| corvus | mnaser: https://review.opendev.org/621479 is the long-term plan for that. i'm hoping we'll be able to start on that later this year | 00:08 |
|---|---|---|
| mnaser | Neat! | 00:09 |
| corvus | as Shrews points out, we're already starting on it really, with his change, and i think we're really close to being ready for the sql reporter change | 00:09 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Refactor build page tabs https://review.opendev.org/675235 | 00:28 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add permalinks to task detail popup https://review.opendev.org/675236 | 00:28 |
| corvus | that's clarkb's other request from yesterday; i think that's all known issues accounted for | 00:28 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Scroll log line anchor into view https://review.opendev.org/675220 | 00:29 |
| *** threestrands has joined #zuul | 00:46 | |
| *** jamesmcarthur has joined #zuul | 01:05 | |
| *** jamesmcarthur has quit IRC | 01:11 | |
| *** jamesmcarthur has joined #zuul | 01:14 | |
| *** wxy-xiyuan has joined #zuul | 01:15 | |
| *** jamesmcarthur has quit IRC | 01:23 | |
| *** jamesmcarthur has joined #zuul | 01:26 | |
| *** tributarian has quit IRC | 01:37 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Refactor build page tabs https://review.opendev.org/675235 | 01:37 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add permalinks to task detail popup https://review.opendev.org/675236 | 01:37 |
| *** jamesmcarthur has quit IRC | 01:40 | |
| *** rlandy|rover|bbl is now known as rlandy|rover | 01:45 | |
| *** rlandy|rover has quit IRC | 01:50 | |
| *** jhesketh has quit IRC | 01:53 | |
| *** bhavikdbavishi has joined #zuul | 02:00 | |
| *** bhavikdbavishi1 has joined #zuul | 02:03 | |
| *** bhavikdbavishi has quit IRC | 02:04 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 02:04 | |
| *** jamesmcarthur has joined #zuul | 02:04 | |
| *** jamesmcarthur has quit IRC | 02:17 | |
| *** jhesketh has joined #zuul | 02:22 | |
| *** jamesmcarthur has joined #zuul | 02:28 | |
| *** bhavikdbavishi has quit IRC | 02:32 | |
| *** igordc has joined #zuul | 03:16 | |
| *** jhesketh has quit IRC | 03:29 | |
| *** bhavikdbavishi has joined #zuul | 03:33 | |
| *** jhesketh has joined #zuul | 03:36 | |
| *** raukadah is now known as chkumar|ruck | 03:49 | |
| *** jamesmcarthur has quit IRC | 04:01 | |
| *** jamesmcarthur has joined #zuul | 04:05 | |
| *** jamesmcarthur has quit IRC | 04:09 | |
| *** jamesmcarthur has joined #zuul | 04:10 | |
| *** jamesmcarthur has quit IRC | 04:19 | |
| *** jamesmcarthur has joined #zuul | 04:23 | |
| *** jamesmcarthur has quit IRC | 04:33 | |
| *** jamesmcarthur has joined #zuul | 04:34 | |
| *** jamesmcarthur has quit IRC | 04:42 | |
| *** jamesmcarthur has joined #zuul | 04:42 | |
| *** jamesmcarthur has quit IRC | 04:48 | |
| *** spsurya has joined #zuul | 05:02 | |
| *** lennyb has quit IRC | 05:26 | |
| *** bjackman has joined #zuul | 05:40 | |
| *** igordc has quit IRC | 05:59 | |
| *** lennyb has joined #zuul | 06:27 | |
| *** ianychoi has quit IRC | 07:01 | |
| *** ianychoi has joined #zuul | 07:01 | |
| *** themroc has joined #zuul | 07:16 | |
| *** tosky has joined #zuul | 07:17 | |
| *** themroc has quit IRC | 07:21 | |
| *** ianychoi has quit IRC | 07:24 | |
| *** ianychoi has joined #zuul | 07:25 | |
| *** themroc has joined #zuul | 07:28 | |
| *** panda has quit IRC | 07:35 | |
| *** panda has joined #zuul | 07:38 | |
| *** threestrands has quit IRC | 08:19 | |
| openstackgerrit | Mark Meyer proposed zuul/zuul master: Rework a cache invalidation issue https://review.opendev.org/674425 | 08:26 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: DNM: Test iptables_rules on Suse https://review.opendev.org/675270 | 08:28 |
| openstackgerrit | Merged zuul/zuul-jobs master: Be consistent about spaces before and after vars https://review.opendev.org/667698 | 09:02 |
| *** electrofelix has joined #zuul | 09:12 | |
| *** AshBullock has joined #zuul | 09:14 | |
| AshBullock | Hey all, I'm interested in passing the comment that triggers a job to the job itself as in zuul vars so that it could be used by our playbooks, does anyone know how feasible this would be? | 09:16 |
| *** dmellado has quit IRC | 09:53 | |
| *** dmellado has joined #zuul | 09:55 | |
| *** AshBullock has quit IRC | 10:28 | |
| *** tobiash has quit IRC | 11:18 | |
| *** tobiash has joined #zuul | 11:20 | |
| *** tobiash has quit IRC | 11:24 | |
| *** tobiash has joined #zuul | 11:26 | |
| chkumar|ruck | Does required projects defined in parent job gets inherited in child job? | 11:43 |
| *** jamesmcarthur has joined #zuul | 11:51 | |
| *** jamesmcarthur has joined #zuul | 11:51 | |
| *** jamesmcarthur has quit IRC | 12:02 | |
| *** jamesmcarthur has joined #zuul | 12:04 | |
| *** rfolco has joined #zuul | 12:12 | |
| *** jamesmcarthur has quit IRC | 12:21 | |
| AJaeger | chkumar|ruck: yes - and you can check this in the log file as well | 12:26 |
| *** rlandy has joined #zuul | 12:28 | |
| *** rlandy is now known as rlandy|rover | 12:28 | |
| *** bjackman has quit IRC | 12:34 | |
| tobiash | Shrews, corvus: any objections against adding cleanup of leaked volumes to nodepool? We have the problem that we leak a percentage of volumes that are created using boot-from-volume (they should get deleted automatically by openstack but that doesn't seem to work reliably). | 12:41 |
| *** jamesmcarthur has joined #zuul | 12:54 | |
| Shrews | tobiash: well, i've been trying to hunt down why that happens. I thought I found it, but apparently was incorrect. What I see in our logs is the first attempt fails because the volume is still in use (Conflict exception), but it never tries again because the znode for it disappears somewhere. I still don't know why that happens. | 13:07 |
| tobiash | Shrews: so nodepool itself tries to delete it? I thought it marks them just as 'delete together with vm' and openstack automagically does that? | 13:08 |
| Shrews | tobiash: oh, you typed 'volume' but i read 'image' *sigh* lack of coffee | 13:09 |
| tobiash | Shrews: well, we also leak images so we have that problem as well ;) | 13:09 |
| tobiash | but the more pressing one for us is volumes atm | 13:09 |
| Shrews | tobiash: I'd be a bit nervous about volume cleanup. Accidentally deleting a volume could be very not good. | 13:11 |
| tobiash | Shrews: the volumes have nodepool specific metadata | 13:11 |
| Shrews | If we can do it safely, I'm all for it | 13:11 |
| tobiash | so we could do it like ports, list them, filter by nodepool metadata and available and delete if we saw them a second time | 13:12 |
| *** themroc has quit IRC | 13:15 | |
| *** themroc has joined #zuul | 13:16 | |
| *** jeliu_ has joined #zuul | 13:19 | |
| *** Diabelko has quit IRC | 13:39 | |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: WIP: Add zuul-operator-functional-openshift job https://review.opendev.org/674355 | 13:44 |
| *** electrofelix has quit IRC | 13:57 | |
| *** tributarian has joined #zuul | 13:58 | |
| corvus | clarkb, mordred, tristanC: would you please review any changes in "topic:console" and go ahead and approve them when they have 2x+2 ? | 14:01 |
| tobiash | corvus: awesome work :) | 14:02 |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: WIP: Add zuul-operator-functional-openshift job https://review.opendev.org/674355 | 14:05 |
| mordred | corvus: stack looks great! | 14:11 |
| corvus | mordred: don't forget https://review.opendev.org/675220 | 14:12 |
| corvus | it's independent | 14:12 |
| mordred | corvus: ++ | 14:21 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: web: refactor the errorsIds into the build action https://review.opendev.org/675350 | 14:28 |
| corvus | tristanC: thanks! | 14:29 |
| *** rlandy|rover is now known as rlandy|rover|mtg | 14:31 | |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: WIP: Add zuul-operator-functional-openshift job https://review.opendev.org/674355 | 14:41 |
| corvus | tristanC: i went to dig up a zuul-preview url for you, and i think there's a bug that allows it to be an open proxy, so i've shut it down for now, sorry | 15:02 |
| tristanC | corvus: that's ok, i'll use my local yarn run then | 15:04 |
| *** themroc has quit IRC | 15:08 | |
| *** rlandy|rover|mtg is now known as rlandy|rover | 15:20 | |
| openstackgerrit | Merged zuul/zuul master: Scroll log line anchor into view https://review.opendev.org/675220 | 15:20 |
| corvus | haha, there's a comment in zuul: https://bugs.executepad.net/zuul/+bug/1078946 | 15:29 |
| corvus | looks like one of the s/launch/execute/ seds got away from someone :) | 15:30 |
| fungi | i can't reach bugs.executepad.net | 15:33 |
| *** spsurya has quit IRC | 15:34 | |
| AJaeger | fungi: s/launch/execute/ - I felt in the same trap, so use bugs.launchpad.net/zuul/+bug/1078946 | 15:34 |
| fungi | d'oh! | 15:34 |
| fungi | so *that's* what corvus meant | 15:34 |
| AJaeger | fungi: the problem is that the comment is wrong | 15:34 |
| AJaeger | Zuul has that broken URL as comment | 15:35 |
| fungi | i thought that looked remarkably like a launchpad url, and wondered if someone was running a private lp instance at that domain at first | 15:35 |
| AJaeger | fungi: for both of us, this URL would have been better: https://opendev.org/zuul/zuul/src/branch/master/tests/unit/test_scheduler.py#L2070 | 15:36 |
| fungi | ahh | 15:36 |
| fungi | now it's making sense | 15:36 |
| corvus | this way you got to have the same experience i did :) | 15:37 |
| corvus | group bonding activity | 15:37 |
| AJaeger | https://review.opendev.org/#/c/445594/ - by pabelanger ;) | 15:37 |
| AJaeger | "Rename zuul-launcher to zuul-executor" | 15:37 |
| fungi | yes, this was a fun trip through zuul history | 15:38 |
| AJaeger | ;) | 15:39 |
| mordred | hahaha | 15:39 |
| clarkb | tobiash: the other day I noticed something weird with zuul's ansible installation management. Let me get a few links together but I was hoping to talk to you about that today to see if I am missing anything | 15:50 |
| tobiash | clarkb: was it failing? I think I also saw some unzip related errors (which are probably due to parallel installation of all envs) | 15:51 |
| tobiash | we should think about serializing those | 15:51 |
| clarkb | tobiash: no I think we use inconsistent installation targets depending on whether or not zuul automanages the venvs or if zuul's venv installer command is used | 15:52 |
| clarkb | tobiash: at https://opendev.org/zuul/zuul/src/branch/master/zuul/lib/ansible.py#L44-L45 we default to /usr/lib/zuul/ansible as the installation target when no other directory is provided. This is the case with the zuul manage_ansible.py tooling | 15:52 |
| tobiash | clarkb: I have to run now, I'll look later at home | 15:53 |
| clarkb | tobiash: but at https://opendev.org/zuul/zuul/src/branch/master/zuul/executor/server.py#L2400-L2408 if the executor config is set to manage_ansible automatically it defaults to os.path.join(state_dir, 'ansible-bin') and the state_dir is /var/lib/zuul by default | 15:54 |
| clarkb | I think this means that if you use the manage_ansible command things get instlaled to /usr/lib/zuul/ansible and if you use [executor] manage_ansible = True things intsall to /var/lib/zuul/ansible-bin | 15:55 |
| clarkb | If this is indeed the case I think it would be safe to update the [executor] manage_ansible case to simply intsall things to /usr/lib/zuul/ansible instead and have a release note that people can clean up /var/lib/zuul/ansible-bin ? | 15:56 |
| clarkb | tobiash: ok, I'm going to push up the change to do ^ and that can collect the discussion | 15:57 |
| openstackgerrit | Clark Boylan proposed zuul/zuul master: Consistent handling of ansible venv install dirs https://review.opendev.org/675403 | 16:26 |
| clarkb | tobiash: ^ I think that explains what I found and cleans it up. And that cleanup should be backward compatible as any override paths are still searched if provided. | 16:27 |
| clarkb | corvus: ^ that is related to my confusion finding the ansible venvs on tuesday | 16:27 |
| clarkb | corvus: did the ara version specifier fix solve the ara on zuul jobs with swift problem? | 16:27 |
| openstackgerrit | Merged zuul/zuul master: Render console in js https://review.opendev.org/674368 | 16:28 |
| *** mattw4 has joined #zuul | 16:29 | |
| openstackgerrit | Merged zuul/zuul master: Usability tweaks for the build page console https://review.opendev.org/675147 | 16:34 |
| *** armstrongs has joined #zuul | 16:35 | |
| *** AshBullock has joined #zuul | 16:37 | |
| *** mattw4 has quit IRC | 16:39 | |
| *** mattw4 has joined #zuul | 16:39 | |
| *** chkumar|ruck is now known as raukadah | 16:42 | |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: WIP: Add zuul-operator-functional-openshift job https://review.opendev.org/674355 | 16:42 |
| AshBullock | Hey all, I'm interested in passing the comment that triggers a job to the job itself as in zuul vars so that it could be used by our playbooks, does anyone know how feasible this would be? | 16:44 |
| *** jangutter has quit IRC | 16:44 | |
| AshBullock | for a gitub pr comment being used as a trigger | 16:45 |
| fungi | i thought someone said that was already provided as a zuul var | 16:47 |
| clarkb | AshBullock: off the top of my head I think you would have to plumb that comment data through zuul from the trigger driver all the way through the scheduler and down into the jobs | 16:47 |
| clarkb | oh /me checks if that is already done | 16:47 |
| fungi | i'll see if i can find documenation | 16:47 |
| fungi | documentation too ;) | 16:47 |
| corvus | i don't think the review comment is, i think just the commit message is | 16:48 |
| corvus | that's a tricky one, because jobs are meant to be independent of their triggers | 16:48 |
| *** tosky has quit IRC | 16:49 | |
| fungi | ahh, okay, so commit message (or pr description presumably?) | 16:49 |
| corvus | so you should get the same result from a job triggered by a patch upload, a 'recheck' comment, a web trigger, an email trigger, etc... | 16:49 |
| corvus | a timer trigger, a zuul trigger.... | 16:49 |
| corvus | you get the idea | 16:49 |
| corvus | AshBullock: can you describe the use case a bit more? | 16:49 |
| clarkb | corvus: ya confirmed the commit message is base64 encoded but I don't see the review comment | 16:51 |
| clarkb | is it known that zuul's zuul status page links to job logs don't work? | 16:56 |
| fungi | clarkb: yes, if you mean ones for builds in sets which have yet to report | 16:56 |
| clarkb | oh ya that explains it | 16:56 |
| clarkb | the sql reporter doesn't report until jobs complete | 16:56 |
| clarkb | *all jobs in buildset | 16:56 |
| fungi | right now it's linking to the result page which won't exist until sql is reported into | 16:56 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add option to report build page https://review.opendev.org/675409 | 16:57 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add release note for Pagure driver https://review.opendev.org/675410 | 16:57 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Move admin-rules setting in tenants doc https://review.opendev.org/675411 | 16:57 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Make auth docs more boring https://review.opendev.org/675412 | 16:57 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Don't capitalize Token in docs https://review.opendev.org/675413 | 16:57 |
| fungi | clarkb: i believe the idea was that we would for now switch those to direct log links and later make sql reporting mandatory and shift it to happen as each build completes | 16:57 |
| corvus | and there's the change to do it :) | 16:57 |
| corvus | also, docs changes since i was in the neighborhood | 16:58 |
| fungi | righteous! | 16:58 |
| AshBullock | Yea, after some digging and checking the zuul vars I saw it is only the commit message that gets passed through. | 16:58 |
| corvus | clarkb, fungi, mordred, tobiash: can you put https://review.opendev.org/675409 at the top of your list? we'll need that merged by monday if we're going to meet my self-imposed deadline to make that switch :) | 16:59 |
| *** hwangbo has joined #zuul | 16:59 | |
| AshBullock | clarkb yea I had looked at the connector for git and it does look non trivial to pass this all the way through the chain | 16:59 |
| clarkb | corvus: yup | 16:59 |
| AshBullock | corvus Essentially we had an idea of controlling the flow of our playbooks based on the contents of the message, so placing a variable into the comment | 17:00 |
| tobiash | corvus: I'll look in a bit | 17:01 |
| corvus | AshBullock: the problem that brings up is that it doesn't match the gating paradigm; if the behavior of the job changes because of manual intervention (ie, text in a comment), then it's not automatically reproducible. whatever you did to make the job run that way when checking it, you would also have to do for gating, and it may not match the way the job runs on the next change through the gate. | 17:04 |
| corvus | AshBullock: so first, i'd suggest being very careful of going down that path -- the more jobs behave deterministically, the more you can automate and gate things. | 17:05 |
| corvus | AshBullock: however, there are use cases where folks may want to run more experimental variants of jobs, with the understanding they could not be used for gating, or deployment, etc.. | 17:06 |
| corvus | AshBullock: we had a long design discussion with tristanC about that, and came up with a solution in https://review.opendev.org/555153 (see the review comments there, but also especially the links to irc discussions). the outcome of that is the system in https://review.opendev.org/639905 | 17:07 |
| *** sshnaidm is now known as sshnaidm|off | 17:10 | |
| *** igordc has joined #zuul | 17:13 | |
| AshBullock | yea completely agree with the fact that jobs should be reproducible tracing back to version control, the use case specifically was to run a development deployment job and to override the target environment without having to update the zuul config. It was an idea to speed up our development workflow | 17:15 |
| fungi | clarkb: regarding your comment on 675409 there's a good reason for the conditional, see reply | 17:17 |
| clarkb | fungi: good point and ya your fix handles that | 17:18 |
| fungi | to be clear, i think it's fine as-is, this is not a performance-critical code path and the conditional is more explicit | 17:18 |
| corvus | AshBullock: gotcha, that matches the use case tristanC was working on pretty closely, so i think once we get that in, you should be able to accomodate it | 17:20 |
| AshBullock | corvus thanks for all the information, I'll keep an eye on this change :) | 17:21 |
| *** jamesmcarthur has quit IRC | 17:22 | |
| *** armstrongs has quit IRC | 17:22 | |
| openstackgerrit | Clark Boylan proposed zuul/zuul master: Consistent handling of ansible venv install dirs https://review.opendev.org/675403 | 17:25 |
| *** AshBullock has quit IRC | 17:27 | |
| tristanC | corvus: AshBullock: actually, we kind of stepped down from that webtrigger proposal and took this git gateway system path: https://pagure.io/software-factory/zuul-gateway | 17:30 |
| tristanC | corvus: AshBullock: it would be nice if this was supported natively in Zuul (parameterized build), but until then, this trick works quite well for our needs | 17:31 |
| corvus | tristanC: yeah, i think we should still work on merging the patches you have up | 17:32 |
| corvus | i think it's a good feature | 17:32 |
| *** jamesmcarthur has joined #zuul | 17:33 | |
| tristanC | corvus: one use-case missing from the current patches is being able to override nodeset too | 17:33 |
| tristanC | iirc, that was the motivation of writting that zuul-gateway | 17:35 |
| corvus | ack; sounds like a good thing to add; i'd have to refresh my memory a bit to know how hard it would be | 17:38 |
| tristanC | corvus: for the web interface, we would need that to display a drop-down: https://review.opendev.org/#/c/605877/ | 17:43 |
| *** ianychoi has quit IRC | 17:43 | |
| *** ianychoi has joined #zuul | 17:47 | |
| *** panda has quit IRC | 17:53 | |
| *** ianychoi_ has joined #zuul | 17:54 | |
| *** panda has joined #zuul | 17:54 | |
| *** ianychoi_ has quit IRC | 17:55 | |
| *** ianychoi_ has joined #zuul | 17:55 | |
| *** ianychoi has quit IRC | 17:58 | |
| corvus | clarkb: can you review topic:console when you have a few mins? | 17:58 |
| clarkb | corvus: ya | 17:58 |
| mordred | corvus: https://review.opendev.org/#/c/675409 ... looking at https://review.opendev.org/#/c/675409/1/zuul/scheduler.py - it looks like it always adds t/{tenant.name} to the webroot ... what about white label? | 18:00 |
| corvus | mordred: see tenant.web-root | 18:00 |
| *** jeliu_ has quit IRC | 18:01 | |
| mordred | corvus: thank you. I was conflating two things in my brainhole :) | 18:01 |
| *** jeliu_ has joined #zuul | 18:02 | |
| mordred | corvus: that has 3x+2 now - any reason to not +A it? | 18:04 |
| fungi | luckily gertty showed me the documentation changes and release notes and tests/fixtures before the implementation | 18:04 |
| fungi | which was probably serendipitous but made it an easy review as a result | 18:04 |
| mordred | fungi: I have this habit of reading zuul changes upside down - it usually helps me but today it did not | 18:04 |
| fungi | you have to wait longer for all the blood to rush to your head | 18:05 |
| mordred | yeah- but once it does boy howdy | 18:06 |
| *** igordc has quit IRC | 18:28 | |
| *** jamesmcarthur has quit IRC | 18:30 | |
| pabelanger | tobiash: have you been running from master for zuul? Curious how you are making how with recent github improvements around github events | 18:37 |
| *** jamesmcarthur has joined #zuul | 18:42 | |
| tobiash | pabelanger: we're running currently from e858b8df (8 days ago) | 18:44 |
| pabelanger | kk, thanks! | 18:45 |
| tobiash | pabelanger: and I didn't notice anything special with github events anymore | 18:45 |
| tobiash | runs smoothly with that regard | 18:45 |
| pabelanger | okay cool, I think we are starting to slow down a little with all the events from ansible/ansible. So looking forward to testing out for next released | 18:45 |
| *** rfolco has quit IRC | 18:51 | |
| tobiash | clarkb: I'm wondering, what the problem with the ansible installation dirs is? The system is meant to either use pre-installed ansible (where the lib dir is correct) or the dynamic variant (where we need to use the state dir because of guaranteed write access). | 18:52 |
| tobiash | it's not supposed to be mixed | 18:52 |
| clarkb | tobiash: the issue for me is lack of consistency and confusion | 18:57 |
| clarkb | why cant the manage ansible command install to the same location and then when you want to find the venvs you arent looking in multiple locations? | 18:58 |
| clarkb | we needed to debug venv contents a couple days ago and I could not find them because code says one thing but really uses another | 18:58 |
| tobiash | clarkb: using the state dir requires a proper zuul configuration and inhibits the possibility to install ansible already into the container image (because you normally need to bind mount a volume there). While installing into the lib dir is only dependent on the install location. | 19:00 |
| *** rfolco has joined #zuul | 19:00 | |
| tobiash | so if we want to make this consistent I'd rather suggest to drop the dynamic installation completely as this is only intended to ease quick start of zuul | 19:01 |
| clarkb | for some reason I thought docker copied over the contents from image to mount but maybe that was an init process? | 19:02 |
| clarkb | or maybe that is volumes vs bind mounts | 19:02 |
| tobiash | no, it doesn't, any existing directory is completely shadowed and unaccessible when bind mounting a volume there | 19:02 |
| tobiash | so there are real technical reasons that those are different directories | 19:03 |
| clarkb | ya its a bind vs volume behavior | 19:03 |
| clarkb | docker volumes do what I describe | 19:03 |
| tobiash | if we want only one directory the only way I see (maybe I'm wrong) is to drop one possiblility of installing ansible (and that would be the dynamic one) | 19:04 |
| clarkb | as an alternative maybe we can pull that logic into the ansible management classes rather than having executor do something different than manage_ansible? | 19:07 |
| clarkb | and add docs (I couldnt find any at least) on where to find the venvs and why the behavior doffered | 19:07 |
| corvus | clarkb: mind if i wip your change for now? | 19:08 |
| tobiash | clarkb: yes, I think that would make sense | 19:08 |
| openstackgerrit | Merged zuul/zuul master: Don't always show expansion option on build console https://review.opendev.org/675163 | 19:08 |
| openstackgerrit | Merged zuul/zuul master: Adjust results headings in build console page https://review.opendev.org/675203 | 19:09 |
| tobiash | corvus: 675409 has 4 time +2, shall we merge it now? | 19:09 |
| clarkb | corvus: sure | 19:09 |
| corvus | tobiash: ++ | 19:10 |
| corvus | there's one more js console change that needs a +3 https://review.opendev.org/675236 | 19:11 |
| clarkb | corvus: ya that one was next on my list but had to step out for a few after reviewing the first 3 | 19:11 |
| tobiash | corvus: do you have an idea why 675039 didn't enter the gate after successful check? | 19:12 |
| clarkb | tobiash: the other thing that came to mind looking at the venv installations was being able to tell manage_ansible.py what path to install to | 19:13 |
| corvus | tobiash: oh i think it's because of the removal of clean check; i think we need to either add something back or add reverify | 19:13 |
| clarkb | I'll take a look at the code again with some of the original intent in mind and see if changes are appropriate | 19:13 |
| tobiash | clarkb: thanks (I thought the installation destination is already overridable but maybe I don't remember correctly) | 19:15 |
| clarkb | tobiash: I think we only load in that path override if doing the runtime install not the manage_ansible.py install | 19:15 |
| tobiash | ah ok | 19:15 |
| clarkb | corvus: left a comment on https://review.opendev.org/#/c/675236/2 | 19:21 |
| corvus | clarkb: good catch | 19:36 |
| clarkb | corvus: fwiw I htink that happens in other makeTaskPath usages too | 19:39 |
| clarkb | oh ya I think you mention that in your esponse on the change | 19:39 |
| corvus | yup | 19:39 |
| *** panda has quit IRC | 19:57 | |
| *** panda has joined #zuul | 19:59 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add permalinks to task detail popup https://review.opendev.org/675236 | 20:01 |
| corvus | mordred, clarkb: ^ | 20:02 |
| clarkb | corvus: is that doing character by character comparison or path entry by path entry comparison? | 20:06 |
| corvus | clarkb: path entry (they are arrays) | 20:07 |
| corvus | clarkb: javascript is not so good at determining array equality :( | 20:08 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: web: extract pure functions from the TaskOutput component https://review.opendev.org/675460 | 20:08 |
| clarkb | corvus: on line 224 is js like C where the condition applies to the next line only if no {}s are used? | 20:11 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Hide "root" variable in job web page https://review.opendev.org/675461 | 20:11 |
| corvus | tristanC: ^ i learned that working on the console stuff | 20:12 |
| corvus | clarkb: yeah. if folks want to adopt the always-use-braces style, i don't object | 20:13 |
| corvus | (that's a rare stylistic choice with real consequences so i'm not going to walk out of the room if we start talking about it :) | 20:13 |
| openstackgerrit | Merged zuul/zuul master: Add option to report build page https://review.opendev.org/675409 | 20:13 |
| clarkb | corvus: I'm fine with it as is. I just wanted ot make sure the behavior was what I expected | 20:14 |
| corvus | (i think, really, the behavior is always that the conditional applies to the next statement, but {} makes a compound statement) (statement is probably not the right word, but i have not paged in all of my compiler theory.) | 20:15 |
| corvus | tristanC: i'd like to merge https://review.opendev.org/630079 soon. do you agree with my proposed change? if so, do you want to implement it or should i? | 20:27 |
| corvus | pabelanger: have you had a chance to test base-minimal-test since https://github.com/ansible/project-config/pull/171 merged? | 20:30 |
| tristanC | corvus: i can look into reviving 630079 in a few hours | 20:31 |
| corvus | tristanC: thanks! | 20:31 |
| openstackgerrit | Merged zuul/zuul master: Use wait for empty update queue before accepting merges https://review.opendev.org/675039 | 20:33 |
| *** jeliu_ has quit IRC | 20:46 | |
| *** bhavikdbavishi has quit IRC | 20:47 | |
| *** jeliu_ has joined #zuul | 20:49 | |
| openstackgerrit | Merged zuul/zuul master: Refactor build page tabs https://review.opendev.org/675235 | 20:54 |
| clarkb | tobiash: looking at the need to configure the path for manage_ansible.py one option (though maybe not a great one) would be to add that then if you want to install to say /usr/lib pass that path to manage_ansible.py and set the config for it | 20:55 |
| clarkb | that does increase users' need to configure things though | 20:55 |
| *** tjgresha has joined #zuul | 20:58 | |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: WIP: Add zuul-operator-functional-openshift job https://review.opendev.org/674355 | 21:07 |
| *** jeliu_ has quit IRC | 21:12 | |
| *** igordc has joined #zuul | 21:13 | |
| *** rfolco has quit IRC | 21:18 | |
| *** jamesmcarthur has quit IRC | 21:19 | |
| corvus | tristanC: can you review https://review.opendev.org/673841 ? | 21:21 |
| openstackgerrit | Merged zuul/zuul master: Add permalinks to task detail popup https://review.opendev.org/675236 | 21:23 |
| mordred | corvus, clarkb: from scrollback - I wouldn't mind adopting always-braces for ifs - largely because it _does_ have impacts, and coming from python if I saw two lines indented but no braces around them I'm pretty sure my brain would take AGES to figure out the logic error | 21:25 |
| mordred | I don't feel absurdly strongly either way - but it's one of those style things that really can help with logic accidents in brace languages | 21:26 |
| corvus | ok. i had been doing always-braces, and only put in a few lines of no-brances; i'll undo those as i see them and stick with always in the future. | 21:27 |
| *** armstrongs has joined #zuul | 21:28 | |
| fungi | steeped in c long enough i'm pretty sure i mostly ignore indentation, but still, can't hurt | 21:31 |
| corvus | ooh, i wonder if we can syntax highlight yaml/json files like http://zuul.opendev.org/t/zuul/build/1bf501bb821f41bba727aa00b09eaa39/log/zuul-info/inventory.yaml | 21:31 |
| fungi | s/ignore/distrust. | 21:32 |
| corvus | also, now that i'm starting to use the logs tab more, i kinda want the log viewer to be in the logs tab, so i can switch to a different logfile without hitting back | 21:32 |
| corvus | (the first few weeks of using this stuff is going to be really fun :) | 21:33 |
| fungi | that seems like a worthwhile optimization | 21:33 |
| *** armstrongs has quit IRC | 21:38 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: web: refactor the errorsIds into the build action https://review.opendev.org/675350 | 21:40 |
| corvus | tristanC: rebased ^ | 21:40 |
| mordred | fungi: I have enough perl in my background that I've learned to distrust almost any syntactic sugar that lets me express more in fewer characters ... | 21:42 |
| mordred | although I do miss the obfuscated perl contests ... | 21:43 |
| fungi | fair | 21:44 |
| mordred | https://www.foo.be/docs/tpj/issues/vol4_3/tpj0403-0017.html <-- the wc script that compiles in both c and perl continues to be fun | 21:46 |
| openstackgerrit | Clark Boylan proposed zuul/zuul master: Improve functionality and docs around ansible installation https://review.opendev.org/675403 | 21:49 |
| Shrews | mordred: i wasn't aware there was a Perl Journal, even within a contest. On a related note, did you see Llinux Journal is closing its doors? | 21:49 |
| clarkb | corvus: tobiash ^ maybe something like that is what we want in the end? I think the biggest thing was all the docs ignored the automagic state dir install so I wans't aware that there were possibly two things happening | 21:50 |
| fungi | Shrews: end of an era | 21:52 |
| mordred | Shrews: I did not see that ... but yeah, end of an era | 22:01 |
| *** panda has quit IRC | 22:03 | |
| *** panda has joined #zuul | 22:03 | |
| *** mattw4 has quit IRC | 22:04 | |
| *** mattw4 has joined #zuul | 22:04 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: web: refactor the errorsIds into the build action https://review.opendev.org/675350 | 22:05 |
| *** mattw4 has quit IRC | 22:14 | |
| *** mattw4 has joined #zuul | 22:14 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Correctly identify failed tasks https://review.opendev.org/675488 | 22:16 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Refactor task result detection https://review.opendev.org/675489 | 22:16 |
| *** panda has quit IRC | 22:16 | |
| corvus | clarkb, tristanC, mordred, tobiash: i thought we were done but we need at least up to 675488 or we'll miss some failed tasks. 489 after it is a simple cleanup. | 22:17 |
| *** panda has joined #zuul | 22:20 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: web: add buildset page https://review.opendev.org/630079 | 22:20 |
| tristanC | corvus: 675488 sounds correct to me, but without test case it's hard to review... | 22:22 |
| tristanC | corvus: the last PS for the buildset page should look like what you described | 22:23 |
| corvus | tristanC: yeah, that's a good candidate for a test | 22:23 |
| corvus | we can copy in a whole bunch of different ansible json output test cases | 22:24 |
| tristanC | corvus: we could also test the whole console component, but the reactdom stuff is not very easy to work with. If we do that, we should look into enzyme to assert rendering | 22:25 |
| tristanC | and then, not worry with the internal procedure | 22:25 |
| corvus | that sounds like it might be worth a look.. maybe a poc patch? | 22:26 |
| corvus | (i like "assert rendering" but i'm worried about "not very easy to work with" :) | 22:26 |
| tristanC | create-react-app only setup ReactTestUtils and ReactDOM, which is limited to look for dom content, see the App.test.jsx | 22:28 |
| openstackgerrit | Merged zuul/zuul master: Provide buildset.uuid in /builds API result https://review.opendev.org/674759 | 22:30 |
| corvus | tristanC: buildset page looks great, with that and the change that just merged ^ we should be able to link back to the buildset page from the build page | 22:32 |
| tristanC | well i'd rather stay as far as possible from the dom, thus unit testing the pure js function sounds like a good first step too | 22:33 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: web: link the buildset page from the build https://review.opendev.org/675493 | 22:38 |
| *** mattw4 has quit IRC | 22:44 | |
| *** mattw4 has joined #zuul | 22:54 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: web: extract pure functions from the TaskOutput component https://review.opendev.org/675460 | 22:55 |
| *** mattw4 has quit IRC | 23:03 | |
| *** mattw4 has joined #zuul | 23:04 | |
| *** mattw4 has quit IRC | 23:52 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!