vs17
runs-on: windows-latest
strategy:
fail-fast: false
matrix:
include:
- {gen: Visual Studio 17 2022, arch: x64, build_type: Debug}
- {gen: Visual Studio 17 2022, arch: x64, build_type: Release}
- {gen: Visual Studio 17 2022, arch: x64, build_type: RelWithDebInfo}
steps:
- name: checkout
uses: actions/checkout@v4
- name: Configure
run: |
cmake -G "${{matrix.gen}}" -A ${{matrix.arch}} -T ClangCL -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_COMPETITION=OFF -B build
- name: Build
run: cmake --build build --config ${{matrix.build_type}} --verbose
- name: Run tests
run: |
cd build
ctest -C ${{matrix.build_type}} -LE explicitonly --output-on-failure
- name: Install
run: |
cmake --install build --config ${{matrix.build_type}}
- name: Test Installation
run: |
cmake -G "${{matrix.gen}}" -A ${{matrix.arch}} -B build_install_test tests/installation_tests/find
cmake --build build_install_test --config ${{matrix.build_type}}
+10 -14
View File
@@ -13,29 +13,25 @@ jobs:
fail-fast: false
matrix:
include:
- {gen: Visual Studio 17 2022, arch: x64}
- {gen: Visual Studio 17 2022, arch: x64, build_type: Debug}
- {gen: Visual Studio 17 2022, arch: x64, build_type: Release}
- {gen: Visual Studio 17 2022, arch: x64, build_type: RelWithDebInfo}
steps:
- name: checkout
uses: actions/checkout@v2
uses: actions/checkout@v4
- name: Configure
run: |
cmake -G "${{matrix.gen}}" -A ${{matrix.arch}} -T ClangCL -DSIMDJSON_DEVELOPER_MODE=ON -DSIMDJSON_COMPETITION=OFF -B build
- name: Build Debug
run: cmake --build build --config Debug --verbose
- name: Build Release
run: cmake --build build --config Release --verbose
- name: Run Release tests
- name: Build
run: cmake --build build --config ${{matrix.build_type}} --verbose
- name: Run tests
run: |
cd build
ctest -C Release -LE explicitonly --output-on-failure
- name: Run Debug tests
run: |
cd build
ctest -C Debug -LE explicitonly --output-on-failure
ctest -C ${{matrix.build_type}} -LE explicitonly --output-on-failure
- name: Install
run: |
cmake --install build --config Release
cmake --install build --config ${{matrix.build_type}}
- name: Test Installation
run: |
cmake -G "${{matrix.gen}}" -A ${{matrix.arch}} -B build_install_test tests/installation_tests/find
cmake --build build_install_test --config Release
cmake --build build_install_test --config ${{matrix.build_type}}
+2 -2
View File
@@ -7,8 +7,8 @@ jobs:
name: windows-vs17
runs-on: windows-latest
steps:
- uses: actions/checkout@v2
- uses: actions/cache@v2
- uses: actions/checkout@v4
- uses: actions/cache@v4
with:
path: dependencies/.cache
key: ${{ hashFiles('dependencies/CMakeLists.txt') }}
+4 -1
View File
@@ -38,7 +38,7 @@ cmake-build-release/
.history/
# Visual Studio artifacts
/VS/
/.vs/
# C/C++ build outputs
.build/
@@ -104,3 +104,6 @@ objs
!.vscode/tasks.json
!.vscode/launch.json
!.vscode/extensions.json
# clangd
.cache
+8 -3
View File
@@ -4,10 +4,15 @@
// List of extensions which should be recommended for users of this workspace.
"recommendations": [
// Syntax
"ms-vscode.cpptools",
"ms-vscode.cmake-tools",
// C++
"llvm-vs-code-extensions.vscode-clangd",
"xaver.clang-format",
// Python
"ms-python.python",
// .github/*
"github.vscode-github-actions",
// cmake
"ms-vscode.cmake-tools",
"twxs.cmake"
],
// List of extensions recommended by VS Code that should not be recommended for users of this workspace.
+44 -1
View File
@@ -3,8 +3,12 @@
{"column": 95 },
{"column": 120 }
],
"cmake.configureArgs": [
"-DSIMDJSON_DEVELOPER_MODE=ON"
],
"files.trimTrailingWhitespace": true,
"files.associations": {
".clangd": "yaml",
"array": "cpp",
"iterator": "cpp",
"chrono": "cpp",
@@ -86,6 +90,45 @@
"vector": "cpp",
"*.ipp": "cpp",
"__functional_base_03": "cpp",
"filesystem": "cpp"
"filesystem": "cpp",
"*.inc": "cpp",
"compare": "cpp",
"concepts": "cpp",
"variant": "cpp",
"__bits": "cpp",
"csignal": "cpp",
"future": "cpp",
"queue": "cpp",
"shared_mutex": "cpp",
"ranges": "cpp",
"span": "cpp",
"__verbose_abort": "cpp",
"charconv": "cpp",
"source_location": "cpp",
"strstream": "cpp",
"typeindex": "cpp",
"*.tcc": "cpp",
"memory_resource": "cpp",
"numbers": "cpp",
"semaphore": "cpp",
"stop_token": "cpp",
"cfenv": "cpp",
"format": "cpp",
"xlocmes": "cpp",
"xlocmon": "cpp",
"xlocnum": "cpp",
"xloctime": "cpp",
"xutility": "cpp",
"coroutine": "cpp",
"xfacet": "cpp",
"xhash": "cpp",
"xiosbase": "cpp",
"xlocale": "cpp",
"xlocbuf": "cpp",
"xlocinfo": "cpp",
"xmemory": "cpp",
"xstring": "cpp",
"xtr1common": "cpp",
"xtree": "cpp"
}
}
+140 -18
View File
@@ -3,7 +3,7 @@ cmake_minimum_required(VERSION 3.14)
project(
simdjson
# The version number is modified by tools/release.py
VERSION 2.0.2
VERSION 4.0.7
DESCRIPTION "Parsing gigabytes of JSON per second"
HOMEPAGE_URL "https://simdjson.org/"
LANGUAGES CXX C
@@ -20,8 +20,14 @@ string(
# ---- Options, variables ----
# These version numbers are modified by tools/release.py
set(SIMDJSON_LIB_VERSION "11.0.0" CACHE STRING "simdjson library version")
set(SIMDJSON_LIB_SOVERSION "11" CACHE STRING "simdjson library soversion")
set(SIMDJSON_LIB_VERSION "27.0.0" CACHE STRING "simdjson library version")
set(SIMDJSON_LIB_SOVERSION "27" CACHE STRING "simdjson library soversion")
option(SIMDJSON_BUILD_STATIC_LIB "Build simdjson_static library along with simdjson (only makes sense if BUILD_SHARED_LIBS=ON)" OFF)
if(SIMDJSON_BUILD_STATIC_LIB AND NOT BUILD_SHARED_LIBS)
message(WARNING "SIMDJSON_BUILD_STATIC_LIB only makes sense if BUILD_SHARED_LIBS is set to ON")
message(WARNING "You might be building and installing a two identical static libraries.")
endif()
option(SIMDJSON_ENABLE_THREADS "Link with thread support" ON)
@@ -37,6 +43,20 @@ if(SIMDJSON_DISABLE_DEPRECATED_API)
)
endif()
if(${CMAKE_VERSION} VERSION_GREATER_EQUAL "3.25.0")
option(SIMDJSON_STATIC_REFLECTION "Enables static reflection (experimental), requires C++26" OFF)
else()
set(SIMDJSON_STATIC_REFLECTION OFF CACHE BOOL "Enables static reflection (experimental)" FORCE)
message(WARNING "SIMDJSON_STATIC_REFLECTION is disabled because your CMake version is below 3.25")
endif()
if(SIMDJSON_STATIC_REFLECTION)
simdjson_add_props(
target_compile_definitions PUBLIC
SIMDJSON_STATIC_REFLECTION=1
)
endif()
option(SIMDJSON_DEVELOPMENT_CHECKS "Enable development-time aids, such as \
checks for incorrect API usage. Enabled by default in DEBUG." OFF)
if(SIMDJSON_DEVELOPMENT_CHECKS)
@@ -49,6 +69,7 @@ endif()
if(is_top_project)
option(SIMDJSON_DEVELOPER_MODE "Enable targets for developing simdjson" OFF)
option(BUILD_SHARED_LIBS "Build simdjson as a shared library" OFF)
option(SIMDJSON_SINGLEHEADER "Disable singleheader generation" ON)
endif()
include(cmake/handle-deprecations.cmake)
@@ -56,8 +77,17 @@ include(cmake/developer-options.cmake)
# ---- simdjson library ----
add_library(simdjson src/simdjson.cpp)
set(SIMDJSON_SOURCES src/simdjson.cpp)
add_library(simdjson ${SIMDJSON_SOURCES})
add_library(simdjson::simdjson ALIAS simdjson)
set(SIMDJSON_LIBRARIES simdjson)
if(SIMDJSON_BUILD_STATIC_LIB)
add_library(simdjson_static STATIC ${SIMDJSON_SOURCES})
add_library(simdjson::simdjson_static ALIAS simdjson_static)
list(APPEND SIMDJSON_LIBRARIES simdjson_static)
endif()
set_target_properties(
simdjson PROPERTIES
@@ -68,7 +98,7 @@ set_target_properties(
)
# FIXME: Use proper CMake integration for exports
if(MSVC AND BUILD_SHARED_LIBS)
if(WIN32 AND BUILD_SHARED_LIBS)
target_compile_definitions(
simdjson
PRIVATE SIMDJSON_BUILDING_WINDOWS_DYNAMIC_LIBRARY=1
@@ -82,7 +112,16 @@ simdjson_add_props(
PRIVATE "$<BUILD_INTERFACE:${PROJECT_SOURCE_DIR}/src>"
)
simdjson_add_props(target_compile_features PUBLIC cxx_std_11)
if(SIMDJSON_STATIC_REFLECTION)
# We would like to require C++26, but no compiler supports that!
# This is a hack:
simdjson_add_props(
target_compile_options PUBLIC
-freflection -fexpansion-statements -stdlib=libc++ -std=c++26
)
else()
simdjson_add_props(target_compile_features PUBLIC cxx_std_11)
endif()
# workaround for GNU GCC poor AVX load/store code generation
if(
@@ -95,6 +134,45 @@ if(
)
endif()
option(SIMDJSON_MINUS_ZERO_AS_FLOAT "Treat -0 as a floating-point value" OFF)
if(SIMDJSON_MINUS_ZERO_AS_FLOAT)
simdjson_add_props(target_compile_definitions PRIVATE SIMDJSON_MINUS_ZERO_AS_FLOAT=1)
endif(SIMDJSON_MINUS_ZERO_AS_FLOAT)
if(CMAKE_SYSTEM_PROCESSOR MATCHES "^(loongarch64)$")
option(SIMDJSON_PREFER_LSX "Prefer LoongArch SX" ON)
include(CheckCXXCompilerFlag)
check_cxx_compiler_flag(-mlasx COMPILER_SUPPORTS_LASX)
check_cxx_compiler_flag(-mlsx COMPILER_SUPPORTS_LSX)
if(COMPILER_SUPPORTS_LASX AND NOT SIMDJSON_PREFER_LSX)
simdjson_add_props(
target_compile_options PRIVATE
-mlasx
)
elseif(COMPILER_SUPPORTS_LSX)
simdjson_add_props(
target_compile_options PRIVATE
-mlsx
)
endif()
endif()
# GCC and Clang have horrendous Debug builds when using SIMD.
# A common fix is to use '-Og' instead.
# bug https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54412
if(
(CMAKE_CXX_COMPILER_ID STREQUAL "GNU" OR
CMAKE_CXX_COMPILER_ID STREQUAL "Clang" OR
CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
)
message(STATUS "Adding -Og to compile flag")
simdjson_add_props(
target_compile_options PRIVATE
$<$<CONFIG:DEBUG>:-Og>
)
endif()
if(SIMDJSON_ENABLE_THREADS)
find_package(Threads REQUIRED)
simdjson_add_props(target_link_libraries PUBLIC Threads::Threads)
@@ -102,17 +180,22 @@ if(SIMDJSON_ENABLE_THREADS)
endif()
simdjson_apply_props(simdjson)
if(SIMDJSON_BUILD_STATIC_LIB)
simdjson_apply_props(simdjson_static)
endif()
# ---- Install rules ----
include(CMakePackageConfigHelpers)
include(GNUInstallDirs)
install(
FILES singleheader/simdjson.h
DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
COMPONENT simdjson_Development
)
if(SIMDJSON_SINGLEHEADER)
install(
FILES singleheader/simdjson.h
DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
COMPONENT simdjson_Development
)
endif()
install(
TARGETS simdjson
@@ -123,7 +206,6 @@ install(
ARCHIVE COMPONENT simdjson_Development
INCLUDES DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
)
configure_file(cmake/simdjson-config.cmake.in simdjson-config.cmake @ONLY)
write_basic_package_version_file(
@@ -149,7 +231,40 @@ install(
EXPORT simdjsonTargets
NAMESPACE simdjson::
DESTINATION "${SIMDJSON_INSTALL_CMAKEDIR}"
COMPONENT example_Development
COMPONENT simdjson_Development
)
if(SIMDJSON_BUILD_STATIC_LIB)
install(
TARGETS simdjson_static
EXPORT simdjson_staticTargets
ARCHIVE COMPONENT simdjson_Development
INCLUDES DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
)
install(
EXPORT simdjson_staticTargets
NAMESPACE simdjson::
DESTINATION "${SIMDJSON_INSTALL_CMAKEDIR}"
COMPONENT simdjson_Development
)
endif()
# pkg-config
include(cmake/JoinPaths.cmake)
join_paths(PKGCONFIG_INCLUDEDIR "\${prefix}" "${CMAKE_INSTALL_INCLUDEDIR}")
join_paths(PKGCONFIG_LIBDIR "\${prefix}" "${CMAKE_INSTALL_LIBDIR}")
if(SIMDJSON_ENABLE_THREADS)
set(PKGCONFIG_CFLAGS "-DSIMDJSON_THREADS_ENABLED=1")
if(CMAKE_THREAD_LIBS_INIT)
set(PKGCONFIG_LIBS_PRIVATE "Libs.private: ${CMAKE_THREAD_LIBS_INIT}")
endif()
endif()
configure_file("simdjson.pc.in" "simdjson.pc" @ONLY)
install(
FILES "${CMAKE_CURRENT_BINARY_DIR}/simdjson.pc"
DESTINATION "${CMAKE_INSTALL_LIBDIR}/pkgconfig"
)
#
@@ -170,11 +285,14 @@ endif()
# ---- Developer mode extras ----
if(is_top_project AND NOT SIMDJSON_DEVELOPER_MODE)
message(STATUS "Building only the library. Advanced users and contributors may want to turn SIMDJSON_DEVELOPER_MODE to ON, e.g., via -D SIMDJSON_DEVELOPER_MODE=ON.")
elseif(SIMDJSON_DEVELOPER_MODE AND NOT is_top_project)
message(AUTHOR_WARNING "Developer mode in simdjson is intended for the developers of simdjson")
endif()
if(NOT SIMDJSON_DEVELOPER_MODE)
message(STATUS "Building only the library. Advanced users may want to turn SIMDJSON_DEVELOPER_MODE to ON, e.g., via -D SIMDJSON_DEVELOPER_MODE=ON.")
return()
elseif(NOT is_top_project)
message(AUTHOR_WARNING "Developer mode is intended for developers of simdjson")
endif()
simdjson_apply_props(simdjson-internal-flags)
@@ -198,6 +316,7 @@ enable_testing()
add_custom_target(all_tests)
add_subdirectory(windows)
include(cmake/CPM.cmake)
add_subdirectory(dependencies) ## This needs to be before tools because of cxxopts
add_subdirectory(tools) ## This needs to be before tests because of cxxopts
@@ -205,8 +324,9 @@ add_subdirectory(tools) ## This needs to be before tests because of cxxopts
# most of the data has been moved to https://github.com/simdjson/simdjson-data
add_subdirectory(jsonexamples)
if(SIMDJSON_SINGLEHEADER)
add_subdirectory(singleheader)
endif()
@@ -215,7 +335,9 @@ add_subdirectory(singleheader)
#
add_subdirectory(tests)
add_subdirectory(examples)
add_subdirectory(benchmark)
if(CMAKE_SIZEOF_VOID_P EQUAL 8) # we only include the benchmarks on 64-bit systems.
add_subdirectory(benchmark)
endif()
add_subdirectory(fuzz)
#
+8 -7
View File
@@ -52,23 +52,24 @@ General Guidelines
Contributors are encouraged to :
- Document their changes. Though we do not enforce a rule regarding code comments, we prefer that non-trivial algorithms and techniques be somewhat documented in the code.
- Follow as much as possible the existing code style. We do not enforce a specific code style, but we prefer consistency.
- Follow as much as possible the existing code style. We do not enforce a specific code style, but we prefer consistency. We avoid contractions (isn't, aren't) in the comments.
- Modify as few lines of code as possible when working on an issue. The more lines you modify, the harder it is for your fellow human beings to understand what is going on.
- Tools may report "problems" with the code, but we never delegate programming to tools: if there is a problem with the code, we need to understand it. Thus we will not "fix" code merely to please a static analyzer if we do not understand.
- Tools may report "problems" with the code, but we never delegate programming to tools: if there is a problem with the code, we need to understand it. Thus we will not "fix" code merely to please a static analyzer.
- Provide tests for any new feature. We will not merge a new feature without tests.
- Run before/after benchmarks so that we can appreciate the effect of the changes on the performance.
Pull Requests
--------------
Pull requests are always invited. However, we ask that you follow these guidelines:
- It is wiser to discuss your ideas first as part of an issue before you start coding. If you omit this step and code first, be prepare to have your code receive scrutiny and be dropped.
- Users should provide a rationale for their changes. Does it improve performance? Does it add a feature? Does it improve maintainability? Does fix a bug? This must be explicitly stated as part of the pull request. Do not propose changes based on taste or intuition. We do not delegate programming to tools: that some tool suggested a code change is not reason enough to change the code.
- It is wise to discuss your ideas first as part of an issue before you start coding. If you omit this step and code first, be prepared to have your code receive scrutiny and be dropped.
- Users should provide a rationale for their changes. Does it improve performance? Does it add a feature? Does it improve maintainability? Does it fix a bug? This must be explicitly stated as part of the pull request. Do not propose changes based on taste or intuition. We do not delegate programming to tools: that some tool suggested a code change is not reason enough to change the code.
1. When your code improves performance, please document the gains with a benchmark using hard numbers.
2. If your code fixes a bug, please be either fix a failing test, or propose a new test.
2. If your code fixes a bug, please either fix a failing test, or propose a new test.
3. Other types of changes must be clearly motivated. We openly discourage changes with no identifiable benefits.
- Changes should be focused and minimal. You should change as few lines of code as possible. Please do not reformat or touch files needlessly.
- New features must be accompanied of new tests, in general.
- New features must be accompanied by new tests, in general.
- Your code should pass our continuous-integration tests. It is your responsibility to ensure that your proposal pass the tests. We do not merge pull requests that would break our build.
- An exception to this would be changes to non-code files, such as documentation and assets, or trivial changes to code, such as comments, where it is encouraged to explicitly ask for skipping a CI run using the `[skip ci]` prefix in your Pull Request title **and** in the first line of the most recent commit in a push. Example for such a commit: `[skip ci] Fixed typo in power_of_ten's docs`
This benefits the project in such a way that the CI pipeline is not burdened by running jobs on changes that don't change any behavior in the code, which reduces wait times for other Pull Requests that do change behavior and require testing.
@@ -91,7 +92,7 @@ We welcome contributions from women and less represented groups. If you need hel
Consider the following points when engaging with the project:
- We discourage arguments from authority: ideas are discusssed on their own merits and not based on who stated it.
- We discourage arguments from authority: ideas are discussed on their own merits and not based on who stated it.
- Be mindful that what you may view as an aggression is maybe merely a difference of opinion or a misunderstanding.
- Be mindful that a collection of small aggressions, even if mild in isolation, can become harmful.
+11 -5
View File
@@ -38,7 +38,7 @@ PROJECT_NAME = simdjson
# could be handy for archiving the generated documentation or if some version
# control system is used.
PROJECT_NUMBER = "2.0.2"
PROJECT_NUMBER = "4.0.7"
# Using the PROJECT_BRIEF tag one can provide an optional one line description
# for a project that appears at the top of each page and should give viewer a
@@ -829,7 +829,7 @@ WARN_LOGFILE =
# spaces. See also FILE_PATTERNS and EXTENSION_MAPPING
# Note: If this tag is empty the current directory is searched.
INPUT = doc include
INPUT = doc include/simdjson include/simdjson/dom include/simdjson/generic
# This tag can be used to specify the character encoding of the source files
# that doxygen parses. Internally doxygen uses the UTF-8 encoding. Doxygen uses
@@ -1246,7 +1246,10 @@ HTML_STYLESHEET =
# list). For an example see the documentation.
# This tag requires that the tag GENERATE_HTML is set to YES.
HTML_EXTRA_STYLESHEET =
HTML_EXTRA_STYLESHEET = theme/doxygen-awesome.css \
theme/doxygen-awesome-sidebar-only.css \
theme/doxygen-awesome-sidebar-only-darkmode-toggle.css
# The HTML_EXTRA_FILES tag can be used to specify one or more extra images or
# other source files which should be copied to the HTML output directory. Note
@@ -1256,7 +1259,10 @@ HTML_EXTRA_STYLESHEET =
# files will be copied as-is; there are no commands or markers available.
# This tag requires that the tag GENERATE_HTML is set to YES.
HTML_EXTRA_FILES =
HTML_EXTRA_FILES = theme/doxygen-awesome-darkmode-toggle.js \
theme/doxygen-awesome-interactive-toc.js \
theme/doxygen-awesome-fragment-copy-button.js \
theme/doxygen-awesome-paragraph-link.js
# The HTML_COLORSTYLE_HUE tag controls the color of the HTML output. Doxygen
# will adjust the colors in the style sheet and background images according to
@@ -1543,7 +1549,7 @@ DISABLE_INDEX = NO
# The default value is: NO.
# This tag requires that the tag GENERATE_HTML is set to YES.
GENERATE_TREEVIEW = NO
GENERATE_TREEVIEW = YES
# The ENUM_VALUES_PER_LINE tag can be used to set the number of enum values that
# doxygen will group on one line in the generated HTML documentation.
+94 -29
View File
@@ -1,3 +1,4 @@
Hacking simdjson
================
@@ -5,6 +6,67 @@ Here is wisdom about how to build, test and run simdjson from within the reposit
If you plan to contribute to simdjson, please read our [CONTRIBUTING](https://github.com/simdjson/simdjson/blob/master/CONTRIBUTING.md) guide.
- [Hacking simdjson](#hacking-simdjson)
- [Build Quickstart](#build-quickstart)
- [Design notes](#design-notes)
- [Developer mode](#developer-mode)
- [Directory Structure and Source](#directory-structure-and-source)
- [Runtime Dispatching](#runtime-dispatching)
- [Regenerating Single-Header Files](#regenerating-single-header-files)
- [Usage (CMake on 64-bit platforms like Linux, FreeBSD or macOS)](#usage-cmake-on-64-bit-platforms-like-linux-freebsd-or-macos)
- [Usage (CMake on 64-bit Windows using Visual Studio 2019 or better)](#usage-cmake-on-64-bit-windows-using-visual-studio-2019-or-better)
- [Various References](#various-references)
Build Quickstart
------------------------------
For non-Windows system,
```bash
cmake -B -D SIMDJSON_DEVELOPER_MODE=ON ..
cmake --build build
ctest --test-dir build
```
It is similar for Visual Studio users, please see the CMake or Visual Studio documentation.
By default the library is built in Release mode.
Assertions and development checks
------------------------------
We do not use conventional `assert` in simdjson. Instead we use the macro
`SIMDJSON_ASSUME`:
```cpp
SIMDJSON_ASSUME(something_that_is_true());
```
Sometimes, you need to do a bit more work that a simple check.
The `SIMDJSON_DEVELOPMENT_CHECKS` macro is true only in Debug mode unless manually set.
It is acceptable to add checks that you would not do in Release mode as long as
they are guarded:
```cpp
#if SIMDJSON_DEVELOPMENT_CHECKS
// do sanity checks here
```
Working with sanitizers
------------------------------
The simdjson library must be memory-safe. We cannot allow buffer overruns.
During development, if you system supports it, we recommend configuring
the project with `-D SIMDJSON_SANITIZE=ON`.
```bash
cmake -B -D SIMDJSON_SANITIZE=ON -D SIMDJSON_DEVELOPER_MODE=ON ..
cmake --build build
ctest --test-dir build
```
Design notes
------------------------------
@@ -52,37 +114,42 @@ simdjson's source structure, from the top level, looks like this:
* **CMakeLists.txt:** The main build system.
* **include:** User-facing declarations and inline definitions (most user-facing functions are inlined).
* simdjson.h: A "main include" that includes files from include/simdjson/. This is equivalent to
* simdjson.h: the `simdjson` namespace. A "main include" that includes files from include/simdjson/. This is equivalent to
the distributed simdjson.h.
* simdjson/*.h: Declarations for public simdjson classes and functions.
* simdjson/*-inl.h: Definitions for public simdjson classes and functions.
* simdjson/*.h: Declarations for public simdjson classes and functions.
* simdjson/*-inl.h: Definitions for public simdjson classes and functions.
* simdjson/internal/*.h: the `simdjson::internal` namespace. Private classes and functions used by the rest of simdjson.
* simdjson/dom.h: the `simdjson::dom` namespace. Includes all public DOM classes.
* simdjson/dom/*.h: Declarations/definitions for individual DOM classes.
* simdjson/arm64|fallback|haswell|icelake|ppc64|westmere.h: `simdjson::<implementation>` namespace. Common implementation-specific tools like number and string parsing, as well as minification.
* simdjson/arm64|fallback|haswell|icelake|ppc64|westmere/*.h: implementation-specific functions such as , etc.
* simdjson/generic/*.h: the bulk of the actual code, written generically and compiled for each implementation, using functions defined in the implementation's .h files.
* simdjson/generic/dependencies.h: dependencies on common, non-implementation-specific simdjson classes. This will be included before including amalgamated.h.
* simdjson/generic/amalgamated.h: all generic ondemand classes for an implementation.
* simdjson/ondemand.h: the `simdjson::ondemand` namespace. Includes all public ondemand classes.
* simdjson/builtin.h: the `simdjson::builtin` namespace. Aliased to the most universal implementation available.
* simdjson/builtin/ondemand.h: the `simdjson::builtin::ondemand` namespace.
* simdjson/arm64|fallback|haswell|icelake|ppc64|westmere/ondemand.h: the `simdjson::<implementation>::ondemand` namespace. On-Demand compiled for the specific implementation.
* simdjson/generic/ondemand/*.h: individual On-Demand classes, generically written.
* simdjson/generic/ondemand/dependencies.h: dependencies on common, non-implementation-specific simdjson classes. This will be included before including amalgamated.h.
* simdjson/generic/ondemand/amalgamated.h: all generic ondemand classes for an implementation.
* **src:** The source files for non-inlined functionality (e.g. the architecture-specific parser
implementations).
* simdjson.cpp: A "main source" that includes all implementation files from src/. This is
equivalent to the distributed simdjson.cpp.
* arm64/|fallback/|haswell/|ppc64/|westmere/: Architecture-specific implementations. All functions are
Each architecture defines its own namespace, e.g. simdjson::haswell.
* generic/: Generic implementations of the simdjson parser. These files may be included and
compiled multiple times, from whichever architectures use them. They assume they are already
enclosed in a namespace, e.g.:
```c++
namespace simdjson {
namespace haswell {
#include "generic/stage1/json_structural_indexer.h"
}
}
```
* *.cpp: other misc. implementations, such as `simdjson::implementation` and the minifier.
* arm64|fallback|haswell|icelake|ppc64|westmere.cpp: Architecture-specific parser implementations.
* generic/*.h: `simdjson::<implementation>` namespace. Generic implementation of the parser, particularly the `dom_parser_implementation`.
* generic/stage1/*.h: `simdjson::<implementation>::stage1` namespace. Generic implementation of the simd-heavy tokenizer/indexer pass of the simdjson parser. Used for the On-Demand interface
* generic/stage2/*.h: `simdjson::<implementation>::stage2` namespace. Generic implementation of the tape creator, which consumes the index from stage 1 and actually parses numbers and string and such. Used for the DOM interface.
Other important files and directories:
* **.drone.yml:** Definitions for Drone CI.
* **.appveyor.yml:** Definitions for Appveyor CI (Windows).
* **.circleci:** Definitions for Circle CI.
* **.github/workflows:** Definitions for GitHub Actions (CI).
* **singleheader:** Contains generated `simdjson.h` and `simdjson.cpp` that we release. The files `singleheader/simdjson.h` and `singleheader/simdjson.cpp` should never be edited by hand.
* **singleheader/amalgamate.py:** Generates `singleheader/simdjson.h` and `singleheader/simdjson.cpp` for release (python script).
* **singleheader/amalgamate.py:** Generates `singleheader/simdjson.h` and `singleheader/simdjson.cpp` for release (python script). If you add a new implementation (e.g., rvv), you need to edit this file (IMPLEMENTATIONS).
* **benchmark:** This is where we do benchmarking. Benchmarking is core to every change we make; the
cardinal rule is don't regress performance without knowing exactly why, and what you're trading
for it. Many of our benchmarks are microbenchmarks. We are effectively doing controlled scientific experiments for the purpose of understanding what affects our performance. So we simplify as much as possible. We try to avoid irrelevant factors such as page faults, interrupts, unnnecessary system calls. We recommend checking the performance as follows:
for it. Many of our benchmarks are microbenchmarks. We are effectively doing controlled scientific experiments for the purpose of understanding what affects our performance. So we simplify as much as possible. We try to avoid irrelevant factors such as page faults, interrupts, unnecessary system calls. We recommend checking the performance as follows:
```bash
mkdir build
cd build
@@ -114,7 +181,7 @@ Other important files and directories:
* `json2json mydoc.json` parses the document, constructs a model and then dumps back the result to standard output.
* `json2json -d mydoc.json` parses the document, constructs a model and then dumps model (as a tape) to standard output. The tape format is described in the accompanying file `tape.md`.
* `minify mydoc.json` minifies the JSON document, outputting the result to standard output. Minifying means to remove the unneeded white space characters.
*`jsonpointer mydoc.json <jsonpath> <jsonpath> ... <jsonpath>` parses the document, constructs a model and then processes a series of [JSON Pointer paths](https://tools.ietf.org/html/rfc6901). The result is itself a JSON document.
* `jsonpointer mydoc.json <jsonpath> <jsonpath> ... <jsonpath>` parses the document, constructs a model and then processes a series of [JSON Pointer paths](https://tools.ietf.org/html/rfc6901). The result is itself a JSON document.
> **Don't modify the files in singleheader/ directly; these are automatically generated.**
@@ -193,7 +260,7 @@ point it gets included (but only once per header). singleheader/simdjson.cpp is
src/simdjson.cpp the same way, except files under generic/ may be included and copy/pasted multiple
times.
### Usage (CMake on 64-bit platforms like Linux, FreeBSD or macOS)
## Usage (CMake on 64-bit platforms like Linux, FreeBSD or macOS)
Requirements: In addition to git, we require a recent version of CMake as well as bash.
@@ -254,9 +321,9 @@ If your compiler does not default on C++11 support or better you may get failing
Note that the name of directory (`build`) is arbitrary, you can name it as you want (e.g., `buildgcc`) and you can have as many different such directories as you would like (one per configuration).
## Usage (CMake on 64-bit Windows using Visual Studio 2019 or better)
### Usage (CMake on 64-bit Windows using Visual Studio 2019)
Recent versions of Visual Studio support CMake natively, [please refer to the Visual Studio documentation](https://learn.microsoft.com/en-us/cpp/build/cmake-projects-in-visual-studio?view=msvc-170).
We assume you have a common 64-bit Windows PC with at least Visual Studio 2019.
@@ -273,19 +340,17 @@ Though having Visual Studio installed is necessary, one can build simdjson using
- `mkdir build`
- `cd build`
- `cmake ..`
- `cmake --build . -config Release`
- `cmake --build . --config Release`
Furthermore, if you have installed LLVM clang on Windows, for example as a component of Visual Studio 2019, you can configure and build simdjson using LLVM clang on Windows using cmake:
- `mkdir build`
- `cd build`
- `cmake -T ClangCL ..`
- `cmake --build . -config Release`
- `cmake --build . --config Release`
### Various References
## Various References
- [How to implement atoi using SIMD?](https://stackoverflow.com/questions/35127060/how-to-implement-atoi-using-simd)
- [Parsing JSON is a Minefield 💣](http://seriot.ch/parsing_json.php)
+1 -1
View File
@@ -186,7 +186,7 @@
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright 2018-2019 The simdjson authors
Copyright 2018-2025 The simdjson authors
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
+18
View File
@@ -0,0 +1,18 @@
Copyright 2018-2025 The simdjson authors
Permission is hereby granted, free of charge, to any person obtaining a copy of
this software and associated documentation files (the "Software"), to deal in
the Software without restriction, including without limitation the rights to
use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
the Software, and to permit persons to whom the Software is furnished to do so,
subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+78 -28
View File
@@ -1,9 +1,7 @@
[![][license img]][license] [![][licensemit img]][licensemit]
![Ubuntu 18.04 CI](https://github.com/simdjson/simdjson/workflows/Ubuntu%2018.04%20CI%20(GCC%207)/badge.svg)
[![Ubuntu 20.04 CI](https://github.com/simdjson/simdjson/workflows/Ubuntu%2020.04%20CI%20(GCC%209)/badge.svg)](https://simdjson.org/plots.html)
![VS16-CI](https://github.com/simdjson/simdjson/workflows/VS16-CI/badge.svg)
![MinGW64-CI](https://github.com/simdjson/simdjson/workflows/MinGW64-CI/badge.svg)
[![][license img]][license] [![Doxygen Documentation](https://img.shields.io/badge/docs-doxygen-green.svg)](https://simdjson.org/api/2.0.0/index.html)
[![Doxygen Documentation](https://img.shields.io/badge/docs-doxygen-green.svg)](https://simdjson.github.io/simdjson/)
simdjson : Parsing gigabytes of JSON per second
===============================================
@@ -26,16 +24,49 @@ This library is part of the [Awesome Modern C++](https://awesomecpp.com) list.
Table of Contents
-----------------
* [Real-world usage](#real-world-usage)
* [Quick Start](#quick-start)
* [Documentation](#documentation)
* [Godbolt](#godbolt)
* [Performance results](#performance-results)
* [Real-world usage](#real-world-usage)
* [Packages](#packages)
* [Bindings and Ports of simdjson](#bindings-and-ports-of-simdjson)
* [About simdjson](#about-simdjson)
* [Funding](#funding)
* [Contributing to simdjson](#contributing-to-simdjson)
* [License](#license)
Real-world usage
----------------
- [Node.js](https://nodejs.org/)
- [ClickHouse](https://github.com/ClickHouse/ClickHouse)
- [Meta Velox](https://velox-lib.io)
- [Google Pax](https://github.com/google/paxml)
- [milvus](https://github.com/milvus-io/milvus)
- [QuestDB](https://questdb.io/blog/questdb-release-8-0-3/)
- [Clang Build Analyzer](https://github.com/aras-p/ClangBuildAnalyzer)
- [Shopify HeapProfiler](https://github.com/Shopify/heap-profiler)
- [StarRocks](https://github.com/StarRocks/starrocks)
- [Microsoft FishStore](https://github.com/microsoft/FishStore)
- [Intel PCM](https://github.com/intel/pcm)
- [WatermelonDB](https://github.com/Nozbe/WatermelonDB)
- [Apache Doris](https://github.com/apache/doris)
- [Dgraph](https://github.com/dgraph-io/dgraph)
- [UJRPC](https://github.com/unum-cloud/ujrpc)
- [fastgltf](https://github.com/spnda/fastgltf)
- [vast](https://github.com/tenzir/vast)
- [ada-url](https://github.com/ada-url/ada)
- [fastgron](https://github.com/adamritter/fastgron)
- [WasmEdge](https://wasmedge.org)
- [RonDB](https://github.com/logicalclocks/rondb)
- [GreptimeDB](https://github.com/GreptimeTeam/greptimedb)
- [mamba](https://github.com/mamba-org/mamba)
If you are planning to use simdjson in a product, please work from one of our releases.
Quick Start
-----------
@@ -43,9 +74,9 @@ The simdjson library is easily consumable with a single .h and .cpp file.
0. Prerequisites: `g++` (version 7 or better) or `clang++` (version 6 or better), and a 64-bit
system with a command-line shell (e.g., Linux, macOS, freeBSD). We also support programming
environments like Visual Studio and Xcode, but different steps are needed.
environments like Visual Studio and Xcode, but different steps are needed. Users of clang++ may need to specify the C++ version (e.g., `c++ -std=c++17`) since clang++ tends to default on C++98.
1. Pull [simdjson.h](singleheader/simdjson.h) and [simdjson.cpp](singleheader/simdjson.cpp) into a
directory, along with the sample file [twitter.json](jsonexamples/twitter.json).
directory, along with the sample file [twitter.json](jsonexamples/twitter.json). You can download them with the `wget` utility:
```
wget https://raw.githubusercontent.com/simdjson/simdjson/master/singleheader/simdjson.h https://raw.githubusercontent.com/simdjson/simdjson/master/singleheader/simdjson.cpp https://raw.githubusercontent.com/simdjson/simdjson/master/jsonexamples/twitter.json
@@ -62,13 +93,14 @@ int main(void) {
ondemand::document tweets = parser.iterate(json);
std::cout << uint64_t(tweets["search_metadata"]["count"]) << " results." << std::endl;
}
```
```
3. `c++ -o quickstart quickstart.cpp simdjson.cpp`
4. `./quickstart`
```
```
100 results.
```
```
Documentation
-------------
@@ -76,10 +108,19 @@ Documentation
Usage documentation is available:
* [Basics](doc/basics.md) is an overview of how to use simdjson and its APIs.
* [Builder](doc/builder.md) is an overview of how to efficiently write JSON strings using simdjson.
* [Performance](doc/performance.md) shows some more advanced scenarios and how to tune for them.
* [Implementation Selection](doc/implementation-selection.md) describes runtime CPU detection and
how you can work with it.
* [API](https://simdjson.org/api/1.0.0/annotated.html) contains the automatically generated API documentation.
* [API](https://simdjson.github.io/simdjson/) contains the automatically generated API documentation.
Godbolt
-------------
Some users may want to browse code along with the compiled assembly. You want to check out the following lists of examples:
* [C++26 reflection example](https://godbolt.org/z/K3Px64TqK)
* [simdjson examples with errors handled through exceptions](https://godbolt.org/z/7G5qE4sr9)
* [simdjson examples with errors without exceptions](https://godbolt.org/z/e9dWb9E4v)
Performance results
-------------------
@@ -107,16 +148,10 @@ speed for [synthetic files over various sizes generated with a script](https://g
For NDJSON files, we can exceed 3 GB/s with [our multithreaded parsing functions](https://github.com/simdjson/simdjson/blob/master/doc/parse_many.md).
Packages
------------------------------
[![Packaging status](https://repology.org/badge/vertical-allrepos/simdjson.svg)](https://repology.org/project/simdjson/versions)
Real-world usage
----------------
- [Microsoft FishStore](https://github.com/microsoft/FishStore)
- [Yandex ClickHouse](https://github.com/yandex/ClickHouse)
- [Clang Build Analyzer](https://github.com/aras-p/ClangBuildAnalyzer)
- [Shopify HeapProfiler](https://github.com/Shopify/heap-profiler)
If you are planning to use simdjson in a product, please work from one of our releases.
Bindings and Ports of simdjson
------------------------------
@@ -137,9 +172,15 @@ We distinguish between "bindings" (which just wrap the C++ code) and a port to a
- [simdjson-go](https://github.com/minio/simdjson-go): Go port using Golang assembly.
- [rcppsimdjson](https://github.com/eddelbuettel/rcppsimdjson): R bindings.
- [simdjson_erlang](https://github.com/ChomperT/simdjson_erlang): erlang bindings.
- [simdjsone](https://github.com/saleyn/simdjsone): erlang bindings.
- [lua-simdjson](https://github.com/FourierTransformer/lua-simdjson): lua bindings.
- [hermes-json](https://hackage.haskell.org/package/hermes-json): haskell bindings.
- [zimdjson](https://github.com/EzequielRamis/zimdjson): Zig port.
- [simdjzon](https://github.com/travisstaloch/simdjzon): Zig port.
- [JSON-Simd](https://github.com/rawleyfowler/JSON-simd): Raku bindings.
- [JSON::SIMD](https://metacpan.org/pod/JSON::SIMD): Perl bindings; fully-featured JSON module that uses simdjson for decoding.
- [gemmaJSON](https://github.com/sainttttt/gemmaJSON): Nim JSON parser based on simdjson bindings.
- [simdjson-java](https://github.com/simdjson/simdjson-java): Java port.
About simdjson
--------------
@@ -148,7 +189,11 @@ The simdjson library takes advantage of modern microarchitectures, parallelizing
instructions, reducing branch misprediction, and reducing data dependency to take advantage of each
CPU's multiple execution cores.
Some people [enjoy reading our paper](https://arxiv.org/abs/1902.08318): A description of the design
Our default front-end is called On-Demand, and we wrote a paper about it:
- John Keiser, Daniel Lemire, [On-Demand JSON: A Better Way to Parse Documents?](http://arxiv.org/abs/2312.17149), Software: Practice and Experience 54 (6), 2024.
Some people [enjoy reading the first (2019) simdjson paper](https://arxiv.org/abs/1902.08318): A description of the design
and implementation of simdjson is in our research article:
- Geoff Langdale, Daniel Lemire, [Parsing Gigabytes of JSON per Second](https://arxiv.org/abs/1902.08318), VLDB Journal 28 (6), 2019.
@@ -165,12 +210,17 @@ For the video inclined, <br />
Funding
-------
The work is supported by the Natural Sciences and Engineering Research Council of Canada under grant
number RGPIN-2017-03910.
The work is supported by the Natural Sciences and Engineering Research Council of Canada under grants
RGPIN-2017-03910 and RGPIN-2024-03787.
[license]: LICENSE
[license img]: https://img.shields.io/badge/License-Apache%202-blue.svg
[licensemit]: LICENSE-MIT
[licensemit img]: https://img.shields.io/badge/License-MIT-blue.svg
Contributing to simdjson
------------------------
@@ -180,9 +230,9 @@ Head over to [CONTRIBUTING.md](CONTRIBUTING.md) for information on contributing
License
-------
This code is made available under the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0.html).
This code is made available under the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0.html) as well as under the MIT License. As a user, you can pick the license you prefer.
Under Windows, we build some tools using the windows/dirent_portable.h file (which is outside our library code): it under the liberal (business-friendly) MIT license.
Under Windows, we build some tools using the windows/dirent_portable.h file (which is outside our library code): it is under the liberal (business-friendly) MIT license.
For compilers that do not support [C++17](https://en.wikipedia.org/wiki/C%2B%2B17), we bundle the string-view library which is published under the [Boost license](http://www.boost.org/LICENSE_1_0.txt). Like the Apache license, the Boost license is a permissive license allowing commercial redistribution.
-69
View File
@@ -1,69 +0,0 @@
# 0.5
## Highlights
Performance
* Faster and simpler UTF-8 validation with the lookup4 algorithm https://github.com/simdjson/simdjson/pull/993
* We improved the performance of simdjson under Visual Studio by about 25%. Users will still get better performance with clang-cl (+30%) but the gap has been reduced. https://github.com/simdjson/simdjson/pull/1031
Code usability
* In `parse_many`, when parsing streams of JSON documetns, we give to the users runtime control as to whether threads are used (via the parser.threaded attribute). https://github.com/simdjson/simdjson/issues/925
* Prefixed public macros to avoid name clashes with other libraries. https://github.com/simdjson/simdjson/issues/1035
* Better documentation regarding package managers (brew, MSYS2, conan, apt, vcpkg, FreeBSD package manager, etc.).
* Better documentation regarding CMake usage.
Standards
* We improved standard compliance with respect to both the JSON RFC 8259 and JSON Pointer RFC 6901. We added the at_pointer method to nodes for standard-compliant JSON Pointer queries. The legacy `at(std::string_view)` method remains but is deprecated since it is not standard-compliant as per RFC 6901.
* We removed computed GOTOs without sacrificing performance thus improving the C++ standard compliance (since computed GOTOs are compiler-specific extensions).
* Better support for C++20 https://github.com/simdjson/simdjson/pull/1050
# 0.4
## Highlights
- Test coverage has been greatly improved and we have resolved many static-analysis warnings on different systems.
- We added a fast (8GB/s) minifier that works directly on JSON strings.
- We added fast (10GB/s) UTF-8 validator that works directly on strings (any strings, including non-JSON).
- The array and object elements have a constant-time size() method.
- Performance improvements to the API (type(), get<>()).
- The parse_many function (ndjson) has been entirely reworked. It now uses a single secondary thread instead of several new threads.
- We have introduced a faster UTF-8 validation algorithm (lookup3) for all kernels (ARM, x64 SSE, x64 AVX).
- C++11 support for older compilers and systems.
- FreeBSD support (and tests).
- We support the clang front-end compiler (clangcl) under Visual Studio.
- It is now possible to target ARM platforms under Visual Studio.
- The simdjson library will never abort or print to standard output/error.
# 0.3
## Highlights
- **Multi-Document Parsing:** Read a bundle of JSON documents (ndjson) 2-4x faster than doing it
individually. [API docs](https://github.com/simdjson/simdjson/blob/master/doc/basics.md#newline-delimited-json-ndjson-and-json-lines) / [Design Details](https://github.com/simdjson/simdjson/blob/master/doc/parse_many.md)
- **Simplified API:** The API has been completely revamped for ease of use, including a new JSON
navigation API and fluent support for error code *and* exception styles of error handling with a
single API. [Docs](https://github.com/simdjson/simdjson/blob/master/doc/basics.md#the-basics-loading-and-parsing-json-documents)
- **Exact Float Parsing:** Now simdjson parses floats flawlessly *without* any performance loss,
thanks to [great work by @michaeleisel and @lemire](https://github.com/simdjson/simdjson/pull/558).
[Blog Post](https://lemire.me/blog/2020/03/10/fast-float-parsing-in-practice/)
- **Even Faster:** The fastest parser got faster! With a [shiny new UTF-8 validator](https://github.com/simdjson/simdjson/pull/387)
and meticulously refactored SIMD core, simdjson 0.3 is 15% faster than before, running at 2.5 GB/s
(where 0.2 ran at 2.2 GB/s).
## Minor Highlights
- Fallback implementation: simdjson now has a non-SIMD fallback implementation, and can run even on
very old 64-bit machines.
- Automatic allocation: as part of API simplification, the parser no longer has to be preallocated--
it will adjust automatically when it encounters larger files.
- Runtime selection API: We've exposed simdjson's runtime CPU detection and implementation selection
as an API, so you can tell what implementation we detected and test with other implementations.
- Error handling your way: Whether you use exceptions or check error codes, simdjson lets you handle
errors in your style. APIs that can fail return simdjson_result<T>, letting you check the error
code before using the result. But if you are more comfortable with exceptions, skip the error code
and cast straight to T, and exceptions will be thrown automatically if an error happens. Use the
same API either way!
- Error chaining: We also worked to keep non-exception error-handling short and sweet. Instead of
having to check the error code after every single operation, now you can *chain* JSON navigation
calls like looking up an object field or array element, or casting to a string, so that you only
have to check the error code once at the very end.
+7
View File
@@ -0,0 +1,7 @@
# Security Policy
## Reporting a Vulnerability
Please use the following contact information for reporting a vulnerability:
- [Daniel Lemire](https://github.com/lemire) - daniel@lemire.me
+19
View File
@@ -4,6 +4,9 @@ add_subdirectory(dom)
include_directories( . linux )
link_libraries(simdjson-windows-headers test-data)
link_libraries(simdjson)
if(SIMDJSON_STATIC_REFLECTION)
add_compile_definitions(SIMDJSON_STATIC_REFLECTION=1)
endif(SIMDJSON_STATIC_REFLECTION)
add_executable(benchfeatures benchfeatures.cpp)
add_executable(get_corpus_benchmark get_corpus_benchmark.cpp)
@@ -26,6 +29,22 @@ if (TARGET benchmark::benchmark)
if(TARGET nlohmann_json)
target_link_libraries(bench_ondemand PRIVATE nlohmann_json)
endif()
if(TARGET boostjson)
target_link_libraries(bench_ondemand PRIVATE boostjson)
endif()
endif()
endif()
include(CheckCXXCompilerFlag)
check_cxx_compiler_flag("-std=c++20" SIMDJSON_COMPILER_SUPPORTS_CXX20)
if(SIMDJSON_STATIC_REFLECTION)
add_subdirectory(static_reflect)
else()
if(SIMDJSON_EXCEPTIONS AND SIMDJSON_COMPILER_SUPPORTS_CXX20)
add_subdirectory(from)
add_subdirectory(car_builder)
endif()
endif(SIMDJSON_STATIC_REFLECTION)
@@ -15,11 +15,11 @@ using namespace json_benchmark;
struct brand {
double cumulative_rating;
uint64_t reviews_count;
simdjson_really_inline bool operator==(const brand &other) const {
simdjson_inline bool operator==(const brand &other) const {
return cumulative_rating == other.cumulative_rating &&
reviews_count == other.reviews_count;
}
simdjson_really_inline bool operator!=(const brand &other) const { return !(*this == other); }
simdjson_inline bool operator!=(const brand &other) const { return !(*this == other); }
};
simdjson_unused static std::ostream &operator<<(std::ostream &o, const brand &b) {
@@ -66,8 +66,8 @@ struct runner : public file_runner<I> {
template<bool threaded>
struct simdjson_dom;
template<typename I> simdjson_really_inline static void amazon_cellphones(benchmark::State &state) {
template<typename I> simdjson_inline static void amazon_cellphones(benchmark::State &state) {
run_json_benchmark<runner<I>, runner<simdjson_dom<UNTHREADED>>>(state);
}
} // namespace amazon_cellphones
} // namespace amazon_cellphones
File diff suppressed because it is too large Load Diff
+22 -155
View File
@@ -473,28 +473,6 @@ static void twitter_count(State& state) {
}
BENCHMARK(twitter_count);
#ifndef SIMDJSON_DISABLE_DEPRECATED_API
SIMDJSON_PUSH_DISABLE_WARNINGS
SIMDJSON_DISABLE_DEPRECATED_WARNING
static void iterator_twitter_count(State& state) {
// Prints the number of results in twitter.json
padded_string json = padded_string::load(TWITTER_JSON);
ParsedJson pj = build_parsed_json(json);
for (simdjson_unused auto _ : state) {
ParsedJson::Iterator iter(pj);
// uint64_t result_count = doc["search_metadata"]["count"];
if (!iter.move_to_key("search_metadata")) { return; }
if (!iter.move_to_key("count")) { return; }
if (!iter.is_integer()) { return; }
int64_t result_count = iter.get_integer();
if (result_count != 100) { return; }
}
}
BENCHMARK(iterator_twitter_count);
SIMDJSON_POP_DISABLE_WARNINGS
#endif // SIMDJSON_DISABLE_DEPRECATED_API
static void twitter_default_profile(State& state) {
// Count unique users with a default profile.
dom::parser parser;
@@ -521,7 +499,7 @@ static void twitter_image_sizes(State& state) {
set<tuple<uint64_t, uint64_t>> image_sizes;
for (dom::object tweet : doc["statuses"]) {
dom::array media;
if (not (error = tweet["entities"]["media"].get(media))) {
if (! (error = tweet["entities"]["media"].get(media))) {
for (dom::object image : media) {
for (auto size : image["sizes"].get_object()) {
image_sizes.emplace(size.value["w"], size.value["h"]);
@@ -564,7 +542,7 @@ static void error_code_twitter_default_profile(State& state) noexcept {
for (dom::element tweet : tweets) {
dom::object user;
if ((error = tweet["user"].get(user))) { return; }
bool default_profile;
bool default_profile{};
if ((error = user["default_profile"].get(default_profile))) { return; }
if (default_profile) {
std::string_view screen_name;
@@ -578,54 +556,6 @@ static void error_code_twitter_default_profile(State& state) noexcept {
}
BENCHMARK(error_code_twitter_default_profile);
#ifndef SIMDJSON_DISABLE_DEPRECATED_API
SIMDJSON_PUSH_DISABLE_WARNINGS
SIMDJSON_DISABLE_DEPRECATED_WARNING
static void iterator_twitter_default_profile(State& state) {
// Count unique users with a default profile.
padded_string json;
auto error = padded_string::load(TWITTER_JSON).get(json);
if (error) { std::cerr << error << std::endl; return; }
ParsedJson pj = build_parsed_json(json);
for (simdjson_unused auto _ : state) {
set<string_view> default_users;
ParsedJson::Iterator iter(pj);
// for (dom::object tweet : doc["statuses"]) {
if (!(iter.move_to_key("statuses") && iter.is_array())) { return; }
if (iter.down()) { // first status
do {
// dom::object user = tweet["user"];
if (!(iter.move_to_key("user") && iter.is_object())) { return; }
// if (user["default_profile"]) {
if (iter.move_to_key("default_profile")) {
if (iter.is_true()) {
if (!iter.up()) { return; } // back to user
// default_users.insert(user["screen_name"]);
if (!(iter.move_to_key("screen_name") && iter.is_string())) { return; }
default_users.emplace(iter.get_string(), iter.get_string_length());
}
if (!iter.up()) { return; } // back to user
}
if (!iter.up()) { return; } // back to status
} while (iter.next()); // next status
}
if (default_users.size() != 86) { return; }
}
}
SIMDJSON_POP_DISABLE_WARNINGS
BENCHMARK(iterator_twitter_default_profile);
#endif // SIMDJSON_DISABLE_DEPRECATED_API
static void error_code_twitter_image_sizes(State& state) noexcept {
// Count unique image sizes
dom::parser parser;
@@ -638,7 +568,7 @@ static void error_code_twitter_image_sizes(State& state) noexcept {
if ((error = doc["statuses"].get(statuses))) { return; }
for (dom::element tweet : statuses) {
dom::array images;
if (not (error = tweet["entities"]["media"].get(images))) {
if (! (error = tweet["entities"]["media"].get(images))) {
for (dom::element image : images) {
dom::object sizes;
if ((error = image["sizes"].get(sizes))) { return; }
@@ -656,91 +586,28 @@ static void error_code_twitter_image_sizes(State& state) noexcept {
}
BENCHMARK(error_code_twitter_image_sizes);
#ifndef SIMDJSON_DISABLE_DEPRECATED_API
SIMDJSON_PUSH_DISABLE_WARNINGS
SIMDJSON_DISABLE_DEPRECATED_WARNING
static void iterator_twitter_image_sizes(State& state) {
// Count unique image sizes
padded_string json;
auto error = padded_string::load(TWITTER_JSON).get(json);
if (error) { std::cerr << error << std::endl; return; }
ParsedJson pj = build_parsed_json(json);
for (simdjson_unused auto _ : state) {
set<tuple<uint64_t, uint64_t>> image_sizes;
ParsedJson::Iterator iter(pj);
// for (dom::object tweet : doc["statuses"]) {
if (!(iter.move_to_key("statuses") && iter.is_array())) { return; }
if (iter.down()) { // first status
do {
// dom::object media;
// not_found = tweet["entities"]["media"].get(media);
// if (!not_found) {
if (iter.move_to_key("entities")) {
if (!iter.is_object()) { return; }
if (iter.move_to_key("media")) {
if (!iter.is_array()) { return; }
// for (dom::object image : media) {
if (iter.down()) { // first media
do {
// for (auto [key, size] : dom::object(image["sizes"])) {
if (!(iter.move_to_key("sizes") && iter.is_object())) { return; }
if (iter.down()) { // first size
do {
iter.move_to_value();
// image_sizes.insert({ size["w"], size["h"] });
if (!(iter.move_to_key("w")) && !iter.is_integer()) { return; }
uint64_t width = iter.get_integer();
if (!iter.up()) { return; } // back to size
if (!(iter.move_to_key("h")) && !iter.is_integer()) { return; }
uint64_t height = iter.get_integer();
if (!iter.up()) { return; } // back to size
image_sizes.emplace(width, height);
} while (iter.next()); // next size
if (!iter.up()) { return; } // back to sizes
}
if (!iter.up()) { return; } // back to image
} while (iter.next()); // next image
if (!iter.up()) { return; } // back to media
}
if (!iter.up()) { return; } // back to entities
}
if (!iter.up()) { return; } // back to status
}
} while (iter.next()); // next status
}
if (image_sizes.size() != 15) { return; };
}
}
BENCHMARK(iterator_twitter_image_sizes);
#endif // SIMDJSON_DISABLE_DEPRECATED_API
#ifndef SIMDJSON_DISABLE_DEPRECATED_API
static void print_json(State& state) noexcept {
// Prints the number of results in twitter.json
static void parse_surrogate_pairs(State& state) {
// NOTE: This mostly exists to show there's a tiny benefit to
// loading and comparing both bytes of "\\u" simultaneously.
// (which should also reduce the compiled code size).
// The repeated surrogate pairs make this easier to measure.
dom::parser parser;
padded_string json;
auto error = padded_string::load(TWITTER_JSON).get(json);
if (error) { std::cerr << error << std::endl; return; }
int code = json_parse(json, parser);
if (code) { cerr << error_message(code) << endl; return; }
const std::string_view data = "\"\\uD834\\uDD1E\\uD834\\uDD1E\\uD834\\uDD1E\\uD834\\uDD1E\\uD834\\uDD1E\\uD834\\uDD1E\\uD834\\uDD1E\\uD834\\uDD1E\\uD834\\uDD1E\\uD834\\uDD1E\"";
padded_string docdata{data};
// we do not want mem. alloc. in the loop.
auto error = parser.allocate(docdata.size());
if (error) {
cout << error << endl;
return;
}
for (simdjson_unused auto _ : state) {
std::stringstream s;
if (!parser.print_json(s)) { cerr << "print_json failed" << endl; return; }
dom::element doc;
if ((error = parser.parse(docdata).get(doc))) {
cerr << "could not parse string" << error << endl;
return;
}
}
}
BENCHMARK(print_json);
#endif // SIMDJSON_DISABLE_DEPRECATED_API
SIMDJSON_POP_DISABLE_WARNINGS
BENCHMARK(parse_surrogate_pairs);
BENCHMARK_MAIN();
+116 -57
View File
@@ -21,10 +21,126 @@ SIMDJSON_PUSH_DISABLE_ALL_WARNINGS
#include <nlohmann/json.hpp>
#endif
#ifdef SIMDJSON_COMPETITION_BOOSTJSON
#include <boost/json.hpp>
#endif
// This has to be last, for reasons I don't yet understand
#include <benchmark/benchmark.h>
SIMDJSON_POP_DISABLE_WARNINGS
#include "json2msgpack/simdjson_ondemand.h"
#include "json2msgpack/simdjson_dom.h"
#include "json2msgpack/yyjson.h"
#include "json2msgpack/rapidjson.h"
#if SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "json2msgpack/sajson.h"
#endif // SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "json2msgpack/nlohmann_json.h"
#include "json2msgpack/boostjson.h"
#include "partial_tweets/simdjson_ondemand.h"
#include "partial_tweets/simdjson_dom.h"
#include "partial_tweets/yyjson.h"
#if SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "partial_tweets/sajson.h"
#endif // SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "partial_tweets/rapidjson.h"
#if SIMDJSON_COMPETITION_SAX
#include "partial_tweets/rapidjson_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "partial_tweets/nlohmann_json.h"
#if SIMDJSON_COMPETITION_SAX
#include "partial_tweets/nlohmann_json_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "partial_tweets/boostjson.h"
#include "distinct_user_id/simdjson_ondemand.h"
#include "distinct_user_id/simdjson_ondemand_json_pointer.h"
#include "distinct_user_id/simdjson_dom.h"
#include "distinct_user_id/simdjson_dom_json_pointer.h"
#include "distinct_user_id/yyjson.h"
#if SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "distinct_user_id/sajson.h"
#endif // SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "distinct_user_id/rapidjson.h"
#if SIMDJSON_COMPETITION_SAX
#include "distinct_user_id/rapidjson_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "distinct_user_id/nlohmann_json.h"
#if SIMDJSON_COMPETITION_SAX
#include "distinct_user_id/nlohmann_json_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "distinct_user_id/boostjson.h"
#include "find_tweet/simdjson_ondemand.h"
#include "find_tweet/simdjson_dom.h"
#include "find_tweet/yyjson.h"
#if SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "find_tweet/sajson.h"
#endif // SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "find_tweet/rapidjson.h"
#if SIMDJSON_COMPETITION_SAX
#include "find_tweet/rapidjson_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "find_tweet/nlohmann_json.h"
#if SIMDJSON_COMPETITION_SAX
#include "find_tweet/nlohmann_json_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "find_tweet/boostjson.h"
#include "top_tweet/simdjson_ondemand.h"
#include "top_tweet/simdjson_dom.h"
#include "top_tweet/yyjson.h"
#if SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "top_tweet/sajson.h"
#endif // SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "top_tweet/rapidjson.h"
#if SIMDJSON_COMPETITION_SAX
#include "top_tweet/rapidjson_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "top_tweet/nlohmann_json.h"
#if SIMDJSON_COMPETITION_SAX
#include "top_tweet/nlohmann_json_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "top_tweet/boostjson.h"
#include "kostya/simdjson_ondemand.h"
#include "kostya/simdjson_dom.h"
#include "kostya/yyjson.h"
#if SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "kostya/sajson.h"
#endif // SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "kostya/rapidjson.h"
#if SIMDJSON_COMPETITION_SAX
#include "kostya/rapidjson_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "kostya/nlohmann_json.h"
#if SIMDJSON_COMPETITION_SAX
#include "kostya/nlohmann_json_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "kostya/boostjson.h"
#include "large_random/simdjson_ondemand.h"
#if SIMDJSON_COMPETITION_ONDEMAND_UNORDERED
#include "large_random/simdjson_ondemand_unordered.h"
#endif // SIMDJSON_COMPETITION_ONDEMAND_UNORDERED
#include "large_random/simdjson_dom.h"
#include "large_random/yyjson.h"
#if SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "large_random/sajson.h"
#endif // SIMDJSON_COMPETITION_ONDEMAND_SAJSON
#include "large_random/rapidjson.h"
#if SIMDJSON_COMPETITION_SAX
#include "large_random/rapidjson_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "large_random/nlohmann_json.h"
#if SIMDJSON_COMPETITION_SAX
#include "large_random/nlohmann_json_sax.h"
#endif // SIMDJSON_COMPETITION_SAX
#include "large_random/boostjson.h"
#include "amazon_cellphones/simdjson_dom.h"
#include "amazon_cellphones/simdjson_ondemand.h"
@@ -32,61 +148,4 @@ SIMDJSON_POP_DISABLE_WARNINGS
#include "large_amazon_cellphones/simdjson_dom.h"
#include "large_amazon_cellphones/simdjson_ondemand.h"
#include "partial_tweets/simdjson_dom.h"
#include "partial_tweets/simdjson_ondemand.h"
#include "partial_tweets/yyjson.h"
#include "partial_tweets/sajson.h"
#include "partial_tweets/rapidjson.h"
#include "partial_tweets/rapidjson_sax.h"
#include "partial_tweets/nlohmann_json.h"
#include "partial_tweets/nlohmann_json_sax.h"
#include "large_random/simdjson_dom.h"
#include "large_random/simdjson_ondemand.h"
#include "large_random/simdjson_ondemand_unordered.h"
#include "large_random/yyjson.h"
#include "large_random/sajson.h"
#include "large_random/rapidjson.h"
#include "large_random/rapidjson_sax.h"
#include "large_random/nlohmann_json.h"
#include "large_random/nlohmann_json_sax.h"
#include "kostya/simdjson_dom.h"
#include "kostya/simdjson_ondemand.h"
#include "kostya/yyjson.h"
#include "kostya/sajson.h"
#include "kostya/rapidjson.h"
#include "kostya/rapidjson_sax.h"
#include "kostya/nlohmann_json.h"
#include "kostya/nlohmann_json_sax.h"
#include "distinct_user_id/simdjson_dom.h"
#include "distinct_user_id/simdjson_dom_json_pointer.h"
#include "distinct_user_id/simdjson_ondemand.h"
#include "distinct_user_id/simdjson_ondemand_json_pointer.h"
#include "distinct_user_id/yyjson.h"
#include "distinct_user_id/sajson.h"
#include "distinct_user_id/rapidjson.h"
#include "distinct_user_id/rapidjson_sax.h"
#include "distinct_user_id/nlohmann_json.h"
#include "distinct_user_id/nlohmann_json_sax.h"
#include "find_tweet/simdjson_dom.h"
#include "find_tweet/simdjson_ondemand.h"
#include "find_tweet/yyjson.h"
#include "find_tweet/sajson.h"
#include "find_tweet/rapidjson.h"
#include "find_tweet/rapidjson_sax.h"
#include "find_tweet/nlohmann_json.h"
#include "find_tweet/nlohmann_json_sax.h"
#include "top_tweet/simdjson_dom.h"
#include "top_tweet/simdjson_ondemand.h"
#include "top_tweet/yyjson.h"
#include "top_tweet/sajson.h"
#include "top_tweet/rapidjson.h"
#include "top_tweet/rapidjson_sax.h"
#include "top_tweet/nlohmann_json.h"
#include "top_tweet/nlohmann_json_sax.h"
BENCHMARK_MAIN();
-30
View File
@@ -169,22 +169,6 @@ BENCHMARK(parse_gsoc)->Repetitions(10)->ComputeStatistics("max", [](const std::v
})->DisplayAggregatesOnly(true);
#ifndef SIMDJSON_DISABLE_DEPRECATED_API
SIMDJSON_PUSH_DISABLE_WARNINGS
SIMDJSON_DISABLE_DEPRECATED_WARNING
static void json_parse(State& state) {
ParsedJson pj;
if (!pj.allocate_capacity(EMPTY_ARRAY.length())) { return; }
for (simdjson_unused auto _ : state) {
auto error = json_parse(EMPTY_ARRAY, pj);
if (error) { return; }
}
}
SIMDJSON_POP_DISABLE_WARNINGS
BENCHMARK(json_parse);
#endif // SIMDJSON_DISABLE_DEPRECATED_API
static void parser_parse_error_code(State& state) {
dom::parser parser;
if (parser.allocate(EMPTY_ARRAY.length())) { return; }
@@ -213,20 +197,6 @@ BENCHMARK(parser_parse_exception);
#endif // SIMDJSON_EXCEPTIONS
#ifndef SIMDJSON_DISABLE_DEPRECATED_API
SIMDJSON_PUSH_DISABLE_WARNINGS
SIMDJSON_DISABLE_DEPRECATED_WARNING
static void build_parsed_json(State& state) {
for (simdjson_unused auto _ : state) {
dom::parser parser = simdjson::build_parsed_json(EMPTY_ARRAY);
if (!parser.valid) { return; }
}
}
SIMDJSON_POP_DISABLE_WARNINGS
BENCHMARK(build_parsed_json);
#endif
static void document_parse_error_code(State& state) {
for (simdjson_unused auto _ : state) {
dom::parser parser;
+40 -4
View File
@@ -164,7 +164,7 @@ struct feature_benchmarker {
}
simdjson_really_inline void run_iterations(size_t iterations, bool stage1_only=false) {
simdjson_inline void run_iterations(size_t iterations, bool stage1_only=false) {
struct7.run_iterations(iterations, stage1_only);
struct7_miss.run_iterations(iterations, stage1_only);
struct7_full.run_iterations(iterations, stage1_only);
@@ -204,10 +204,13 @@ struct feature_benchmarker {
}
// Rate of 1-7-structural misses per 8-structural flip
double struct1_7_miss_rate(BenchmarkStage stage) const {
#if SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
return 1;
#else
if (!has_events()) { return 1; }
return struct7_miss[stage].best.branch_misses() - struct7[stage].best.branch_misses() / double(struct7_miss.stats->blocks_with_1_structural_flipped);
#endif
}
// Extra cost of an 8-15 structural block over a 1-7 structural block
double struct8_15_cost(BenchmarkStage stage) const {
return cost_per_block(stage, struct15, struct15.stats->blocks_with_8_structurals, struct7);
@@ -218,8 +221,12 @@ struct feature_benchmarker {
}
// Rate of 8-15-structural misses per 8-structural flip
double struct8_15_miss_rate(BenchmarkStage stage) const {
#if SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
return 1;
#else
if (!has_events()) { return 1; }
return double(struct15_miss[stage].best.branch_misses() - struct15[stage].best.branch_misses()) / double(struct15_miss.stats->blocks_with_8_structurals_flipped);
#endif
}
// Extra cost of a 16+-structural block over an 8-15 structural block (actual varies based on # of structurals!)
@@ -232,10 +239,15 @@ struct feature_benchmarker {
}
// Rate of 16-structural misses per 16-structural flip
double struct16_miss_rate(BenchmarkStage stage) const {
#if SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
return 1;
#else
if (!has_events()) { return 1; }
return double(struct23_miss[stage].best.branch_misses() - struct23[stage].best.branch_misses()) / double(struct23_miss.stats->blocks_with_16_structurals_flipped);
#endif
}
// Extra cost of having UTF-8 in a block
double utf8_cost(BenchmarkStage stage) const {
return cost_per_block(stage, utf8, utf8.stats->blocks_with_utf8, struct7_full);
@@ -246,10 +258,13 @@ struct feature_benchmarker {
}
// Rate of UTF-8 misses per UTF-8 flip
double utf8_miss_rate(BenchmarkStage stage) const {
#if SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
return 1;
#else
if (!has_events()) { return 1; }
return double(utf8_miss[stage].best.branch_misses() - utf8[stage].best.branch_misses()) / double(utf8_miss.stats->blocks_with_utf8_flipped);
#endif
}
// Extra cost of having escapes in a block
double escape_cost(BenchmarkStage stage) const {
return cost_per_block(stage, escape, escape.stats->blocks_with_escapes, struct7_full);
@@ -260,10 +275,15 @@ struct feature_benchmarker {
}
// Rate of escape misses per escape flip
double escape_miss_rate(BenchmarkStage stage) const {
#if SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
return 1;
#else
if (!has_events()) { return 1; }
return double(escape_miss[stage].best.branch_misses() - escape[stage].best.branch_misses()) / double(escape_miss.stats->blocks_with_escapes_flipped);
#endif
}
double calc_expected_feature_cost(BenchmarkStage stage, const benchmarker& file) const {
// Expected base ns/block (empty)
json_stats& stats = *file.stats;
@@ -300,7 +320,6 @@ struct feature_benchmarker {
double calc_expected(BenchmarkStage stage, const benchmarker& file) const {
return calc_expected_feature_cost(stage, file) + calc_expected_miss_cost(stage, file);
}
void print(const option_struct& options) const {
printf("\n");
printf("Features in ns/block (64 bytes):\n");
@@ -359,6 +378,22 @@ struct feature_benchmarker {
}
};
#if SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
void print_file_effectiveness(BenchmarkStage stage, const char* filename, const benchmarker& results, const feature_benchmarker& features) {
double actual = results[stage].best.elapsed_ns() / double(results.stats->blocks);
double calc = features.calc_expected(stage, results);
double calc_misses = features.calc_expected_misses(stage, results);
double calc_miss_cost = features.calc_expected_miss_cost(stage, results);
printf(" | %-8s ", benchmark_stage_name(stage));
printf("| %-15s ", filename);
printf("| %8.3g ", features.calc_expected_feature_cost(stage, results));
printf("| %8.3g ", calc_miss_cost);
printf("| %8.3g ", calc);
printf("| %8.3g ", actual);
printf("| %+8.3g ", actual - calc);
printf("| %13llu ", (long long unsigned)(calc_misses));
}
#else
void print_file_effectiveness(BenchmarkStage stage, const char* filename, const benchmarker& results, const feature_benchmarker& features) {
double actual = results[stage].best.elapsed_ns() / double(results.stats->blocks);
double calc = features.calc_expected(stage, results);
@@ -382,6 +417,7 @@ void print_file_effectiveness(BenchmarkStage stage, const char* filename, const
}
printf("|\n");
}
#endif
int main(int argc, char *argv[]) {
// Read options
+618
View File
@@ -0,0 +1,618 @@
#!/bin/bash
# JSON Parsing Compilation Benchmark: Reflection Usage vs Manual Parsing
# Compares compilation times when ACTUALLY USING reflection for parsing vs manual parsing
# This measures the compile-time cost of reflection-based automatic deserialization
set -e
echo "=== simdjson Reflection Usage Compilation Benchmark ==="
echo "Measuring compilation impact of ACTUALLY USING reflection for parsing"
echo "Starting at: $(date)"
echo
echo "╔════════════════════════════════════════════════════════════════════════════╗"
echo "║ METHODOLOGY ║"
echo "╠════════════════════════════════════════════════════════════════════════════╣"
echo "║ ║"
echo "║ FAIR COMPARISON STRATEGY: ║"
echo "║ This benchmark compares two DIFFERENT approaches to parsing the same JSON: ║"
echo "║ ║"
echo "║ • MANUAL PARSING: Traditional simdjson with explicit .get() calls ║"
echo "║ - Uses doc[\"field\"].get(variable) for each field ║"
echo "║ - No reflection involved ║"
echo "║ ║"
echo "║ • REFLECTION PARSING: Automatic deserialization with reflection ║"
echo "║ - Uses doc.get<MyStruct>() for automatic field mapping ║"
echo "║ - Relies on compile-time reflection to generate parsing code ║"
echo "║ ║"
echo "║ WHAT WE'RE MEASURING: ║"
echo "║ • Compile-time cost of reflection-based automatic deserialization ║"
echo "║ • Template instantiation overhead for reflection parsing ║"
echo "║ • Code generation complexity from using reflection features ║"
echo "║ ║"
echo "║ TEST SCENARIOS: ║"
echo "║ ║"
echo "║ 1. SIMPLE STRUCT: Basic fields (string, int, bool) ║"
echo "║ - Measures baseline reflection overhead ║"
echo "║ ║"
echo "║ 2. NESTED STRUCT: Multiple levels of nested objects ║"
echo "║ - Measures reflection complexity scaling ║"
echo "║ ║"
echo "║ 3. COMPLEX STRUCT: Arrays, optional fields, mixed types ║"
echo "║ - Measures real-world reflection usage impact ║"
echo "║ ║"
echo "║ WHY THIS IS MEANINGFUL: ║"
echo "║ • Shows actual cost of using reflection features ║"
echo "║ • Measures compile-time code generation overhead ║"
echo "║ • Helps developers understand reflection's compilation impact ║"
echo "║ • Compares equivalent functionality implemented two different ways ║"
echo "║ ║"
echo "╚════════════════════════════════════════════════════════════════════════════╝"
echo
# Configuration
ITERATIONS=10
JOBS=4
# ───────────────────────────── BOX-PRINT HELPER ────────────────────────────
BOX_WIDTH=74 # characters between the pipes
print_box_line() { # usage: print_box_line "text"
printf "║ %-*s ║\n" "${BOX_WIDTH}" "$1"
}
# Function to test if a compiler supports reflection with debug output
test_reflection_support() {
local compiler="$1"
echo " → Testing compiler: $compiler"
if [ ! -x "$compiler" ] && ! command -v "$compiler" >/dev/null 2>&1; then
echo " → Compiler not found or not executable"
return 1
fi
# Check compiler version first
echo " → Compiler version: $("$compiler" --version 2>/dev/null | head -n1 || echo "version check failed")"
# Simple test: check if compiler accepts reflection flags
local test_file=$(mktemp /tmp/reflection_test_XXXXXX.cpp)
cat > "$test_file" << 'EOF'
int main() {
return 0;
}
EOF
echo " → Testing basic reflection flags..."
local test_exe=$(mktemp /tmp/reflection_test_XXXXXX)
local basic_result=$("$compiler" -freflection -fexpansion-statements -std=c++26 "$test_file" -o "$test_exe" 2>&1)
local basic_exit_code=$?
if [ $basic_exit_code -ne 0 ]; then
echo " → Basic flags FAILED with exit code $basic_exit_code"
echo " → Error output: $basic_result"
rm -f "$test_file" "$test_exe"
return 1
fi
echo " → Basic flags: OK"
rm -f "$test_exe"
# Test reflection syntax
echo " → Testing reflection syntax..."
cat > "$test_file" << 'EOF'
struct Test {
int x;
};
int main() {
auto refl = ^^Test;
return 0;
}
EOF
local syntax_result=$("$compiler" -freflection -fexpansion-statements -std=c++26 "$test_file" -o "$test_exe" 2>&1)
local syntax_exit_code=$?
if [ $syntax_exit_code -eq 0 ]; then
echo " → Reflection syntax: OK"
echo " → ✓ REFLECTION SUPPORT CONFIRMED"
rm -f "$test_file" "$test_exe"
return 0
else
echo " → Reflection syntax FAILED with exit code $syntax_exit_code"
echo " → Error output: $syntax_result"
rm -f "$test_file" "$test_exe"
return 1
fi
}
# Find a compiler with reflection support
echo "Searching for clang++ with reflection support..."
REFLECTION_CXX=""
REFLECTION_CC=""
# List of potential clang++ locations to check
POTENTIAL_COMPILERS=(
"/usr/local/bin/clang++"
"/opt/clang/bin/clang++"
"/usr/bin/clang++"
"clang++"
)
# If CXX is already set, test it first
if [ -n "$CXX" ]; then
echo "Testing user-specified compiler: $CXX"
if test_reflection_support "$CXX"; then
REFLECTION_CXX="$CXX"
echo "✓ User-specified compiler supports reflection: $CXX"
else
echo "✗ User-specified compiler does not support reflection: $CXX"
echo "Will search for alternative..."
fi
fi
# If we don't have a working compiler yet, search for one
if [ -z "$REFLECTION_CXX" ]; then
for compiler in "${POTENTIAL_COMPILERS[@]}"; do
echo "Testing: $compiler"
if test_reflection_support "$compiler"; then
REFLECTION_CXX="$compiler"
echo "✓ Found reflection-enabled compiler: $compiler"
break
else
echo "✗ No reflection support: $compiler"
fi
done
fi
# Check if we found a working compiler
if [ -z "$REFLECTION_CXX" ]; then
echo
echo "╔════════════════════════════════════════════════════════════════════════════╗"
echo "║ ERROR ║"
echo "╠════════════════════════════════════════════════════════════════════════════╣"
echo "║ ║"
echo "║ No clang++ compiler with reflection support found! ║"
echo "║ ║"
echo "║ This benchmark requires a compiler that supports C++26 reflection. ║"
echo "║ ║"
echo "║ Options: ║"
echo "║ 1. Use the Docker container: ./p2996/run_docker.sh ║"
echo "║ 2. Build clang with reflection from: https://github.com/bloomberg/clang-p2996 ║"
echo "║ 3. Set CXX environment variable to point to reflection-enabled clang++ ║"
echo "║ ║"
echo "║ Example: CXX=/path/to/reflection-clang++ ./benchmark_script.sh ║"
echo "║ ║"
echo "╚════════════════════════════════════════════════════════════════════════════╝"
exit 1
fi
# Set the compilers
export CXX="$REFLECTION_CXX"
# Find corresponding C compiler
if [ -n "$CC" ]; then
REFLECTION_CC="$CC"
elif [ "$REFLECTION_CXX" = "/usr/local/bin/clang++" ]; then
REFLECTION_CC="/usr/local/bin/clang"
elif [ "$REFLECTION_CXX" = "/opt/clang/bin/clang++" ]; then
REFLECTION_CC="/opt/clang/bin/clang"
elif [ "$REFLECTION_CXX" = "/usr/bin/clang++" ]; then
REFLECTION_CC="/usr/bin/clang"
else
REFLECTION_CC="clang"
fi
export CC="$REFLECTION_CC"
echo
echo "Using reflection-enabled compiler: $($CXX --version | head -n1)"
echo "Using C compiler: $($CC --version | head -n1)"
echo
# Function to create manual parsing test
create_manual_parsing_test() {
local test_name="$1"
local struct_complexity="$2"
cat > "${test_name}_manual.cpp" << 'EOF'
#include <simdjson.h>
#include <iostream>
#include <string>
#include <vector>
#include <optional>
// Test structures
struct Person {
std::string name;
int age;
bool active;
};
struct Address {
std::string street;
std::string city;
int zipcode;
};
struct Employee {
Person person;
Address address;
std::vector<std::string> skills;
std::optional<std::string> department;
double salary;
};
// Manual parsing functions
bool parse_person_manual(simdjson::ondemand::value& val, Person& person) {
auto obj = val.get_object();
if (obj.error()) return false;
for (auto field : obj) {
std::string_view key = field.unescaped_key();
if (key == "name") {
std::string_view name_val;
if (field.value().get(name_val)) return false;
person.name = name_val;
} else if (key == "age") {
if (field.value().get(person.age)) return false;
} else if (key == "active") {
if (field.value().get(person.active)) return false;
}
}
return true;
}
bool parse_address_manual(simdjson::ondemand::value& val, Address& address) {
auto obj = val.get_object();
if (obj.error()) return false;
for (auto field : obj) {
std::string_view key = field.unescaped_key();
if (key == "street") {
std::string_view street_val;
if (field.value().get(street_val)) return false;
address.street = street_val;
} else if (key == "city") {
std::string_view city_val;
if (field.value().get(city_val)) return false;
address.city = city_val;
} else if (key == "zipcode") {
if (field.value().get(address.zipcode)) return false;
}
}
return true;
}
bool parse_employee_manual(simdjson::ondemand::document& doc, Employee& employee) {
auto obj = doc.get_object();
if (obj.error()) return false;
for (auto field : obj) {
std::string_view key = field.unescaped_key();
if (key == "person") {
auto person_val = field.value();
if (!parse_person_manual(person_val, employee.person)) return false;
} else if (key == "address") {
auto addr_val = field.value();
if (!parse_address_manual(addr_val, employee.address)) return false;
} else if (key == "skills") {
auto skills_array = field.value().get_array();
if (skills_array.error()) return false;
for (auto skill : skills_array) {
std::string_view skill_val;
if (skill.get(skill_val)) return false;
employee.skills.emplace_back(skill_val);
}
} else if (key == "department") {
std::string_view dept_val;
if (!field.value().get(dept_val)) {
employee.department = dept_val;
}
} else if (key == "salary") {
if (field.value().get(employee.salary)) return false;
}
}
return true;
}
int main() {
simdjson::ondemand::parser parser;
std::string json_str = R"({
"person": {
"name": "John Doe",
"age": 30,
"active": true
},
"address": {
"street": "123 Main St",
"city": "Anytown",
"zipcode": 12345
},
"skills": ["C++", "JSON", "Programming"],
"department": "Engineering",
"salary": 85000.50
})";
simdjson::ondemand::document doc;
auto error = parser.iterate(simdjson::pad(json_str)).get(doc);
if (error) {
std::cerr << "Parse error" << std::endl;
return 1;
}
Employee employee;
if (!parse_employee_manual(doc, employee)) {
std::cerr << "Manual parsing failed" << std::endl;
return 1;
}
std::cout << "Manual parsing successful: " << employee.person.name
<< ", age " << employee.person.age << std::endl;
return 0;
}
EOF
}
# Function to create reflection parsing test
create_reflection_parsing_test() {
local test_name="$1"
local struct_complexity="$2"
cat > "${test_name}_reflection.cpp" << 'EOF'
#include <simdjson.h>
#include <iostream>
#include <string>
#include <vector>
#include <optional>
// Test structures (same as manual version)
struct Person {
std::string name;
int age;
bool active;
};
struct Address {
std::string street;
std::string city;
int zipcode;
};
struct Employee {
Person person;
Address address;
std::vector<std::string> skills;
std::optional<std::string> department;
double salary;
};
int main() {
simdjson::ondemand::parser parser;
std::string json_str = R"({
"person": {
"name": "John Doe",
"age": 30,
"active": true
},
"address": {
"street": "123 Main St",
"city": "Anytown",
"zipcode": 12345
},
"skills": ["C++", "JSON", "Programming"],
"department": "Engineering",
"salary": 85000.50
})";
simdjson::ondemand::document doc;
auto error = parser.iterate(simdjson::pad(json_str)).get(doc);
if (error) {
std::cerr << "Parse error" << std::endl;
return 1;
}
// Use reflection-based automatic deserialization
Employee employee;
auto result = doc.get<Employee>();
if (result.error()) {
std::cerr << "Reflection parsing failed" << std::endl;
return 1;
}
employee = result.value();
std::cout << "Reflection parsing successful: " << employee.person.name
<< ", age " << employee.person.age << std::endl;
return 0;
}
EOF
}
# Function to time compilation of parsing approach
time_parsing_compilation() {
local description="$1"
local test_file="$2"
local use_reflection="$3"
local iteration="$4"
echo "[$iteration] $description"
# Clean build
rm -rf build_parsing_test
mkdir build_parsing_test
cd build_parsing_test
# Copy test file
cp "../$test_file" .
echo " Configuring..."
if [ "$use_reflection" = "true" ]; then
cmake -DCMAKE_CXX_COMPILER="$CXX" \
-DSIMDJSON_DEVELOPER_MODE=ON \
-DSIMDJSON_STATIC_REFLECTION=ON \
-DBUILD_SHARED_LIBS=OFF \
../.. >/dev/null 2>&1
else
cmake -DCMAKE_CXX_COMPILER="$CXX" \
-DSIMDJSON_DEVELOPER_MODE=ON \
-DSIMDJSON_STATIC_REFLECTION=OFF \
-DBUILD_SHARED_LIBS=OFF \
../.. >/dev/null 2>&1
fi
echo " Building simdjson..."
cmake --build . --target simdjson >/dev/null 2>&1
echo " Compiling parsing test..."
# Time just the test compilation
start_time=$(date +%s.%N)
"$CXX" -std=c++17 -I../../include "$test_file" -L. -lsimdjson -o parsing_test >/dev/null 2>&1
end_time=$(date +%s.%N)
# Calculate time duration
time_taken=$(echo "$end_time $start_time" | awk '{printf "%.3f", $1 - $2}')
echo " Completed in: ${time_taken}s"
cd ..
rm -rf build_parsing_test
echo "$time_taken"
}
# Create test files
echo "Creating test files..."
create_manual_parsing_test "complex" "complex"
create_reflection_parsing_test "complex" "complex"
# Arrays to store times
times_manual=""
times_reflection=""
echo
echo "=== MANUAL PARSING COMPILATION ==="
echo "Testing traditional simdjson parsing with explicit .get() calls"
echo
for i in $(seq 1 $ITERATIONS); do
time_result=$(time_parsing_compilation "Compiling manual parsing test" "complex_manual.cpp" "false" "$i" | tail -n1)
times_manual="$times_manual $time_result"
done
echo
echo "=== REFLECTION PARSING COMPILATION ==="
echo "Testing automatic deserialization with doc.get<Struct>()"
echo
for i in $(seq 1 $ITERATIONS); do
time_result=$(time_parsing_compilation "Compiling reflection parsing test" "complex_reflection.cpp" "true" "$i" | tail -n1)
times_reflection="$times_reflection $time_result"
done
echo
echo "╔════════════════════════════════════════════════════════════════════════════╗"
echo "║ REFLECTION USAGE COMPILATION RESULTS ║"
echo "╠════════════════════════════════════════════════════════════════════════════╣"
echo "║ ║"
echo "║ MANUAL PARSING (explicit .get() calls): ║"
count=1
for t in $times_manual; do
if [[ "$t" =~ ^[0-9]+\.?[0-9]*$ ]]; then
line=$(printf "Run %2d: %7.3f seconds" "$count" "$t")
print_box_line "$line"
count=$((count + 1))
fi
done
echo "║ ║"
echo "║ REFLECTION PARSING (automatic doc.get<Struct>()): ║"
count=1
for t in $times_reflection; do
if [[ "$t" =~ ^[0-9]+\.?[0-9]*$ ]]; then
line=$(printf "Run %2d: %7.3f seconds" "$count" "$t")
print_box_line "$line"
count=$((count + 1))
fi
done
echo "╚════════════════════════════════════════════════════════════════════════════╝"
echo
echo "╔════════════════════════════════════════════════════════════════════════════╗"
echo "║ ANALYSIS SUMMARY ║"
echo "╠════════════════════════════════════════════════════════════════════════════╣"
echo "║ ║"
# Calculate averages and percentages - filter to only numeric values first
manual_numbers=""
reflection_numbers=""
for t in $times_manual; do
if [[ "$t" =~ ^[0-9]+\.?[0-9]*$ ]]; then
manual_numbers="$manual_numbers $t"
fi
done
for t in $times_reflection; do
if [[ "$t" =~ ^[0-9]+\.?[0-9]*$ ]]; then
reflection_numbers="$reflection_numbers $t"
fi
done
if [ -n "$manual_numbers" ] && [ -n "$reflection_numbers" ]; then
manual_avg=$(echo "$manual_numbers" | awk '{sum=0; for(i=1;i<=NF;i++) sum+=$i; print sum/NF}')
reflection_avg=$(echo "$reflection_numbers" | awk '{sum=0; for(i=1;i<=NF;i++) sum+=$i; print sum/NF}')
overhead=$(echo "$reflection_avg $manual_avg" | awk '{printf "%.3f", $1 - $2}')
if [ $(echo "$manual_avg > 0" | awk '{print ($1 > 0)}') -eq 1 ]; then
percent=$(echo "$reflection_avg $manual_avg" | awk '{printf "%.1f", ($1 - $2) / $2 * 100}')
else
percent="0"
fi
print_box_line "MANUAL PARSING RESULTS:"
print_box_line "$(printf "Average compilation time: %.3fs" "$manual_avg")"
print_box_line ""
print_box_line "REFLECTION PARSING RESULTS:"
print_box_line "$(printf "Average compilation time: %.3fs" "$reflection_avg")"
print_box_line ""
print_box_line "REFLECTION OVERHEAD:"
print_box_line "$(printf "Additional time: %.3fs (%+.1f%%)" "$overhead" "$percent")"
print_box_line ""
else
echo "║ ERROR: Could not extract valid timing data ║"
echo "║ Manual times: $times_manual"
echo "║ Reflection times: $times_reflection"
echo "║ ║"
fi
echo "╠════════════════════════════════════════════════════════════════════════════╣"
echo "║ INTERPRETATION ║"
echo "╠════════════════════════════════════════════════════════════════════════════╣"
echo "║ ║"
echo "║ WHAT THESE RESULTS SHOW: ║"
echo "║ ║"
echo "║ • COMPILE-TIME COST: How much longer reflection parsing takes to compile ║"
echo "║ - Higher % = more expensive template instantiation and codegen ║"
echo "║ ║"
echo "║ • CODE GENERATION OVERHEAD: Reflection creates parsing code at compile ║"
echo "║ time, which requires more template processing than manual parsing ║"
echo "║ ║"
echo "║ • DEVELOPER TRADE-OFF: Reflection provides automatic deserialization ║"
echo "║ but at the cost of increased compilation time ║"
echo "║ ║"
echo "║ EVALUATION: ║"
echo "║ • Low overhead (0-20%): Reflection is compile-time efficient ║"
echo "║ • Medium overhead (20-50%): Noticeable but potentially acceptable ║"
echo "║ • High overhead (50%+): Significant compilation cost for reflection ║"
echo "║ ║"
echo "║ REAL-WORLD IMPACT: ║"
echo "║ • Small projects: Absolute time matters more than percentage ║"
echo "║ • Large projects: Percentage overhead compounds across many files ║"
echo "║ • CI/CD pipelines: Longer builds affect development velocity ║"
echo "║ ║"
echo "╚════════════════════════════════════════════════════════════════════════════╝"
# Clean up test files
rm -f complex_manual.cpp complex_reflection.cpp
echo
echo "Completed at: $(date)"
+18 -13
View File
@@ -308,7 +308,7 @@ struct benchmarker {
return all_stages_without_allocation.iterations;
}
simdjson_really_inline void run_iteration(bool stage1_only, bool hotbuffers=false) {
simdjson_inline void run_iteration(bool stage1_only, bool hotbuffers=false) {
// Allocate dom::parser
collector.start();
dom::parser parser;
@@ -384,7 +384,7 @@ struct benchmarker {
loop << all_loop_count;
}
simdjson_really_inline void run_iterations(size_t iterations, bool stage1_only, bool hotbuffers=false) {
simdjson_inline void run_iterations(size_t iterations, bool stage1_only, bool hotbuffers=false) {
for (size_t i = 0; i<iterations; i++) {
run_iteration(stage1_only, hotbuffers);
}
@@ -423,7 +423,7 @@ struct benchmarker {
stage.instructions() / static_cast<double>(stats->structurals),
stage.instructions() / static_cast<double>(stage.cycles())
);
#if !SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
// NOTE: removed cycles/miss because it is a somewhat misleading stat
printf("%s%-13s: %7.0f branch misses (%6.2f%%) - %.0f cache misses (%6.2f%%) - %.2f cache references\n",
prefix,
@@ -434,6 +434,7 @@ struct benchmarker {
percent(stage.cache_misses(), all_stages_without_allocation.cache_misses()),
stage.cache_references()
);
#endif
}
}
@@ -444,7 +445,7 @@ struct benchmarker {
return 100.0 * a / b;
}
void print(bool tabbed_output) const {
void print(bool tabbed_output, bool stage1_only) const {
if (tabbed_output) {
char* filename_copy = reinterpret_cast<char*>(malloc(strlen(filename)+1));
SIMDJSON_PUSH_DISABLE_WARNINGS
@@ -502,24 +503,28 @@ struct benchmarker {
stats->blocks_with_16_structurals_flipped, percent(stats->blocks_with_16_structurals_flipped, stats->blocks));
}
printf("\n");
printf("All Stages (excluding allocation)\n");
print_aggregate("| " , all_stages_without_allocation.best);
// frequently, allocation is a tiny fraction of the running time so we omit it
if(allocate_stage.best.elapsed_sec() > 0.01 * all_stages_without_allocation.best.elapsed_sec()) {
printf("|- Allocation\n");
print_aggregate("| ", allocate_stage.best);
if(!stage1_only) {
printf("All Stages (excluding allocation)\n");
print_aggregate("| " , all_stages_without_allocation.best);
// frequently, allocation is a tiny fraction of the running time so we omit it
if(allocate_stage.best.elapsed_sec() > 0.01 * all_stages_without_allocation.best.elapsed_sec()) {
printf("|- Allocation\n");
print_aggregate("| ", allocate_stage.best);
}
}
printf("|- Stage 1\n");
print_aggregate("| ", stage1.best);
printf("|- Stage 2\n");
print_aggregate("| ", stage2.best);
if(!stage1_only) {
printf("|- Stage 2\n");
print_aggregate("| ", stage2.best);
}
if (collector.has_events()) {
double freq1 = (stage1.best.cycles() / stage1.best.elapsed_sec()) / 1000000000.0;
double freq2 = (stage2.best.cycles() / stage2.best.elapsed_sec()) / 1000000000.0;
double freqall = (all_stages_without_allocation.best.cycles() / all_stages_without_allocation.best.elapsed_sec()) / 1000000000.0;
double freqmin = min(freq1, freq2);
double freqmax = max(freq1, freq2);
if((freqall < 0.95 * freqmin) or (freqall > 1.05 * freqmax)) {
if((freqall < 0.95 * freqmin) || (freqall > 1.05 * freqmax)) {
printf("\nWarning: The processor frequency fluctuates in an expected way!!!\n"
"Range for stage 1 and stage 2 : [%.3f GHz, %.3f GHz], overall: %.3f GHz.\n",
freqmin, freqmax, freqall);
+14
View File
@@ -0,0 +1,14 @@
# Executable
add_executable(benchmark_car_builder benchmark_car_builder.cpp)
# Compile for C++20.
target_compile_features(benchmark_car_builder PRIVATE cxx_std_20)
# Check if -march=native is supported
include(CheckCXXCompilerFlag)
check_cxx_compiler_flag("-march=native" SIMDJSON_SUPPORTS_MARCH_NATIVE)
if(SIMDJSON_SUPPORTS_MARCH_NATIVE)
target_compile_options(benchmark_car_builder PRIVATE -march=native)
endif()
target_include_directories(benchmark_car_builder PRIVATE ${CMAKE_CURRENT_LIST_DIR}/..)
@@ -0,0 +1,127 @@
#include "event_counter.h"
#include <random>
#include <vector>
#include <simdjson.h>
event_collector collector;
struct Car {
std::string make;
std::string model;
int64_t year; // We deliberately do not include the tire pressure.
};
std::vector<Car> generate_random_cars(size_t count) {
static const std::vector<std::string> makes = {"Toyota", "Honda", "Ford",
"BMW", "Mazda"};
static const std::vector<std::string> models = {"Camry", "Civic", "Focus",
"320i", "3"};
static thread_local std::mt19937 rng{std::random_device{}()};
std::uniform_int_distribution<int> make_dist(0, makes.size() - 1);
std::uniform_int_distribution<int> model_dist(0, models.size() - 1);
std::uniform_int_distribution<int64_t> year_dist(2000, 2025);
std::uniform_real_distribution<double> pressure_dist(30.0, 45.0);
std::vector<Car> cars;
cars.reserve(count);
for (size_t i = 0; i < count; ++i) {
Car car;
car.make = makes[make_dist(rng)];
car.model = models[model_dist(rng)];
car.year = year_dist(rng);
cars.push_back(std::move(car));
}
return cars;
}
std::string_view serialize(simdjson::builder::string_builder &sb,
const std::vector<Car> &cars) {
sb.clear();
sb.start_array();
for (const auto &car : cars) {
sb.start_object();
sb.append_key_value("make", car.make);
sb.append_comma();
sb.append_key_value("model", car.model);
sb.append_comma();
sb.append_key_value("year", car.year);
sb.end_object();
}
sb.end_array();
std::string_view result;
if (sb.view().get(result)) {
return ""; // unexpected (error)
}
return result;
}
double pretty_print(const std::string &name, size_t num_chars,
std::pair<event_aggregate, size_t> result) {
const auto &agg = result.first;
size_t N = result.second;
num_chars *= N;
printf("%-40s : %8.2f ns %8.2f GB/s", name.c_str(),
agg.elapsed_ns() / num_chars, num_chars / agg.elapsed_ns());
if (collector.has_events()) {
printf(" %8.2f GHz %8.2f cycles/char %8.2f ins./char %8.2f i/c",
agg.cycles() / agg.elapsed_ns(), agg.cycles() / num_chars,
agg.instructions() / num_chars, agg.instructions() / agg.cycles());
}
printf("\n");
return num_chars / agg.elapsed_ns();
}
template <class function_type>
std::pair<event_aggregate, size_t>
bench(const function_type &&function, size_t min_repeat = 100,
size_t min_time_ns = 40'000'000, size_t max_repeat = 10000000) {
size_t N = min_repeat;
if (N == 0) {
N = 1;
}
event_aggregate warm_aggregate{};
for (size_t i = 0; i < N; i++) {
std::atomic_thread_fence(std::memory_order_acquire);
collector.start();
function();
std::atomic_thread_fence(std::memory_order_release);
event_count allocate_count = collector.end();
warm_aggregate << allocate_count;
if ((i + 1 == N) && (warm_aggregate.total_elapsed_ns() < min_time_ns) &&
(N < max_repeat)) {
N *= 10;
}
}
event_aggregate aggregate{};
for (size_t i = 0; i < 10; i++) {
std::atomic_thread_fence(std::memory_order_acquire);
collector.start();
for (size_t i = 0; i < N; i++) {
function();
}
std::atomic_thread_fence(std::memory_order_release);
event_count allocate_count = collector.end();
aggregate << allocate_count;
}
return {aggregate, N};
}
void run_benchmarks() {
std::vector<Car> source = generate_random_cars(100000);
simdjson::builder::string_builder sb;
size_t volume = serialize(sb, source).size();
pretty_print("string_builder", volume, bench([&source, &sb]() -> size_t {
return serialize(sb, source).size();
}));
}
int main() {
for (size_t trial = 0; trial < 3; trial++) {
printf("Trial %zu:\n", trial + 1);
run_benchmarks();
printf("\n");
}
return EXIT_SUCCESS;
}
+29
View File
@@ -0,0 +1,29 @@
#pragma once
#ifdef SIMDJSON_COMPETITION_BOOSTJSON
#include "distinct_user_id.h"
namespace distinct_user_id {
struct boostjson {
bool run(simdjson::padded_string &json, std::vector<uint64_t> &result) {
auto root = boost::json::parse(json);
for (const auto &tweet : root.at("statuses").as_array()) {
result.push_back(tweet.at("user").at("id").to_number<uint64_t>());
if (tweet.as_object().if_contains("retweeted_status")) {
result.push_back(tweet.at("retweeted_status").at("user").at("id").to_number<uint64_t>());
}
}
return true;
}
};
BENCHMARK_TEMPLATE(distinct_user_id, boostjson)->UseManualTime();
} // namespace distinct_user_id
#endif // SIMDJSON_COMPETITION_BOOSTJSON
@@ -46,7 +46,7 @@ struct runner : public file_runner<I> {
struct simdjson_dom;
template<typename I> simdjson_really_inline static void distinct_user_id(benchmark::State &state) {
template<typename I> simdjson_inline static void distinct_user_id(benchmark::State &state) {
run_json_benchmark<runner<I>, runner<simdjson_dom>>(state);
}
+3 -2
View File
@@ -46,13 +46,14 @@ struct rapidjson : rapidjson_base {
};
BENCHMARK_TEMPLATE(distinct_user_id, rapidjson)->UseManualTime();
#if SIMDJSON_COMPETITION_ONDEMAND_INSITU
struct rapidjson_insitu : rapidjson_base {
bool run(simdjson::padded_string &json, std::vector<uint64_t> &result) {
return rapidjson_base::run(doc.ParseInsitu<kParseValidateEncodingFlag>(json.data()), result);
return rapidjson_base::run(doc.ParseInsitu<kParseValidateEncodingFlag|kParseInsituFlag>(json.data()), result);
}
};
BENCHMARK_TEMPLATE(distinct_user_id, rapidjson_insitu)->UseManualTime();
#endif // SIMDJSON_COMPETITION_ONDEMAND_INSITU
} // namespace partial_tweets
#endif // SIMDJSON_COMPETITION_RAPIDJSON
+4 -2
View File
@@ -9,12 +9,14 @@ namespace distinct_user_id {
struct sajson {
size_t ast_buffer_size{0};
size_t *ast_buffer{nullptr};
simdjson_really_inline std::string_view get_string_view(const ::sajson::value &obj, std::string_view key) {
~sajson() { free(ast_buffer); }
simdjson_inline std::string_view get_string_view(const ::sajson::value &obj, std::string_view key) {
auto val = obj.get_value_of_key({key.data(), key.length()});
if (val.get_type() != ::sajson::TYPE_STRING) { throw "field is not a string"; }
return { val.as_cstring(), val.get_string_length() };
}
simdjson_really_inline uint64_t get_str_uint64(const ::sajson::value &obj, std::string_view key) {
simdjson_inline uint64_t get_str_uint64(const ::sajson::value &obj, std::string_view key) {
// Since sajson only supports 53-bit numbers, and IDs in twitter.json can be > 53 bits, we read the corresponding id_str and parse that.
auto val = obj.get_value_of_key({key.data(), key.length()});
if (val.get_type() != ::sajson::TYPE_STRING) { throw "field not a string"; }
+10 -2
View File
@@ -44,17 +44,25 @@ struct yyjson_base {
struct yyjson : yyjson_base {
bool run(simdjson::padded_string &json, std::vector<uint64_t> &result) {
return yyjson_base::run(yyjson_read(json.data(), json.size(), 0), result);
yyjson_doc *doc = yyjson_read(json.data(), json.size(), 0);
bool b = yyjson_base::run(doc, result);
yyjson_doc_free(doc);
return b;
}
};
BENCHMARK_TEMPLATE(distinct_user_id, yyjson)->UseManualTime();
#if SIMDJSON_COMPETITION_ONDEMAND_INSITU
struct yyjson_insitu : yyjson_base {
bool run(simdjson::padded_string &json, std::vector<uint64_t> &result) {
return yyjson_base::run(yyjson_read_opts(json.data(), json.size(), YYJSON_READ_INSITU, 0, 0), result);
yyjson_doc *doc = yyjson_read_opts(json.data(), json.size(), YYJSON_READ_INSITU, 0, 0);
bool b = yyjson_base::run(doc, result);
yyjson_doc_free(doc);
return b;
}
};
BENCHMARK_TEMPLATE(distinct_user_id, yyjson_insitu)->UseManualTime();
#endif // SIMDJSON_COMPETITION_ONDEMAND_INSITU
} // namespace distinct_user_id
+1 -1
View File
@@ -218,7 +218,7 @@ int main(int argc, char *argv[]) {
if (!options.verbose) { progress.erase(); }
for (size_t i=0; i<options.files.size(); i++) {
benchmarkers[i]->print(options.tabbed_output);
benchmarkers[i]->print(options.tabbed_output, options.stage1_only);
delete benchmarkers[i];
}
+2 -3
View File
@@ -25,12 +25,11 @@ int main(int argc, char *argv[]) {
exit(1);
}
const char *filename = argv[1];
auto v = simdjson::padded_string::load(filename);
if (v.error()) {
simdjson::padded_string p;
if (simdjson::padded_string::load(filename).get(p)) {
std::cerr << "Could not load the file " << filename << std::endl;
return EXIT_FAILURE;
}
const simdjson::padded_string& p = v.value_unsafe();
if (test_baseline) {
std::wclog << "Baseline: Getline + normal parse... " << std::endl;
std::cout << "Gigabytes/second\t"
+1 -1
View File
@@ -41,7 +41,7 @@ using stat_t = struct stat_s;
simdjson_really_inline void simdjson_process_atom(stat_t &s,
simdjson_inline void simdjson_process_atom(stat_t &s,
simdjson::dom::element element) {
if (element.is<int64_t>()) {
s.integer_count++;
+57 -8
View File
@@ -1,6 +1,15 @@
#ifndef __EVENT_COUNTER_H
#define __EVENT_COUNTER_H
#ifndef SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
#ifdef __aarch64__
// on ARM, we use just cycles and instructions
#define SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS 1
#else
// elsewhere, we try to use four counters.
#define SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS 0
#endif
#endif
#include <cassert>
#include <cctype>
#ifndef _MSC_VER
@@ -25,11 +34,15 @@
#include <string>
#include <vector>
#include "linux-perf-events.h"
#ifdef __linux__
#include "linux-perf-events.h"
#include <libgen.h>
#endif
#if __APPLE__ && __aarch64__
#include "apple/apple_arm_events.h"
#endif
#include "simdjson.h"
using std::string;
@@ -46,6 +59,12 @@ struct event_count {
event_count(const event_count& other): elapsed(other.elapsed), event_counts(other.event_counts) { }
// The types of counters (so we can read the getter more easily)
#if SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
enum event_counter_types {
CPU_CYCLES,
INSTRUCTIONS
};
#else
enum event_counter_types {
CPU_CYCLES,
INSTRUCTIONS,
@@ -53,15 +72,16 @@ struct event_count {
CACHE_REFERENCES,
CACHE_MISSES
};
#endif
double elapsed_sec() const { return duration<double>(elapsed).count(); }
double elapsed_ns() const { return duration<double, std::nano>(elapsed).count(); }
double cycles() const { return static_cast<double>(event_counts[CPU_CYCLES]); }
double instructions() const { return static_cast<double>(event_counts[INSTRUCTIONS]); }
#if !SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
double branch_misses() const { return static_cast<double>(event_counts[BRANCH_MISSES]); }
double cache_references() const { return static_cast<double>(event_counts[CACHE_REFERENCES]); }
double cache_misses() const { return static_cast<double>(event_counts[CACHE_MISSES]); }
#endif
event_count& operator=(const event_count& other) {
this->elapsed = other.elapsed;
this->event_counts = other.event_counts;
@@ -102,12 +122,15 @@ struct event_aggregate {
}
double elapsed_sec() const { return total.elapsed_sec() / iterations; }
double total_elapsed_ns() const { return total.elapsed_ns(); }
double elapsed_ns() const { return total.elapsed_ns() / iterations; }
double cycles() const { return total.cycles() / iterations; }
double instructions() const { return total.instructions() / iterations; }
#if !SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
double branch_misses() const { return total.branch_misses() / iterations; }
double cache_references() const { return total.cache_references() / iterations; }
double cache_misses() const { return total.cache_misses() / iterations; }
#endif
};
struct event_collector {
@@ -116,33 +139,59 @@ struct event_collector {
#if defined(__linux__)
LinuxEvents<PERF_TYPE_HARDWARE> linux_events;
event_collector(bool quiet = false) : linux_events(vector<int>{
event_collector() : linux_events(vector<int>{
#if SIMDJSON_SIMPLE_PERFORMANCE_COUNTERS
PERF_COUNT_HW_CPU_CYCLES,
PERF_COUNT_HW_INSTRUCTIONS,
#else
PERF_COUNT_HW_CPU_CYCLES,
PERF_COUNT_HW_INSTRUCTIONS,
PERF_COUNT_HW_BRANCH_MISSES,
PERF_COUNT_HW_CACHE_REFERENCES,
PERF_COUNT_HW_CACHE_MISSES
}, quiet) {}
#endif
}) {}
bool has_events() {
return linux_events.is_working();
}
#elif __APPLE__ && __aarch64__
AppleEvents apple_events;
performance_counters diff;
event_collector() : diff(0) {
apple_events.setup_performance_counters();
}
bool has_events() {
return apple_events.setup_performance_counters();
}
#else
event_collector(simdjson_unused bool _quiet = false) {}
event_collector() {}
bool has_events() {
return false;
}
#endif
simdjson_really_inline void start() {
simdjson_inline void start() {
#if defined(__linux)
linux_events.start();
#elif __APPLE__ && __aarch64__
if(has_events()) { diff = apple_events.get_counters(); }
#endif
start_clock = steady_clock::now();
}
simdjson_really_inline event_count& end() {
simdjson_inline event_count& end() {
time_point<steady_clock> end_clock = steady_clock::now();
#if defined(__linux)
linux_events.end(count.event_counts);
#elif __APPLE__ && __aarch64__
if(has_events()) {
performance_counters end = apple_events.get_counters();
diff = end - diff;
}
count.event_counts[0] = diff.cycles;
count.event_counts[1] = diff.instructions;
count.event_counts[2] = diff.missed_branches;
count.event_counts[3] = 0;
count.event_counts[4] = 0;
#endif
count.elapsed = end_clock - start_clock;
return count;
+30
View File
@@ -0,0 +1,30 @@
#pragma once
#ifdef SIMDJSON_COMPETITION_BOOSTJSON
#include "find_tweet.h"
namespace find_tweet {
struct boostjson {
using StringType=std::string;
bool run(simdjson::padded_string &json, uint64_t find_id, std::string &result) {
auto root = boost::json::parse(json);
for (const auto &tweet : root.at("statuses").as_array()) {
if (tweet.at("id") == find_id) {
result = tweet.at("text").as_string();
return true;
}
}
return false;
}
};
BENCHMARK_TEMPLATE(find_tweet, boostjson)->UseManualTime();
} // namespace find_tweet
#endif // SIMDJSON_COMPETITION_BOOSTJSON
+1 -1
View File
@@ -33,7 +33,7 @@ struct runner : public file_runner<I> {
struct simdjson_dom;
template<typename I> simdjson_really_inline static void find_tweet(benchmark::State &state) {
template<typename I> simdjson_inline static void find_tweet(benchmark::State &state) {
run_json_benchmark<runner<I>, runner<simdjson_dom>>(state);
}
+1 -1
View File
@@ -13,7 +13,7 @@ struct nlohmann_json {
auto root = nlohmann::json::parse(json.data(), json.data() + json.size());
for (auto tweet : root["statuses"]) {
if (tweet["id"] == find_id) {
result = tweet["text"];
result = to_string(tweet["text"]);
return true;
}
}
+3 -2
View File
@@ -40,13 +40,14 @@ struct rapidjson : rapidjson_base {
};
BENCHMARK_TEMPLATE(find_tweet, rapidjson)->UseManualTime();
#if SIMDJSON_COMPETITION_ONDEMAND_INSITU
struct rapidjson_insitu : rapidjson_base {
bool run(simdjson::padded_string &json, uint64_t find_id, std::string_view &result) {
return rapidjson_base::run(doc.ParseInsitu<kParseValidateEncodingFlag>(json.data()), find_id, result);
return rapidjson_base::run(doc.ParseInsitu<kParseValidateEncodingFlag|kParseInsituFlag>(json.data()), find_id, result);
}
};
BENCHMARK_TEMPLATE(find_tweet, rapidjson_insitu)->UseManualTime();
#endif // SIMDJSON_COMPETITION_ONDEMAND_INSITU
} // namespace find_tweet
#endif // SIMDJSON_COMPETITION_RAPIDJSON
+4 -2
View File
@@ -11,12 +11,14 @@ struct sajson {
size_t ast_buffer_size{0};
size_t *ast_buffer{nullptr};
simdjson_really_inline std::string_view get_string_view(const ::sajson::value &obj, std::string_view key) {
~sajson() { free(ast_buffer); }
simdjson_inline std::string_view get_string_view(const ::sajson::value &obj, std::string_view key) {
auto val = obj.get_value_of_key({key.data(), key.length()});
if (val.get_type() != ::sajson::TYPE_STRING) { throw "field is not a string"; }
return { val.as_cstring(), val.get_string_length() };
}
simdjson_really_inline uint64_t get_str_uint64(const ::sajson::value &obj, std::string_view key) {
simdjson_inline uint64_t get_str_uint64(const ::sajson::value &obj, std::string_view key) {
// Since sajson only supports 53-bit numbers, and IDs in twitter.json can be > 53 bits, we read the corresponding id_str and parse that.
auto val = obj.get_value_of_key({key.data(), key.length()});
if (val.get_type() != ::sajson::TYPE_STRING) { throw "field not a string"; }
+10 -2
View File
@@ -36,17 +36,25 @@ struct yyjson_base {
struct yyjson : yyjson_base {
bool run(simdjson::padded_string &json, uint64_t find_id, std::string_view &result) {
return yyjson_base::run(yyjson_read(json.data(), json.size(), 0), find_id, result);
yyjson_doc *doc = yyjson_read(json.data(), json.size(), 0);
bool b = yyjson_base::run(doc, find_id, result);
yyjson_doc_free(doc);
return b;
}
};
BENCHMARK_TEMPLATE(find_tweet, yyjson)->UseManualTime();
#if SIMDJSON_COMPETITION_ONDEMAND_INSITU
struct yyjson_insitu : yyjson_base {
bool run(simdjson::padded_string &json, uint64_t find_id, std::string_view &result) {
return yyjson_base::run(yyjson_read_opts(json.data(), json.size(), YYJSON_READ_INSITU, 0, 0), find_id, result);
yyjson_doc *doc = yyjson_read_opts(json.data(), json.size(), YYJSON_READ_INSITU, 0, 0);
bool b = yyjson_base::run(doc, find_id, result);
yyjson_doc_free(doc);
return b;
}
};
BENCHMARK_TEMPLATE(find_tweet, yyjson_insitu)->UseManualTime();
#endif // SIMDJSON_COMPETITION_ONDEMAND_INSITU
} // namespace find_tweet
+15
View File
@@ -0,0 +1,15 @@
# Executable
add_executable(from_benchmark from_benchmark.cpp)
# Compile for C++20.
if(CMAKE_CXX_STANDARD LESS 20)
target_compile_features(from_benchmark PRIVATE cxx_std_20)
endif()
# Check if -march=native is supported
include(CheckCXXCompilerFlag)
check_cxx_compiler_flag("-march=native" SIMDJSON_SUPPORTS_MARCH_NATIVE)
if(SIMDJSON_SUPPORTS_MARCH_NATIVE)
target_compile_options(from_benchmark PRIVATE -march=native)
endif()
target_include_directories(from_benchmark PRIVATE ${CMAKE_CURRENT_LIST_DIR}/..)
+77
View File
@@ -0,0 +1,77 @@
#ifndef BENCHMARK_HELPERS_H
#define BENCHMARK_HELPERS_H
#include "event_counter.h"
#include <atomic>
event_collector collector;
template <class function_type>
std::pair<event_aggregate, size_t>
bench(const function_type &&function, size_t min_repeat = 10,
size_t min_time_ns = 40'000'000, size_t max_repeat = 10000000) {
size_t N = min_repeat;
if (N == 0) {
N = 1;
}
event_aggregate warm_aggregate{};
for (size_t i = 0; i < N; i++) {
std::atomic_thread_fence(std::memory_order_acquire);
collector.start();
function();
std::atomic_thread_fence(std::memory_order_release);
event_count allocate_count = collector.end();
warm_aggregate << allocate_count;
if ((i + 1 == N) && (warm_aggregate.total_elapsed_ns() < min_time_ns) &&
(N < max_repeat)) {
N *= 10;
}
}
event_aggregate aggregate{};
for (size_t i = 0; i < 10; i++) {
std::atomic_thread_fence(std::memory_order_acquire);
collector.start();
for (size_t i = 0; i < N; i++) {
function();
}
std::atomic_thread_fence(std::memory_order_release);
event_count allocate_count = collector.end();
aggregate << allocate_count;
}
return {aggregate, N};
}
double pretty_print(const std::string &name, size_t num_chars,
std::pair<event_aggregate, size_t> result) {
const auto &agg = result.first;
size_t N = result.second;
num_chars *= N;
printf("%-40s : %8.2f ns %8.2f GB/s", name.c_str(),
agg.elapsed_ns() / num_chars, num_chars / agg.elapsed_ns());
if (collector.has_events()) {
printf(" %8.2f GHz %8.2f cycles/char %8.2f ins./char %8.2f i/c",
agg.cycles() / agg.elapsed_ns(), agg.cycles() / num_chars,
agg.instructions() / num_chars, agg.instructions() / agg.cycles());
}
printf("\n");
return num_chars / agg.elapsed_ns();
}
double pretty_print_array(const std::string &name, size_t num_chars, size_t num_elements,
std::pair<event_aggregate, size_t> result) {
const auto &agg = result.first;
size_t N = result.second;
num_chars *= N;
printf("%-40s : %8.2f ns/char %8.2f ns/value %8.2f GB/s", name.c_str(),
agg.elapsed_ns() / num_chars, agg.elapsed_ns() / num_elements, num_chars / agg.elapsed_ns());
if (collector.has_events()) {
printf(" %8.2f GHz %8.2f cycles/char %8.2f ins./char %8.2f ins./value %8.2f i/c",
agg.cycles() / agg.elapsed_ns(), agg.cycles() / num_chars,
agg.instructions() / num_chars, agg.instructions() / num_elements, agg.instructions() / agg.cycles());
}
printf("\n");
return num_chars / agg.elapsed_ns();
}
#endif
+146
View File
@@ -0,0 +1,146 @@
#ifndef CAR_HELPERS_H
#define CAR_HELPERS_H
#include <iomanip>
#include <random>
#include <sstream>
#include <vector>
#include <simdjson.h>
struct Car {
std::string make;
std::string model;
int64_t year; // We deliberately do not include the tire pressure.
};
using namespace simdjson;
std::string random_car_json() {
static const std::vector<std::string> makes = {"Toyota", "Honda", "Ford",
"BMW", "Mazda"};
static const std::vector<std::string> models = {"Camry", "Civic", "Focus",
"320i", "3"};
static thread_local std::mt19937 rng{std::random_device{}()};
std::uniform_int_distribution<int> make_dist(0, makes.size() - 1);
std::uniform_int_distribution<int> model_dist(0, models.size() - 1);
std::uniform_int_distribution<int> year_dist(2000, 2025);
std::uniform_real_distribution<double> pressure_dist(30.0, 45.0);
std::ostringstream oss;
oss << R"({ "make": ")" << makes[make_dist(rng)] << R"(", "model": ")"
<< models[model_dist(rng)] << R"(", "year": )" << year_dist(rng)
<< R"(, "tire_pressure": [ )" << std::fixed << std::setprecision(1)
<< pressure_dist(rng) << ", " << pressure_dist(rng) << R"( ] })";
return oss.str();
}
std::string generate_car_json_stream(size_t N) {
std::ostringstream oss;
for (size_t i = 0; i < N; ++i) {
oss << random_car_json();
oss << "\n";
}
return oss.str();
}
std::string generate_car_json_array(size_t N) {
std::ostringstream oss;
oss << "[\n";
for (size_t i = 0; i < N; ++i) {
oss << random_car_json();
if (i < N - 1) {
oss << ",";
}
oss << "\n";
}
oss << "]";
return oss.str();
}
// We want to maximize C++ portability, but this is not needed with static
// reflection (C++26).
template <>
simdjson_inline simdjson_result<Car>
simdjson::ondemand::document::get() & noexcept {
ondemand::object obj;
auto error = get_object().get(obj);
if (error) {
return error;
}
Car car;
if ((error = obj["make"].get_string(car.make))) {
return error;
}
if ((error = obj["model"].get_string(car.model))) {
return error;
}
if ((error = obj["year"].get_int64().get(car.year))) {
return error;
}
return car;
}
template <>
simdjson_inline simdjson_result<Car> simdjson::ondemand::value::get() noexcept {
ondemand::object obj;
auto error = get_object().get(obj);
if (error) {
return error;
}
Car car;
if ((error = obj["make"].get_string(car.make))) {
return error;
}
if ((error = obj["model"].get_string(car.model))) {
return error;
}
if ((error = obj["year"].get_int64().get(car.year))) {
return error;
}
return car;
}
template <>
simdjson_inline simdjson_result<Car>
simdjson::ondemand::document_reference::get() & noexcept {
ondemand::object obj;
auto error = get_object().get(obj);
if (error) {
return error;
}
Car car;
if ((error = obj["make"].get_string(car.make))) {
return error;
}
if ((error = obj["model"].get_string(car.model))) {
return error;
}
if ((error = obj["year"].get_int64().get(car.year))) {
return error;
}
return car;
}
template <>
simdjson_inline simdjson_result<Car>
simdjson::ondemand::document_reference::get() && noexcept {
ondemand::object obj;
auto error = get_object().get(obj);
if (error) {
return error;
}
Car car;
if ((error = obj["make"].get_string(car.make))) {
return error;
}
if ((error = obj["model"].get_string(car.model))) {
return error;
}
if ((error = obj["year"].get_int64().get(car.year))) {
return error;
}
return car;
}
#endif
+132
View File
@@ -0,0 +1,132 @@
// We are going to assume C++20.
#include <cstdint>
#include <cstdio>
#include <cstdlib>
#include <simdjson.h>
#include <string>
#include "benchmark_helpers.h"
#include "car_helpers.h"
using namespace simdjson;
void run_benchmarks() {
printf("Running benchmarks on tiny input...\n");
simdjson::padded_string json =
R"({ "make": "Toyota", "model": "Camry", "year": 2018, "tire_pressure": [ 40.1, 39.9 ] })"_padded;
volatile size_t dummy = 0;
ondemand::parser parser;
// Classic On-Demand
pretty_print("simdjson classic", json.size(),
bench([&json, &dummy, &parser]() {
ondemand::document doc = parser.iterate(json);
Car car = doc.get<Car>();
dummy = dummy + car.year;
}));
// Benchmark simdjson::from() without parser
pretty_print("simdjson::from<Car>() (no parser)", json.size(),
bench([&json, &dummy]() {
Car car = simdjson::from(json);
dummy = dummy + car.year;
}));
// Benchmark simdjson::from() with parser
pretty_print("simdjson::from<Car>() (with parser)", json.size(),
bench([&json, &dummy, &parser]() {
Car car = simdjson::from(parser, json);
dummy = dummy + car.year;
}));
}
void run_array_benchmarks() {
printf("Running benchmarks on large array...\n");
size_t N = 1'000'000;
simdjson::padded_string json = generate_car_json_array(N);
volatile size_t dummy = 0;
ondemand::parser parser;
// Classic On-Demand
pretty_print_array("simdjson classic", json.size(), N,
bench([&json, &dummy, &parser]() {
dummy = 0;
ondemand::document doc = parser.iterate(json);
ondemand::array array = doc.get_array();
for (auto value : array) {
Car car = value.get<Car>();
dummy = dummy + car.year;
}
}));
// from with array
pretty_print_array("simdjson::from(json).array()", json.size(), N,
bench([&json, &dummy, &parser]() {
dummy = 0;
for (auto value : simdjson::from(json).array()) {
Car car = value.get<Car>();
dummy = dummy + car.year;
}
}));
#if SIMDJSON_SUPPORTS_RANGES_FROM
// Benchmark simdjson::from() without parser (EXPERIMENTAL)
pretty_print_array("simdjson::from<Car>() (experimental, no parser)",
json.size(), N, bench([&json, &dummy]() {
dummy = 0;
for (Car car :
simdjson::from(json) | simdjson::as<Car>()) {
dummy = dummy + car.year;
}
}));
// Benchmark simdjson::from() with parser (EXPERIMENTAL)
pretty_print_array("simdjson::from<Car>() (experimental, with parser)",
json.size(), N, bench([&json, &dummy, &parser]() {
dummy = 0;
for (Car car : simdjson::from(parser, json) |
simdjson::as<Car>()) {
dummy = dummy + car.year;
}
}));
#endif // SIMDJSON_SUPPORTS_RANGES_FROM
}
void run_stream_benchmarks() {
printf("Running stream benchmarks...\n");
simdjson::padded_string json = generate_car_json_stream(1'000'000);
volatile size_t dummy = 0;
ondemand::parser parser;
// Classic On-Demand
pretty_print("simdjson classic", json.size(),
bench([&json, &dummy, &parser]() {
ondemand::document_stream stream = parser.iterate_many(json);
for (auto doc : stream) {
Car car = doc.get<Car>();
dummy = dummy + car.year;
}
}));
// from_many
pretty_print("simdjson with thread local", json.size(),
bench([&json, &dummy, &parser]() {
ondemand::document_stream stream =
ondemand::parser::get_parser().iterate_many(json);
for (auto doc : stream) {
Car car = doc.get<Car>();
dummy = dummy + car.year;
}
}));
}
int main() {
for (size_t trial = 0; trial < 3; trial++) {
printf("Trial %zu:\n", trial + 1);
run_array_benchmarks();
run_stream_benchmarks();
run_benchmarks();
printf("\n");
}
return EXIT_SUCCESS;
}
+5 -1
View File
@@ -9,7 +9,11 @@ simdjson_never_inline
double bench(std::string filename, simdjson::padded_string& p) {
std::chrono::time_point<std::chrono::steady_clock> start_clock =
std::chrono::steady_clock::now();
simdjson::padded_string::load(filename).value_unsafe().swap(p);
auto error = simdjson::padded_string::load(filename).get(p);
if(error) {
std::cerr << simdjson::error_message(error) << std::endl;
std::abort();
}
std::chrono::time_point<std::chrono::steady_clock> end_clock =
std::chrono::steady_clock::now();
std::chrono::duration<double> elapsed = end_clock - start_clock;
+104
View File
@@ -0,0 +1,104 @@
#pragma once
#ifdef SIMDJSON_COMPETITION_BOOSTJSON
#include "json2msgpack.h"
namespace json2msgpack {
struct boostjson2msgpack {
inline std::string_view to_msgpack(const boost::json::value &root, uint8_t *buf) {
buff = buf;
recursive_processor(root);
return std::string_view(reinterpret_cast<char *>(buf), size_t(buff - buf));
}
private:
uint8_t *buff{};
inline void write_double(const double d) noexcept {
*buff++ = 0xcb;
::memcpy(buff, &d, sizeof(d));
buff += sizeof(d);
}
inline void write_byte(const uint8_t b) noexcept {
*buff = b;
buff++;
}
inline void write_uint32(const uint32_t w) noexcept {
::memcpy(buff, &w, sizeof(w));
buff += sizeof(w);
}
inline void write_string(const std::string & str) {
write_byte(0xdb);
write_uint32(uint32_t(str.size()));
::memcpy(buff, str.data(), str.size());
buff += str.size();
}
inline void recursive_processor(const boost::json::value &element) {
switch(element.kind()) {
case boost::json::kind::array: {
write_byte(0xdd);
const auto &array = element.as_array();
write_uint32(static_cast<uint32_t>(array.size()));
for (const auto &child : array) {
recursive_processor(child);
}
} break;
case boost::json::kind::object: {
write_byte(0xdf);
const auto &object = element.as_object();
write_uint32(static_cast<uint32_t>(object.size()));
for (const auto &child : object) {
write_string(child.key_c_str());
recursive_processor(child.value());
}
} break;
case boost::json::kind::int64:
case boost::json::kind::uint64:
case boost::json::kind::double_:
write_double(element.to_number<double>());
break;
case boost::json::kind::string:
write_string(element.as_string().c_str());
break;
case boost::json::kind::bool_:
write_byte(0xc2 + element.as_bool());
break;
case boost::json::kind::null:
write_byte(0xc0);
break;
default:
printf("unexpected\n");
break;
}
}
};
struct boostjson {
using StringType=std::string;
boostjson2msgpack parser{};
bool run(simdjson::padded_string &json, char *buffer, std::string_view &result) {
auto root = boost::json::parse(json);
result = parser.to_msgpack(root, reinterpret_cast<uint8_t *>(buffer));
return true;
}
};
BENCHMARK_TEMPLATE(json2msgpack, boostjson)->UseManualTime();
} // namespace json2msgpack
#endif // SIMDJSON_COMPETITION_BOOSTJSON
+48
View File
@@ -0,0 +1,48 @@
#pragma once
#include "json_benchmark/file_runner.h"
namespace json2msgpack {
using namespace json_benchmark;
template <typename I> struct runner : public file_runner<I> {
std::string_view result;
std::unique_ptr<char[]> buffer;
bool setup(benchmark::State &state) {
bool isok = this->load_json(state, TWITTER_JSON);
if (isok) {
// Let us allocate a sizeable buffer.
buffer = std::unique_ptr<char[]>(new char[this->json.size() * 4 + 1024]);
}
return isok;
}
bool before_run(benchmark::State &state) {
if (!file_runner<I>::before_run(state)) {
return false;
}
// Clear the buffer.
::memset(buffer.get(), 0, this->json.size() * 4 + 1024);
return true;
}
bool run(benchmark::State &) {
return this->implementation.run(this->json, buffer.get(), result);
}
template <typename R>
bool diff(benchmark::State &state, runner<R> &reference) {
return diff_results(state, result.size(), reference.result.size(), diff_flags::NONE);
}
};
struct simdjson_ondemand;
template <typename I>
simdjson_inline static void json2msgpack(benchmark::State &state) {
run_json_benchmark<runner<I>, runner<simdjson_ondemand>>(state);
}
} // namespace json2msgpack

Some files were not shown because too many files have changed in this diff Show More