| *** JasonCL has joined #zuul | 00:03 | |
| *** JasonCL has quit IRC | 00:04 | |
| *** JasonCL has joined #zuul | 00:05 | |
| *** dtruong has quit IRC | 00:05 | |
| *** dtruong has joined #zuul | 00:06 | |
| *** JasonCL has quit IRC | 00:06 | |
| *** dtruong has quit IRC | 00:12 | |
| *** dtruong has joined #zuul | 00:12 | |
| mnaser | so ill dig into this now | 00:14 |
|---|---|---|
| mordred | mnaser: tools/install-js-tools.sh is a script thatwill install node and yarn - although if you have a recent node and not yarn you can just lift the yarn lines out | 00:14 |
| mnaser | mordred: im almost tempted just to angular-ify that little box too but i think that'll probably not be as easy as i think it is | 00:14 |
| mordred | mnaser: npm run start will spawn the status page via webpack-dev-server with it configured to use the live zuul status.json | 00:15 |
| mnaser | im not sure how jquery+angular might sit well together | 00:15 |
| mordred | the live openstack status.json that is | 00:15 |
| mordred | mnaser: yah - part of me is thinking just getting the status refactor done would get us in a better place - but I have a hunch that'll be slightly more work :) | 00:15 |
| * mnaser is in npm install waiting room | 00:17 | |
| mnaser | npm ERR! package.json Non-registry package missing package.json: graphitejs@https://github.com/prestontimmons/graphitejs/archive/master.tar.gz. | 00:19 |
| mnaser | come onnnn | 00:19 |
| *** dkranz has quit IRC | 00:19 | |
| mnaser | mordred: any workarounds for that? | 00:20 |
| *** dkranz has joined #zuul | 00:20 | |
| mnaser | i cant `npm install` because of that | 00:21 |
| mnaser | ok, it works somehow now magically | 00:22 |
| mnaser | ill stop the noise | 00:22 |
| mordred | mnaser: oh - yah - we're set up to use yarn instead of npm for install ... | 00:23 |
| mordred | mnaser: so it's possible npm install might or might not work ... but if it does, cool | 00:23 |
| *** JasonCL has joined #zuul | 00:29 | |
| *** xinliang has quit IRC | 00:37 | |
| *** JasonCL has quit IRC | 00:46 | |
| *** xinliang has joined #zuul | 00:49 | |
| *** xinliang has quit IRC | 00:49 | |
| *** xinliang has joined #zuul | 00:49 | |
| *** JasonCL has joined #zuul | 00:50 | |
| *** JasonCL has quit IRC | 00:59 | |
| *** JasonCL has joined #zuul | 01:00 | |
| *** JasonCL has quit IRC | 01:00 | |
| *** JasonCL has joined #zuul | 01:03 | |
| *** JasonCL has quit IRC | 01:09 | |
| *** JasonCL has joined #zuul | 01:17 | |
| *** JasonCL has quit IRC | 01:19 | |
| *** JasonCL has joined #zuul | 01:22 | |
| *** JasonCL has quit IRC | 01:26 | |
| *** JasonCL has joined #zuul | 01:31 | |
| *** rlandy has quit IRC | 01:33 | |
| *** JasonCL has quit IRC | 01:37 | |
| *** JasonCL has joined #zuul | 01:43 | |
| *** JasonCL has quit IRC | 01:44 | |
| *** Diabelko has joined #zuul | 02:13 | |
| Diabelko | hello | 02:13 |
| Diabelko | is there any way to kill whole buildset and/or job without zuul trying to run it again? | 02:14 |
| clarkb | Diabelko: if you kill it during the run or post-run phases it won't be retried | 02:16 |
| clarkb | a new patchset (with gerrit at least) will stop running all the job and start new ones for the new patchset | 02:16 |
| clarkb | I don't think there is a more direct way than that | 02:17 |
| Diabelko | mmmm, my job hanged (destination VM got rebooted) and it was still trying to keep ansible/ssh connection | 02:18 |
| Diabelko | it was already in the run phase | 02:18 |
| Diabelko | when I killed it, it started again | 02:18 |
| Diabelko | is there any specific process I need to kill? maybe I killed it 'too far' at the process tree? | 02:18 |
| clarkb | Diabelko: I typically kill the ansible-playbook process | 02:19 |
| clarkb | and that seems to do what I expect | 02:19 |
| Diabelko | ok, let me try that | 02:19 |
| *** JasonCL has joined #zuul | 02:20 | |
| *** harlowja has quit IRC | 02:21 | |
| Diabelko | ok, I killed ansible-playbook process and it went straight to post | 02:22 |
| Diabelko | thanks | 02:22 |
| *** JasonCL has quit IRC | 02:25 | |
| *** swest has quit IRC | 02:29 | |
| *** swest has joined #zuul | 02:43 | |
| *** JasonCL has joined #zuul | 03:50 | |
| *** JasonCL has quit IRC | 03:59 | |
| *** jaianshu has joined #zuul | 04:21 | |
| *** swest has quit IRC | 05:23 | |
| *** harlowja has joined #zuul | 05:25 | |
| *** dkranz has quit IRC | 05:56 | |
| *** dkranz has joined #zuul | 05:56 | |
| *** swest has joined #zuul | 06:00 | |
| *** harlowja has quit IRC | 06:00 | |
| *** swest has quit IRC | 06:05 | |
| *** swest has joined #zuul | 06:06 | |
| *** harlowja has joined #zuul | 06:19 | |
| *** sshnaidm|afk has quit IRC | 06:47 | |
| *** jaianshu has quit IRC | 07:38 | |
| *** jaianshu has joined #zuul | 07:39 | |
| *** harlowja has quit IRC | 08:08 | |
| *** tobasco|away is now known as tobasco | 08:25 | |
| *** electrofelix has joined #zuul | 08:28 | |
| *** threestrands_ has quit IRC | 08:48 | |
| *** JasonCL has joined #zuul | 09:20 | |
| *** JasonCL has quit IRC | 09:25 | |
| *** hashar has joined #zuul | 09:49 | |
| *** mhu has joined #zuul | 09:52 | |
| *** jaianshu_ has joined #zuul | 09:58 | |
| *** jaianshu has quit IRC | 10:02 | |
| *** JasonCL has joined #zuul | 10:41 | |
| *** JasonCL has quit IRC | 10:45 | |
| *** JasonCL has joined #zuul | 11:11 | |
| *** JasonCL has quit IRC | 11:20 | |
| *** JasonCL has joined #zuul | 11:42 | |
| *** jaianshu_ has quit IRC | 11:44 | |
| *** JasonCL has quit IRC | 11:52 | |
| *** JasonCL has joined #zuul | 12:38 | |
| *** JasonCL has quit IRC | 12:41 | |
| *** JasonCL has joined #zuul | 12:41 | |
| *** JasonCL has quit IRC | 12:53 | |
| *** JasonCL has joined #zuul | 13:10 | |
| *** rlandy has joined #zuul | 13:31 | |
| *** JasonCL has quit IRC | 13:59 | |
| *** JasonCL has joined #zuul | 14:23 | |
| *** JasonCL has quit IRC | 14:27 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 14:45 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 14:49 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 14:54 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 15:02 |
| *** JasonCL has joined #zuul | 15:05 | |
| *** JasonCL has quit IRC | 15:07 | |
| *** JasonCL has joined #zuul | 15:10 | |
| *** JasonCL has quit IRC | 15:11 | |
| tobiash | mordred: ^ | 15:11 |
| *** JasonCL has joined #zuul | 15:12 | |
| clarkb | that reminds me reviewing the js topic is on my todo list | 15:19 |
| clarkb | tobiash: looks like tests are not passing | 15:21 |
| tobiash | clarkb: yepp noticed that | 15:22 |
| *** JasonCL has quit IRC | 15:22 | |
| openstackgerrit | Merged openstack-infra/nodepool master: Revert "Add /label-list to the webapp" https://review.openstack.org/548424 | 15:26 |
| openstackgerrit | Merged openstack-infra/nodepool master: Consolidate status results https://review.openstack.org/548134 | 15:27 |
| openstackgerrit | Merged openstack-infra/nodepool master: Consolidate node_list, add generic filter https://review.openstack.org/548135 | 15:27 |
| openstackgerrit | Merged openstack-infra/nodepool master: webapp: use content detection for return https://review.openstack.org/548440 | 15:30 |
| *** JasonCL has joined #zuul | 15:30 | |
| openstackgerrit | Merged openstack-infra/nodepool master: Add label-list webapp endpoint https://review.openstack.org/548442 | 15:32 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 15:57 |
| corvus | hi -- i'm semi-back today and am going to try to start catching up on things so i'll be really back next week | 16:29 |
| corvus | i'm declaring irc bankruptcy though, so if you pinged me with something important in the last week or so, please do so again :) | 16:30 |
| clarkb | corvus: probably the most important bits are that the zuul javascript things landed then had to be worked around then monty got everything deployed and happy yesterday. There is also a thread on zuul list related to that | 16:35 |
| clarkb | otherwise I think most of us have just been crawling back to normality as a snails pace (at least that is how I have felt about my return home) | 16:35 |
| corvus | clarkb: at 15:19 you say reviewing js stuff is on your list -- is there more stuff? | 16:35 |
| clarkb | corvus: ya topic:javascript though its largely refactoring and improving now rather than "make it work again" | 16:35 |
| corvus | cool, sonds great :) | 16:36 |
| corvus | sounds | 16:36 |
| mordred | indeed | 16:40 |
| mordred | corvus, clarkb: there is a bug in the currently deployed status page that mnaser and I were looking at yesterday - still no fix - but it's not a fatal bug or anything | 16:40 |
| clarkb | mordred: the (x) on the search field thing? | 16:41 |
| mordred | yup | 16:41 |
| mnaser | yes, i got sucked into some other stuff in $work to put out fires | 16:41 |
| mordred | mnaser: no worries - we'll sort it out eventually | 16:41 |
| mnaser | mordred: i'll put aside an hour just to see whats going on because im genuinely curious heh | 16:41 |
| mordred | mnaser: I'm taking some opportunities to dig in slightly further in to some places to make sure I understand exactly what's going on | 16:41 |
| mnaser | mordred: my theory is that for some reason, the span z-index is not sitting on top of the textbox | 16:42 |
| mnaser | mordred: because if you notice, the pointer doesnt change when you hover, even though it has cursor: pointer; in css afaik | 16:42 |
| mordred | mnaser: yah. I didn't have those words fully, but that was somewhat what my brain was thinking | 16:42 |
| mordred | mnaser: yes - I did notice that too | 16:42 |
| mnaser | i tried hacking z-index:999999; in css but that didn't do it either so bleh | 16:42 |
| corvus | i'm super-excited we're now in a position to accept js patches from mnaser :) | 16:43 |
| mnaser | corvus: yeah, the work that mordred did is *really* painful so im glad he was able to get through it all | 16:43 |
| mnaser | so many times i've had things to convert to webpack or some sort of js dependency management that i'd just give up because it wasn't something you could 'ease' into | 16:44 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/job.html page to display job details https://review.openstack.org/535545 | 16:48 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/projects.html web page https://review.openstack.org/537870 | 16:48 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Fix indentation and renable the eslint rule https://review.openstack.org/545671 | 16:48 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Shift html templates into components https://review.openstack.org/551327 | 16:48 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 16:52 |
| mordred | corvus: ^^ the last patch there is a refactor of a few things - even if you don't do an immediate deep review - would you mind glancing it over to make sure it doesn't make the js structure less understandable? (there are a couple of improvements and I think it actually clarifies at least one thing, but I've learned enough about how everything hangs together I want to be careful to not move the goalposts too | 16:54 |
| mordred | much) | 16:54 |
| openstackgerrit | Matthieu Huin proposed openstack-infra/zuul-jobs master: Publish build status to MQTT https://review.openstack.org/551329 | 16:55 |
| mordred | the primary issue it solves is that we see a flash of unexpanded template when loading the dashboard pages. the secondary issue is that there was confusion the other day about variable names due to there being both angular and webpack templating going on in the same file | 16:55 |
| mordred | corvus: oh - also - welcome semi-back! | 16:59 |
| *** harlowja has joined #zuul | 17:00 | |
| corvus | mordred: i don't understand the plumbing completely and may not be capable of it today, but my general impression of the change is favorable and that it makes things more intuitive. | 17:00 |
| mordred | corvus: ok. good- that's the main feedback I was looking for - the details of the plumbing can be a topic for later | 17:01 |
| mordred | mnaser: I had a hunch that the x issue had something to do with having wrapped the status page in an angular controller ... but I just went back and checked in the old version of the stack where we had things broken out, and it turns out the x was broken by the very first webpack patch | 17:04 |
| mordred | mnaser: which is still huge - but you can see the original version here: https://review.openstack.org/#/c/538099/12 before we squashed a bunch of the changes | 17:04 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 17:12 |
| mordred | mnaser: BOOM! Found it | 17:17 |
| mnaser | mordred: nice! i'm curious | 17:17 |
| mordred | mnaser: bootstrap version | 17:17 |
| mnaser | oh that's very likely | 17:17 |
| mnaser | i guess we bumped versions? | 17:17 |
| mordred | mnaser: the old versoin was using bootstrap 3.1.1 - but in the new code I wasn't strict enough and it pulled in 3.3 | 17:18 |
| mnaser | i guess we can pin back to 3.1.1 till we figure everything out | 17:18 |
| mordred | yup | 17:18 |
| mnaser | or maybe hop into 4.0.0 too | 17:18 |
| mnaser | oh 4.0.0 looks pretty | 17:20 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 17:21 |
| tobiash | clarkb: responded to your comments ^ | 17:21 |
| *** electrofelix has quit IRC | 17:22 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/jobs/{job_name} route https://review.openstack.org/550978 | 17:30 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/projects routes https://review.openstack.org/550979 | 17:30 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/pipelines route https://review.openstack.org/541521 | 17:30 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/job.html page to display job details https://review.openstack.org/535545 | 17:30 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: add /{tenant}/projects.html web page https://review.openstack.org/537870 | 17:30 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Fix indentation and renable the eslint rule https://review.openstack.org/545671 | 17:30 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Shift html templates into components https://review.openstack.org/551327 | 17:30 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Use bootstrap 3.1.1 instead of 3.3.7 https://review.openstack.org/551338 | 17:30 |
| mordred | mnaser, tobiash, corvus, clarkb: 551338 should fix the x issue - and I rebased the rest of the stack on it so it's safe to go ahead and land | 17:36 |
| tobiash | clarkb, mordred, corvus: 551301 is green now :) | 17:45 |
| corvus | tobiash: thanks for picking that up! | 17:46 |
| mordred | tobiash: woot | 17:46 |
| tobiash | my first idea was a multinode test, but that isn't even needed as long as we connect to localhost via the interface ip :) | 17:48 |
| corvus | tobiash: yeah, that's the thought process i went through too :) | 17:48 |
| *** harlowja has quit IRC | 17:54 | |
| mordred | y'all are smarter than I am | 17:55 |
| *** harlowja has joined #zuul | 17:56 | |
| tobiash | well, I wouldn't have managed to do this crazy js refactoring... | 17:57 |
| tobiash | mordred: I think there are a few minor glitches on 537870 | 17:58 |
| tobiash | mordred: and 535545 too | 17:59 |
| *** harlowja has quit IRC | 18:00 | |
| clarkb | does paramiko really require you to open the socket yourself? | 18:01 |
| mordred | tobiash: yah - those are a bit better in the components patch - but maybe I should reorganize that stack so that the component patch comes first ... | 18:01 |
| clarkb | clearly its been a while since I had to paramiko | 18:01 |
| openstackgerrit | Merged openstack-infra/zuul master: Use bootstrap 3.1.1 instead of 3.3.7 https://review.openstack.org/551338 | 18:02 |
| tobiash | mordred: ah, that fixes one of the glitches: https://review.openstack.org/#/c/551327/2/web/templates/project.html | 18:03 |
| *** hashar is now known as hasharDinner | 18:03 | |
| corvus | clarkb: i +2d 551301 and left +W for you | 18:04 |
| tobiash | mordred: found one glitch in 551327 | 18:04 |
| mordred | tobiash: the other div issues are related to the opening div being at the end of the body line | 18:05 |
| mordred | welll - not the 551327 one | 18:05 |
| tobiash | ah, ok | 18:05 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Shift html templates into components https://review.openstack.org/551327 | 18:08 |
| clarkb | tobiash: rereviewed, just a couple things mostly to respond to your responses | 18:09 |
| tobiash | clarkb: yah, /etc/resolv.conf sounds better and generic enough | 18:09 |
| tobiash | clarkb: regarding the keep jobdir, what do you suggest? | 18:10 |
| tobiash | an env var ZUUL_REMOTE_KEEP and set this in ci? | 18:11 |
| clarkb | ya that might be a good way to do it | 18:11 |
| *** dmellado has quit IRC | 18:16 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 18:17 |
| tobiash | clarkb: that should address your comments ^ | 18:17 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add new tox-remote job https://review.openstack.org/551301 | 18:19 |
| *** dmsimard is now known as dmsimard|afk | 18:25 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Refactor load sensors into drivers https://review.openstack.org/549275 | 18:30 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: WIP: Add cgroup support to ram sensor https://review.openstack.org/549506 | 18:30 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul master: Use re2 for change_matcher https://review.openstack.org/536389 | 18:33 |
| SpamapS | corvus: ^ uses released py35 compatible fb-re2 now | 18:33 |
| SpamapS | guessing we aren't going to be able to do that for 3.0 though | 18:34 |
| SpamapS | Which presents a problem for adding it as-is in 3.0.x or 3..1 | 18:34 |
| SpamapS | I wonder if we could make it a mode flag for 3.x and then flip the default in a 4.x release somewhere down the road. | 18:36 |
| *** dmellado has joined #zuul | 18:42 | |
| clarkb | looks like zuul.o.o status page is broken by bootstrap | 18:47 |
| clarkb | mordred: mnaser ^ | 18:47 |
| mnaser | o | 18:47 |
| clarkb | Typeerror: e[r] is undefined | 18:48 |
| clarkb | in bootstrap:54 | 18:48 |
| mnaser | mordred: is it possible we switched bootstrap css but not the js? | 18:48 |
| mnaser | wait | 18:48 |
| mnaser | GET http://zuul.openstack.org/vendor.472d16695ac5d771265e.js 404 (Not Found) | 18:48 |
| mnaser | clarkb: do you know if whatever generation process didn't probably run cleanly by any chance? | 18:48 |
| clarkb | I don't | 18:49 |
| clarkb | but I confirm the same 404 | 18:49 |
| mordred | clarkbm, mnaser uhoh- lemme look | 18:53 |
| mrhillsman | corvus clarkb mordred i can try out the guide | 18:53 |
| mrhillsman | is it in-tree? | 18:54 |
| mordred | clarkb: OH - crap - I grok the bug | 18:54 |
| mordred | clarkb: for now, restarting zuul-web will fix it | 18:54 |
| clarkb | mordred: are we caching the files and not serving them when they change ? | 18:55 |
| mordred | clarkb: we're doing a directory listing and registering static routes - at startup time | 18:55 |
| mordred | clarkb: so when the hashed files update, there are no routes for the new files | 18:56 |
| tobiash | are the hashed files hashed by content or something random?? | 18:58 |
| clarkb | probably worth updating to just serve all the files in that dir? | 18:58 |
| mordred | clarkb: yah - I'll get a fixed one | 18:59 |
| tobiash | or do they only change on web changes? | 18:59 |
| mordred | tobiash: they're by content - to facilitate cache-busting | 18:59 |
| tobiash | so currently every change to zuul-web will break your installation | 18:59 |
| mordred | tobiash: yes. until we fix this at least | 19:00 |
| mordred | tobiash: we could also turn off the hashed filenames for now until we get a better fix | 19:00 |
| *** harlowja has joined #zuul | 19:00 | |
| mnaser | that might make deployments complicated depending on how caching works | 19:00 |
| mnaser | but yeah. 'until we get a better fix' | 19:00 |
| mordred | yah. it's really most applicable for folks serving the files directly from apache or whatnot | 19:01 |
| tobiash | mordred: inotify on the static path? | 19:01 |
| tobiash | and then a reload? | 19:01 |
| tobiash | or a signaled reload and sighup via puppet | 19:01 |
| mordred | clarkb: ok - I restarted zuul-web - status should be working again - I'll work now on a real fix so that it does't happen again | 19:07 |
| mrhillsman | where's that getting started guide? https://docs.openstack.org/infra/zuul/admin/zuul-from-scratch.html ? | 19:07 |
| clarkb | mordred: thanks | 19:07 |
| clarkb | mrhillsman: https://docs.openstack.org/infra/zuul/admin/quick-start.html is the quick start then zuul from scratch is more in depth aiui | 19:08 |
| mrhillsman | ok cool | 19:08 |
| clarkb | SpamapS: it may be worth considering making that change now simply to avoid any transitions later | 19:10 |
| clarkb | SpamapS: a few of us early adopters might have to cope but that should be minimal | 19:10 |
| tobiash | I think the zuul-from-scratch is intended to be tried for starters | 19:10 |
| tobiash | and more complete | 19:10 |
| tobiash | the quick start misses many things | 19:10 |
| clarkb | we do continue to use negative lookahead in places to say things like don't run this job on the driverbranches branch and not on ocata and so on. I think those should all be solveable by doing the branch based configs properly | 19:13 |
| mrhillsman | tobiash quite possibly, i think of the quickstart as someone has heard about and have a general idea of the concept(s) zuul facilitates (attended a talk, read code/docs) and are ready to just install | 19:16 |
| mrhillsman | i'm going to help as i can with trying both and giving some feedback since i consider myself still quite new to zuul | 19:18 |
| tobiash | mrhillsman: yah, the quick start is more of a general overview and the from scratch is the walk through how to set it up for the first time | 19:18 |
| SpamapS | clarkb: I did some spot checking and there are a few regexes in infra that would need to be re-done as negative matches | 19:18 |
| SpamapS | clarkb: but I agree, if we *can* do it before release, I'd like to | 19:18 |
| SpamapS | Just figured the train was a little too far down the tracks to pull that switch. | 19:19 |
| clarkb | maybe? I know corvus has been using this pre release time to make those jumps | 19:20 |
| clarkb | because less painful than waiting down the road | 19:20 |
| corvus | mrhillsman: yeah, the zuul-from-scratch is the thing i mentioned earlier | 19:20 |
| SpamapS | clarkb: yeah that was the intention, but the fb-re2 guys took forever to cut 1.0.6 :-/ | 19:21 |
| SpamapS | I'm actually pretty excited about there being a good re2 python3 module now just in general | 19:21 |
| SpamapS | I've always disliked the built in re module. | 19:21 |
| mrhillsman | corvus: ah ok, i will go through that then | 19:22 |
| corvus | SpamapS: we're still waiting on one more release blocker thing, we may be able to squeeze it in | 19:22 |
| SpamapS | corvus: oh sweet. Is that release blocker the js stuff? | 19:22 |
| mordred | SpamapS: log streaming | 19:22 |
| * mordred glares at self | 19:23 | |
| corvus | SpamapS: it's (mostly) landed, though i think we're learning we need to land a little more cleanup.. it's ^ that | 19:23 |
| corvus | SpamapS: i think what we may need to do for the re2 thing is to land the negative matcher first, ask early adopters to port, then land the re2 change itself | 19:24 |
| corvus | basically, same thing we'd do after release, but with a much abbreviated timetable | 19:24 |
| SpamapS | corvus: yeah that makes sense. | 19:24 |
| SpamapS | We coud actually land it where the negative matchers only use re2 also. | 19:25 |
| corvus | SpamapS: wfm | 19:25 |
| SpamapS | That way people can phase in and know it will work.. | 19:25 |
| SpamapS | Ok I'll put that on my urgent todo list for next week.. | 19:25 |
| corvus | SpamapS: cool, thanks. i should be able to pitch in by then too. | 19:25 |
| SpamapS | woot | 19:26 |
| SpamapS | http://automationbroker.io/ <-- attending a talk on this.... .might be a group to collaborate on as a parallel effort similar to zuul but container focused. | 19:27 |
| SpamapS | s/on/with/ | 19:27 |
| corvus | should we normalize the words 'irrelevant' and 'ignore'? ie, use 'ignore-files' instead of irrelevant files, or 'irrelevant-branches' instead of 'ignore-branches' ? | 19:27 |
| SpamapS | yeah one is imperative and one is declarative | 19:28 |
| SpamapS | I'd say we're aiming at declaring, not commanding.. | 19:28 |
| corvus | SpamapS: yeah, that's what i was thinking when i went with the (admittedly awkward) irrelevant-files | 19:29 |
| SpamapS | so I'd lean toward normalizing on irrelevant. But at the same time, people are used to ignore. | 19:29 |
| corvus | SpamapS: 'irrelevant-files' is the existing thing... you mean they're used to ignore from other things? | 19:29 |
| SpamapS | corvus: correct, I think people are used to the word ignore from other things in the automation space.. But I personally think irrelevant-files and irrelevant-branches is more clear as to what will be done with that field.. | 19:30 |
| corvus | SpamapS: okay, let's be the vanguard here and stick with irrelevant :) | 19:31 |
| SpamapS | :-D | 19:32 |
| AJaeger | corvus: ignore is easier to type - but that might be an argument for irrelevant ;) | 19:32 |
| corvus | i guess "ignored-foo" is more declarative and easier to type, if we wanted to open up a third option. | 19:33 |
| corvus | that seems like a cop out though :) | 19:33 |
| corvus | AJaeger: i'll score your argument as a point for both. :) | 19:34 |
| Shrews | i like "unregarded" :-P | 19:34 |
| AJaeger | corvus: exactly | 19:34 |
| corvus | Shrews: neglected | 19:34 |
| SpamapS | f-these-branches: ? | 19:35 |
| SpamapS | f-these-files: | 19:35 |
| Shrews | SpamapS: oh, that reminds me of this thing I saw today: https://alcohollick.com/tfw.html | 19:36 |
| corvus | f stands for fun | 19:36 |
| * AJaeger ignores this relevant discussion now and wishes everybody a fun weekend ;) | 19:47 | |
| *** sshnaidm|afk has joined #zuul | 19:51 | |
| *** sshnaidm|afk has quit IRC | 19:52 | |
| *** sshnaidm|afk has joined #zuul | 19:53 | |
| clarkb | mordred: the early route stuff in your js stack is not controversial right? I should just approve things if they look good? | 19:54 |
| SpamapS | Shrews: :-D | 19:57 |
| clarkb | "Maybe stay inside" <- dublin weather service | 19:57 |
| mordred | clarkb: yah - to my knowledge those are non-controversial | 20:03 |
| clarkb | mordred: I left a minor nit on the first one, probably not worth respinning but wanted you to see it | 20:03 |
| clarkb | mordred: in the routes what does the .* do in {project_name:.*} ? I notice it wasn't part of the get job but is part of get project? | 20:05 |
| clarkb | appears to state that any character is allowed in a project name | 20:06 |
| clarkb | butI guess we don't allow any character in job names? | 20:06 |
| mordred | well -but we allow /'s in project names | 20:07 |
| mordred | same thing in the get_key thing | 20:07 |
| *** adam_g has quit IRC | 20:07 | |
| clarkb | oh right openstack/nova | 20:07 |
| clarkb | so if we don't explicitly allow it there it won't do the right routing | 20:07 |
| mordred | you wan to GET /openstack/git.openstack.org/openstack-infra/zuul.pub ... so /{tenant}/{project:.*}.pub' puts openstack in tenant and git.openstack.org/openstack-infra/zuul in project | 20:08 |
| *** adam_g has joined #zuul | 20:08 | |
| mordred | exactly | 20:08 |
| clarkb | mordred: https://review.openstack.org/#/c/550979/3 that has some thoughts about error hanlding and reporting which may affect the rest of the stack if you can take a look really quick | 20:15 |
| *** threestrands has joined #zuul | 20:22 | |
| mordred | clarkb: I'm glad you just reviewed that stuff - it'll give you good context for the next patch I'm about to push up | 20:23 |
| mordred | clarkb: hrm. interesting point re: exceptions, gearman and 404 vs. 500 | 20:24 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Serve the static files more dynamically https://review.openstack.org/551366 | 20:24 |
| mordred | clarkb: ^^ there's a thing to consider for the static file issue of earlier - the unittests in zuul should be testing that it works - but also I tested synthetically with http://paste.openstack.org/show/696914/ | 20:26 |
| clarkb | cool I'll get back to the reviews in a bit but digging up lunch now | 20:26 |
| mordred | clarkb: lunch is a good plan | 20:27 |
| clarkb | curious to see what you have to say about 500s vs 404s there | 20:27 |
| mordred | clarkb: yah. me too - I'm mulling currently :) | 20:27 |
| corvus | mordred, clarkb: iirc, there was a change recently to return some specific 'project not found' in json... maybe do the same for tenant? | 21:08 |
| clarkb | mordred: corvus https://review.openstack.org/#/c/541521/14 comments again there. Mostly I'm just noticing things as I go through the stack that probably apply to the whole stack but also aren't fatal flaws | 21:12 |
| clarkb | I'm not approving things in case that feedback is useful enough to deserve respinning some things | 21:14 |
| mordred | clarkb: ++ on 541521 json/python/json comment | 21:22 |
| mordred | corvus: yah - that's probably a good call | 21:22 |
| *** threestrands has quit IRC | 21:22 | |
| clarkb | I'm now somewhat lost in a sea of js and html | 21:26 |
| clarkb | so less likely to be insightful at this point :) | 21:26 |
| clarkb | for the templating thing to make sure I understand right, instead of having a template already loaded in the html which cause the unexpanded flash we load the template with the js so we don't show anything until we are able to load and expand the template? | 21:27 |
| mordred | clarkb: yah- the template basically gets stored as string constant in the js as part of the webpack compile | 21:28 |
| mordred | clarkb: it'll take that require('') and expand it to just be the string containing the html template in the output | 21:29 |
| clarkb | ok I don't know that I'll vote on the js changes but I think I grok the direction they are going in and seems reasonable | 21:30 |
| clarkb | https://review.openstack.org/#/c/551366/ failed unittests too btw | 21:30 |
| mordred | clarkb: cool. that's the important thing - being able to read them and understand what's going on so that if you did have to start debugging you wouldn't be lost or anything | 21:31 |
| mordred | clarkb: and darn | 21:31 |
| mordred | ZOMG. if I could learn to type words | 21:32 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Serve the static files more dynamically https://review.openstack.org/551366 | 21:32 |
| mordred | the variable is static_path ... not static_dir | 21:33 |
| mordred | I think python should know I use the words path and dir interchangably :) | 21:33 |
| clarkb | looks like you properly sanitize the paths which is my only concern with /{path:.*} :) | 21:35 |
| mordred | clarkb: :) | 21:36 |
| clarkb | if we want to be extra careful we can pass it through os.path.realpath() too but I think that isn't necessary unless you put bad symlinks in the web content dir | 21:38 |
| mnaser | "Serve the static files more dynamically" | 21:42 |
| mnaser | thats a fun commit message | 21:42 |
| mnaser | :) | 21:42 |
| mordred | mnaser: I'm glad you like it :) | 21:42 |
| mnaser | and "This is really just an excuse to write that commit subject." right below, aha | 21:42 |
| clarkb | and webpack is a js lib that I can go read about somewhere? | 21:42 |
| clarkb | I've inferred what it is doing based on diffs and brwoser debugging but at some point I should go read on it proper | 21:43 |
| mnaser | clarkb: webpack is kinda like dependency management on steriods | 21:43 |
| corvus | and build-time pre-processing | 21:43 |
| mnaser | think dependency management + 'compiler' for the web | 21:43 |
| *** dtruong has quit IRC | 21:43 | |
| clarkb | it also appears to manage how things are loaded into the browser as runtime? | 21:44 |
| clarkb | and sets up the debug maps and all that | 21:44 |
| corvus | yeah, that's the 'pack' part of it... | 21:44 |
| mnaser | yes, it can do things like load stuff on demand too | 21:44 |
| mnaser | which is probably not a big value in zuul, but bigger apps | 21:44 |
| corvus | mordred: do we have a job that emits that block diagram? | 21:44 |
| corvus | the one with all the boxes for each module? that really helped me grok what webpack was doing | 21:44 |
| corvus | (also, we really should generate that at least on all js changes, it's awesome) | 21:45 |
| corvus | there's an npm directory here, but it's empty http://logs.openstack.org/66/551366/1/check/zuul-tox-py35/bd23498/npm/ | 21:47 |
| corvus | i wonder if it's supposed to be there | 21:48 |
| corvus | oh yeah we do, here | 21:50 |
| corvus | clarkb: http://logs.openstack.org/38/551338/1/gate/nodejs-npm-run-lint/40f06b6/npm/reports/bundle.html | 21:50 |
| corvus | so those are the bundles that are made | 21:51 |
| clarkb | we have a jquery.zuul.js that is distinct from jquery.js? | 21:52 |
| clarkb | that is an interesting graph though | 21:52 |
| corvus | clarkb: yeah, i think that's the zuul jquery plugin that implements the status page | 21:52 |
| corvus | clarkb: also https://docs.openstack.org/infra/zuul/developer/javascript.html may be a useful reference (and we should update it if it's missing things) | 21:53 |
| corvus | clarkb: it has a section on webpack! https://docs.openstack.org/infra/zuul/developer/javascript.html#webpack-asset-management | 21:54 |
| clarkb | cool docs. I do wonder what npm run build doesn't do that npm run build:docker does do (from a functional perspective) | 21:56 |
| clarkb | I guess that is so you don't have to have particular versions of things and it can install what it wants for you in the container? | 21:57 |
| clarkb | I tend to use the virtualenv thing for nodejs because epython | 21:57 |
| clarkb | (but does similar isolation) | 21:57 |
| corvus | clarkb: yeah, i believe the difference is without docker you need to run install-js-tools.sh which adds extra apt sources, etc | 21:59 |
| corvus | not necessarily something all devs will want to do on their workstation | 22:00 |
| mordred | clarkb, corvus: yes, that is the difference | 22:03 |
| clarkb | I really wish that bwrap or a similar tool was more common for use cases like that | 22:04 |
| clarkb | (because the idea of handing root to everything is :( but bwrap can operate without root and just use the host as a base to isolate) | 22:05 |
| mordred | https://webpack.js.org is the main webpack site - and I've found their docs complete and easy to deal with (the getting-started guide was initial very helpful to me) | 22:05 |
| clarkb | I guess the plan with gnome is to make bwrap more ubiquitous so maybe that will happen eventually | 22:05 |
| mordred | I hope so | 22:05 |
| *** hasharDinner has quit IRC | 22:06 | |
| mordred | the other important piece of the chain (that's mostly transparent behind webpack) is Babel - which is what is transpiling ES6 down to javascript browsers can be expected to understanded | 22:07 |
| mordred | mostly important to know about just to know that's happening, not because you have to interface with it | 22:07 |
| clarkb | are we actually using any es6? | 22:10 |
| clarkb | I guess libs like angular may not be js proper and expect a transpile? | 22:10 |
| mordred | oh - we're using a bunch of es6 | 22:11 |
| mordred | import, for (foo of bar), let are all es6 just off the top of my head | 22:12 |
| mordred | oh - and there's a template string form I think is in use somewhere | 22:12 |
| *** zhuli has quit IRC | 22:26 | |
| *** zhuli has joined #zuul | 22:26 | |
| *** gregdek has quit IRC | 22:28 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Use arrow functions for http callbacks https://review.openstack.org/551399 | 22:29 |
| clarkb | heh you took that as a challenge to add more es6 :)? | 22:29 |
| mordred | clarkb: ^^ and in fact, that fixes an ugly in the components patch - and is es6 | 22:29 |
| *** gregdek has joined #zuul | 22:30 | |
| mordred | clarkb: almost - I was double-checking the arrow function sntax because I thought we already had one but couldn't find it. we don't - but then I noticed that the description of the feature includes a description of a problem I had to work around earlier today | 22:30 |
| *** JasonCL has quit IRC | 23:08 | |
| *** rlandy has quit IRC | 23:10 | |
| *** JasonCL has joined #zuul | 23:14 | |
| *** JasonCL has quit IRC | 23:23 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!