Wednesday, 2026-05-20

opendevreviewOpenStack Proposal Bot proposed openstack/project-config master: Normalize projects.yaml  https://review.opendev.org/c/openstack/project-config/+/98932002:58
opendevreviewMerged openstack/project-config master: Normalize projects.yaml  https://review.opendev.org/c/openstack/project-config/+/98932005:54
cschwedeAnybode else seen the 404 errors from multiple centos9-stream mirrors? Looks multiple of them are affected (mirror.iad3.openmetal.opendev.org, mirror.iad3.raxflex.opendev.org, mirror.bhs1.ovh.opendev.org - that were the ones I did see in the logs09:18
cschwedeExample failure: https://zuul.opendev.org/t/openstack/build/2c632ac783e14db39714d7233122f36d10:08
dpawlikfungi clarkb hey folks o/ is possible to re-trigger script related to sync centos stream? 10:43
ykarelalso if someone can check why mirror sync logs not included for last 9 months https://static.opendev.org/mirror/logs/rsync-mirrors/?C=M;O=D11:02
*** darmach1002 is now known as darmach10011:09
fricklerafaict the last sync from rsync.stream.centos.org ran without issues, maybe there was an issue with upstream? the next sync should happen soon anyway, we can recheck after that maybe12:00
ykarelyes it's synched now12:20
ykarelfrickler, do you know why https://static.opendev.org/mirror/logs/rsync-mirrors/?C=M;O=D not updated for 9 months? or those pushed to some other location? 12:23
fricklerI don't know that right away, I checked the logs directly on the mirror-update server12:41
fungii'll look at why the logs haven't been getting copied into afs12:47
fungii see it, the comment at https://opendev.org/opendev/system-config/src/commit/1c81a5b/playbooks/roles/mirror-update/files/publish-mirror-logs#L26-L27 indicates that we just expect people to look at the rw volume not a ro replica12:54
fungibut the static.opendev.org vhost simply browses ro replicas not rw12:55
fungiso at some point ~9mos ago someone did a vos release of the mirror.logs volume, but it's not regularly released12:56
fungii guess we could alias /mirror/logs/ to /afs/.openstack.org/mirror/logs/ to shadow the stale ro replica with the rw version12:57
fungihttps://review.opendev.org/c/opendev/system-config/+/989377 Shadow mirror.logs RO replica with RW original [NEW]13:05
fungithat's an option for addressing it13:05
fungian alternative would be to have a mirror-logs.opendev.org site with its docroot at /afs/.openstack.org/mirror/logs/13:06
opendevreviewMauricio Harley proposed openstack/project-config master: Add #openstack-pqc IRC channel  https://review.opendev.org/c/openstack/project-config/+/98877013:51
opendevreviewStephen Finucane proposed openstack/pbr master: Add classifiers for latest Python  https://review.opendev.org/c/openstack/pbr/+/98938814:00
opendevreviewStephen Finucane proposed openstack/pbr master: Remove more unused code  https://review.opendev.org/c/openstack/pbr/+/98938914:00
opendevreviewStephen Finucane proposed openstack/pbr master: Remove need for setup.cfg, setup.py  https://review.opendev.org/c/openstack/pbr/+/98939014:00
opendevreviewStephen Finucane proposed openstack/pbr master: Hook in LocalSDist override  https://review.opendev.org/c/openstack/pbr/+/98939114:00
opendevreviewStephen Finucane proposed openstack/pbr master: Populate pyproject.toml file  https://review.opendev.org/c/openstack/pbr/+/98939214:00
stephenfinclarkb: fungi: That middle patch should obviate the need to keep stub setup.cfg/setup.py files around for pbr-using projects ☝️14:04
sean-k-mooneyas an aside rdo is kidn fo dead but it currently depned on setup.py because it was nto ported to use python -m build14:06
sean-k-mooneythat is not a reason to not impove thing in pbr14:06
sean-k-mooneyjust notting that that is one of the reason we didnt delete the stub setup.py/setup.cfg files when we were mvoign to pyproject.toml14:07
sean-k-mooneyas long as we dont have to maintain team (update them) having them in the repo as extra compat does not hurt14:07
sean-k-mooneybut eveually that will proably break in setuptool and i dont htink we shoudl do any heroic to keep it working if setuptool moves away form it14:08
stephenfinsean-k-mooney: this came up in the #openstack-tc channel yesterday, but pbr's setup.cfg support is entirely implemented in pbr14:09
stephenfinthe only way it will break is if someone is using a setuptools' specific option rather than pbr's shim14:09
sean-k-mooneyack14:09
opendevreviewMerged openstack/project-config master: Add #openstack-pqc IRC channel  https://review.opendev.org/c/openstack/project-config/+/98877014:09
fungiright, but if there's no setup.py in a project, you also can't run it14:09
stephenfinfungi: what's "it"?14:09
fungiautomation like rdo's that's directly running setup.py still14:10
sean-k-mooneyfungi: the spec files/mock command used in rdo have conditional logic based on the presence of a setup.py to use it 14:10
stephenfinah, yes14:10
sean-k-mooneyand no supprot for python -m build or pyproject.toml14:10
sean-k-mooneyemma and i were workign on bumping rdo to use setuptools 80 last weeak instead of 6914:10
fungithat said, while i like the idea of continuing to provide backwards compatibility for older setuptools, i think that ship has sailed because the versions of python we're using now in projects probably lock us into newer setuptools that have removed support for directly running setup.py, or will soon if not14:10
sean-k-mooneybecause it did not supprot the string form of the lisce and a bughc of toher things14:11
stephenfinsean-k-mooney: having very recently proposed a patch against fedora's git-review package, there are new macros that use pyproject.toml and infer all their metadata from there14:11
stephenfinso they can be reworked but I don't imagine anyone will do that now14:11
sean-k-mooneyoh they can14:11
sean-k-mooneyim just sayign that non of our midstream (rdo) or downstram (brew) builds do today14:11
stephenfinhttps://src.fedoraproject.org/rpms/git-review/pull-request/7 fwiw14:12
sean-k-mooneyso deleteing setup.py form nova woudl break all our rdo and downstream ablity to build form master if that was actlly a thing that exsited today14:12
sean-k-mooneythats going to be less of an issue in the future as we are in the process of retoolign all fo that14:13
sean-k-mooneyjust noting that thre may be other distos that depend on it14:13
sean-k-mooneyi.e. i dont knwo what ubuntu/debian do14:13
sean-k-mooneyim sure they coudl adapt but they may not have14:13
stephenfinThere's no need to delete setup.py/setup.cfg from any project but at least it should be possible now14:13
sean-k-mooneyyep so that the compat level im advocating for14:14
sean-k-mooneydont delete it but dont depend on it either14:14
sean-k-mooneyall oru upstream toolign shoudl use the more modern way to build14:14
fungiin the case of debian/ubuntu, the makefiles (rules) which build packages generally use debhelper wrapper scripts which are general purpose and, e.g., the dh_python wrapper uses standard tools like pyproject-build that can automatically handle both old and new style packaging14:15
sean-k-mooneyyep on the rpm side this is largely handel by macros too14:16
sean-k-mooneybut at least rdo defiens therfe own14:17
sean-k-mooneyfedora's proably already hadnel it fine14:17
sean-k-mooneycentos may even but rdo has mostly been dormant for 18 months now14:17
fungithis is what the makefile for oslo.messaging in debian looks like, ftr: https://salsa.debian.org/openstack-team/oslo/python-oslo.messaging/-/blob/debian/gazpacho/debian/rules?ref_type=heads14:19
fungiso no specific setuptools logic baked in there14:19
sean-k-mooneyhttps://github.com/softwarefactory-project/DLRN/blob/60269d3dcd059716e87386d2a074ecf90dfe8c98/scripts/build_srpm.sh#L38 14:27
sean-k-mooneyis what i was referitn too14:27
sean-k-mooneythe deloran server that build the srpms via mock to then buidl the final rpms use that shell scritp to actully do the build14:28
sean-k-mooneythe sepc files themeself dont really care14:29
sean-k-mooneybut there are thing like that in the pipeline that do14:29
sean-k-mooneyand yes that coudl jsut be ```/usr/bin/mock $MOCKOPTS --chroot "python3 -m build -s /var/tmp/pkgsrc"```14:30
fungias long as the build package is installed in the module search path for that environment14:33
sean-k-mooneyyep mock shoudl provide that although i think you need to give mock a set of deps14:34
sean-k-mooneyin anycase as was noted in the rdo email we are lookign at doign direct source builds in the future14:34
sean-k-mooneyso no rpm step at all14:34
sean-k-mooneythe tool we will be useing will use konflux https://konflux-ci.dev/docs/mintmaker/support/#python-dependencies14:36
sean-k-mooneywhich basiclly suprpot anyting with a requirements.txt14:36
sean-k-mooneyor that can be converted to that14:37
sean-k-mooneythat also incluse supprot for https://peps.python.org/pep-0621/ and https://peps.python.org/pep-0723/14:39
opendevreviewMonty Taylor proposed openstack/project-config master: Import drizzle revival and add a bindep-rs support project  https://review.opendev.org/c/openstack/project-config/+/98941715:40
opendevreviewMonty Taylor proposed openstack/project-config master: Import drizzle revival and add a bindep-rs support project  https://review.opendev.org/c/openstack/project-config/+/98941715:57
opendevreviewMonty Taylor proposed openstack/project-config master: Import drizzle revival and add a bindep-rs support project  https://review.opendev.org/c/openstack/project-config/+/98941718:01
opendevreviewSabbir Ahmed proposed openstack/project-config master: Add starlingx/app-machine-operator project  https://review.opendev.org/c/openstack/project-config/+/98944618:39
opendevreviewSabbir Ahmed proposed openstack/project-config master: Add starlingx/app-machine-operator project  https://review.opendev.org/c/openstack/project-config/+/98944619:17
*** skandix4692 is now known as skandix46922:16

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!