Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
* Improve write time
* Print warning if mpi4py is not found in openpmd-pipe
* ADIOS1: Remove task from IO queue if it fails with exception
* `setup.py`: Extra CMake Arg Control
* EMSCRIPTEN: Skip Py Extension
* g++/clang++: Add `-Wsign-compare`
* python/RecordComponent: fix `-Wsign-compare`
* Patch MSVC pybind11 bug
Improve patch comment
* Python: Do Not Strip Symbols In Debug
Avoid stripping symbols for Python debug builds, so we can see lines
in coredumps and debugger runs.
* CI: NVHPC New Apt Repo
Update the NVHPC install instructions to the latest and greatest.
Fix failing CI (dependency install).
Also upgrade to 21.11 as on Perlmutter.
* clang-format & pre-commit & python include
* run pre-commit
* ps_make_timer_name_: add memory leak suppression
* Fix: Missing HDF5 Include
This tries to fix a compile error on Maxwell (DESY).
```
/home/diederse/software/hipace_4GPU/hipace/build/_deps/fetchedopenpmd-src/src/IO/HDF5/HDF5IOHandler.cpp: In member function 'virtual void openPMD::HDF5IOHandlerImpl::readAttribute(openPMD::Writable*, openPMD::Parameter<openPMD::Operation::READ_ATT>&)':
/home/diederse/software/hipace_4GPU/hipace/build/_deps/fetchedopenpmd-src/src/IO/HDF5/HDF5IOHandler.cpp:1586:17: error: 'H5free_memory' was not declared in this scope
1586 | H5free_memory(m1);
| ^~~~~~~~~~~~~
/home/diederse/software/hipace_4GPU/hipace/build/_deps/fetchedopenpmd-src/src/IO/HDF5/HDF5IOHandler.cpp:1609:17: error: 'H5free_memory' was not declared in this scope
1609 | H5free_memory(m1);
| ^~~~~~~~~~~~~
/home/diederse/software/hipace_4GPU/hipace/build/_deps/fetchedopenpmd-src/src/IO/HDF5/HDF5IOHandler.cpp:1628:17: error: 'H5free_memory' was not declared in this scope
1628 | H5free_memory(m2);
| ^~~~~~~~~~~~~
```
* Increase reference count also in other load_chunk overload
Don't know if it is necessary, but looks like we forgot it earlier
Add comment
* Spack: Remove Old Files
* Spack: v0.17.1
* Update Env Files
* Python Iteration: Fix __repr__ (time)
Small numbers, as common for iterations, were flushed to zero
in `std::to_string(double)` of the representation of `Iteration`
variables:
```
step __repr__
100 <openPMD.Iteration at t = '0.000000 s'>
was: 8.687655225973454e-14 * 1.0
```
* CI: Update CUDA repo key
Nvidia has made changes in the signing keys.
https://forums.developer.nvidia.com/t/notice-cuda-linux-repository-key-rotation/212771
* CI: Switch to Mamba
* Doc: Update HDF5 Coll. Metadata Versions
Update the documentation about HDFFV-11260 and the newly released
fixes in release lines.
* HDF5 Coll. MD Reads: Simplify Wording
* Tiny wording improvement and a +
* Remove deprecated debug parameter in ADIOS2
* HDF5IOHandler: Clang-Format/Clang-Tidy
```
/home/runner/work/openPMD-api/openPMD-api/src/IO/HDF5/HDF5IOHandler.cpp:1835:21:
warning: repeated branch in conditional chain [bugprone-branch-clone]
isLegacyLibSplashAttr =
^
```
* `conda.yaml`: add `pre-commit`
* Fix: Python Variant Issue on Conda
* Test: Demonstrate Pattern Issue
As in #1173
* JSON: Improve File Open Error Message
Include path to file
* Upon parsing, store each iteration's filename
If the padding is inconsistent, a later Iteration::open() needs the
original filename. Trying to compute the filename from the expansion
pattern will lead to wrong filenames.
* Backport: fix-open-iteration (#5)
* Pass-through flushing parameters
* CI fixes
* Test
* Don't flush when opening an iteration
* CI fixes
* FlushLevel: Use default base class in NVC++
* clang-tidy: Define member defaults of Writable in-class
* Update include/openPMD/IO/AbstractIOHandler.hpp
Co-authored-by: Axel Huebl <[email protected]>
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* Release 0.14.5 conflicting backports (#6)
* Don't safeguard empty strings while reading
* Update TODO Comment
* Add missing update for VEC_CLONG_DOUBLE
* Ignore deprecated AccessType in SerialIOTest
* SerialIOTest: Fix GCC Pragma Check (#1260)
There was a typo in `__GNUC_MINOR__`, causing the test not to work on all GCC versions.
Follow-up to #1213
* Clearly fail when users select a wrong backend
* ICC: Unreachable end of function
* Fix use after free in ADIOS1IOHandler
* Remove unavailable_backend_test
Hard to trigger without explicit backend selection, and test coverage is
present on dev anyway
Co-authored-by: Axel Huebl <[email protected]>
* Version: 0.14.5
including Changelog
Co-authored-by: Jean Luca Bez <[email protected]>
Co-authored-by: Franz Pöschel <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>- Loading branch information
4 people
authored
Jun 7, 2022
1 parent
756bffc
commit ce25526