You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The build-windows-msys2-mingw CI target is now failing on tests that involve the autotest/gdrivers/data/hdf5/u8be.h5 file, since MSYS2 has updated their mingw-w64-x86_64-hdf5 package from 1.14.3-2 to 1.14.5-1
GDAL autotest test failures seem to suggest that the file can be opened, but that the array of type H5T_STD_U8BE (unsigned integer 8 bit big endian (note of myself: why does HDF5 need to distinguish little and big endian for a single-byte data type...)):
2024-10-31T01:42:31.2420877Z 26: FAILED hdf5.py::test_hdf5_3 - RuntimeError: `HDF5:"data/hdf5/u8be.h5"://TestArray' does not exist in the file system, and is not recognized as a supported dataset name.
2024-10-31T01:42:31.2423598Z 26: FAILED hdf5.py::test_hdf5_4 - RuntimeError: `HDF5:"data/hdf5/u8be.h5"://TestArray' does not exist in the file system, and is not recognized as a supported dataset name.
2024-10-31T01:42:31.2426093Z 26: FAILED hdf5multidim.py::test_hdf5_multidim_basic - AssertionError: assert [] == ['TestArray']
A local Linux build of hdf5 1.14.5 doesn't show any difference in the output of the h5dump utility compared to earlier versions, so it might be either an issue with the API, or a Windows specific issue, or a msys2 packaging one...
The text was updated successfully, but these errors were encountered:
The build-windows-msys2-mingw CI target is now failing on tests that involve the autotest/gdrivers/data/hdf5/u8be.h5 file, since MSYS2 has updated their mingw-w64-x86_64-hdf5 package from 1.14.3-2 to 1.14.5-1
Last good build: https://github.com/OSGeo/gdal/actions/runs/11599287020/job/32297101088 (OSGeo/GDAL master)
First failed build: https://github.com/rouault/gdal/actions/runs/11604312201/job/32312870698 (a branch of mine on ~ same master + unrelated changes)
msys2 related update is at msys2/MINGW-packages#22342 . My investigations at msys2/MINGW-packages#22342 (comment)
Also reported upstream as HDFGroup/hdf5#5052
GDAL autotest test failures seem to suggest that the file can be opened, but that the array of type H5T_STD_U8BE (unsigned integer 8 bit big endian (note of myself: why does HDF5 need to distinguish little and big endian for a single-byte data type...)):
A local Linux build of hdf5 1.14.5 doesn't show any difference in the output of the h5dump utility compared to earlier versions, so it might be either an issue with the API, or a Windows specific issue, or a msys2 packaging one...
The text was updated successfully, but these errors were encountered: