15:00:21 <sdoran> #startmeeting Ansible Core Public IRC Meeting
15:00:21 <zodbot> Meeting started Thu Dec  3 15:00:21 2020 UTC.
15:00:21 <zodbot> This meeting is logged and archived in a public location.
15:00:21 <zodbot> The chair is sdoran. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:21 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:21 <zodbot> The meeting name has been set to 'ansible_core_public_irc_meeting'
15:00:56 <mkrizek> \o
15:01:18 <sdoran> #chair mkrizek
15:01:18 <zodbot> Current chairs: mkrizek sdoran
15:01:23 <felixfontein> hi!
15:01:29 <sdoran> #chair felixfontein
15:01:29 <zodbot> Current chairs: felixfontein mkrizek sdoran
15:02:17 <felixfontein> how's the AZP migration going?
15:03:19 <felixfontein> I saw that some PRs already have AZP results (like https://github.com/ansible/ansible/pull/72831)
15:03:38 <sdoran> It's done. Just ironing out a few things here and there.
15:03:44 <felixfontein> and others (older ones) are waiting for "CI" to run (next to shippable results which are already there)
15:03:58 <felixfontein> cool :)
15:04:20 <sdoran> #info Ansible Core CI has moved to Azure Pipelines. New commits need to be pushed to PR branches in order to trigger a CI run on the new platform.
15:04:29 <felixfontein> I guess next step will be collection shippable -> AZP transition?
15:04:40 <sdoran> Yup. Let me know when you're ready. :)
15:04:43 <felixfontein> does closing+reopening also works?
15:05:00 <sdoran> That's a good point. I'm not sure.
15:05:02 <sdoran> Let me try.
15:05:47 <felixfontein> I think gundalow suggested to start with community.crypto, at least on the community side
15:05:51 <sdoran> Looks like that works too. Sweet.
15:06:00 <felixfontein> that's nicer than pushing empty commits :)
15:06:04 <sdoran> #info Closing/reopening a PR will also trigger a CI run in AZP.
15:06:33 <sdoran> Agreed.
15:06:42 <mkrizek> I believe just commenting `/azp run` works too
15:07:01 <sdoran> I think the problem with that is it won't clear the Shippable status.
15:07:14 <felixfontein> mkrizek: I'm now trying that too
15:07:17 <sdoran> Which is fine if the Shippable run passed. But if the Shippable run failed, it'll block merging.
15:07:29 <felixfontein> mkrizek: 'Commenter does not have sufficient privileges for PR 72697 in repo ansible/ansible'
15:07:58 <mkrizek> ah, that makes sense
15:08:16 <mkrizek> then nevermind :)
15:08:46 <felixfontein> #info People with commit rights can also add a comment with `/azp run`
15:09:11 <felixfontein> that will be useful in collections (assuming it works there as well)
15:09:31 <felixfontein> sdoran: is shippable disabled, and that's why it won't run again?
15:09:32 <sdoran> mkrizek: Are there any other AZP bot commands folks should know about?
15:09:41 <sdoran> felixfontein: Correct. We disable Shippable.
15:09:44 <sdoran> *d
15:10:17 <felixfontein> cool!
15:10:24 <mkrizek> https://docs.microsoft.com/en-us/azure/devops/pipelines/repos/github?view=azure-devops&tabs=yaml#comment-triggers
15:10:33 <mkrizek> these are from AZP
15:10:37 <sdoran> mkrizek: Do `/rebuild` and `/rebuild_failed` still work in AZP?
15:10:42 <sdoran> Oh neat. I did not know that.
15:11:20 <felixfontein> depends on whether the bot code has been adjusted, I think
15:11:25 <mkrizek> `/rebuild`, `/rebuild_failed` and `/rebuild_merge` *should* work the same as with shippable, if not please ping me :)
15:11:34 <sdoran> 👍
15:12:15 <sdoran> #info Azure Pipelines has some comment triggers. https://docs.microsoft.com/en-us/azure/devops/pipelines/repos/github?view=azure-devops&tabs=yaml#comment-triggers
15:12:21 <felixfontein> mkrizek: I'll try `/rebuild` :)
15:12:38 <felixfontein> if you want to know how the `/azp help` output looks: https://github.com/ansible/ansible/pull/72697#issuecomment-738066290
15:15:25 <sdoran> I should have made that a topic. :)
15:15:34 <sdoran> Anyway, we have one item on the agenda today.
15:16:43 <sdoran> #topic BaseFileCache. Related to https://github.com/ansible/ansible/pull/57646.
15:17:04 <sdoran> Is briantist around?
15:21:00 <sdoran> Anyone have any thoughts on this?
15:21:05 <sdoran> Here is the comment from briantist: https://github.com/ansible/community/issues/570#issuecomment-732417395
15:21:53 <briantist> hi
15:22:18 <felixfontein> mkrizek: looks like `/rebuild` does not help for triggering AZP if it hasn't run before, though I guess that's probably expected? https://github.com/ansible/ansible/pull/72625#issuecomment-738067823
15:22:38 <mkrizek> yes, the run needs to exist
15:22:56 <briantist> sdoran: I'm here, is this the core meeting? I thought it was a few hours later, maybe I got the time wrong
15:23:08 <sdoran> Yup, this is the meeting.
15:23:23 <felixfontein> briantist: the times on Tuesday are different than on Thursday
15:23:45 <briantist> ah got it, that's where I messed up
15:24:54 <briantist> In the last meeting someone linked to this older PR, so the discussion there is probably relevant too: https://github.com/ansible/ansible/pull/57646
15:25:08 <sdoran> I was reading through that.
15:25:36 <sdoran> Looks like that PR just stalled after a few ideas were discussed.
15:26:00 <sdoran> We got a PR yesterday (which we'll go over in triage later today) fixing the keys error, it seems. https://github.com/ansible/ansible/pull/72789
15:26:20 <Shrews> sdoran: i think we were waiting for shertel or bcoca to return to comment on the cache stuff
15:26:39 <Shrews> iirc
15:26:41 <sdoran> Seems like it. is Sloane out today? I know Brian is.
15:26:49 <sdoran> They know the cache code best.
15:27:15 <gundalow> (lag) Yes, community.crypto first collection repo for AZP (cc felixfontein sdoran)
15:27:55 <Shrews> i'm not sure when either return
15:28:10 <briantist> if they're both still out for this meeting it probably makes sense to punt the issue again til next(++) time I suppose?
15:28:56 <sdoran> briantist: I think it'd be fine it you wanted to open a PR taking into account the feedback and suggestions in #57646 since it seems that PR has stalled.
15:29:05 <felixfontein> gundalow: sdoran: I'm looking forward to it :) I won't have time today though to help
15:31:00 <briantist> I might, but I wanted to better understand the comments there. The original author was going to do the same thing I proposed, but it seems they preferred a different change in a different place, and I don't fully understand the reason. I  may just have to read it more closely, we'll see. If I do, I'll just link the PR in the agenda item too?
15:32:15 <gundalow> felixfontein: I think it's on my plate to more to AZP, so you are safe :)
15:33:01 <shertel> I'm here today. I was just looking at the caching PR.
15:33:17 <sdoran> #chair shertel
15:33:17 <zodbot> Current chairs: felixfontein mkrizek sdoran shertel
15:33:26 <shertel> I'm going to open a new PR to address the last requested changes
15:33:45 <shertel> the class that the PR modifies also impacts the FactCache, which isn't affected by that bug
15:33:53 <sdoran> Hey shertel! Welcome back.
15:33:56 <shertel> thanks!
15:36:30 <sdoran> shertel: So does briantist need to open a PR? Seems like you plan to take care of it.
15:38:09 <briantist> so if I understand it correctly, the reason the warning happens spuriously is because I happen to be using it for inventory caching (not fact caching)? And the fix is more like this gist you posted in the original PR? https://gist.github.com/s-hertel/50f72d8f521378225a3cdc0b42644509
15:39:38 <shertel> briantist: Yes. If you want to take a stab at it, that was what I had in mind. The vars manager handles the fact cache KeyErrors appropriately, but inventory doesn't
15:40:47 <sdoran> briantist: Is there anything more you'd like to discuss?
15:41:02 <shertel> the CachePluginAdjudicator gives inventory plugins a uniform way to access file-backed and database-backed caches, so it would be the right place to fix this
15:41:13 <briantist> I see thanks shertel. If you were going to put in a PR anyway, then by all means! if you do would you mind @ mentioning me? I just want to follow along! If you're not able to I'll give it a try eventually (starting to get a bit of a backlog myself 😛)
15:41:35 <sdoran> For the purpose of this meeting, I think we're in agreement with where/how this should be fixed.
15:41:56 <briantist> in any case nothing else from me sdoran , thanks for the ping I would have missed this otherwise :)
15:41:58 <shertel> If you want to take a stab at it, I'd also be happy to review. But yes, otherwise, I'll open a PR in a bit, today and @ you
15:42:01 <sdoran> Further discussion can happen in #ansible-devel if you have specific questions, briantist.
15:42:15 <sdoran> Sure thing.
15:42:20 <sdoran> #topic open floor
15:42:42 <briantist> that sounds great shertel , I definitely won't have a chance to put one in that soon. Thank you!
15:42:55 <shertel> +1
15:45:27 <sdoran> I'll go ahead and end the meeting in five minutes if there are no further topics to discuss.
15:46:50 <sdoran> #info A big thanks to mkrizek, mattclay, and relrod for working so hard on the transition to Azure Pipelines and making it go smoothly.
15:48:27 <mkrizek> #info ... and a big thanks to sdoran for the same too ;-)
15:48:40 <sdoran> 😊
15:51:19 <sdoran> Thank you everyone for attending.
15:51:20 <sdoran> #endmeeting