fungi | tonyb: sorry, set what up? | 00:46 |
---|---|---|
tonyb | As an admin got the Google Search Console | 00:46 |
tonyb | fungi: at least from what I can see right now | 00:47 |
tonyb | Oh that's not true, someone has | 00:47 |
fungi | for openstack.org oe wiki.o.o? | 00:48 |
tonyb | openstack.org | 00:49 |
tonyb | host -t any openstack.org | grep google-site | 00:49 |
fungi | i can ask foundation webdevs about it, but yeah they have the base domain set up, no idea about subdomains tho | 00:49 |
tonyb | If you "own" the basedomain you also own the subdomains | 00:50 |
tonyb | Not important right now | 00:55 |
*** dviroel_ is now known as dviroel | 14:23 | |
simondodsley | I have a patch in (https://review.opendev.org/c/openstack/requirements/+/903099) to update requirements/upper-contraints.txt but it seems like the current `urllib3` version is get to a version 1 z-release lower than the new requirement I'm adding. How do I get `urllib3` version bumped? What is `generate-constraints` job that seems to do this automatically? | 16:21 |
fungi | simondodsley: constraints are a resolved self-consistent set of transitive requirements, so if some package in the set depends on an older version of urllib3 then the resulting constraints set will include that version | 16:24 |
fungi | basically you'll need to figure out what depends on urllib3 and then work out what needs to be done to make it support newer urllib3 | 16:25 |
simondodsley | looks like all the current dependencies would be ok with the .z bump. Can I do that in my patch then? | 16:26 |
fungi | simondodsley: also the change you linked is for py-pure-client not urllib3? | 16:26 |
fungi | or maybe i'm misunderstanding what you say you're trying to do | 16:27 |
simondodsley | yep - urllib3 is a dependency in py-pure-client, but the version we specificy is higher than the pinned version | 16:27 |
simondodsley | can i bump urllib3 in my patch? | 16:27 |
fungi | oh, are you referring to a test failure? | 16:28 |
simondodsley | yes | 16:28 |
simondodsley | pinned at 1.26.16 and we need 1.26.17 | 16:29 |
fungi | https://zuul.opendev.org/t/openstack/build/cfb396f6793f40a5928fa3c0bda68e55 says you're conflicting on python-dateutil requirements, but maybe the resolver is unable to get past that | 16:29 |
simondodsley | that is an old one. We just updated py-pure-client to 1.45.0 | 16:29 |
fungi | can you link to the job failure you're talking about then? | 16:30 |
fungi | sorry, i'm in a meeting right now so not able to dig much | 16:30 |
simondodsley | i jus rebased - i'll drop the failure here as soon as it is visible | 16:30 |
simondodsley | fungi: https://zuul.opendev.org/t/openstack/build/15199b466bde4b37a1063f3ca76698ed | 16:31 |
fungi | simondodsley: yeah, i think you can just increase it in upper-constraints.txt but i thought that file normally got automatically updated by a periodic (daily) job | 16:35 |
simondodsley | fungi: thanks | 16:36 |
frickler | fungi: it is a weekly job, but the results are always unmergeable due to several things openstack projects are lagging behind in adopting, like sqla 2.0, so it needs manual intervention to filter out the mergeable version bumps, which happens very rarely | 16:38 |
frickler | tonyb has been starting some work to improve on the upgrade job to at least handle differing constraints for different python versions, but that's also not ready yet | 16:39 |
fungi | simondodsley: yeah, looks like https://review.opendev.org/c/openstack/requirements/+/903802/4/upper-constraints.txt#177 is proposing to update it to 1.26.18 | 16:40 |
fungi | so i would just try setting it to that in your change | 16:40 |
simondodsley | :+1 | 16:40 |
fungi | frickler: thanks, that explains why i wasn't finding it in the builds list for periodic pipeline, i needed to look at periodic-weekly | 16:41 |
iurygregory | Hey Infra team o/, I'm trying to recover my pypi account (got locked out due to the 2FA), https://github.com/pypi/support/issues/2708 , , I've received and email from ewdurbin " In order to verify the request, please push a branch with the name <Some Code> to the public source code repository for the ironic-prometheus-exporter project and respond to this email. " any idea how can I do this? =) | 19:06 |
iurygregory | The branch does not need to include any commits or changes and can be deleted after it’s been verified. | 19:06 |
fungi | iurygregory: we're having our weekly meeting right now, but i can take a look in an hour | 19:06 |
iurygregory | fungi, ack no rush =) | 19:07 |
fungi | iurygregory: so "the public source code repository" is being determined based on the "homepage" link on https://pypi.org/project/ironic-prometheus-exporter/ i guess? | 20:04 |
iurygregory | fungi, I think so =) | 20:05 |
iurygregory | it doesn't say I have to push the branch, so maybe we can ask release team to create one with the name they asked me? | 20:06 |
iurygregory | or maybe I need to do, not sure lol | 20:07 |
tonyb | Could we create $branchname pointing to an essentially empty repo with a READEME that explains why its there .... and removes it once you have your account? | 20:08 |
fungi | yeah, i'm trying to figure out how *we'd* authenticate you since you don't even seem to use a registered nick in irc. something you can only do via gerrit would suffice | 20:08 |
tonyb | iurygregory: worth explaining to pypi that the repo is managed and it'd be created by tooling | 20:09 |
iurygregory | fungi, my nick is registered afaik .-. | 20:12 |
iurygregory | I've authenticated again to see if fixes =) | 20:12 |
iurygregory | tonyb, ok | 20:12 |
fungi | oftc -- [iurygregory] user has identified and verified with services | 20:13 |
fungi | now it's there ;) | 20:13 |
iurygregory | my bad I probably lost connection and forgot to do the identify | 20:13 |
fungi | iurygregory: one way to do that would be to temporarily amend the gerrit acl for ironic-prometheus-exporter to grant a group you're in permission to create/delete a branch and push directly to it, and then have the ptl +1 that change (for expediency, propose a revert of the change too so we can fast-track approval of it when you're done) | 20:15 |
iurygregory | JayF, fyi ^ | 20:16 |
iurygregory | fungi, I think it makes sense | 20:16 |
fungi | that also gets us some clear authentication because you're using your gerrit account to propose the acl change | 20:17 |
iurygregory | agree, do you have info about what I should add to the acl? | 20:18 |
fungi | iurygregory: there are a few ways to go about it. is the name of the branch you need to add secret in some way? | 20:19 |
fungi | otherwise i think we can scope the creation/deletion to that temporary branch name for added safety | 20:20 |
iurygregory | the name they asked is kinda a password that google generates for you if you accept things in the browser lol | 20:20 |
iurygregory | 16 characthers letters and numbers | 20:21 |
tonyb | Ultimately they want that branch to be public so I don't think it can be secret right? | 20:22 |
iurygregory | yeah | 20:23 |
iurygregory | since they will check the repo to find the branch | 20:23 |
iurygregory | The branch does not need to include any commits or changes and can be deleted after it’s been verified. After that, I’ll reset 2FA for your account and issue a password reset. | 20:24 |
tonyb | iurygregory: So I think you'd want to do something like: https://opendev.org/openstack/project-config/src/commit/60de765fa0c8dac1654b11ded4b3986db0ec681c/gerrit/acls/openstack/ironic.config#L4-L10 | 20:29 |
tonyb | but it'd be refs/heads/$branch_name | 20:29 |
fungi | it looks like it already uses the ironic.config acl according to https://opendev.org/openstack/project-config/src/branch/master/gerrit/projects.yaml#L3720-L3726 | 20:30 |
iurygregory | yeah it uses the ironic one =) | 20:30 |
tonyb | and also include create = group ... and delete = group ... | 20:31 |
opendevreview | Brian Rosmaita proposed openstack/project-config master: Deprecate cinderlib https://review.opendev.org/c/openstack/project-config/+/903260 | 20:31 |
iurygregory | tonyb, I was about to ask if extra options would be required =) | 20:31 |
fungi | so you'll either be changing that acl or you can temporarily give - project: openstack/ironic-prometheus-exporter its own dedicated acl and then remove it again when you're done | 20:31 |
iurygregory | I think the first approach is ok, since we only need to change one file | 20:33 |
tonyb | iurygregory: It (the first approach) is a little larger hammer in that any ironic-core could create/delete $branch_name in any ironic project | 20:34 |
iurygregory | oh right | 20:34 |
tonyb | iurygregory: where as the idea fungi suggested is more targeted. | 20:34 |
tonyb | iurygregory: Having said that personally I'm okay with the bigger hammer | 20:35 |
iurygregory | ok, so I would need to create a new file for ironic-prometheus-exporter and also update in projects.yaml? | 20:35 |
tonyb | iurygregory: Yup you could create ironic-prometheus-exporter.config that starts with something like [access] inheritFrom = ironic.config | 20:37 |
tonyb | and adds the hunk you need for your branch | 20:37 |
opendevreview | Brian Rosmaita proposed openstack/project-config master: Restore cinderlib zuul jobs https://review.opendev.org/c/openstack/project-config/+/904035 | 20:37 |
tonyb | at least I think that's the syntax for inheritFrom | 20:38 |
fungi | and temporary remove the acl-file override for it in projects.yaml | 20:39 |
fungi | also, if it seems like we don't entirely know what we're doing, we don't. this is an unusual circumstance and it may take a couple of tries to get things right since there's no perfect example to crib from | 20:40 |
fungi | gerrit's acl language is a bit complicated | 20:40 |
fungi | but i'm happy to help work with you on it | 20:41 |
opendevreview | Iury Gregory Melo Ferreira proposed openstack/project-config master: Temporary update IPE ACL https://review.opendev.org/c/openstack/project-config/+/904036 | 20:44 |
iurygregory | I've pushed based on what I understood from our conversation =) | 20:44 |
iurygregory | not sure if is 100% correct | 20:44 |
tonyb | iurygregory: Looks like a pretty solid attempt | 20:45 |
iurygregory | just wondering, after we get approved, how I would create the branch? in a normal repo in github I would just do "git checkout -b <name> origin/master" and "git push origin <name>" | 20:49 |
fungi | iurygregory: you do it through the gerrit webui (or optionally through the rest api). also i left you some trview comments | 20:50 |
iurygregory | fungi, tks! | 20:51 |
opendevreview | Iury Gregory Melo Ferreira proposed openstack/project-config master: Temporary update IPE ACL https://review.opendev.org/c/openstack/project-config/+/904036 | 20:52 |
opendevreview | Merged openstack/project-config master: Temporary update IPE ACL https://review.opendev.org/c/openstack/project-config/+/904036 | 21:38 |
iurygregory | "Error 403 (Forbidden): not permitted: create on refs/heads/Tv8HwD5dt65yMHGE" .-. | 21:41 |
iurygregory | Endpoint: /projects/*/branches/* | 21:42 |
iurygregory | any ideas? .-. | 21:42 |
fungi | iurygregory: in the webui? | 21:42 |
iurygregory | yeah | 21:42 |
iurygregory | I'm wondering if the revision is required | 21:42 |
iurygregory | aha | 21:43 |
tonyb | https://stackoverflow.com/questions/29511087/permission-to-create-branches-with-gerrit | 21:44 |
tonyb | I guess you need to use git directly with the gerrit remote | 21:45 |
JayF | git checkout -b secretcode; git push gerrit secretcode | 21:45 |
JayF | should do the trick | 21:45 |
JayF | please confirm this, someone else | 21:45 |
JayF | and please under no circumstances run git push g.errit without a fourth argument :) | 21:45 |
iurygregory | it worked, it needs the revision sha-1 otherwise it fails | 21:46 |
fungi | odd, create normally works in the webui, but it's possible you need permission on the path parent, i.e. refs/heads | 21:46 |
fungi | aha, yes you need to tell it what commit to branch from | 21:46 |
iurygregory | yeah | 21:46 |
tonyb | fungi: the stackoverflow link says you need "owner" to use the web-UI | 21:46 |
iurygregory | which makes sense (but the ui highlited the button create after I added the name lol) | 21:46 |
fungi | unfortunately, stackoverflow is notoriously full of people who are confused and have no idea what they're talking about | 21:47 |
tonyb | fungi: fair | 21:47 |
fungi | "owner" is just a default role | 21:47 |
opendevreview | Iury Gregory Melo Ferreira proposed openstack/project-config master: Revert "Temporary update IPE ACL" https://review.opendev.org/c/openstack/project-config/+/904012 | 21:48 |
iurygregory | revert in place, I've marked W-1 so we avoid merging before they tell me things are verified in pypi | 21:49 |
iurygregory | Tks everyone! | 21:49 |
tonyb | https://opendev.org/openstack/ironic-prometheus-exporter/src/branch/Tv8HwD5dt65yMHGE \o/ | 21:49 |
fungi | which is also my point on the openstack-discuss ml thread about web forums full of people asking questions and only getting answers from other people who are equally clueless | 21:49 |
tonyb | iurygregory: ... You'll remove that branch before you also remove your -W right ;P | 21:52 |
iurygregory | tonyb, correct =) | 21:52 |
JayF | but we need a souvenir of this strange and clever account recovery process :D | 21:54 |
tonyb | JayF: we have changes 904036 and 904012 for that. | 21:57 |
tonyb | They'll forever be a part of our shared history now :) | 21:58 |
JayF | lol | 21:58 |
fungi | our souvenir is the git history which will allow us to more easily repeat the process the next time someone loses their phone, yep | 21:58 |
JayF | I am just glad that Iury cares enough to actually jump through all the flaming hoops; we have a long-tail of pypi package owners who still haven't :D | 22:00 |
tonyb | Yeah | 22:02 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!