| *** HenryG has quit IRC | 00:01 | |
| *** HenryG has joined #openstack-infra | 00:02 | |
| *** tetsuro has joined #openstack-infra | 00:07 | |
| *** ryohayakawa has joined #openstack-infra | 00:09 | |
| *** jamesmcarthur has quit IRC | 00:09 | |
| *** jamesmcarthur has joined #openstack-infra | 00:11 | |
| *** jamesmcarthur has quit IRC | 00:12 | |
| *** jamesmcarthur has joined #openstack-infra | 00:12 | |
| *** jamesmcarthur has quit IRC | 00:13 | |
| *** _erlon_ has quit IRC | 00:14 | |
| *** jamesmcarthur has joined #openstack-infra | 00:18 | |
| *** jamesmcarthur has quit IRC | 00:20 | |
| *** Lucas_Gray has quit IRC | 00:24 | |
| *** Lucas_Gray has joined #openstack-infra | 00:27 | |
| *** jamesmcarthur has joined #openstack-infra | 00:30 | |
| *** gyee has quit IRC | 00:30 | |
| *** jamesmcarthur has quit IRC | 00:32 | |
| *** jamesmcarthur has joined #openstack-infra | 00:37 | |
| *** jamesmcarthur has quit IRC | 00:38 | |
| *** jamesmcarthur has joined #openstack-infra | 00:44 | |
| *** rlandy has quit IRC | 00:50 | |
| *** tetsuro_ has joined #openstack-infra | 00:54 | |
| *** tetsuro_ has quit IRC | 00:56 | |
| *** tetsuro_ has joined #openstack-infra | 00:56 | |
| *** tetsuro has quit IRC | 00:58 | |
| *** tetsuro_ has quit IRC | 00:59 | |
| *** tetsuro has joined #openstack-infra | 00:59 | |
| *** hashar has joined #openstack-infra | 01:29 | |
| *** rfolco has quit IRC | 01:39 | |
| *** Lucas_Gray has quit IRC | 01:43 | |
| *** dave-mccowan has joined #openstack-infra | 01:45 | |
| *** Lucas_Gray has joined #openstack-infra | 01:49 | |
| *** jamesmcarthur has quit IRC | 01:59 | |
| *** d34dh0r53 has quit IRC | 02:01 | |
| *** d34dh0r53 has joined #openstack-infra | 02:04 | |
| *** tetsuro_ has joined #openstack-infra | 02:11 | |
| *** dave-mccowan has quit IRC | 02:13 | |
| *** apetrich has quit IRC | 02:14 | |
| *** tetsuro has quit IRC | 02:15 | |
| *** aarents has quit IRC | 02:22 | |
| *** aarents has joined #openstack-infra | 02:23 | |
| *** ramishra has joined #openstack-infra | 02:24 | |
| *** mtreinish has joined #openstack-infra | 02:39 | |
| *** rcernin has joined #openstack-infra | 03:07 | |
| *** markvoelker has joined #openstack-infra | 03:10 | |
| *** tetsuro_ has quit IRC | 03:12 | |
| *** markvoelker has quit IRC | 03:15 | |
| *** Lucas_Gray has quit IRC | 03:31 | |
| *** psachin has joined #openstack-infra | 03:33 | |
| *** jamesmcarthur has joined #openstack-infra | 03:37 | |
| *** jamesmcarthur has quit IRC | 03:42 | |
| *** markvoelker has joined #openstack-infra | 03:46 | |
| *** hashar has quit IRC | 03:50 | |
| *** markvoelker has quit IRC | 03:51 | |
| *** tkajinam has quit IRC | 03:51 | |
| *** tkajinam has joined #openstack-infra | 03:52 | |
| *** markvoelker has joined #openstack-infra | 03:53 | |
| *** yamamoto has quit IRC | 04:02 | |
| *** markvoelker has quit IRC | 04:05 | |
| *** vishalmanchanda has joined #openstack-infra | 04:15 | |
| *** ramishra has quit IRC | 04:16 | |
| *** ociuhandu has joined #openstack-infra | 04:18 | |
| *** yamamoto has joined #openstack-infra | 04:19 | |
| *** ociuhandu has quit IRC | 04:23 | |
| *** jamesmcarthur has joined #openstack-infra | 04:28 | |
| *** udesale has joined #openstack-infra | 04:33 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-infra | 04:33 | |
| *** ykarel|away has joined #openstack-infra | 04:43 | |
| *** ramishra has joined #openstack-infra | 04:44 | |
| *** ykarel|away is now known as ykarel | 04:46 | |
| *** raukadah is now known as chkumar|rover | 04:54 | |
| *** jamesmcarthur has quit IRC | 04:55 | |
| *** lbragstad_ has joined #openstack-infra | 05:04 | |
| *** lbragstad has quit IRC | 05:07 | |
| *** jamesmcarthur has joined #openstack-infra | 05:15 | |
| *** ykarel has quit IRC | 05:24 | |
| *** ykarel has joined #openstack-infra | 05:29 | |
| *** lmiccini has joined #openstack-infra | 05:29 | |
| *** udesale_ has joined #openstack-infra | 05:45 | |
| *** udesale has quit IRC | 05:47 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/project-config master: Normalize projects.yaml https://review.opendev.org/744822 | 06:07 |
|---|---|---|
| *** eolivare has joined #openstack-infra | 06:13 | |
| *** Ajohn has quit IRC | 06:15 | |
| *** udesale_ has quit IRC | 06:24 | |
| *** dklyle has quit IRC | 06:31 | |
| *** jcapitao has joined #openstack-infra | 06:32 | |
| *** udesale has joined #openstack-infra | 06:52 | |
| *** jamesmcarthur has quit IRC | 06:58 | |
| *** rcernin has quit IRC | 06:58 | |
| *** rcernin_ has joined #openstack-infra | 06:59 | |
| *** jamesmcarthur has joined #openstack-infra | 06:59 | |
| *** nightmare_unreal has joined #openstack-infra | 07:00 | |
| *** ryohayakawa has quit IRC | 07:02 | |
| *** zxiiro has quit IRC | 07:02 | |
| *** jamesmcarthur has quit IRC | 07:03 | |
| *** rcernin_ has quit IRC | 07:05 | |
| *** apetrich has joined #openstack-infra | 07:06 | |
| *** rcernin has joined #openstack-infra | 07:06 | |
| *** slaweq has joined #openstack-infra | 07:07 | |
| *** tosky has joined #openstack-infra | 07:38 | |
| openstackgerrit | Merged openstack/project-config master: Normalize projects.yaml https://review.opendev.org/744822 | 07:41 |
| *** jpena|off is now known as jpena | 07:54 | |
| *** gfidente|afk is now known as gfidente | 08:03 | |
| *** dtantsur|afk is now known as dtantsur | 08:04 | |
| *** pkopec has joined #openstack-infra | 08:11 | |
| *** lucasagomes has joined #openstack-infra | 08:15 | |
| *** jamesmcarthur has joined #openstack-infra | 08:17 | |
| *** xek has joined #openstack-infra | 08:21 | |
| *** tosky has quit IRC | 08:26 | |
| *** jamesmcarthur has quit IRC | 08:27 | |
| *** tosky has joined #openstack-infra | 08:27 | |
| *** ralonsoh has joined #openstack-infra | 08:30 | |
| *** andrewbonney has joined #openstack-infra | 08:50 | |
| *** derekh has joined #openstack-infra | 08:51 | |
| *** tetsuro has joined #openstack-infra | 08:54 | |
| *** sshnaidm|afk is now known as sshnaidm | 08:56 | |
| *** rcernin has quit IRC | 09:04 | |
| *** rcernin has joined #openstack-infra | 09:04 | |
| *** rcernin has quit IRC | 09:05 | |
| *** rcernin has joined #openstack-infra | 09:05 | |
| *** tetsuro has quit IRC | 09:11 | |
| *** ociuhandu has joined #openstack-infra | 09:23 | |
| *** Lucas_Gray has joined #openstack-infra | 09:33 | |
| *** ralonsoh has quit IRC | 09:50 | |
| *** tkajinam has quit IRC | 10:15 | |
| *** rcernin has quit IRC | 10:23 | |
| *** xek has quit IRC | 10:24 | |
| *** jamesmcarthur has joined #openstack-infra | 10:24 | |
| *** jamesmcarthur has quit IRC | 10:31 | |
| *** hemna has quit IRC | 10:38 | |
| *** hemna has joined #openstack-infra | 10:38 | |
| *** ociuhandu has quit IRC | 10:54 | |
| *** ociuhandu has joined #openstack-infra | 10:55 | |
| *** yamamoto has quit IRC | 10:56 | |
| *** ociuhandu has quit IRC | 11:00 | |
| *** yamamoto has joined #openstack-infra | 11:01 | |
| *** lmiccini has quit IRC | 11:04 | |
| *** ralonsoh has joined #openstack-infra | 11:08 | |
| *** hashar has joined #openstack-infra | 11:08 | |
| *** markvoelker has joined #openstack-infra | 11:16 | |
| *** lmiccini has joined #openstack-infra | 11:19 | |
| *** markvoelker has quit IRC | 11:25 | |
| *** rcernin has joined #openstack-infra | 11:25 | |
| *** rcernin has quit IRC | 11:30 | |
| *** ociuhandu has joined #openstack-infra | 11:36 | |
| *** jpena is now known as jpena|lunch | 11:39 | |
| *** xek has joined #openstack-infra | 11:40 | |
| *** jcapitao is now known as jcapitao_lunch | 11:42 | |
| *** ociuhandu has quit IRC | 11:42 | |
| *** ociuhandu has joined #openstack-infra | 11:42 | |
| *** eolivare has quit IRC | 11:51 | |
| *** rfolco has joined #openstack-infra | 11:51 | |
| openstackgerrit | Aurelien Lourot proposed openstack/project-config master: Mirror keystone-kerberos and ceph-iscsi charms to GitHub https://review.opendev.org/744890 | 11:54 |
| *** aedc has joined #openstack-infra | 11:56 | |
| *** aedc has quit IRC | 11:56 | |
| *** xek has quit IRC | 11:56 | |
| *** rcernin has joined #openstack-infra | 11:56 | |
| *** yamamoto has quit IRC | 11:58 | |
| *** rcernin has quit IRC | 12:02 | |
| *** eolivare has joined #openstack-infra | 12:03 | |
| *** rlandy has joined #openstack-infra | 12:03 | |
| *** hashar has quit IRC | 12:13 | |
| *** rcernin has joined #openstack-infra | 12:17 | |
| *** rcernin has quit IRC | 12:22 | |
| *** derekh has quit IRC | 12:24 | |
| *** yamamoto has joined #openstack-infra | 12:26 | |
| *** yamamoto has quit IRC | 12:36 | |
| *** udesale has quit IRC | 12:37 | |
| *** udesale has joined #openstack-infra | 12:38 | |
| *** zxiiro has joined #openstack-infra | 12:40 | |
| *** jpena|lunch is now known as jpena | 12:46 | |
| openstackgerrit | Sean McGinnis proposed openstack/pbr master: Drop py27 support and fix compatiblity with virtualenv 20.x+ https://review.opendev.org/744793 | 12:52 |
| AJaeger | smcginnis: I don't think we can drop py27 support at all, we have no way to limit install for older releases AFAIK ^ | 12:56 |
| AJaeger | fungi, clarkb , please double check ^ | 12:57 |
| *** rcernin has joined #openstack-infra | 13:00 | |
| smcginnis | AJaeger: We have some odd failures now trying to run under py27. Maybe someone can figure that out to keep it going, but I'm afraid we may be forced into it at this point. | 13:02 |
| smcginnis | Previous job failures was "ImportError: No module named posixpath", which from what little I could find on that was some sort of compatibility problem between virtualenv, setuptools, and py27. | 13:03 |
| *** Adri2000 has quit IRC | 13:04 | |
| *** jcapitao_lunch is now known as jcapitao | 13:06 | |
| *** yamamoto has joined #openstack-infra | 13:07 | |
| AJaeger | Oh fun ;( | 13:07 |
| *** derekh has joined #openstack-infra | 13:09 | |
| *** priteau has joined #openstack-infra | 13:09 | |
| *** Adri2000 has joined #openstack-infra | 13:11 | |
| *** iurygregory has quit IRC | 13:12 | |
| *** yamamoto has quit IRC | 13:15 | |
| *** rcernin has quit IRC | 13:15 | |
| *** jamesmcarthur has joined #openstack-infra | 13:16 | |
| *** xek has joined #openstack-infra | 13:18 | |
| *** iurygregory has joined #openstack-infra | 13:18 | |
| *** udesale has quit IRC | 13:21 | |
| *** udesale has joined #openstack-infra | 13:22 | |
| openstackgerrit | Merged openstack/project-config master: Move non-voting neutron tempest jobs to separate graph https://review.opendev.org/743729 | 13:22 |
| openstackgerrit | Aurelien Lourot proposed openstack/project-config master: Mirror keystone-kerberos and ceph-iscsi charms to GitHub https://review.opendev.org/744890 | 13:29 |
| openstackgerrit | Sean McGinnis proposed openstack/pbr master: Drop py27 support and fix compatiblity with virtualenv 20.x+ https://review.opendev.org/744793 | 13:45 |
| *** kaisers has joined #openstack-infra | 13:47 | |
| *** dave-mccowan has joined #openstack-infra | 13:49 | |
| openstackgerrit | Sean McGinnis proposed openstack/pbr master: Drop py27 support and fix compatiblity with virtualenv 20.x+ https://review.opendev.org/744793 | 13:58 |
| clarkb | AJaeger: smcginnis correct we cannot drop pbr's python2 support without breaking stable | 13:59 |
| clarkb | I commented similarly on sean mooney's change a while back | 13:59 |
| smcginnis | clarkb: OK. Then someone is going to have to figure out how to get it to actually work. It's completely broken right now. | 13:59 |
| *** dave-mccowan has quit IRC | 14:00 | |
| clarkb | that may be true too, but pbr is a setup requires installed by easy_install which means we dont really get to control its versions in the wild. that means anyone trying to useit with a python2 project, like openstack stable branches will break if we drop python2 support | 14:00 |
| *** ociuhandu has quit IRC | 14:04 | |
| openstackgerrit | Oleksandr Kozachenko proposed openstack/project-config master: Add openstack/barbican in required project list of vexxhost https://review.opendev.org/744909 | 14:04 |
| *** ociuhandu has joined #openstack-infra | 14:05 | |
| *** xek has quit IRC | 14:06 | |
| *** ociuhandu has quit IRC | 14:10 | |
| *** jamesmcarthur has quit IRC | 14:18 | |
| *** sshnaidm is now known as sshnaidm|afk | 14:24 | |
| *** redrobot has joined #openstack-infra | 14:32 | |
| *** lmiccini has quit IRC | 14:34 | |
| *** dklyle has joined #openstack-infra | 14:37 | |
| *** lmiccini has joined #openstack-infra | 14:49 | |
| *** xek has joined #openstack-infra | 14:54 | |
| fungi | smcginnis: clarkb: AJaeger: sean-k-mooney: perhaps this is a repugnant idea, but we could start publishing py3-only releases of pbr under a different name from earlier py2-supporting releases, then projects reference one or the other | 14:54 |
| fungi | more likely this is something pep 517/518 backend handling would address, by making sure the correct pbr version was selected based on the interpreter in use | 14:56 |
| *** psachin has quit IRC | 14:57 | |
| *** hashar has joined #openstack-infra | 15:01 | |
| clarkb | fungi: aiui those pep's don't help because easy_install is used not pip | 15:01 |
| clarkb | the upstream solution to this problem is pyproject.toml specifying packaging dependencies with rich version control | 15:01 |
| clarkb | splitting the package names seems reasonable, but also it shouldn't be that difficult to fix this issue in pbr, iirc its a test only problem | 15:02 |
| clarkb | if nothing else we can simply run a couple tests less under python2 | 15:02 |
| clarkb | the test suite creates a venv to install a fake package into then check the results. We need to use appropriate virtualenv and api depending on the python version. We can select that with python version constraints like sdk does forexample. We can drop the test under python2. We have options | 15:03 |
| fungi | yep, or run different tests written for py2 | 15:03 |
| * clarkb runs the test suite locally | 15:05 | |
| sean-k-mooney | fungi: im not sure why that would be required | 15:06 |
| sean-k-mooney | fungi: is there an issue you are trying to workaround by that approch | 15:07 |
| sean-k-mooney | oh is this realted to the lower constriats issues? | 15:08 |
| sean-k-mooney | and the recnet chagne to virtualenv and pip? | 15:08 |
| fungi | sean-k-mooney: separate, i think | 15:12 |
| fungi | that was just a regression in pip where it failed to match on dist names with a "." them when looking to see what's installed | 15:13 |
| clarkb | I'm working on a patch | 15:14 |
| sean-k-mooney | ya im aware of that issue too | 15:15 |
| clarkb | if we drop the constraints and build a python27 specific list of deps then we end up with 2 test failures | 15:15 |
| clarkb | one is virtualenv api change which we can fix in the test the other I'm not sure yet | 15:15 |
| openstackgerrit | Merged openstack/project-config master: Add openstack/barbican in required project list of vexxhost https://review.opendev.org/744909 | 15:15 |
| *** ykarel is now known as ykarel|away | 15:21 | |
| *** chkumar|rover is now known as raukadah | 15:26 | |
| *** tosky has quit IRC | 15:31 | |
| *** lmiccini has quit IRC | 15:31 | |
| *** jamesmcarthur has joined #openstack-infra | 15:31 | |
| *** aedc has joined #openstack-infra | 15:34 | |
| *** aedc has quit IRC | 15:36 | |
| openstackgerrit | Clark Boylan proposed openstack/pbr master: Fix python2.7 testing https://review.opendev.org/744929 | 15:36 |
| clarkb | its not clear to me if the second test issue I've got is related to suse's python setup or not so starting with ^ | 15:36 |
| clarkb | in the test for setup develop we modify pythonpath and that breaks the process running under that path's ability to import posixpath on my system | 15:37 |
| clarkb | the virtualenv issue is addressed though so if upstream CI doesn't have the python path issues I think we're good | 15:37 |
| clarkb | oh even better sfinucane wrote https://review.opendev.org/#/c/719964/9 | 15:40 |
| clarkb | smcginnis: ^ I've rechecked that change so that we can get log files for the failing tests | 15:41 |
| *** lbragstad_ is now known as lbragstad | 15:42 | |
| clarkb | I'm guessing its a problem in https://review.opendev.org/#/c/719964/9/tools/integration.sh the changes there | 15:42 |
| dtantsur | folks, opensuse jobs started failing on bifrost on bindep state: https://zuul.opendev.org/t/openstack/build/962c4506bec64d3c85350f98f37cad35/log/job-output.txt#475 | 15:43 |
| dtantsur | any ideas? | 15:43 |
| fungi | "Zypper run command failed with return code 8." | 15:44 |
| fungi | that's... helpful | 15:44 |
| clarkb | https://zuul.opendev.org/t/openstack/build/962c4506bec64d3c85350f98f37cad35/console#1/0/15/opensuse-15 provides much more info | 15:44 |
| *** slaweq_ has joined #openstack-infra | 15:45 | |
| fungi | yeah, that's where i headed after seeing it swallowed the stdout | 15:45 |
| *** slaweq has quit IRC | 15:45 | |
| *** slaweq_ is now known as slaweq | 15:45 | |
| clarkb | it couldn't find a package | 15:46 |
| clarkb | possible we're mirroring bad state again due to upstream mirror problems? | 15:46 |
| fungi | i love the "Abort, retry, ignore?" in there. classic | 15:46 |
| fungi | reminds me of when dos couldn't read my floppy | 15:47 |
| fungi | i'm checking the rsync logs now | 15:47 |
| fungi | looks like we're successfully mirroring from mirror.us.leaseweb.net/opensuse/update/leap/15.1/ | 15:50 |
| *** jcapitao has quit IRC | 15:50 | |
| fungi | also mirror.us.leaseweb.net/opensuse/distribution/leap/15.1/repo/ | 15:50 |
| fungi | so we're expecting to find oss/suse/x86_64/libJudy1-1.0.5-lp151.2.2.x86_64.rpm there i guess | 15:51 |
| dtantsur | Tried in a podman container, ended up with libJudy1-1.0.5-lp152.4.1.x86_64.rpm | 15:51 |
| dtantsur | hmmm, I guess I'm on 15.2 already | 15:51 |
| fungi | https://mirror.us.leaseweb.net/opensuse/distribution/leap/15.1/repo/oss/suse/x86_64/libJudy1-1.0.5-lp151.2.2.x86_64.rpm is also a 404 | 15:52 |
| clarkb | the label in use there is opensuse-15 whihc should be latest opensuse | 15:52 |
| fungi | oh, that may be the wrong path | 15:52 |
| clarkb | I wonder if that means our image builds have broken | 15:52 |
| clarkb | but I'd still expect the 15.1 mirror to work, it is likely a packge ordering sync issue | 15:53 |
| clarkb | we do have a way to verify the package lists against indexes and dep lists with rpm mirrors | 15:53 |
| fungi | https://mirror.us.leaseweb.net/opensuse/distribution/leap/15.1/repo/oss/ | 15:53 |
| fungi | i don't see a suse there | 15:53 |
| clarkb | fungi: ya so broken package index? | 15:54 |
| clarkb | https://mirror.us.leaseweb.net/opensuse/distribution/leap/15.1/repo/oss/x86_64/libJudy1-1.0.5-lp151.2.2.x86_64.rpm is there | 15:55 |
| fungi | maybe | 15:55 |
| *** dave-mccowan has joined #openstack-infra | 15:55 | |
| clarkb | could also be a mirror config on the client side I suppose, constructing package paths relative to the mirror root wrong | 15:55 |
| *** pkopec has quit IRC | 15:55 | |
| fungi | https://mirror.us.leaseweb.net/opensuse/distribution/leap/15.1/repo/oss/x86_64/ exists | 15:55 |
| fungi | yeah, there's a https://mirror.us.leaseweb.net/opensuse/distribution/leap/15.1/repo/oss/x86_64/libJudy1-1.0.5-lp151.2.2.x86_64.rpm | 15:56 |
| fungi | similarly, we have https://mirror.ca-ymq-1.vexxhost.opendev.org/opensuse/distribution/leap/15.1/repo/oss/x86_64/libJudy1-1.0.5-lp151.2.2.x86_64.rpm | 15:57 |
| fungi | so removing the "suse" from the path it tried to fetch would have worked, i think | 15:57 |
| *** xek_ has joined #openstack-infra | 15:57 | |
| *** ykarel|away has quit IRC | 15:58 | |
| *** xek has quit IRC | 16:00 | |
| *** lucasagomes has quit IRC | 16:00 | |
| fungi | i wonder if opensuse reorganized their repositories? | 16:02 |
| fungi | but yeah i'm leaning toward it being an index problem or a bad mirror | 16:04 |
| clarkb | on my local tumbleweed intsall http://download.opensuse.org/tumbleweed/repo/oss/ is the zypper repo config | 16:04 |
| fungi | we set our baseurl to {{ package_mirror }}/tumbleweed/repo/oss/ | 16:04 |
| clarkb | there isn't another value to set sub urls | 16:04 |
| clarkb | so yes I thik this is a suse mirroring/index problem | 16:04 |
| fungi | or {{ package_mirror }}/distribution/leap/$releasever/repo/oss/ | 16:04 |
| fungi | in this particular case | 16:04 |
| fungi | https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/configure-mirrors/tasks/mirror/Suse.yaml#L3 | 16:05 |
| fungi | https://download.opensuse.org/distribution/leap/15.1/repo/oss/ similarly has no "suse" directory | 16:06 |
| clarkb | I was just going to link that :) | 16:06 |
| fungi | gonna try picking apart the index | 16:07 |
| fungi | pretty simple. INDEX.gz has "./x86_64/libJudy1-1.0.5-lp151.2.2.x86_64.rpm" in the official site | 16:07 |
| clarkb | the index for download.opensuse.org shows no suse/ prefix | 16:08 |
| fungi | right | 16:08 |
| clarkb | our mirror is the same | 16:08 |
| fungi | i agree | 16:08 |
| clarkb | possible it would work now and whatever that was was a temporary index bug? | 16:08 |
| *** hamalq has joined #openstack-infra | 16:08 | |
| fungi | yeah, i'm not seeing why it would try to grab that path at all | 16:09 |
| *** hamalq has quit IRC | 16:10 | |
| *** hamalq has joined #openstack-infra | 16:11 | |
| *** jpena is now known as jpena|off | 16:12 | |
| *** eolivare has quit IRC | 16:20 | |
| openstackgerrit | Clark Boylan proposed openstack/pbr master: Add support for virtualenv 20.x https://review.opendev.org/719964 | 16:21 |
| clarkb | smcginnis: sfinucane ^ that updates the integration.sh script to fix what I think is at least part of the problem (we need to source the venv in orderto operate within it under process forks since stestr will fork) | 16:21 |
| *** dave-mccowan has quit IRC | 16:26 | |
| *** markvoelker has joined #openstack-infra | 16:27 | |
| *** tosky has joined #openstack-infra | 16:29 | |
| openstackgerrit | Sean McGinnis proposed openstack/project-config master: Gerritbot: only comment on stable:follows-policy repos https://review.opendev.org/744947 | 16:30 |
| *** udesale has quit IRC | 16:32 | |
| fungi | weshay|ruck: so the more thorough caching hasn't completely prevented dockerhub's throttling i guess? | 16:37 |
| fungi | and idea if it has reduced the failure rate at least? | 16:38 |
| fungi | er, any idea | 16:38 |
| *** dtantsur is now known as dtantsur|afk | 16:39 | |
| * weshay|ruck looks again | 16:39 | |
| clarkb | the email seemd to indicate it may be related to all the auth requests? | 16:40 |
| clarkb | we're only caching the layers themselves so its possible that those extra requests are still a problem. Also the jobs were fetching somelayers many times | 16:40 |
| clarkb | (and each fetch results in requests to dockerhub even if cached) | 16:40 |
| fungi | ahh, yep. so maybe reducing the request volume by half and returning images much faster, but not an overwhelming (like order of magnitude) reduction in requests | 16:42 |
| weshay|ruck | logstash is telling me it's worse than it was, but I suspect the old data may be purged | 16:43 |
| weshay|ruck | thought I had some screenshots of the original queries I posted | 16:44 |
| clarkb | weshay|ruck: yes, we only have a 7 day retention there | 16:44 |
| weshay|ruck | aye.. I don't think it's as bad.. | 16:44 |
| *** gfidente is now known as gfidente|afk | 16:44 | |
| weshay|ruck | still enough to have issues landing patches :) | 16:44 |
| *** gyee has joined #openstack-infra | 16:48 | |
| clarkb | I'm goingto pop out on a bike ride now. The latest ps for the pbr fix is happy so far | 16:58 |
| clarkb | I'll check on it when I get back | 16:58 |
| *** derekh has quit IRC | 17:00 | |
| *** hamalq has quit IRC | 17:10 | |
| *** hamalq has joined #openstack-infra | 17:11 | |
| *** Lucas_Gray has quit IRC | 17:12 | |
| *** yamamoto has joined #openstack-infra | 17:14 | |
| *** hamalq has quit IRC | 17:18 | |
| *** ociuhandu has joined #openstack-infra | 17:18 | |
| *** yamamoto has quit IRC | 17:18 | |
| openstackgerrit | Jeremy Stanley proposed openstack/pbr master: Update python requires packaging metadata for package https://review.opendev.org/715917 | 17:19 |
| *** sshnaidm|afk is now known as sshnaidm | 17:19 | |
| *** nightmare_unreal has quit IRC | 17:20 | |
| *** ociuhandu has quit IRC | 17:22 | |
| *** andrewbonney has quit IRC | 17:26 | |
| *** priteau has quit IRC | 17:38 | |
| *** hamalq has joined #openstack-infra | 17:44 | |
| *** hamalq has joined #openstack-infra | 17:46 | |
| *** dciabrin has quit IRC | 17:57 | |
| *** dciabrin has joined #openstack-infra | 17:58 | |
| fungi | clarkb: so close. looks like it's failing devstack jobs | 18:00 |
| fungi | AssertionError: Failed process args=(['python', 'update.py', '/opt/stack/new/python-swiftclient'],), kwargs={'cwd': '/opt/stack/new/requirements', 'stderr': -1, 'stdin': -1, 'stdout': -1}, returncode=1 | 18:04 |
| fungi | AttributeError: module 'importlib._bootstrap' has no attribute 'SourceFileLoader' | 18:05 |
| fungi | legacy devstack-gate job setup too | 18:07 |
| *** ralonsoh has quit IRC | 18:12 | |
| *** aedc has joined #openstack-infra | 18:30 | |
| *** aedc has quit IRC | 18:30 | |
| fungi | so the error is arising when pbr tries to import pkg_resources, and it seems to depend on an internal which is not present in the importlib available for that stdlib build | 18:45 |
| fungi | mewer setuptools issue maybe? | 18:48 |
| fungi | newer | 18:48 |
| fungi | oh, this seems to be old setuptools trying to do pre-python3.6 things with importlib on python3.6 | 18:52 |
| fungi | at least based on web search results i'm finding for that particular exception | 18:52 |
| fungi | though this is in a virtualenv | 18:55 |
| fungi | yeesh, we're getting invalid syntax exceptions from setuptools too | 18:56 |
| fungi | https://zuul.opendev.org/t/openstack/build/140ab2af4279400bb1f2089969bf722f/log/job-output.txt#7081-7127 | 18:57 |
| *** stevebaker has joined #openstack-infra | 18:57 | |
| fungi | https://pypi.org/project/setuptools/0.7.2/ Released: Jun 9, 2013 | 18:59 |
| fungi | umm... | 18:59 |
| fungi | oh, hah, this is test_pip_versions(setuptools-minimum) | 19:00 |
| fungi | okay, so basically there's just no way some of these tests are going to run on newer python | 19:00 |
| fungi | well, run successfully anyway | 19:00 |
| fungi | they need to be tested on much older platforms/interpreters | 19:00 |
| clarkb | we can probably just skip based on the pyth9n verson | 19:03 |
| clarkb | that change already does that for another test | 19:03 |
| fungi | well, i also question the utility of running those tests at all since the job only runs on bionic/3.6 | 19:03 |
| clarkb | oh good point | 19:04 |
| clarkb | unless its something we want to update to work on python3 I dont think we need it | 19:04 |
| fungi | i understand the desire to test that pbr will work with really old pip/setuptools | 19:05 |
| *** xek_ has quit IRC | 19:06 | |
| fungi | but i think we'd need different jobs (or at least different platforms) to do that | 19:07 |
| clarkb | ++ | 19:07 |
| clarkb | did you want to work on a new ps or should I try that after some lunch? | 19:07 |
| fungi | i'll see what needs skipping and whip up a new rev for now | 19:11 |
| fungi | but those jobs also flat out need replacing | 19:11 |
| clarkb | I think the intent of those jobs is that pbr installs openstack packages in devstack successfully then does some additionally checking with is own integration tests? | 19:12 |
| clarkb | its been so long so I had to page pbr things in | 19:12 |
| fungi | right | 19:12 |
| fungi | same here really | 19:12 |
| *** hashar has quit IRC | 19:32 | |
| fungi | one of these things is not like the other... | 19:34 |
| fungi | https://zuul.opendev.org/t/openstack/build/140ab2af4279400bb1f2089969bf722f/log/job-output.txt#7446-7508 | 19:34 |
| fungi | that's got me stumped | 19:34 |
| clarkb | fungi: looking at the test case more closely I think what we want to do is update the scenarios list. Like we don't need trusty but bionic and focal may be reasonable as would rhel8 | 19:34 |
| fungi | yep, the bulk of the failures are test_pip_versions (every scenario except pip-latest) and test_lts_venv_default_versions (every scenario except WheezyPrecise because that one is skipped) | 19:36 |
| clarkb | fungi: that test seems to list all of the repodirs that devstack manages the ninstals all of them. it then fails on swiftclient for a reason I don't understand yet | 19:36 |
| fungi | right, that's basically where i got to with it | 19:37 |
| clarkb | its the python update.py script that is failing | 19:38 |
| clarkb | I don't know what that is yet | 19:38 |
| clarkb | I think that is the requirements sync thing but we've since relaxed those rules? | 19:38 |
| clarkb | this feels like pbr's test suite wasn't updated when the requirements syncing rules changes (and I'll be honest I don't quit egrok the new rules) | 19:38 |
| fungi | anyway, i'll first try to get the scenario lists for test_pip_versions and test_lts_venv_default_versions updated, though i'm torn on whether to delete tests that won't work with modern python or just comment them out, or try to work out what interpreter versions *can* support them and set pyver-based skips beyond just the py3support one | 19:38 |
| *** tosky has quit IRC | 19:39 | |
| fungi | i'm inclined to say that unit tests are the logical way for pbr to maintain support for platforms with older pip/setuptools, and just rip these out of the integration testing | 19:40 |
| clarkb | fungi: I would delete the broken ones then add bionic, focal, and rhel8 cases | 19:40 |
| clarkb | those + latest should be reasonable coverage | 19:40 |
| fungi | yeah, working on it now | 19:40 |
| fungi | i guess that also allows ripping out some of this python3-based skipping | 19:40 |
| clarkb | ya that update.py is from openstack/requirements | 19:41 |
| clarkb | does that script evenexist anymore | 19:42 |
| clarkb | yes it does | 19:42 |
| clarkb | and futures is a python2 only dep and update.py doesn't seem to know how to handle that | 19:43 |
| clarkb | given the new requirements rules is more accurate to just install the package and pass in upper-constraints? | 19:44 |
| fungi | yeah, i think so | 19:44 |
| clarkb | in that case I think that is how you/I/we should update that test case | 19:44 |
| fungi | swift has basically forked upper constraints anyway, i believe, to have their own python 2.7 constraints | 19:45 |
| clarkb | yup | 19:45 |
| clarkb | it also does an unncessary git clone if we're not going to modify the repo | 19:45 |
| clarkb | so we can rip that out then modify PIP_CMD to inclue -c $PATH_TO_UPPER_CONSTRAINTS | 19:46 |
| clarkb | fungi: should I work on that while you do the pip scenarios or would you like to do both? | 19:46 |
| fungi | you can work on that if you want and then we can combine. it's taking me a while to look up the right package versions | 19:48 |
| fungi | or i can work on that when i'm done with the scenarios | 19:48 |
| *** jamesmcarthur has quit IRC | 19:49 | |
| clarkb | k I'll take a look at it its appears pretty invovled | 19:49 |
| *** jamesmcarthur has joined #openstack-infra | 19:49 | |
| *** jamesmcarthur has quit IRC | 19:54 | |
| clarkb | fungi: I think I'm ready to push should I do that then you rebaes or do you want me to rebase on your ps? | 20:00 |
| fungi | i'm pushing now | 20:00 |
| clarkb | k I'll rebase | 20:00 |
| openstackgerrit | Jeremy Stanley proposed openstack/pbr master: Add support for virtualenv 20.x https://review.opendev.org/719964 | 20:00 |
| fungi | clarkb: ^ | 20:00 |
| fungi | hopefully rebases clean | 20:00 |
| openstackgerrit | Clark Boylan proposed openstack/pbr master: Add support for virtualenv 20.x https://review.opendev.org/719964 | 20:03 |
| clarkb | yup was clean | 20:03 |
| clarkb | there is a possibility I'll have permissions issues with my changes | 20:03 |
| fungi | i guess we'll find out | 20:03 |
| clarkb | I'm not sure if the clone into a tmpdir is necessary for htat reason but if we can avoid it it should speed tests up so we'll try avoiding it first | 20:04 |
| fungi | this is turning into quite the pbr-neglected-ci-updates bundle | 20:04 |
| clarkb | probably a good thing considering how out of date things were | 20:04 |
| clarkb | now we'll be testing more real world situations | 20:05 |
| fungi | yep. also i felt bad rechecking some of my more trivial pbr changes for months to see if someone else had fixed the jobs yet rather than rolling up my sleeves | 20:05 |
| clarkb | there is a pep8 error. checking nit now | 20:14 |
| fungi | might have been me, i forgot to check it | 20:14 |
| clarkb | or me I didn't check either :) | 20:14 |
| clarkb | was me | 20:15 |
| openstackgerrit | Clark Boylan proposed openstack/pbr master: Add support for virtualenv 20.x https://review.opendev.org/719964 | 20:16 |
| fungi | AttributeError: 'FlakesChecker' object has no attribute 'CONSTANT' | 20:16 |
| fungi | i can't even run it locally anyway | 20:16 |
| clarkb | fungi: yup I have the same issue | 20:16 |
| clarkb | so I'm relying on zuul | 20:16 |
| clarkb | going to look at hte local error next | 20:16 |
| fungi | i guess i need to apply constraints | 20:16 |
| clarkb | oh ya that is probably it | 20:17 |
| clarkb | I wish we'd put upper caps on things when they just don't work | 20:17 |
| clarkb | but we lost that with the changes to syncing? | 20:17 |
| fungi | looks like this should have been fixed by pyflakes 1.4, but we're running waaaay old pyflakes | 20:20 |
| fungi | https://github.com/PyCQA/pyflakes/pull/80 | 20:20 |
| *** jamesmcarthur has joined #openstack-infra | 20:21 | |
| fungi | so it's been fixed in pyflakes for around 4 years | 20:21 |
| clarkb | ah if we bump the hacking value (then fix all the problems that exposes) we may be good? | 20:21 |
| * clarkb tries that | 20:21 | |
| fungi | pbr is just set for waaay old pyflakes (maybe because of way old hacking>=1.1.0,<1.2.0) | 20:21 |
| clarkb | ya | 20:21 |
| clarkb | I've removed the cap and am trying it now to see what it complains about | 20:21 |
| smcginnis | That's right, I hit that too and forgot to go back and check. | 20:22 |
| clarkb | the pain is actually minimal I'll fix that up and push a follow on change | 20:22 |
| openstackgerrit | Oleksandr Kozachenko proposed openstack/project-config master: Add openstack/barbican-tempest-plugin to vexxhost https://review.opendev.org/744982 | 20:23 |
| fungi | yeah, it doesn't seem to be blocking other changes so no need to squash | 20:24 |
| *** vishalmanchanda has quit IRC | 20:25 | |
| openstackgerrit | Clark Boylan proposed openstack/pbr master: Update hacking https://review.opendev.org/744983 | 20:26 |
| fungi | i guess it needs newer python than bionic provides to show up, so i was seeing it because 3.8 is my current default | 20:35 |
| clarkb | ya I have 3.8 too | 20:35 |
| *** gyee has quit IRC | 20:36 | |
| *** gyee has joined #openstack-infra | 20:38 | |
| *** armax has quit IRC | 20:46 | |
| *** smcginni1 has joined #openstack-infra | 20:47 | |
| *** smcginnis has quit IRC | 20:50 | |
| *** smcginni1 is now known as smcginnis | 20:50 | |
| openstackgerrit | Merged openstack/project-config master: Add openstack/barbican-tempest-plugin to vexxhost https://review.opendev.org/744982 | 20:51 |
| *** ociuhandu has joined #openstack-infra | 20:51 | |
| smcginnis | pbr-installation-devstack failed | 20:52 |
| clarkb | looking at it, not entirely unexecpted | 20:53 |
| clarkb | (we made some fairly large changes to it) | 20:53 |
| smcginnis | Yeah | 20:53 |
| clarkb | its the use of constraints because we are installing things that a constrained | 20:54 |
| clarkb | so we need to do the thing where we remove the thing being installed from constraints | 20:54 |
| * clarkb is working on that | 20:55 | |
| clarkb | first checking how devstack deals with that | 20:55 |
| *** ociuhandu has quit IRC | 20:56 | |
| *** armax has joined #openstack-infra | 20:58 | |
| clarkb | I think in this case because we install to different venvs each time we can simply remove the current package from the constraints file | 20:59 |
| *** slaweq has quit IRC | 21:18 | |
| openstackgerrit | Clark Boylan proposed openstack/pbr master: Add support for virtualenv 20.x https://review.opendev.org/719964 | 21:19 |
| openstackgerrit | Clark Boylan proposed openstack/pbr master: Update hacking https://review.opendev.org/744983 | 21:19 |
| clarkb | lets see if ^ is any better | 21:19 |
| *** xek has joined #openstack-infra | 21:19 | |
| *** gyee has quit IRC | 21:24 | |
| *** tosky has joined #openstack-infra | 21:25 | |
| *** gyee has joined #openstack-infra | 21:25 | |
| smcginnis | clarkb: Just commented. I see a potential issue there. | 21:27 |
| smcginnis | Could also address in a follow up if we see this rev works. | 21:27 |
| clarkb | smcginnis: ah ya I think it may work as is because constraints are a layer of security not necessarily required, but thats a good point. re python2 I think it would be good to run a old devstack + python2 job to ensure we don't break stable too | 21:29 |
| clarkb | but I wasn't going to try and solve that immediately | 21:29 |
| clarkb | if this change works as is I'll do a followon otherwise incorporate better name checks in a new ps | 21:30 |
| smcginnis | ++ | 21:32 |
| clarkb | if project_name != line.split('===')[0]: <- I think that should work | 21:33 |
| *** rcernin has joined #openstack-infra | 21:35 | |
| *** rcernin has quit IRC | 21:36 | |
| *** rcernin has joined #openstack-infra | 21:36 | |
| *** yamamoto has joined #openstack-infra | 21:45 | |
| *** xek has quit IRC | 21:47 | |
| fungi | yeah, or use edit-constraints | 21:49 |
| clarkb | fungi: I was going to do that but it requires installing requirements first and meh | 21:49 |
| clarkb | that said https://opendev.org/openstack/glance_store/src/branch/master/setup.cfg#L2 is broken and breaking us :( | 21:50 |
| clarkb | our constraints have glance-store not glance_store which I think means the normalized form is glance-store not glance_store? | 21:50 |
| fungi | oh, yeah great point | 21:50 |
| clarkb | https://pypi.org/project/glance-store/ that seems to agree | 21:50 |
| clarkb | yay we're finding bugs in other packages now too :) | 21:50 |
| fungi | integration tests do that | 21:51 |
| clarkb | I've got a simple workaround for that | 21:51 |
| clarkb | new patch shortly | 21:51 |
| *** adam_g has quit IRC | 21:52 | |
| clarkb | or atleast I think there is a way to normalize the name via setuptools | 21:52 |
| clarkb | trying to figure it out | 21:52 |
| *** markmcclain has quit IRC | 21:53 | |
| fungi | ooh, yeah there be dragons | 21:53 |
| fungi | pip.util.normalize_name() used to be an option... checking to see if it still is | 21:55 |
| fungi | AttributeError: module 'pip' has no attribute 'util' | 21:56 |
| fungi | maybe packaging took that over now | 21:56 |
| clarkb | pkg_resources.safe_name() | 21:57 |
| fungi | packaging.utils.canonicalize_name('foo-bar.baz') | 21:57 |
| fungi | 'foo-bar-baz' | 21:57 |
| clarkb | what is packaging? | 21:58 |
| fungi | interesting that packaging.utils.canonicalize_name() changes . to - but pkg_resources.safe_name() leaves . alone | 21:59 |
| clarkb | also won't that break oslo.db? | 21:59 |
| clarkb | yes I think that is a bug in packaging | 21:59 |
| clarkb | https://pypi.org/project/oslo.db/ no - there | 21:59 |
| fungi | packaging is supposed to be the new library that is supplanting pkg_resources and which utility functions for pip are being extracted into | 21:59 |
| clarkb | (in the rendered version not just the url) | 21:59 |
| clarkb | fungi: gotcha well its broken I think :) | 21:59 |
| fungi | seems so, yes | 21:59 |
| fungi | https://pypi.org/project/packaging/ "Core utilities for Python packages. The packaging project includes the following: version handling, specifiers, markers, requirements, tags, utilities." | 22:00 |
| fungi | https://github.com/pypa/packaging | 22:01 |
| fungi | anyway, any time someone wants to call into a pip or setuptools internal function, they're told to stick it in packaging instead | 22:01 |
| openstackgerrit | Clark Boylan proposed openstack/pbr master: Add support for virtualenv 20.x https://review.opendev.org/719964 | 22:01 |
| openstackgerrit | Clark Boylan proposed openstack/pbr master: Update hacking https://review.opendev.org/744983 | 22:01 |
| clarkb | fungi: maybe we should file a bug? | 22:01 |
| clarkb | because pypi seems to say the canonical name doesn't replace the . with - | 22:01 |
| fungi | https://github.com/pypa/packaging/blob/master/docs/utils.rst#reference | 22:02 |
| clarkb | oslo-db redirects to oslo.db | 22:02 |
| clarkb | so either a bug in packages or warehouse | 22:02 |
| fungi | wow, it even lists oslo.concurrency -> oslo-concurrency as an example in their docs | 22:02 |
| clarkb | I still think its wrong since pypi sort of sets the expectations there | 22:02 |
| clarkb | eg if it is a warehouse bug then we're likeyl stuck with it | 22:02 |
| clarkb | and other tools should accomodate | 22:03 |
| *** rcernin has quit IRC | 22:03 | |
| clarkb | I'm going to step out for a bit while that latest iteration churns. I think we're very close | 22:04 |
| fungi | https://www.python.org/dev/peps/pep-0503/#id4 | 22:04 |
| fungi | yeesh | 22:04 |
| fungi | er, i meant https://www.python.org/dev/peps/pep-0503/#normalized-names | 22:04 |
| fungi | re.sub(r"[-_.]+", "-", name).lower() | 22:05 |
| clarkb | why does pypi.org get it wrong then | 22:05 |
| clarkb | I mean thats my gripe. The library and spec can say all they want but if the canonical server of the packages disagrees... | 22:05 |
| fungi | i asked in #pypa-dev just now | 22:09 |
| fungi | no idea if anyone's paying attention | 22:09 |
| fungi | i may ask on the distutils-sig ml they're making noise about closing down (because who would want to use mailing lists when they have a nice shiny discord webforum now?) | 22:10 |
| clarkb | when we hvae an answer we might want to normalize all the package names in setup.cfg | 22:10 |
| clarkb | just for our own sanity | 22:10 |
| fungi | or is it discourse? i can't keep the two straight | 22:10 |
| *** yamamoto has quit IRC | 22:11 | |
| clarkb | discord is irc voip chat tool | 22:13 |
| clarkb | discourse is forum tool | 22:13 |
| *** rcernin has joined #openstack-infra | 22:17 | |
| *** rlandy is now known as rlandy|bbl | 22:17 | |
| fungi | discord is also a proprietary webforum/chat saas thing | 22:28 |
| *** rcernin has quit IRC | 22:40 | |
| *** rcernin has joined #openstack-infra | 22:40 | |
| *** rcernin has quit IRC | 22:40 | |
| *** rcernin has joined #openstack-infra | 22:44 | |
| sean-k-mooney | discord was created for pc gaming to have a conventint tool to use for voice chat while in game | 22:45 |
| sean-k-mooney | they rounded it out by adding text and then later video change with the ablity to do screen sharing | 22:45 |
| *** artom has quit IRC | 22:45 | |
| sean-k-mooney | its basically a cross between mumble/teamspeck and slack | 22:46 |
| *** artom has joined #openstack-infra | 22:46 | |
| clarkb | it hsa the wonderful feature of being able to control volume for each participant on your client | 22:46 |
| clarkb | I'm always sad more tools don't have this (I think jitsi actually has it too) | 22:46 |
| sean-k-mooney | yes | 22:46 |
| clarkb | looks like the pbr integration test job will pass \o/ | 22:47 |
| clarkb | ok really popping out now for a few | 22:47 |
| sean-k-mooney | its been long requested to allow you to run your own server to use it in enterpises | 22:47 |
| sean-k-mooney | as in many many people have said to the discord amintianer i would liek to pay you to be able to run this on my own servers but they want to keep it as software as a service | 22:48 |
| sean-k-mooney | clarkb: i was talkign to smcginnis about transitive bin-dep dependicnes. | 22:48 |
| sean-k-mooney | i was wondering if pbr only only used when creating the package or if its also executed when installing the package | 22:49 |
| sean-k-mooney | the reason i ask is i was wondering it it would make sense to be abel to have it run bindeps pip install | 22:50 |
| sean-k-mooney | but i also dont know if that could work | 22:50 |
| sean-k-mooney | basically i was wondering if pbr runs during the pip install could it check if the installed package had a bin-dep file and then optional install the packages. | 22:52 |
| sean-k-mooney | so something like "bin-dep-install=true pip install os-brick" | 22:53 |
| sean-k-mooney | i.e. if a env var is set have pbr do the install | 22:53 |
| sean-k-mooney | i think pbr is only used on the inital build stepp so when the wheel is created it wont run after that point so i assuem this wont work regardless of if it is a good idea or not | 22:55 |
| fungi | sean-k-mooney: pbr is run by pip, so won't generally have access to install system packages unless you're running pip as root | 22:58 |
| fungi | but also, bindep doesn't install anything. it's a tool to list what distro package dependencies you're missing | 22:58 |
| *** jamesmcarthur has quit IRC | 22:59 | |
| sean-k-mooney | ya i know it just list the packages although i kind of assuem to knwo what is missing it need to be able to list the packages so had some awareness of the packagemanager | 22:59 |
| fungi | and usually those deps are going to be install-time deps for whatever you're installing with pip, so the best you might hope for is a better error message saying the package wasn't installed because you're missing mysql-dev or whatever | 22:59 |
| *** tkajinam has joined #openstack-infra | 23:00 | |
| fungi | the poster child for bindep, what actually caused it to be created at all, was to find out if you have all the distro packages needed to run tox -e pyXX for nova | 23:01 |
| sean-k-mooney | the issue i have with listing transitive deps for optional libs is that it makes thing like nova have very long bindep files | 23:01 |
| sean-k-mooney | well many fo the packages in nova bindeps are not need for the unit tests | 23:02 |
| fungi | so... *if* you shipped the bindep files inside the python packages (which we don't because where would they get installed to?) you'd still need to recursively install python packages to collect and aggregate their bindep files recursively | 23:02 |
| sean-k-mooney | what i was suggesting to smcginnis was that we add more profiles to bindep.txt | 23:03 |
| sean-k-mooney | so that if we do add transitive depencies we put them into a profile | 23:04 |
| fungi | and yeah, not all the entries in a given bindep file are going to be necessary for any one purpose. the design expects that yuo'll tag dependencies for different scenarios (running unit tests, building docs, whatever) | 23:04 |
| sean-k-mooney | so that the os-brick once would be say listed under [cinder] | 23:04 |
| sean-k-mooney | since os-brick is not need unless you use nova with cinder | 23:04 |
| fungi | then how do you identify what packages you need to be able to run os-brick unit tests? install cinder first? | 23:05 |
| sean-k-mooney | no so os-brick an list the packages it need for its unit tests | 23:05 |
| fungi | oh, you mean os-brick's binary distro package dependencies listed in nova's bindep.txt | 23:05 |
| sean-k-mooney | yes | 23:05 |
| fungi | yeah, tagging the entries in nova's bindep.txt with specific profiles for different test cases is the intended design | 23:06 |
| fungi | similar to "extras" in python package metadata | 23:06 |
| sean-k-mooney | well they shoudl not be needed for any unit test. maybe fucntional at a streach | 23:06 |
| sean-k-mooney | * nova unit tests | 23:07 |
| sean-k-mooney | calls to os-brick shoudl really be mocked out | 23:07 |
| sean-k-mooney | anyway it sound liek there isnt an easy way to make bindep work transitivly with other python packages | 23:08 |
| sean-k-mooney | im going to call it a night o/ | 23:09 |
| fungi | no, not unless someone wants to file a pep to get bindep lists added as first-class metadata for python packages so that they're available from the pypi/warehouse api | 23:09 |
| fungi | otherwise anything you put *inside* a python package needs you to install that package to find out what the contents are | 23:09 |
| fungi | or at least retrieve and unpack the package | 23:09 |
| fungi | which is basically all wheel installation is | 23:10 |
| *** ociuhandu has joined #openstack-infra | 23:10 | |
| *** markmcclain has joined #openstack-infra | 23:12 | |
| *** adam_g has joined #openstack-infra | 23:12 | |
| *** tosky has quit IRC | 23:14 | |
| *** ociuhandu has quit IRC | 23:16 | |
| fungi | sean-k-mooney: for a bit of irony, this was just posted to distutils-sig moments ago: https://mail.python.org/archives/list/distutils-sig@python.org/thread/MSS42UYQ7FJWHID54FXSW5M5KCMK7ZQI/ | 23:27 |
| *** guilhermesp has joined #openstack-infra | 23:29 | |
| *** irclogbot_1 has quit IRC | 23:34 | |
| *** irclogbot_3 has joined #openstack-infra | 23:37 | |
| *** yoctozepto3 has joined #openstack-infra | 23:44 | |
| *** yoctozepto has quit IRC | 23:45 | |
| *** yoctozepto3 is now known as yoctozepto | 23:45 | |
| *** rlandy|bbl is now known as rlandy | 23:53 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!