* Update basic.md to document JSON pointer for On Demand.
* Add automatic rewind for at_pointer
* Remove DOM examples in basics.md and update documentation reflecting addition of at_pointer automatic rewinding.
* Review
* Add test
* Add document_stream constructors and iterate_many
* Attempt to implement streaming.
* Kind of fixed next() for getting next document
* Temporary save.
* Putting in working order.
* Add working doc_index and add function next_document()
* Attempt to implement streaming.
* Re-anchoring json_iterator after a call to stage 1
* I am convinced it should be a 'while'.
* Add source() with test.
* Add truncated_bytes().
* Fix casting issues.
* Fix old style cast.
* Fix privacy issue.
* Fix privacy issues.
* Again
* .
* Add more tests. Add error() for iterator class.
* Fix source() to not included whitespaces between documents.
* Fixing CI.
* Fix source() for multiple batches. Add new tests.
* Fix batch_start when document has leading spaces. Add new tests for that.
* Add new tests.
* Temporary save.
* Working hacky multithread version.
* Small fix in header files.
* Correct version (not working).
* Adding a move assignment to ondemand::parser.
* Fix attempt by changing std::swap.
* Moving DEFAULT_BATCH_SIZE and MINIMAL_BATCH_SIZE.
* Update doc and readme tests.
* Update basics.md
* Update readme_examples tests.
* Fix exceptions in test.
* Partial setup for amazon_cellphones.
* Benchmark with vectors.
* Benchmark with maps
* With vectors again.
* Fix for weighted average.
* DOM benchmark.
* Fix typos. Add On Demand benchmark.
* Add large amazon_cellphones benchmark for DOM
* Add benchmark for On demand.
* Fix broken read_me test.
* Add parser.threaded to enable/disable thread usage.
Co-authored-by: Daniel Lemire <lemire@gmail.com>
* upload corpus to https://www.pauldreik.se/ from the x64 github action job (keep the github action cache)
* drop the github action cache (which was not working anyway) for power fuzzer and download the fuzz corpus from https://www.pauldreik.se/ instead
* resurrect arm64 fuzzing on drone CI, downloading the fuzz corpus from https://www.pauldreik.se/
* update the fuzzing documentation
* Adds compile-test for Visual Studio + ARM and turn developer mode throughout CI.
* Correcting YAML error.
* Disabling google benchmarks under Windows ARM.
* Turning off exceptions under ARM.
* Remove CMP0025 policy
This policy is already set to NEW by the minimum required version.
* Use HOMEPAGE_URL in the project call
* Use VERSION in the project call
* Detect if this is the top project
* Port simdjson-user-cmakecache to a CMake script
* Create a developer mode
The SIMDJSON_DEVELOPER_MODE option set to ON will enable targets that
are only useful for developers of simdjson.
* Consolidate root CML commands into logical sections
* Warn about intended use of developer mode
* Prettify the just_ascii test
* Remove redundant CMake variables
* Inline CML contents from include and src
* Raise minimum CMake requirement to 3.14
* Define proper install rules
* Restore thread support variable
* Add BUILD_SHARED_LIBS as a top level only option
* Force developer mode to be on in CI
* Include flags earlier in developer mode
* Set CMAKE_BUILD_TYPE conditionally
CMAKE_BUILD_TYPE is used only by single configuration generators and is
otherwise completely ignored.
* Remove useless static/shared options
simdjson now uses the CMake builtin BUILD_SHARED_LIBS to switch the
built artifact's type.
* Remove unused CMAKE_MODULE_PATH variable
* Refactor implementation switching into a module
* Factor exception option out into a module
* Reformat simdjson-flags.cmake
* Rename simdjson-flags to developer-options
* Accumulate properties into an include module
This is done this way to avoid using utility targets that must be
exported and installed, which could potentially be misused by users of
the library.
* Port impl definitions to props
* Port exception options to props
* Lift normal options to the top
* Port developer options to props
* Remove simdjson-flags from benchmark
* Document the developer mode in HACKING
* Fix include path in installed config file
* Fix formatting of prop commands
* Fix tests that include .cpp files
* Change GCC AVX fixes back to compile options
* Deprecate SIMDJSON_BUILD_STATIC
* Always link fuzz targets to simdjson
* Install CMake from simdjson's debian repo
* Add gnupg for apt-key
* Make sure ASan link flags come first
* Pass CI env variable to cmake invocation
* Install package for apt-add-repository
* Remove return() from flush macro
* Use directory level commands instead of props
* Restore the github repository variable
* Set developer mode unconditionally for checkperf
The CI env variable is only set in the CI and this target is always run
in developer mode.
* Attempt to fix ODR violation in parsing checks
These tests were compiling the simdjson.cpp file again and linking to
the simdjson library target causes ODR violations.
Instead of linking to the target, just inherit its props.
* Move variables before the source dir
* Mark props to be flushed after adding more
* Use props for every command for the library
* Use keyword form for linking libs
* Handle deprecation of SIMDJSON_JUST_LIBRARY
* Handle deprecations in a separate module
Co-authored-by: friendlyanon <friendlyanon@users.noreply.github.com>
* This gives the CMake install the necessarily information (and flags) to know
whether we have a Windows DLL and in such cases how to handle the linkage.
* This adds a little test to see if we can compiler with very strict flags.
* Trimming a leftover old-style cast.
* More cleaning.
* A few more pedantic casts.
* first try
* use ubuntu 20.04, do the fuzzing
* new try at power fuzz
* hard code clang version
* setting env variables does not seem to work
* use fuzzer-no-link
* switch to Debian Buster for power fuzz
* use non-sanitizer build for power
* me not like yaml
* fix bad syntax
* add ndjson fuzzer
* reproduce #1310 in the newly added unit test
Had to replace the input, because:
1)
the fuzzer uses the first part of the input to determine
the batch_size to use, so that has to be cut off
2)
the master now protects against low values of batch_size
I also made the test not return early, so the error is triggered.
Forks that would like to contribute via PRs from feature branches
needlessly run CI on those branches on top of the PRs.
This is a waste of resources.
Co-authored-by: friendlyanon <friendlyanon@users.noreply.github.com>
* Fix Cirrus CI
CMake is not installed at this point yet.
* Add caching in Circle CI
* Add caching to the MinGW Github workflows
* Fix Circle CI config
Co-authored-by: friendlyanon <friendlyanon@users.noreply.github.com>
This builds the CI fuzzers with the intended clang version. It also allows users to set the clang version locally,
in case they need to.
It also switches the CI fuzzers to use an optimized sanitizer build, to do something oss-fuzz doesn't and get more done in the short time the CI fuzzer runs.
The images used by github already come with a sufficiently recent CMake
version instead of a very outdated one. See:
https://github.com/actions/virtual-environments
Co-authored-by: friendlyanon <friendlyanon@users.noreply.github.com>
* add definitions for is_number and tie (by lemire)
* add fuzzer for element
* update fuzz documentation
* fix UB in creating an empty padded string
* don't bother null terminating padded_string, it is done by the std::memset already
* refactor fuzz data splitting into a separate class
This refactors the dynamic check of which implementations are supported at runtime.
It also reduces duplicated effort in the CI fuzzing job, the differential fuzzers don't need to run with different values of SIMDJSON_FORCE_IMPLEMENTATION.
There is also a convenience script to run the fuzzers locally, to quickly check that the fuzzers still build, run and no easy to find bugs are there. It should be handy not only when developing the fuzzers, but also when modifying simdjson.
Daniel Lemire
Nicolas Boyer
Paul Dreik
friendlyanon
John Keiser