Overview
Request 1093400 accepted
- Make python3-clang noarch again and fix conflicts/provides to
point to %python3_sitelib instead of %python3_sitearch.
- Add patch to fix segmentation fault on aarch64 - boo#1212264:
* llvm14-fix-segfault-on-aarch64.patch
- Created by aaronpuchert
- In state accepted
- Supersedes 1093168
There is an installcheck
failure due to a conflict between python3-clang13
with python3-clang14
, though both have
%package -n python3-clang%{_sonum} # ... Conflicts: %{python3_sitearch}/clang/ Provides: %{python3_sitearch}/clang/
Notably the former is noarch
, but that's not only wrong (the package installs into {%_libdir}
, but it also causes one of the architectures to randomly win the above Conflicts/Provides pair. I've fixed that in llvm14
, but forgot to port it back.
This is a temporary fallout during the migration from py3.10 to py3.11
The provides was not in sync at the time installxheck ran - one mentioned the py3.10 party, the other the 3.11 path.
This will settle
Request History
aaronpuchert created request
- Make python3-clang noarch again and fix conflicts/provides to
point to %python3_sitelib instead of %python3_sitearch.
- Add patch to fix segmentation fault on aarch64 - boo#1212264:
* llvm14-fix-segfault-on-aarch64.patch
factory-auto added opensuse-review-team as a reviewer
Please review sources
factory-auto accepted review
Check script succeeded
licensedigger accepted review
ok
staging-bot added openSUSE:Factory:Staging:adi:11 as a reviewer
Being evaluated by staging project "openSUSE:Factory:Staging:adi:11"
staging-bot accepted review
Picked "openSUSE:Factory:Staging:adi:11"
darix accepted review
Accepted review for by_group opensuse-review-team request 1093400 from user staging-bot
dimstar_suse accepted review
Staging Project openSUSE:Factory:Staging:adi:11 got accepted.
dimstar_suse approved review
Staging Project openSUSE:Factory:Staging:adi:11 got accepted.
dimstar_suse accepted request
Staging Project openSUSE:Factory:Staging:adi:11 got accepted.
probably not wrong - but this requires all llvm versions to do this in step-lock (otoh: the provides: is a pure string, technically seen without much meaning other than the string being itself)
Couldn't there be an issue where the
i586
build ends up producing thenoarch
package, and then the Conflicts/Provides points to/usr/lib
instead of/usr/lib64
? Or similar in PowerPC/ARM land.Nevertheless, I have now submitted requests for all LLVM versions (except those that I've submitted delete requests for).