| ianw | i can't see away around this, i think we need to disable to job in zuul-client and do a release before zuul gate will be unbroken | 00:04 |
|---|---|---|
| clarkb | a release that doesn't pin requests? | 00:05 |
| ianw | yeah. because our tox role is always going to try and do an initial install of zuul-client from pypi | 00:05 |
| clarkb | and the next install doesn't bump up requests? | 00:06 |
| clarkb | (sorry today has been weird and I'm catching up) | 00:06 |
| ianw | the problem is that first install doesn't work, due to the pin | 00:06 |
| ianw | we could cap urllib in zuul i guess | 00:07 |
| fungi | or in zuul-client | 00:12 |
| fungi | oh, but that would have to be released | 00:13 |
| fungi | yeah | 00:13 |
| ianw | fungi: I don't think that will work, because it will try to isntall it from pypi always first | 00:13 |
| fungi | yup | 00:13 |
| ianw | or, force merge it and release ... i'm taking opinions :) | 00:13 |
| openstackgerrit | Ian Wienand proposed zuul/zuul master: requirements: temporarily pin urblib3 for zuul-client https://review.opendev.org/762593 | 00:15 |
| openstackgerrit | Ian Wienand proposed zuul/zuul master: requirements: temporarily pin urblib3 for zuul-client https://review.opendev.org/762593 | 00:16 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-client master: requirements: uncap requests https://review.opendev.org/762591 | 00:16 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 00:16 |
| ianw | zuul actually probably eneds to pin requests as well | 00:22 |
| fungi | or make the zuul-client job for zuul non-voting while the problem is addressed | 00:26 |
| ianw | or maybe not. as long as the urllib3 in the tox environment isn't >=1.26.0, then zuul-client should be able to install | 00:26 |
| ianw | ok, i think that swizzle will work | 00:36 |
| openstackgerrit | Ian Wienand proposed zuul/zuul master: Revert "requirements: temporarily pin urblib3 for zuul-client" https://review.opendev.org/762595 | 00:37 |
| *** Goneri has quit IRC | 00:38 | |
| *** zenkuro has quit IRC | 01:03 | |
| *** armstrongs has joined #zuul | 01:11 | |
| *** armstrongs has quit IRC | 01:20 | |
| *** polls45 has joined #zuul | 01:45 | |
| openstackgerrit | Merged zuul/zuul master: requirements: temporarily pin urblib3 for zuul-client https://review.opendev.org/762593 | 02:32 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 02:40 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 02:59 |
| *** rlandy|rover has quit IRC | 03:01 | |
| openstackgerrit | Merged zuul/zuul-client master: requirements: uncap requests https://review.opendev.org/762591 | 03:04 |
| *** bhavikdbavishi has joined #zuul | 03:05 | |
| *** bhavikdbavishi1 has joined #zuul | 03:12 | |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 03:13 |
| *** bhavikdbavishi has quit IRC | 03:13 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:13 | |
| *** rfolco has quit IRC | 03:14 | |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 03:30 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 03:41 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 03:48 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 03:49 |
| *** bhavikdbavishi has quit IRC | 03:54 | |
| *** bhavikdbavishi has joined #zuul | 03:54 | |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 04:07 |
| *** wuchunyang has joined #zuul | 04:08 | |
| *** wuchunyang has quit IRC | 04:12 | |
| *** bhavikdbavishi has quit IRC | 04:27 | |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 04:27 |
| *** bhavikdbavishi has joined #zuul | 04:28 | |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 04:40 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 04:55 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 04:58 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 05:04 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 05:10 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 05:25 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 05:32 |
| *** evrardjp has quit IRC | 05:33 | |
| *** evrardjp has joined #zuul | 05:33 | |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 05:34 |
| *** bhavikdbavishi1 has joined #zuul | 05:55 | |
| *** bhavikdbavishi has quit IRC | 05:56 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 05:56 | |
| *** tosky has joined #zuul | 06:24 | |
| openstackgerrit | Merged zuul/zuul master: Enable ANSI rendering via react-ansi https://review.opendev.org/739444 | 06:49 |
| *** rpittau|afk is now known as rpittau | 07:20 | |
| *** zenkuro has joined #zuul | 07:20 | |
| *** bhavikdbavishi has quit IRC | 07:30 | |
| *** bhavikdbavishi has joined #zuul | 07:53 | |
| *** bhavikdbavishi has quit IRC | 08:01 | |
| *** jcapitao has joined #zuul | 08:07 | |
| *** hashar has joined #zuul | 08:17 | |
| *** zenkuro has quit IRC | 08:21 | |
| *** zenkuro has joined #zuul | 08:21 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Drop python-passlib from run-buildset-registry https://review.opendev.org/762444 | 08:22 |
| openstackgerrit | Antoine Musso proposed zuul/nodepool master: Docs: correct static node uniqueness constraint https://review.opendev.org/756412 | 08:23 |
| hashar | ianw: I have fixed an easy conflict in a Nodepool document change you W+1 earlier ( https://review.opendev.org/#/c/756412/2 ) :] | 08:25 |
| *** zenkuro has quit IRC | 08:29 | |
| *** zenkuro has joined #zuul | 08:29 | |
| hashar | tobiash: danke schon! | 08:34 |
| tobiash | no problem :) | 08:35 |
| *** bhavikdbavishi has joined #zuul | 08:39 | |
| *** jpena|off is now known as jpena | 08:41 | |
| *** bhavikdbavishi1 has joined #zuul | 08:42 | |
| *** bhavikdbavishi has quit IRC | 08:43 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 08:43 | |
| *** nils has joined #zuul | 09:27 | |
| openstackgerrit | Merged zuul/nodepool master: Docs: correct static node uniqueness constraint https://review.opendev.org/756412 | 09:44 |
| *** bhavikdbavishi has quit IRC | 10:06 | |
| danpawlik | tobiash: if you are checking the docs, pls check also this one :) https://review.opendev.org/#/c/756671/ | 10:09 |
| danpawlik | cc ianw ^^ | 10:09 |
| tobiash | corvus, clarkb: how is opendev zuul accessing gerrit? Are you combining ssh stream events with http everything else? | 11:15 |
| tobiash | our zuul is still ssh only and I'm thinking about switching to http | 11:15 |
| avass | tobiash: isn't http only used to report line comments and kinds like that? I could be wrong | 11:28 |
| tobiash | it's at least needed for reporting line comments | 11:29 |
| avass | the documentation for the gerrit driver could need an update | 11:30 |
| *** zenkuro has quit IRC | 11:31 | |
| *** zenkuro has joined #zuul | 11:32 | |
| avass | looking at the documentation my understanding is http is only used for reporting | 11:34 |
| *** hashar is now known as hasharLunch | 11:35 | |
| *** jcapitao is now known as jcapitao_lunch | 11:35 | |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Add option to install kubernetes with kind https://review.opendev.org/740935 | 11:39 |
| avass | ianw: re 740935 | 11:39 |
| *** zenkuro has quit IRC | 11:41 | |
| avass | ianw: I feel like using 'latest' is better since it's less work to maintain to update that now and then. And it removes the expectation that the role will default to a specific version. | 11:41 |
| avass | which could break CI for someone if they rely on the role never changing default version to install | 11:41 |
| avass | and from what I've seen so far that default version has been whatever the latest version was when the role was added. So this feels less arbirary | 11:46 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: revoke-sudo: delete cloud-config sudoers file https://review.opendev.org/762198 | 11:52 |
| *** rfolco has joined #zuul | 11:59 | |
| *** jpena is now known as jpena|lunch | 12:28 | |
| *** Goneri has joined #zuul | 12:31 | |
| *** rlandy has joined #zuul | 12:34 | |
| *** rlandy is now known as rlandy|rover | 12:36 | |
| *** jcapitao_lunch is now known as jcapitao | 12:49 | |
| *** hasharLunch is now known as hashar | 12:54 | |
| openstackgerrit | Slawek Kaplonski proposed zuul/zuul-jobs master: [multi-node-bridge] Add script to configure connectivity https://review.opendev.org/762650 | 13:05 |
| *** jpena|lunch is now known as jpena | 13:33 | |
| openstackgerrit | Slawek Kaplonski proposed zuul/zuul-jobs master: [multi-node-bridge] Add script to configure connectivity https://review.opendev.org/762650 | 13:39 |
| *** polls45 has joined #zuul | 13:49 | |
| *** polls45 has quit IRC | 13:49 | |
| openstackgerrit | Slawek Kaplonski proposed zuul/zuul-jobs master: [multi-node-bridge] Add script to configure connectivity https://review.opendev.org/762650 | 14:06 |
| openstackgerrit | Merged zuul/zuul-jobs master: revoke-sudo: delete cloud-config sudoers file https://review.opendev.org/762198 | 14:09 |
| avass | this page: https://zuul.opendev.org/t/zuul/status/change/762650,3 could probably be narrower. We started reporting that and since the name of the job and the results are so far apart it's not easy to see which job failed | 14:24 |
| openstackgerrit | Merged zuul/zuul master: UI: Don't show "Invalid Date" for SKIPPED builds in builds table https://review.opendev.org/758532 | 14:32 |
| *** rpittau is now known as rpittau|afk | 14:49 | |
| corvus | tobiash: yes ssh for stream-events and http for reporting | 14:51 |
| corvus | avass: ++ re change status page | 14:53 |
| corvus | fungi, tobiash, ianw, zbr: i have some questions about https://review.opendev.org/739444 -- i see at least 3 outstanding issues that could potentially warrant not merging that change yet: 1) there are some concerns about dark/light backgrounds and contrast level (the dark bg is not very dark meaning the block is low-contast; and also we're putting small dark-bg areas inside of large light-bg areas). 2) the | 15:07 |
| corvus | output on the summary/console page doesn't match the text logs page (the text logs are not colorized; there was a change to do that, but it was way too slow, so it may not even be possible to do in the future). 3) it drops the horizontal scrolling. we had a whole discussion about that elsewhere, but there wasn't even any mention of that in the commit message or reviews. | 15:07 |
| corvus | 2/3 of those were raised in reviews but not addressed. | 15:08 |
| corvus | are we sure that's ready to merge? | 15:08 |
| zbr | imho, none of the mentioned issues is critical. | 15:09 |
| corvus | zbr: does that mean they won't be addressed? | 15:09 |
| zbr | i am more than happy to look into addressing them. | 15:09 |
| zbr | but we are not yet sure how and we do need feedback from users for some decisions. | 15:10 |
| corvus | i mean, the change itself isn't critical, so i don't feel like this is the sort of change where we have to merge it before we're satisfied | 15:10 |
| corvus | if we roll this out, people are going to be very unhappy if we revert it | 15:10 |
| zbr | i seen several people liking the dark background and mentioning that most CI do the same. | 15:10 |
| zbr | corvus: do we have a way to perform A/B testing with zuul? | 15:11 |
| corvus | "most ci do the same" is never going to convince me. make arguments on the merits. | 15:11 |
| zbr | some of the UI changes need "live" testing. | 15:11 |
| corvus | zbr: i don't have the capacity to do a/b testing, if you would like to and have the resources you can. | 15:12 |
| corvus | zbr: if you're saying we should merge that change now in order to collect feedback, what specifically are you looking to learn? | 15:12 |
| zbr | i personally do not care much about white vs black, i can survive with both. the reason it was dark is that this is the default config of the react-ansi, but I bet it has an invert option. | 15:12 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 15:13 |
| zbr | corvus: what is your no1 concern? i can start working on a follow-up for it | 15:14 |
| fungi | the argument i was reacting to is that there is an army of users demanding we have ansi colorization in the console view. i can only assume they all found this review naturally and left those +1 votes, not that they were dishonestly encouraged by the change author soliciting them | 15:14 |
| fungi | i'm trying to have good faith and not assume that people who propose changes in this community go out and seek otherwise disinterested reviewers to pump up +1 votes on their behalf in order to make a change seem more important than it is | 15:16 |
| zbr | i disagree here, gerrit users do not usually vote blindly (on generic online surveys I would have agreed with you that people may be less careful with their anwers) | 15:16 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 15:16 |
| zbr | fungi: true that I asked others to review but AFAIK, I always mentioned that I would not mind if they -1. | 15:17 |
| corvus | zbr: those concerns i had are in order: 1) sufficient contrast (i want to make sure the text is readable by people without perfect vision). 2) can we support this in the full-screen text logs? 3) let's be intentional about horizontal scrolling. if we agree to drop it, fine, but this feels like a backdoor workaround after we decided to leave it in place and make it a user toggle. | 15:18 |
| zbr | corvus: ok, starting to work on that right now. | 15:19 |
| zbr | i am aware that blue does not look ok on dark background and that is visible even in the screenshot i included, didn't try to hide it. | 15:20 |
| zbr | keep in mind that these are defaults, and changing CSS is not hard, i already know to which resources to look for good contrast themes. | 15:21 |
| corvus | zbr, fungi: fwiw, i'm not generally swayed by brigaded +1 comments. i do not think soliciting them is helpful, but if someone does it nonetheless, i'm not going to be persuaded. real user feedback is welcome, but should come with some useful information. | 15:21 |
| zbr | corvus: fungi: do you want me to try first to keep the white background, to minimize the differences? | 15:22 |
| corvus | zbr: i would be okay with either a white bg, or a dark one if it has been sufficiently evaluated to determine the contrast is optimal. | 15:23 |
| corvus | the live logs are white-on-black, and the black is black (not medium gray) | 15:24 |
| zbr | i will add an extra task that prints sample text so we can properly evaluate contrast. in fact I likely can provide a custom CSS to test changes before we propose. | 15:24 |
| zbr | corvus: if you ask an UX designer, they will tell you that purse black/white is not optimal for reading, there is such a thing as too big contrast. | 15:25 |
| corvus | zbr: yeah, i'm familiar with that. i don't know off the top of my head what is optimal | 15:25 |
| zbr | lucky for us there are plenty of good blog articles around the subject. no need to make our own research. | 15:26 |
| corvus | maybe the live logs are too contrasty? and maybe 739444 is not contrasty enough? | 15:26 |
| zbr | just to pick a theme that passed the test of time. | 15:26 |
| corvus | zbr: i think that's a great idea. either research from folks who have studied this (maybe even books or peer-reviewed articles!). or some samples we can ask a bunch of folks to evaluate. or both. | 15:27 |
| zbr | let me try to to come with a set today, if not it will be tomorrow. | 15:27 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 15:28 |
| *** ianychoi has joined #zuul | 15:30 | |
| zbr | i raised bug at https://github.com/RaoHai/react-ansi/issues/291 and now I starting to look at the code. | 15:33 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 15:38 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Add option to install kubernetes with kind https://review.opendev.org/740935 | 15:44 |
| avass | tobiash: fixed ^ | 15:44 |
| fungi | background color choice has some compelling arguments in either direction. light background with dark text is more consistent to what we already have and to what the rest of the interface still looks like (aside from the live console stream). dark background with light text is relevant *if* there's ansi color rendering going on, because those colors were likely chosen for contrast on dark/black background | 15:47 |
| fungi | terminals (though i know some folks do dark fg/light bg in their xterms, so not always) | 15:47 |
| corvus | fungi: ++ i'm generally convinced that the scales tip in favor of light-on-dark for ansi colors | 15:48 |
| corvus | (barely on topic: i used to make alternate versions of presentty presentations that were dark-on-light in case i got into a place where the projection was bad and that was necessary. i had to change several colors to make it work, so i've definitely seen that in action. happily, i've never had to use those presentations) | 15:50 |
| *** bhavikdbavishi has joined #zuul | 15:50 | |
| fungi | yeah, light magenta or light green on white is basically as illegible as blue on black | 15:56 |
| fungi | s/light/bold/ | 15:56 |
| fungi | at least in the 16-color palette | 15:56 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 15:57 |
| fungi | also if i'd spotted the change in scrolling behavior i wouldn't have +2'd it | 15:57 |
| corvus | what's worse is a lot of the newer versions of that palette make blue-on-black even harder. iirc under gnome-terminal i have to tell it to use the "linux" palette to get something halfway legible. | 15:58 |
| corvus | (i use gnome-terminal for presentty; otherwise urxvt) | 15:59 |
| clarkb | gnome tango is the color scheme I find readabel in a terminal iirc | 15:59 |
| fungi | interesting. i like to presentty with crt because it can lend a nice nostalgic feel (so long as you don't crank up the effects very far) | 15:59 |
| avass | I never felt like spending the time on getting urxwt to work the way I wanted | 15:59 |
| *** bhavikdbavishi1 has joined #zuul | 16:00 | |
| avass | so I'm using konsole which is default on manjaro kde. but that doesn't rewrap text when you resize the window :( | 16:00 |
| *** bhavikdbavishi has quit IRC | 16:02 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 16:02 | |
| fungi | crt shares konsole's backend renderer i think | 16:02 |
| corvus | avass: oof, i'm constantly resizing windows :/ | 16:05 |
| fungi | huh, crt does rewrap text on resize, i wonder why konsole doesn't | 16:05 |
| fungi | i would have expected that to be up to the qtermwidget backend | 16:06 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 16:07 |
| fungi | one of the reasons i like it is i'm particular about my column/row count and there's independent x and y scaling for its font renderer so i can force a consistent terminal geometry at any window size/resolution | 16:08 |
| pabelanger | ^ should be passing now for upload-container-image | 16:08 |
| pabelanger | lots of dancing trying to get container registry running under podman | 16:09 |
| pabelanger | so ended up doing multinode for it, and keeping docker command for it | 16:09 |
| avass | corvus: yeah me too since i use a lot of splits in i3 | 16:10 |
| avass | but I spend most of the time in vim anyway so it's not annoying enough to switch | 16:11 |
| fungi | for me it's vim/mutt/weechat/wyrd/gertty all in tmux | 16:14 |
| pabelanger | corvus: avass: tobiash: friday review if you are up for it: https://review.opendev.org/762581/ that is upload-container-image role, along with tests | 16:26 |
| pabelanger | promote I don't expect to right now, I still need to setup an intermediate registry | 16:27 |
| corvus | pabelanger: thanks! will look soon | 16:30 |
| corvus | pabelanger: promote is basically the one we can't test since it relies pretty heavily on specific upstream registry behavior | 16:30 |
| corvus | but at least it's simple to test manually/locally | 16:31 |
| clarkb | I'm responding to questions about doing different configs for jobs running on tags. For the stuff that landed in zuul to make that possible what does zuul do if a tag maps to multiple branches? does it just take the one that sorts first? | 16:33 |
| clarkb | corvus: ^ I think you wrote those changes | 16:33 |
| *** hashar has quit IRC | 16:33 | |
| corvus | If a tag item is enqueued, we look up the branches which contain the commit referenced by the tag. If any of those branches match a branch matcher, the matcher is considered to have matched. | 16:34 |
| corvus | This means that if a release job is defined on multiple branches, the branch variant from each branch the tagged commit is on will be used. | 16:34 |
| corvus | 16:34 | |
| corvus | clarkb: ^ | 16:34 |
| corvus | clarkb: commit 04ac8287b62f5d66356977d971c2518836615086 | 16:34 |
| clarkb | thanks | 16:34 |
| openstackgerrit | zbr proposed zuul/zuul master: WIP: Improve ANSI contrast https://review.opendev.org/762682 | 16:35 |
| avass | pabelanger: I don't see any obvious errors | 16:43 |
| fungi | clarkb: oh, i bet there's a parallel openstack-discuss ml thread going which i haven't gotten to yet, been discussing options in #openstack-release | 16:45 |
| fungi | maybe we should sync those up | 16:45 |
| clarkb | ya it was an openstack-discuss thread | 16:46 |
| fungi | i'll try to catch up on e-mail real quick, and you might want to skim the scrollback in that irc channel | 16:46 |
| openstackgerrit | zbr proposed zuul/zuul master: Bump mypy https://review.opendev.org/762187 | 16:47 |
| zbr | can I get some help consolidating the identation on our js/jsx files? | 16:51 |
| zbr | we almost have a 2 char ident, but is not enforced so we had several places where it accidentable diverged. | 16:52 |
| zbr | https://review.opendev.org/#/c/749702/1 is one of the patches. | 16:52 |
| zbr | the last one enables enforcement in eslint, to avoid regression. | 16:53 |
| tobiash | zbr: the second needs a rebase | 16:53 |
| zbr | i know | 16:53 |
| zbr | but lets fix them first, and i rebase, i got tired of rebasing the last one. | 16:53 |
| zbr | because these files are also modified by other patches, it is much easier to do a gradual fix. | 16:54 |
| zbr | so we avoid conflicts. | 16:54 |
| tobiash | k | 16:54 |
| *** jcapitao has quit IRC | 16:59 | |
| pabelanger | corvus: avass: thanks | 17:04 |
| pabelanger | yah for promote, I suspect we might have multiple roles, depending on upstream | 17:04 |
| pabelanger | next up, is to figure out ipv6 issue with zuul-registry and podman | 17:16 |
| pabelanger | then I can drop docker all together | 17:16 |
| avass | pabelanger: lgtm, fiy tobiash left a comment about the missing mode on a copy task | 17:18 |
| pabelanger | looking | 17:19 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 17:20 |
| zbr | do we have some code in zuul that looks at the console and that can generate gerrit lint-comments? | 17:21 |
| avass | zbr: the tox role does that | 17:22 |
| zbr | if we do i wanted to enable zuul to to recognize gihub-style annotations because many linters can already produce them, including ansible-lint. | 17:22 |
| pabelanger | q: I know this has been asked before, but cannot remember the answer. Why do we not have an 'auto recheck' / 'auto enquque' for changes, if a depends-on is updated while child change is still running? | 17:22 |
| avass | zbr: https://zuul-ci.org/docs/zuul/reference/jobs.html#leaving-file-comments | 17:22 |
| clarkb | its worth noting that github has come under fire for the way they do that stuff. We might want to double check our approach (though sticking it in ansible seems like it should make it much safer) | 17:23 |
| avass | I believe github is already supported as well | 17:23 |
| zbr | clarkb: we still have to translate what a cli produces to that file_comments stuff. | 17:24 |
| zbr | does it already work with pep8 format? if so it would be enough. | 17:24 |
| zbr | the key aspect here is to avoid having to change the tool (linter) in order to enable us to produce annotations. | 17:24 |
| pabelanger | zbr: we haven't really enabled github comments yet for zuul.a.c, aside from the default ones that zuul configure errors generates | 17:25 |
| zbr | these are not the same kind of commetns, and in fact you don't have to enable them, any cli tool that produce output in specified format is recognized. | 17:26 |
| avass | anyone for ensure-kubernetes with kind? https://review.opendev.org/#/c/740935/ | 17:26 |
| zbr | i wonder if there is a way to even disable them.... | 17:26 |
| avass | oh you mean that github picks it up automatically? | 17:26 |
| zbr | yes | 17:27 |
| avass | uh | 17:27 |
| zbr | you can produce them with echo | 17:27 |
| pabelanger | TIL | 17:27 |
| zbr | and they apply to the entire codebase, so you can produce comment to file that was not touched by the tested change, quite useful | 17:27 |
| avass | yay | 17:28 |
| zbr | for example when upgrading a linter or its configuration, you can easily get errors (or warnings) from various places. | 17:28 |
| corvus | zbr: if you would like that behavior in zuul, you can write an ansible role that greps the console log text and outputs the yaml for zuul | 17:28 |
| corvus | the console log is written to the executor work directory, so it's available. the logical place for such a role would be the first post-run playbook | 17:29 |
| zbr | corvus: yep, a simple ansible module that greps result.stdout of specific task and returns a file_comments dictionary. | 17:29 |
| zbr | and this way i even avoid the need to access the real console. | 17:29 |
| corvus | either way | 17:30 |
| zbr | because I already know which task is likely to produce annotations (tox run) | 17:30 |
| *** nils has quit IRC | 17:30 | |
| zbr | is there a special trick we can use to leave comment on commit message? | 17:31 |
| corvus | zbr: note that as avass points out, we already have a module for processing pep8 messages from tox runs | 17:32 |
| zbr | i found it! in that case I only need to add new feature to it. | 17:34 |
| corvus | zbr: i have no idea if it will work, but if you want to try, you could try "/COMMIT_MSG" as the path | 17:35 |
| zbr | gh format is much safer than pep8, less likely to match any false positive. | 17:36 |
| zbr | ok i will that in mind. | 17:36 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 17:42 |
| *** jpena is now known as jpena|off | 17:46 | |
| pabelanger | k, https://review.opendev.org/762581/ is passing again | 18:00 |
| openstackgerrit | zbr proposed zuul/zuul-jobs master: WIP: Enable detection of github annotations https://review.opendev.org/762698 | 18:10 |
| *** sean-k-mooney1 has joined #zuul | 18:41 | |
| *** sean-k-mooney has quit IRC | 18:42 | |
| pabelanger | tristanC: thanks | 18:52 |
| pabelanger | https://review.opendev.org/762581/ has 2 +2, could I ask for a +W from a 3rd person? | 18:52 |
| pabelanger | I know it is friday, was hoping to test out quay.io publish today | 18:53 |
| *** ikhan has joined #zuul | 19:03 | |
| corvus | pabelanger: the test seems more complicated than necessary | 19:12 |
| corvus | (i don't think the test framework for upload-container-image needs to do anything different than upload-docker-image) | 19:14 |
| pabelanger | the issue was, I couldn't get podman to work with buildx | 19:21 |
| pabelanger | which is why I kept the registry as docker | 19:21 |
| pabelanger | I'm working on new PS to try again | 19:21 |
| corvus | i don't understand what buildx has to do with it | 19:21 |
| pabelanger | multiarch images uses it | 19:22 |
| corvus | i think those are 2 separate issues | 19:22 |
| pabelanger | let me push up a new change ripping out the 2nd node, which should expose the issue | 19:22 |
| corvus | does build-container-image work with multiarch? | 19:23 |
| pabelanger | no | 19:23 |
| pabelanger | we only test single arch right now | 19:23 |
| corvus | didn't think so; i think it's fine for upload-container-image not to work with multiarch either | 19:23 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Rename utils-tasks/run-docker-registry.yaml https://review.opendev.org/762711 | 19:27 |
| pabelanger | IIRC, podman will bomb on running registry:2 image | 19:28 |
| corvus | pabelanger: we don't need podman to run the registry, we can have docker run it | 19:29 |
| pabelanger | will a node work with both podman and docker installed? | 19:29 |
| pabelanger | I though they conflicted with each other | 19:29 |
| corvus | i'm not familiar with that being a problem; i think i have both installed on my workstation | 19:30 |
| pabelanger | k, if that is the case, lets test | 19:30 |
| corvus | i believe the test job is currently written to specify that docker should run the publication registry; so no change should be needed there | 19:31 |
| pabelanger | I thought at one point podman updated the docker cmd to symlink to podman | 19:31 |
| corvus | neither appears to be a symlink on my ws | 19:31 |
| fungi | might depend on how it's installed... i recall if you install the distro package for "docker" on rhel/fedora you get a docker command which is actually podman? | 19:32 |
| fungi | but i thought that came down to just how the distro packaging for it was done | 19:33 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Simplify testing for upload-container-image role https://review.opendev.org/762711 | 19:34 |
| pabelanger | if that works, I'll squash into parent | 19:35 |
| corvus | yeah something like that is what i was thinking | 19:43 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Simplify testing for upload-container-image role https://review.opendev.org/762711 | 19:48 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Simplify testing for upload-container-image role https://review.opendev.org/762711 | 19:57 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Simplify testing for upload-container-image role https://review.opendev.org/762711 | 19:57 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 20:03 |
| pabelanger | corvus: have to run, but ^ should do what you ask. We still need podman config for insecure registry however | 20:03 |
| pabelanger | guess docker doesn't need it by default | 20:04 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Create upload-container-image role https://review.opendev.org/762581 | 20:04 |
| corvus | pabelanger: yeah, i think docker treats 127.0.0.1 as implied insecure | 20:04 |
| pabelanger | will check back in a few hours | 20:05 |
| pabelanger | thanks all | 20:05 |
| corvus | pabelanger: ty | 20:05 |
| *** bhavikdbavishi has quit IRC | 20:20 | |
| fungi | or rather, implied secure (no need to check the ssl cert) | 20:21 |
| corvus | indeed; i was trying to say that the option with the word 'insecure' in the name was enabled by implication :) | 20:22 |
| fungi | yep, got it | 20:39 |
| *** rfolco has quit IRC | 21:10 | |
| avass | corvus, pabelanger: catching up, podman and docker can both run as long ad you use cgroups v1 | 21:19 |
| *** rlandy|rover has quit IRC | 21:19 | |
| avass | what's the benefit of running a merger instead of an extra executor? | 21:21 |
| avass | the merger is dedicated to merging and does not waste cpu/disk/ram on ansible? | 21:22 |
| clarkb | avass: we run extra mergers that are much smaller than executors | 21:23 |
| clarkb | so ya its basically that, you can scale out merging without adding unnecessary resource overhead | 21:23 |
| avass | ah, since we only need those during restarts and no ansible jobs would be running then anyway | 21:23 |
| clarkb | in poendev I think we worked it out to about 80 nodes per 8GB ram x 8vcpu executor but we still would have slow merging after that | 21:23 |
| clarkb | so we add smaller cheaper mergers just to ahndle the merging | 21:24 |
| *** sean-k-mooney2 has joined #zuul | 21:54 | |
| *** sean-k-mooney1 has quit IRC | 21:56 | |
| ianw | corvus: i don't think the change modifies the horizontal scrolling over what we have now? it seems to me that it's in a box you can only move horizontally by scrolling to the bottom of the page. it's been like this on the opendev.org site for quite a long time, i'm not sure if there has been a release with it like that though | 22:03 |
| avass | ianw: wanna take a look at: https://review.opendev.org/#/c/740935/ again? :) | 22:08 |
| corvus | ianw: pretty sure it looked different when i compared the same output live and via the preview. | 22:08 |
| corvus | ianw: istr that the patch caused it to wrap text | 22:11 |
| corvus | i can load it up again and make screenshots if we're not in agreement that it's different | 22:11 |
| *** hamalq has joined #zuul | 22:30 | |
| avass | we just migrated to k8s, added zookeeper tls, and updated to the latest version of zuul :) | 22:42 |
| fungi | congrats! | 22:57 |
| corvus | \o/ | 23:07 |
| corvus | it is slightly windy and rainy here which has caused the power to go out | 23:08 |
| corvus | so i will start shutting down | 23:08 |
| fungi | enjoy your battery-powered entertainment | 23:08 |
| fungi | or looking out the window | 23:08 |
| avass | or you know, reading a book or something | 23:09 |
| fungi | true, some of those have really long battery life | 23:09 |
| clarkb | we've had stormy weather here too, but no power outages thankfully | 23:19 |
| *** corvus has quit IRC | 23:22 | |
| *** rfolco has joined #zuul | 23:22 | |
| *** jeblair has joined #zuul | 23:25 | |
| *** jeblair is now known as corvus | 23:26 | |
| corvus | wow, the power outage lasted *literally* 15 seconds longer than my ups battery. | 23:26 |
| *** rfolco has quit IRC | 23:26 | |
| fungi | we're going to need a bigger ups | 23:27 |
| corvus | it's probably time for new batteries; it's been like 6 years | 23:27 |
| corvus | though really, i'd like to get a whole house battery + solar | 23:27 |
| fungi | i replaced a lot of my battery backups with smaller straight dc power supplies, lots more battery capacity when it's not wasted on dc->ac->dc conversions | 23:28 |
| clarkb | my home networking setup is super low power and the fios termination has its own ups too. We had a 4 hour outage a while back and everything stayed up whcih emant I could stick to the laptop and be fine | 23:28 |
| fungi | but yeah, my fantasy is whole house generator+battery | 23:29 |
| fungi | we get enough outages here it would be worthwhile | 23:29 |
| fungi | especially during/after major storms | 23:29 |
| avass | I can't remember the last time we had a power outage | 23:29 |
| fungi | would love to not have to worry about losing everything in my big freezer the next time we get a multi-day power outage | 23:30 |
| avass | uuh, I don't I've ever had a power outage that lasterd more than 1-2 hours | 23:30 |
| corvus | nice thing here is that it's easy to be fully self sufficient with solar | 23:31 |
| avass | but we also don't really have a lot of storms here either | 23:31 |
| fungi | the whole island chain here can end up without power for days if the lines from the mainland get severed | 23:31 |
| avass | during the winter we don't have a lot of sun either :) | 23:32 |
| clarkb | avass: that is our problem here as well. The sun goes away for like 4 months | 23:33 |
| corvus | avass: when i was young, a finnish person explained to me that's why all the best programmers are from finland. | 23:33 |
| fungi | we get an okay amount of sun, but i could probably get a decent amount of power generation out of tidal and wind | 23:34 |
| avass | heh | 23:34 |
| avass | I'm technically half finnish.. sooo... ;) | 23:35 |
| corvus | ('best programmer' in that case apparently meant 'able to make the most impressive 640k demos') | 23:35 |
| fungi | i've seen some really good c64 demos, i bet they were finnish | 23:36 |
| avass | I don't think I even know what that means | 23:36 |
| clarkb | fungi: did you have to type them in from a magazine to see them? | 23:37 |
| *** hamalq has quit IRC | 23:37 | |
| fungi | clarkb: no, no, there were disks | 23:38 |
| fungi | and bulletin boards | 23:38 |
| corvus | avass: https://en.wikipedia.org/wiki/Demoscene | 23:38 |
| *** hamalq has joined #zuul | 23:38 | |
| corvus | avass, fungi: particularly https://en.wikipedia.org/wiki/Future_Crew was memorable | 23:38 |
| fungi | oh, yep! | 23:39 |
| corvus | (finnish and c64 to start, pc later) | 23:39 |
| avass | oh yeah I think I've seen that before | 23:42 |
| *** tosky has quit IRC | 23:43 | |
| corvus | they should put 'second reality' on the unesco cultural heritage list :) | 23:43 |
| avass | oh that was made in 1993 | 23:50 |
| fungi | those were my university days | 23:51 |
| corvus | was mind-blowing that could all be calculated/rendered in real-time, and in such a small executable | 23:52 |
| fungi | also inspiring | 23:52 |
| corvus | on a 33mhz machine | 23:52 |
| corvus | i love that it's 10 mins of demo plus another 5 of greetz | 23:57 |
| avass | fungi: well it was two years before I was born :) | 23:57 |
| clarkb | with a turbo button? | 23:57 |
| fungi | i wish i was born with a turbo button | 23:58 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!