| *** tosky has quit IRC | 00:50 | |
| *** brinzhang has joined #openstack-requirements | 00:52 | |
| *** brinzhang has quit IRC | 00:57 | |
| *** brinzhang has joined #openstack-requirements | 00:57 | |
| *** brinzhang_ has joined #openstack-requirements | 01:20 | |
| *** brinzhang has quit IRC | 01:23 | |
| *** vishalmanchanda has joined #openstack-requirements | 01:33 | |
| *** udesale has joined #openstack-requirements | 04:32 | |
| *** evrardjp has quit IRC | 05:36 | |
| *** evrardjp has joined #openstack-requirements | 05:36 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements master: Updated from generate-constraints https://review.opendev.org/713573 | 06:48 |
|---|---|---|
| openstackgerrit | Matthew Thode proposed openstack/requirements master: Updated from generate-constraints https://review.opendev.org/713573 | 07:20 |
| *** mwhahaha has quit IRC | 08:09 | |
| *** mwhahaha has joined #openstack-requirements | 08:10 | |
| *** rpittau|afk is now known as rpittau | 08:11 | |
| *** e0ne has joined #openstack-requirements | 08:16 | |
| *** tosky has joined #openstack-requirements | 08:21 | |
| *** ralonsoh has joined #openstack-requirements | 08:53 | |
| *** yoctozepto has quit IRC | 09:07 | |
| *** yoctozepto0 has joined #openstack-requirements | 09:16 | |
| *** yoctozepto9 has joined #openstack-requirements | 09:27 | |
| *** yoctozepto0 has quit IRC | 09:27 | |
| *** e0ne has quit IRC | 09:36 | |
| *** e0ne has joined #openstack-requirements | 09:36 | |
| *** dtantsur|afk is now known as dtantsur | 09:39 | |
| *** vishalmanchanda has quit IRC | 10:03 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements stable/stein: update constraint for keystoneauth1 to new release 3.13.2 https://review.opendev.org/713818 | 10:20 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements master: update constraint for oslo.service to new release 2.1.0 https://review.opendev.org/713819 | 10:21 |
| *** vishalmanchanda has joined #openstack-requirements | 10:22 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements master: update constraint for taskflow to new release 4.1.0 https://review.opendev.org/713820 | 10:22 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements master: update constraint for oslo.vmware to new release 3.2.1 https://review.opendev.org/713821 | 10:23 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements master: update constraint for tooz to new release 2.2.0 https://review.opendev.org/713822 | 10:23 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements master: update constraint for oslo.policy to new release 3.0.1 https://review.opendev.org/713823 | 10:26 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements master: update constraint for oslo.log to new release 4.1.0 https://review.opendev.org/713824 | 10:26 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements master: update constraint for oslo.versionedobjects to new release 2.0.1 https://review.opendev.org/713825 | 10:27 |
| *** yoctozepto9 is now known as yoctozepto | 10:37 | |
| *** hberaud has quit IRC | 11:19 | |
| *** dtantsur is now known as dtantsur|afk | 11:20 | |
| *** rpittau is now known as rpittau|bbl | 11:30 | |
| *** ccamacho has quit IRC | 11:38 | |
| *** ccamacho has joined #openstack-requirements | 11:44 | |
| *** udesale_ has joined #openstack-requirements | 12:19 | |
| *** udesale has quit IRC | 12:20 | |
| *** e0ne_ has joined #openstack-requirements | 12:20 | |
| *** e0ne has quit IRC | 12:21 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements stable/rocky: update constraint for python-glanceclient to new release 2.13.2 https://review.opendev.org/713866 | 12:40 |
| *** hberaud has joined #openstack-requirements | 12:41 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements stable/rocky: update constraint for glance_store to new release 0.26.2 https://review.opendev.org/713868 | 12:45 |
| *** rpittau|bbl is now known as rpittau | 12:59 | |
| *** udesale_ has quit IRC | 14:00 | |
| rm_work | prometheanfire: I don't understand your comment on https://review.opendev.org/#/c/712187/ ? | 14:42 |
| prometheanfire | rm_work: if there is a known bad release it should be masked in global-requirements | 15:18 |
| prometheanfire | oh, wrong review | 15:20 |
| prometheanfire | rm_work: still -1 but because rocky is ancient (thought it was master it seems) | 15:21 |
| rm_work | err so it wasn't a bad release :D | 15:21 |
| rm_work | but it is now, lol | 15:21 |
| rm_work | really no idea how to resolve this | 15:22 |
| rm_work | infra said talk to requirements... | 15:22 |
| rm_work | but i don't know if this is fixable here | 15:22 |
| rm_work | we can't pin setuptools | 15:22 |
| rm_work | and we can't bump a package in U-C rocky? | 15:22 |
| prometheanfire | ya, that's an infra thing, go to there channel? | 15:22 |
| rm_work | which means our gates are essentially broken with no fix possible | 15:22 |
| prometheanfire | an exception can probably be made for this | 15:22 |
| rm_work | hmm k | 15:23 |
| prometheanfire | but generally setuptools is managed by infra | 15:23 |
| rm_work | yeah this affects a few projects I think | 15:23 |
| rm_work | they said it's not likely feasible to pin that, I think | 15:23 |
| rm_work | I will ask again tho... | 15:23 |
| prometheanfire | only airship and a couple of x/PROJECTNAMEs | 15:23 |
| prometheanfire | ya | 15:23 |
| prometheanfire | given that rocky is ancient at this point I can see there reluctance | 15:24 |
| smcginnis | prometheanfire: Rocky hasn't completely switched over to extended maintenance yet. | 15:27 |
| smcginnis | So we could still make a change like this I guess. Though only an exceptional case that we wouldn't want to normally do. | 15:28 |
| smcginnis | That said, risky changing requirements now that most projects have already transitioned to EM. | 15:29 |
| smcginnis | But yeah. | 15:29 |
| prometheanfire | smcginnis: ya, there are not many good options :| | 15:29 |
| rm_work | Well, it's looking like either we land this change in stable/rocky requirements, or... We drop support and stop running tests on rocky | 15:35 |
| rm_work | OR I guess we could just ... kill the requirements testing, and bump our own local lower-constraint | 15:35 |
| prometheanfire | lol | 15:37 |
| prometheanfire | I did change my vote to +2 | 15:38 |
| smcginnis | Yeah, that seems like the least bad path at this point. | 15:38 |
| smcginnis | We should know soon enough. | 15:39 |
| prometheanfire | soon? | 15:40 |
| rm_work | soooooon | 15:40 |
| rm_work | i mean, if ya'll want to be real assholes and push back super hard, I guess I can *reluctantly* have no choice but to drop queens/rocky maintenance :D | 15:41 |
| rm_work | so, you know... totally up to ya'll ^_^ | 15:41 |
| smcginnis | Hah | 15:42 |
| *** vishalmanchanda has quit IRC | 15:43 | |
| prometheanfire | hehehe | 15:43 |
| rm_work | hmm, probably shouldn't have said that in a logged channel tho ;) | 15:43 |
| *** evrardjp has quit IRC | 17:36 | |
| *** evrardjp has joined #openstack-requirements | 17:36 | |
| *** rpittau is now known as rpittau|afk | 18:06 | |
| openstackgerrit | Sorin Sbarnea proposed openstack/requirements master: Bump virtualenv===20.0.12 https://review.opendev.org/713963 | 18:53 |
| openstackgerrit | Merged openstack/requirements stable/rocky: Bump MarkupSafe to 1.1.1 due to setuptools change https://review.opendev.org/712187 | 19:34 |
| openstackgerrit | Merged openstack/requirements stable/stein: update constraint for keystoneauth1 to new release 3.13.2 https://review.opendev.org/713818 | 19:34 |
| openstackgerrit | Merged openstack/requirements master: update constraint for tooz to new release 2.2.0 https://review.opendev.org/713822 | 19:36 |
| openstackgerrit | Merged openstack/requirements master: update constraint for oslo.log to new release 4.1.0 https://review.opendev.org/713824 | 19:40 |
| openstackgerrit | Matthew Thode proposed openstack/requirements master: Updated from generate-constraints https://review.opendev.org/713573 | 20:10 |
| openstackgerrit | Merged openstack/requirements master: update constraint for oslo.versionedobjects to new release 2.0.1 https://review.opendev.org/713825 | 20:14 |
| *** ralonsoh has quit IRC | 20:19 | |
| openstackgerrit | Merged openstack/requirements stable/rocky: update constraint for python-glanceclient to new release 2.13.2 https://review.opendev.org/713866 | 20:28 |
| *** e0ne_ has quit IRC | 21:04 | |
| *** e0ne has joined #openstack-requirements | 21:08 | |
| *** e0ne has quit IRC | 21:22 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/requirements master: update constraint for tripleo-common to new release 12.2.0 https://review.opendev.org/713997 | 22:01 |
| openstackgerrit | Merged openstack/requirements master: update constraint for taskflow to new release 4.1.0 https://review.opendev.org/713820 | 22:10 |
| *** zigo has quit IRC | 22:22 | |
| *** zigo has joined #openstack-requirements | 22:32 | |
| fungi | rm_work: prometheanfire: it's not so much that the infra team controls whether or not setuptools is pinned, it's that the python packaging ecosystem is designed not to really give you a means to do that. there's no way (with traditional python packaging) for package installation to have any control over what version of setuptools is used because setuptools is assumed to already be present and | 22:52 |
| fungi | tools like virtualenv are just going to give you whatever the latest setuptools is | 22:52 |
| rm_work | yeah, sorry didn't mean to imply it was a malicious choice to not support that or something, lol | 22:53 |
| rm_work | just that infra would/could not | 22:53 |
| fungi | so basically anywhere we pin to dependencies which don't work with latest setuptools, we're in essence declaring bankruptcy on being able to use traditional python packaging | 22:53 |
| fungi | the very bleeding edge latest release of virtualenv added a feature to be able to say what "seed packages" should be included when creating a virtualenv, so it might be possible to leverage that to ask it to use an older setuptools. that said, i don't know how you'd go about orchestrating virtualenv options from tox. our job entry point there is a layer removed from virtualenv | 22:55 |
| fungi | i think you could maybe do it by precreating the virtualenv and then telling tox to use that and not make one of its own | 22:56 |
| fungi | but that starts to get fragile if you expect developers and users to mimic the same pattern on their systems | 22:57 |
| rm_work | yeah | 22:57 |
| rm_work | i believe this bump was the correct option | 22:57 |
| fungi | that workaround would be more or less the same as creating the virtualenv, using pip to downgrade setuptools in it, then telling tox to use that | 22:58 |
| prometheanfire | fungi: ya, looks like we bundle one with our python installs | 22:59 |
| prometheanfire | /usr/lib64/python3.6/ensurepip/_bundled/setuptools-40.6.2-py2.py3-none-any.whl | 22:59 |
| fungi | but also dropping jobs and/or deviating from global upper-constraints.txt are in the cards where extended maintenance branches are concerned. this is a prime example of why it's hard to maintain serious support for stable branches in our ecosystem long-term | 22:59 |
| fungi | prometheanfire: i don't think virtualenv eve uses that | 23:00 |
| rm_work | yeah, but to deviate from u-c requires just dropping the requirements job, right? | 23:00 |
| fungi | rm_work: and possibly forking the constraints list into your repo | 23:00 |
| prometheanfire | fungi: I'm talking about the python package itself | 23:00 |
| prometheanfire | for gentoo | 23:01 |
| rm_work | ah, yes | 23:01 |
| prometheanfire | I'm surprised we bundle at all | 23:01 |
| * prometheanfire may go make a bug about it just to see what happens | 23:01 | |
| fungi | prometheanfire: that's typical of all python except on debian derivatives where they separate it into a different package | 23:01 |
| fungi | prometheanfire: it's used by the venv module | 23:01 |
| prometheanfire | we have a setuptools package though | 23:01 |
| fungi | right, i mean debian actually builds a separate python3-venv package which separates the venv module so that they can build the embedded setuptools et cetera copies from other source packages in the archive rather than using the copies provided by upstream python interpreter build process | 23:02 |
| prometheanfire | ah | 23:02 |
| fungi | so that the version of setuptools provided with the python3-venv package is the same version of setuptools provided in theor python3-setuptools package | 23:03 |
| fungi | and gets the same security fixes backported and so on | 23:03 |
| prometheanfire | we just rebuild all of python :D | 23:04 |
| fungi | it what they do so they won't need to manage security patches for more than one version of setuptools in the archive | 23:04 |
| fungi | and it's convenient insofar as you at least don't need to install python3-pip and python3-setuptools packages to get python3-venv to work, so pip and setuptools don't clutter up your system-wide import path | 23:06 |
| fungi | and python3-wheel and so on | 23:06 |
| prometheanfire | isn't venv a builtin for 3 anyway? | 23:07 |
| prometheanfire | /usr/lib64/python3.6/venv/__init__.py | 23:08 |
| fungi | it's in the stdlib, but debian notoriously breaks venv and ensurepip out of the stdlib collection, along with a few other things that have some onerous build dependencies | 23:08 |
| fungi | they also package the bulk of stdlib separate from the interpreter | 23:09 |
| prometheanfire | don't be debian, gotit :P | 23:10 |
| fungi | so python3-minimal just gets you the interpreter and builtins, and python-3 is a "virtual" package depending on that and libpython3-stdlib | 23:10 |
| fungi | s/python-3/python3/ | 23:10 |
| fungi | well, they have good reasons for why they're doing it that way, but yes it's an unpopular choice with a lot of users and upstream developers alike | 23:10 |
| prometheanfire | to do a lot of other things right (deterministic builds) | 23:11 |
| fungi | well, and for ease of portability to other platforms which may not be able to satisfy all the build dependencies, and to reduce workload on the security team, and so on | 23:12 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!