Package: src:h5py
Version: 3.10.0-1
Severity: serious
Tags: sid trixie

armhf: h5py's tests expose unaligned memory accesses during the build,
this not seen with with the 3.9.0 version. You don't see this on the
Debian buildds itself, because they ignore these misalignments.

Already happens with Python 3.11.

[...] build/h5py/_debian_h5py_serial/tests/test_dims_dimensionproxy.py::TestItems::test_empty
PASSED [ 53%] build/h5py/_debian_h5py_serial/tests/test_dtype.py::TestVlen::test_compound PASSED [ 53%] build/h5py/_debian_h5py_serial/tests/test_dtype.py::TestVlen::test_compound_vlen_bool
Fatal Python error: Bus error

Current thread 0xf7959020 (most recent call first):
File "/<<PKGBUILDDIR>>/.pybuild/cpython3_3.11_h5py_serial/build/h5py/_debian_h5py_serial/_hl/dataset.py",
line 841 in __getitem__
File "/<<PKGBUILDDIR>>/.pybuild/cpython3_3.11_h5py_serial/build/h5py/_debian_h5py_serial/tests/test_dtype.py",
line 60 in test_compound_vlen_bool
File "/usr/lib/python3.11/unittest/case.py", line 579 in _callTestMethod
File "/usr/lib/python3.11/unittest/case.py", line 623 in run
File "/usr/lib/python3.11/unittest/case.py", line 678 in __call__
File "/usr/lib/python3/dist-packages/_pytest/unittest.py", line 333
in runtest
File "/usr/lib/python3/dist-packages/_pytest/runner.py", line 169 in pytest_runtest_call
File "/usr/lib/python3/dist-packages/pluggy/_callers.py", line 77 in _multicall
File "/usr/lib/python3/dist-packages/pluggy/_manager.py", line 115 in _hookexec
File "/usr/lib/python3/dist-packages/pluggy/_hooks.py", line 493 in
__call__
File "/usr/lib/python3/dist-packages/_pytest/runner.py", line 262 in <lambda>
File "/usr/lib/python3/dist-packages/_pytest/runner.py", line 341 in from_call
File "/usr/lib/python3/dist-packages/_pytest/runner.py", line 261 in call_runtest_hook
File "/usr/lib/python3/dist-packages/_pytest/runner.py", line 222 in call_and_report
File "/usr/lib/python3/dist-packages/_pytest/runner.py", line 133 in runtestprotocol
File "/usr/lib/python3/dist-packages/_pytest/runner.py", line 114 in pytest_runtest_protocol
File "/usr/lib/python3/dist-packages/pluggy/_callers.py", line 77 in _multicall
File "/usr/lib/python3/dist-packages/pluggy/_manager.py", line 115 in _hookexec
File "/usr/lib/python3/dist-packages/pluggy/_hooks.py", line 493 in
__call__
File "/usr/lib/python3/dist-packages/_pytest/main.py", line 350 in pytest_runtestloop
File "/usr/lib/python3/dist-packages/pluggy/_callers.py", line 77 in _multicall
File "/usr/lib/python3/dist-packages/pluggy/_manager.py", line 115 in _hookexec
File "/usr/lib/python3/dist-packages/pluggy/_hooks.py", line 493 in
__call__
File "/usr/lib/python3/dist-packages/_pytest/main.py", line 325 in _main
File "/usr/lib/python3/dist-packages/_pytest/main.py", line 271 in wrap_session
File "/usr/lib/python3/dist-packages/_pytest/main.py", line 318 in pytest_cmdline_main
File "/usr/lib/python3/dist-packages/pluggy/_callers.py", line 77 in _multicall
File "/usr/lib/python3/dist-packages/pluggy/_manager.py", line 115 in _hookexec
File "/usr/lib/python3/dist-packages/pluggy/_hooks.py", line 493 in
__call__
File "/usr/lib/python3/dist-packages/_pytest/config/__init__.py",
line 169 in main
File "/usr/lib/python3/dist-packages/_pytest/config/__init__.py",
line 192 in console_main
File "/usr/lib/python3/dist-packages/pytest/__main__.py", line 5 in
<module>
File "<frozen runpy>", line 88 in _run_code
File "<frozen runpy>", line 198 in _run_module_as_main

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 2024-01-17 09:33, Matthias Klose wrote:

Package: src:h5py
Version: 3.10.0-1
Severity: serious
Tags: sid trixie

armhf: h5py's tests expose unaligned memory accesses during the build, this not seen with with the 3.9.0 version. You don't see this on the Debian buildds itself, because they ignore these misalignments.

While it is something to fix for performance issues, I hardly see how it
is a serious issue.

It is not possible to deactivate alignment fixups on a armhf kernel. The
Debian arm64 kernels (and not just the buildds) therefore have the COMPAT_ALIGNMENT_FIXUPS kernel option enabled to reproduce the same
behaviour for 32-bit processes.

--
Aurelien Jarno GPG: 4096R/1DDD8C9B aurelien@aurel32.net http://aurel32.net

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Processing control commands:

forwarded 1061063 https://github.com/h5py/h5py/issues/1927

Bug #1061063 [src:h5py] armhf: h5py's tests expose unaligned memory accesses during the build
Set Bug forwarded-to-address to 'https://github.com/h5py/h5py/issues/1927'.

--
1061063: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1061063
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Source: h5py
Followup-For: Bug #1061063
Control: forwarded 1061063 https://github.com/h5py/h5py/issues/1927

The problem was raised upstream at
https://github.com/h5py/h5py/issues/1927

Makes it difficult to test if we can't reproduce it on all armhf
environments.

A patch was suggested for the upstream report but has already been
applied (fix-unaligned-access.patch).

I'm not sure that we can do much more than that, since we can't
reproduce the bug on debian armhf systems.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Control: severity -1 important
Control: tags -1 + patch

hickle has been auto-removed from testing due to #1060804.
I'm lowering the severity so that h5py can migrate.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Processing control commands:

severity -1 important

Bug #1061063 [src:h5py] armhf: h5py's tests expose unaligned memory accesses during the build
Severity set to 'important' from 'serious'

tags -1 + patch

Bug #1061063 [src:h5py] armhf: h5py's tests expose unaligned memory accesses during the build
Added tag(s) patch.

--
1061063: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1061063
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)