The Python 3.7 rebuild in the f29-python side tag is currently:
- 2168 built - ~170 FTBFS - ~120 blocked by dependencies that FTBFS
Ideally, this would all get fixed before the merge, but that's unrealistic, there are packages that are FTBFS for unrelated reasons, etc.
What is the criterion to merge the side tag?
The critpath installs fine (except stuff that cannot be installed due to yum-deprecated vs yum collisions and gnome-software that has some broken dependency on PackageKit).
Stuff that doesn't build and sounds important:
- ceph (also FTBFS in regular rawhide) - libreoffice (also FTBFS in regular rawhide) - orca (blocked by pyatspi, looking at it ATM) - kernel-tools
"MH" == Miro Hrončok mhroncok@redhat.com writes:
MH> What is the criterion to merge the side tag?
I think ultimately that's up to you, but certainly the idea is to balance disruption (breaking builds or rawhide usability) against holding back progress (the approved python 3.7 feature).
It would certainly be nice to have less disruption than 170 FTBFS packages and 120 broken dependencies, but then if some of those packages are broken already then they shouldn't count against the total unless somehow they're "more broken" after the tag is merged.
I think setting a deadline for the merge (which would have to be pretty soon if the mass rebuild is still on July 11) and sending regular broken package lists and calls for help for help would be the only thing you can really do besides just hitting the button and merging it.
- J<
On 28.6.2018 15:42, Miro Hrončok wrote:
The Python 3.7 rebuild in the f29-python side tag is currently:
- 2168 built - ~170 FTBFS - ~120 blocked by dependencies that FTBFS
Ideally, this would all get fixed before the merge, but that's unrealistic, there are packages that are FTBFS for unrelated reasons, etc.
What is the criterion to merge the side tag?
OK. I'll ask releng to merge on Monday.
python-devel@lists.fedoraproject.org