| yoctozepto | clarkb, fungi, dansmith: I actually like the new behaviour, it helps weed out missed comments; I guess it all boils down to user preference; I am more in the gerrit upstream camp (based on their feedback) - be thorough with reviews | 09:13 |
|---|---|---|
| dansmith | yoctozepto: except if the person doesn't, either because they're a part-time contributor that doesn't know about the treadmill, or refuse to do the treadmill, or use gertty and don't see them, everyone else has the noise | 14:23 |
| dansmith | yesterday I looked at several patches where I was like "Jeez bob, we already saw that you wanted a variable name changed and decided not to, why bring it up again? oh wait, that's old.. not unresolved, just old." | 14:24 |
| spotz | I'll have to look at some older reviews, I mainly did newer stuff the last few days | 14:54 |
| gmann | #startmeeting tc | 15:00 |
| opendevmeet | Meeting started Thu Jan 27 15:00:13 2022 UTC and is due to finish in 60 minutes. The chair is gmann. Information about MeetBot at http://wiki.debian.org/MeetBot. | 15:00 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 15:00 |
| opendevmeet | The meeting name has been set to 'tc' | 15:00 |
| gmann | #topic Roll call | 15:00 |
| gmann | o/ | 15:00 |
| lourot | o/ | 15:00 |
| gmann | tc-members meeting time | 15:00 |
| dansmith | o/ | 15:00 |
| Guest292 | O/ | 15:00 |
| jungleboyj | o/ | 15:00 |
| *** Guest292 is now known as diablo_rojo_phone | 15:01 | |
| gmann | not present today: Radosław Piliszek (yoctozepto), Belmiro Moreira (belmoreira) | 15:01 |
| ricolin | o/ | 15:01 |
| gmann | let's start | 15:01 |
| gmann | #link https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee | 15:01 |
| gmann | ^^ today agenda | 15:01 |
| gmann | #topic Follow up on past action items | 15:02 |
| gmann | nothing from previous meeting | 15:02 |
| gmann | #topic Gate health check | 15:02 |
| gmann | any news on gate | 15:02 |
| spotz | oo/ | 15:03 |
| dansmith | I have only had a few things through the gate this past week, but they all went quite smooth | 15:03 |
| fungi | activity levels have been on the rise since we're aroubd feature freeze for some projects | 15:03 |
| gmann | yeah | 15:03 |
| gmann | stable/train cap with older tempest is merged now. | 15:04 |
| diablo_rojo_phone | Hopefully smooth continues to be the theme | 15:04 |
| gmann | I think centos8-stream issue still not resolved ? | 15:04 |
| fungi | the next few weeks will be our first real-world load test of zuul's persistent state and multi-scheduler capabilities | 15:04 |
| fungi | seeing how it handles the openstack rush | 15:04 |
| fungi | it's been doing well up to this point, but that was over the holiday lull | 15:05 |
| fungi | if something doesn't seem right, please don't hesitate to give folks a heads up in #opendev or #openstack-infra | 15:05 |
| gmann | seems passing but few retry failure #link https://zuul.openstack.org/builds?job_name=tempest+tempest-full-py3-centos-8-stream&job_name=tempest-integrated-compute-centos-8-stream&skip=0 | 15:06 |
| fungi | also we're on a new version of gerrit (3.4) since monday, so aside from already observed ui behavior and visual changes, there's the potential for uncovering regressions like the one we found with signed tags | 15:06 |
| gmann | I will monitor those | 15:06 |
| gmann | anything else on gate updates? | 15:08 |
| jungleboyj | Seems like it has been working ok for us. | 15:08 |
| gmann | ok | 15:08 |
| gmann | #topic Z Release Cycle Name | 15:09 |
| gmann | nomination are closed for naming and election is started on 25th | 15:09 |
| gmann | I have started the poll and hope all tc-members received it, if not please let me know | 15:09 |
| gmann | #link https://governance.openstack.org/tc/reference/release-naming.html#polls | 15:09 |
| jungleboyj | ++ | 15:09 |
| gmann | poll will end on 1st feb or early if all TC members vote | 15:09 |
| diablo_rojo_phone | Received! | 15:10 |
| spotz | voted! | 15:10 |
| gmann | till now only 4 tc-members casted vote, | 15:10 |
| jungleboyj | I will cast my vote in 23 hours after my Twitter poll closes. :-) | 15:10 |
| gmann | jungleboyj: ack, what name people preferring in your poll? | 15:11 |
| jungleboyj | Zenith right now. | 15:11 |
| jungleboyj | https://twitter.com/jungleboyj/status/1486344910693441541 | 15:11 |
| gmann | tc-members: please cast vote if not yet done. so that we can proceed for the next step of legal checks soon | 15:12 |
| diablo_rojo_phone | I feel like that's a bad choice cause it implies we are at the top and going to head downward now? | 15:12 |
| dansmith | diablo_rojo_phone: omicron and openstack have peaked | 15:12 |
| dansmith | :P | 15:12 |
| diablo_rojo_phone | Lol lol lol | 15:12 |
| gmann | :) | 15:13 |
| jungleboyj | *Sigh* I hadn't thought of it that way. | 15:13 |
| gmann | I will vote this time and zombie is my fav :) | 15:13 |
| diablo_rojo_phone | jungleboyj: nor your Twitter voters lol :) | 15:13 |
| jungleboyj | People are worried that that is going to bring the 'OpenStack is Dead' jokes. | 15:14 |
| spotz | I thought there were some good conversation on this channel a few days ago, like zombie | 15:14 |
| jungleboyj | diablo_rojo_phone: Right. | 15:14 |
| jungleboyj | I think of the old TVs that I used to love. | 15:14 |
| gmann | more it bring, more it get clarity | 15:14 |
| diablo_rojo_phone | Perhaps it's just dansmith and I, jungleboyj | 15:14 |
| diablo_rojo_phone | Never seen the show lol | 15:15 |
| dansmith | I'm not really worried, I think zombie has the same sort of "past its prime" connotation, even though I expect it to win | 15:15 |
| diablo_rojo_phone | Technically as a zombie it would make OpenStack 'living dead' or 'undead' | 15:15 |
| dansmith | not worried about zenith I mean | 15:15 |
| spotz | Yeah I doubt my cchoicce will win, even then legal.. | 15:15 |
| dansmith | diablo_rojo_phone: what zombies have you seen that don't smell like something past its prime in the fridge? :) | 15:16 |
| gmann | diablo_rojo_phone: yeah | 15:16 |
| jungleboyj | *Sigh* | 15:16 |
| gmann | anything else on this topic? | 15:16 |
| dansmith | no :) | 15:17 |
| spotz | Sure, constructive probably not:) | 15:17 |
| gmann | #topic Z cycle Technical Elections | 15:17 |
| gmann | Its time for the PTL and TC election | 15:17 |
| gmann | jungleboyj: and I had volunteer for election official along with existing one. if anyone would like to help please raise patch in election repo | 15:18 |
| gmann | I have also proposed the election dates #link https://review.opendev.org/c/openstack/election/+/825017 | 15:18 |
| spotz | belmeiro and I can't if re-running | 15:18 |
| gmann | Please review and let me know your feedback. or if it looks ok then any election official can approve it | 15:18 |
| gmann | spotz: yeah | 15:18 |
| diablo_rojo_phone | I can take a look today. | 15:19 |
| gmann | spotz: diablo_rojo_phone can either of you add me in core list in gerrit group this is merged now https://review.opendev.org/c/openstack/election/+/825009 | 15:19 |
| gmann | diablo_rojo_phone: thanks | 15:19 |
| jungleboyj | Have it open in a tab as well. | 15:20 |
| gmann | I think ianychoi[m], andymccr and hspease also around to help | 15:20 |
| spotz | I used to know where to do it:) | 15:20 |
| gmann | this will be combined election again and for 5 TC seats | 15:20 |
| gmann | spotz: https://review.opendev.org/admin/groups/208108b62745877449d8e6fce45f44e3cc013e8b,members | 15:21 |
| spotz | done! | 15:21 |
| gmann | jungleboyj: if you can also raise patch for election official, we can start the process/meetings | 15:22 |
| gmann | spotz: thanks | 15:22 |
| gmann | anything else on election? | 15:22 |
| fungi | i'm always happy to answer process or tooling questions too, if the officials need to ask any | 15:22 |
| jungleboyj | gmann: Ok. Will work on that. | 15:22 |
| gmann | thanks | 15:22 |
| jungleboyj | fungi: Thanks! I will try to not pester you too much. | 15:22 |
| gmann | #topic Adjutant need PTLs and maintainers | 15:23 |
| gmann | I think I was supposed to remove this from agenda. we will iterate it after election if no leader | 15:23 |
| gmann | #topic Open Reviews | 15:23 |
| gmann | #link https://review.opendev.org/q/projects:openstack/governance+is:open | 15:24 |
| gmann | this need one more vote #link https://review.opendev.org/c/openstack/governance/+/826243 | 15:24 |
| gmann | ricolin: I replied on the project health tool patch 3link https://review.opendev.org/c/openstack/governance/+/810037/5..6/tools/project_health_check.py#b351 | 15:25 |
| gmann | #link https://review.opendev.org/c/openstack/governance/+/810037/5..6/tools/project_health_check.py#b351 | 15:25 |
| ricolin | gmann, thanks will check | 15:25 |
| gmann | thanks | 15:25 |
| gmann | I think all other reviews are in good shape. | 15:25 |
| gmann | I will check FIPs goal today | 15:26 |
| ade_lee__ | gmann, thanks -- I think the fips goal is only missing reviews from you and ricolin | 15:26 |
| spotz | 7 votes on FIPs | 15:26 |
| gmann | ade_lee__: ack. I opened it yesterday but missed it | 15:26 |
| gmann | I will do right after meeting | 15:27 |
| ricolin | +1 | 15:27 |
| ade_lee__ | thanks - assuming this gets approved, what happens next? | 15:27 |
| gmann | #topic Renaming all `master` branches on all repos to `main` | 15:27 |
| lourot | Hey o/ so at Canonical for example we have ~80 openstack repos (essentially charms). We could create a `main` branch ourselves, but we wouldn't have permissions to delete `master`. Also not sure how we would make `main` the default branch in Gitea. That's why we wanted to see if there were plans for a coordinated effort or something | 15:27 |
| gmann | ade_lee__: you need to propose to goal selected in goal/selected folder and we will see the schedule based on current active goals or so | 15:28 |
| gmann | ade_lee__: #link https://governance.openstack.org/tc/goals/#selecting-goals | 15:28 |
| ade_lee__ | gmann, ok - thanks | 15:28 |
| gmann | lourot: if we do that, we need to do for all openstack repos | 15:29 |
| gmann | anyone remember what was discussed before when master/slave replacement happened in github | 15:29 |
| fungi | lourot: the opendev sysadmins have been looking for projects interested in creating new repositories with a different default branch name so that we can make sure things are working as expected, renaming existing branches will be much more involved and premature if there's no confirmation that things actually work with a different default branch name to begin with | 15:30 |
| dansmith | is it already decided that we're making the switch? | 15:30 |
| gmann | I do not think so | 15:30 |
| dansmith | okay good :) | 15:31 |
| fungi | i don't recall anyone deciding anything in that regard, no | 15:31 |
| jungleboyj | I don't remember discussing this. | 15:31 |
| fungi | from opendev's perspective, we want to make sure the collaboratory is able to support projects who want different default branch names, and confirm our tooling supports that correctly | 15:31 |
| spotz | Besides what fungi mentiioned I think we've also been waiting for the git community to make their final decision | 15:31 |
| gmann | yeah, it is lot of work and lot of configuration/scripts/tooling updates not just repo rename | 15:31 |
| dansmith | yeah, allowing new projects to start with main if they want is fine | 15:31 |
| dansmith | making sure that's possible I mean | 15:32 |
| fungi | spotz: well, opendev has been waiting for the git community to make a decision (or not) in order to inform any potential change in opendev's default for new repositories, that doesn't mean that it's necessarily the deciding factor for openstack's repositories | 15:32 |
| spotz | We know changing OpenStack and all the projeccts would be an undertaking. So we need to be sure it will work in our systems and those we rely on before we can make a decision | 15:32 |
| spotz | As much as I'd like us to change, the ops in me doesn't want us to break:) | 15:33 |
| dansmith | or we could decide not to even if it's possible, and then only worry about if new stuff will work with a different branch | 15:33 |
| dansmith | not to change existing projects I mean | 15:33 |
| lourot | do you know what would happen today if I created project-config/governance reviews for importing a new project from GitHub where the default and unique branch would be `main`. Would the import fail? | 15:33 |
| gmann | even before trying in new repo I think we still need updates and to know what all it can break like using master for all and main for new repo in automative scripts/tooling | 15:33 |
| fungi | only changing some repos in openstack could be challenging, particularly for integration testing where jobs assume branch names will match between different repositories | 15:34 |
| gmann | yeah | 15:34 |
| fungi | lourot: that's what we want to test | 15:34 |
| fungi | (in opendev i mean) | 15:34 |
| gmann | is it final in git community that no once can use master ? | 15:34 |
| gmann | *no one | 15:35 |
| fungi | lourot: currently the answer as to what will happen on import is unknown | 15:35 |
| gmann | and in integration testing/tooling too, it s unknown | 15:35 |
| fungi | gmann: no, the git community has only so far made an effort to make sure that people can use any default branch name they want | 15:35 |
| gmann | knowing what all break and need change itself need more time to audit/try | 15:35 |
| gmann | fungi: ok | 15:35 |
| fungi | the most recent word i found on their mailing list was over a year ago indicating they had not themselves decided to change the built-in default for the default branch name | 15:36 |
| fungi | #link https://lore.kernel.org/git/xmqqv9d1yzta.fsf@gitster.c.googlers.com/ | 15:36 |
| lourot | I think the default in GitHub now being `main`, the moment when someone will try to get that imported to Gitea/Gerrit will come sooner or later naturally | 15:36 |
| fungi | lourot: github is a proprietary and predatory leech on open source, i wouldn't want to follow their lead just because | 15:37 |
| fungi | but i would like to test what an import does in that case, yes | 15:37 |
| lourot | it's not my point, just saying that someone will try soon to get a repo with `main` imported | 15:37 |
| lourot | because it's the default there | 15:38 |
| lourot | at least it's our workflow: we create a temp repo on github and we create then project-config/governance reviews to get that imported | 15:38 |
| fungi | we have a configuration option where projects can indicate the default branch name they want, and have done some rudimentary testing, but nothing heavy with typical ci jobs et cetera | 15:38 |
| gmann | lourot: can please you try that and we can know what are things we need to change or need effort ? | 15:38 |
| gmann | and then we test the integration testing thing | 15:39 |
| lourot | gmann, yes, would you be OK if I do that with a dummy repo? I wouldn't want to slow down things on a project we'd need fast | 15:39 |
| fungi | yes, do keep in mind that renaming back from main to master if it doesn't work out for logistics/integration testing reasons could be even harder, so definitely don't try it with something you need working soon | 15:40 |
| gmann | lourot: I am ok for testing dummy | 15:40 |
| fungi | though odds are we could do it with a manual push --force from one of the gerrit sysadmins | 15:40 |
| gmann | fungi: I do not think we should try to move until we are very much sure | 15:40 |
| lourot | perfect, this looks like an action point for me then :) | 15:40 |
| fungi | right, i meant for whatever repo you were testing with | 15:41 |
| gmann | and more than that I am concerned about the effort it need vs how strongly we want/need to change it | 15:41 |
| fungi | lourot: let's sync up in #opendev when you're ready and we'll work through it | 15:41 |
| fungi | i'm happy to help | 15:41 |
| lourot | fungi, thanks! | 15:41 |
| gmann | opendevmeet: fungi thanks | 15:42 |
| opendevmeet | gmann: Error: "fungi" is not a valid command. | 15:42 |
| * fungi is so a valid command! | 15:42 | |
| gmann | opendevmeet: fungi: thanks | 15:42 |
| opendevmeet | gmann: Error: "fungi:" is not a valid command. | 15:42 |
| jungleboyj | he he. | 15:42 |
| gmann | lourot: fungi: thanks | 15:42 |
| gmann | :) | 15:42 |
| lourot | fungi is not a valid command! | 15:42 |
| fungi | hah | 15:42 |
| gmann | and after that we will re-discuss on this | 15:42 |
| gmann | lourot: you want to keep it in agenda or add it back once we are ready with import experiment ? | 15:43 |
| * spotz puts in a patch for a fungi command:) | 15:43 | |
| dansmith | spotz: good luck, I expect that'll be a lot of code | 15:43 |
| lourot | gmann, I'll put things back on the agenda when we got some results with the experiment, thanks! | 15:44 |
| spotz | hehe | 15:44 |
| jungleboyj | dansmith: Hmmm, I think you are going to need a Lenovo Supercomputer to run that AI. | 15:44 |
| gmann | lourot: sure, thanks for brining it here. | 15:44 |
| * jungleboyj goes to the configurator | 15:44 | |
| gmann | last thing | 15:44 |
| gmann | dansmith: is ready with the ffu testing in grenade | 15:44 |
| gmann | #link https://review.opendev.org/c/openstack/grenade/+/826101 | 15:45 |
| dansmith | yeah, although we need to stop calling it ffu I think.. I've been using that term, but it means something else | 15:45 |
| gmann | +1 | 15:45 |
| dansmith | this is really guaranteeing that people can skip specific releases, so I think skip-level is a better term | 15:45 |
| gmann | yeah skip-level much better | 15:45 |
| dansmith | so yeah, I'd like to have a meeting with anyone interested in moving us forward with this tick-tock release model to address the cycle length concerns without slowing us down | 15:45 |
| gmann | yeah, this is good input for upgrade question we had in that discussion | 15:46 |
| dansmith | get a proposal together so we can try to get it in place with kinda smoke testing now (like this) and then hopefully make the AArdvark release where we start | 15:46 |
| fungi | right, in the old discussions where we were debating potential models, we did use "skip-level upgrades" to refer to this sort of solution (an alternative to the fast-forward upgrading model we settled on at the time) | 15:47 |
| dansmith | yeah, and we've always been anti-skip-level for good reason, | 15:47 |
| dansmith | and definitely skipping any random versions is not what we're talking about here, | 15:47 |
| dansmith | just a specific "you can skip odd-lettered releases" (yeah I said it) | 15:48 |
| gmann | and for 'release cadence' meeting will be video call (a separate call from TC meeting), what is best time for all interested in that topic to meet? | 15:48 |
| dansmith | ++ | 15:49 |
| gmann | remember our next TC meeting is on 3rd video call | 15:49 |
| gmann | may be right after that for an hr? | 15:49 |
| dansmith | sure wfm | 15:49 |
| dansmith | is nobody else interested in this? | 15:49 |
| spotz | I can definitelly leave the room open if y'all want it | 15:50 |
| jungleboyj | ++ | 15:50 |
| gmann | spotz: we can restart as separate meeting itself | 15:50 |
| dansmith | I think at least belmiro was, and he's not here | 15:50 |
| spotz | Maybe get reps of from the projects who aren't doing the normal release cadence? | 15:50 |
| gmann | or continue as we might finish TC meeting soon as per agenda | 15:50 |
| dansmith | well, I'm not sure this would really apply to them | 15:51 |
| spotz | Oh I meant the video room next week | 15:51 |
| dansmith | that's something we need to discuss I guess | 15:51 |
| spotz | dansmith: ok | 15:51 |
| spotz | Just trying to think of who might be interested | 15:52 |
| gmann | let me propose the time on ML for 3rd feb at 16UTC (right after TC meeting) | 15:53 |
| dansmith | we'll need to get buy-in from projects anyway, | 15:53 |
| dansmith | we more need to make sure this is something the TC is going to push forward and what the plan is exactly, and then we can start asking projects to agree Ithink | 15:53 |
| gmann | yeah, hope we can conclude it as TC stand on it | 15:54 |
| gmann | and one more thing before we close. | 15:55 |
| gmann | we will check yoga tracer progress in next meeting on 3rd Feb, please check your assigned item #link https://etherpad.opendev.org/p/tc-yoga-tracker | 15:55 |
| gmann | that is all from me today, anything else to discuss ? | 15:55 |
| spotz | Not from me | 15:56 |
| jungleboyj | Not from me. | 15:56 |
| gmann | let's close the meeting. thanks all for joining | 15:56 |
| gmann | #endmeeting | 15:56 |
| opendevmeet | Meeting ended Thu Jan 27 15:56:45 2022 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 15:56 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/tc/2022/tc.2022-01-27-15.00.html | 15:56 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/tc/2022/tc.2022-01-27-15.00.txt | 15:56 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/tc/2022/tc.2022-01-27-15.00.log.html | 15:56 |
| diablo_rojo_phone | Just a heads up | 15:56 |
| spotz | Thanks gmann and everyone! | 15:56 |
| diablo_rojo_phone | that the PTG signup for teams will go out next week | 15:56 |
| gmann | diablo_rojo_phone: oh, please go ahead | 15:56 |
| jungleboyj | Cool. | 15:57 |
| gmann | diablo_rojo_phone: thanks and ack. | 15:57 |
| gmann | will start TC PTG preparation | 15:58 |
| jungleboyj | ++ | 15:58 |
| diablo_rojo_phone | Once we get the list of teams set I will generate the etherpads | 16:00 |
| diablo_rojo_phone | so if we coordinate via email thread and poll that might be good enough for now? | 16:00 |
| gmann | diablo_rojo_phone: ack. sounds good | 16:00 |
| clarkb | fungi: lourot: One point of clarification is that we added support for non master HEAD to jeepyb and our gitea management and the zuul config allows you to set it as well. This means we expect it to work but no one has done it in our system out side of testing | 16:15 |
| clarkb | fungi: lourot: you have to explicitly tell the import process that HEAD is not master too | 16:16 |
| clarkb | if will not infer this from the import process because git doesn't transfer that data on push | 16:16 |
| fungi | yep, i thought i covered all that but it's good to clarify | 16:17 |
| clarkb | ya it was mentioned but I wanted to hammer that point home so it doesn't get missed and accidnetally imported as normal :) | 16:20 |
| clarkb | then we'd have to start over :) | 16:20 |
| fungi | right | 16:28 |
| opendevreview | Merged openstack/election master: Add CIVS opt in information in kickoff voting emails https://review.opendev.org/c/openstack/election/+/807271 | 16:40 |
| opendevreview | Merged openstack/election master: remove unicode from code https://review.opendev.org/c/openstack/election/+/769653 | 16:42 |
| lourot | clarkb, ack, thanks! | 16:54 |
| diablo_rojo_phone | Something else I should have mentioned during the TC meeting- there is a newsletter going out next week. so if any wants to suggest things I need them here( https://etherpad.opendev.org/p/newsletter-openstack-news) by EOD Monday | 19:43 |
| ade_lee__ | gmann, looks like somethings going on with the linters on https://zuul.opendev.org/t/openstack/build/0317357503424a68a141470cf77081b4 | 21:53 |
| ade_lee__ | gmann, something about "Unknown repository skyline/skyline-apiserver as part of skyline-apiserver in skyline | 21:53 |
| ade_lee__ | Unknown repository skyline/skyline-console as part of skyline-console in skyline" | 21:53 |
| fungi | that's almost certainly fallout from the skyline project rename into openstack | 21:55 |
| fungi | i thought there was a corresponding change to update them in governance already | 21:56 |
| fungi | ade_lee__: yep, either rebase on https://review.opendev.org/826243 or wait for it to merge | 21:56 |
| fungi | apparently it's stuck waiting for the skyline ptl to approve it. is that person aware? | 21:57 |
| fungi | also seems like a bit of unwarranted bureaucracy. that change is simply a reflection of the current state of the repositories, nothing that really demands acknowledgement | 21:58 |
| gmann | ade_lee__: yeah, we need to wait fir this to merge https://review.opendev.org/c/openstack/governance/+/826243 | 23:40 |
| gmann | tc-members: who ever 4 members have not voted on Z release name, please do. 5 members voted and 4 are not yet | 23:46 |
| dansmith | not me | 23:46 |
| dansmith | I mean, I have voted, it's not me in the "who ever 4" :) | 23:46 |
| gmann | ack :) | 23:46 |
| gmann | I think mnaser or Belmiro might not seen the poll yet | 23:47 |
| gmann | and I think jungleboyj will vote soon after twitter poll | 23:47 |
| gmann | diablo_rojo_phone: ricolin yoctozepto in case you have not voted ^^ | 23:49 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!