| -@gerrit:opendev.org- Ian Wienand proposed: [zuul/zuul-jobs] 840566: add-gpgkey: trust incoming key https://review.opendev.org/c/zuul/zuul-jobs/+/840566 | 03:42 | |
| -@gerrit:opendev.org- Ian Wienand proposed: [zuul/zuul-jobs] 840566: add-gpgkey: trust incoming key https://review.opendev.org/c/zuul/zuul-jobs/+/840566 | 04:05 | |
| -@gerrit:opendev.org- Ian Wienand proposed: [zuul/zuul-jobs] 840566: add-gpgkey: trust incoming key https://review.opendev.org/c/zuul/zuul-jobs/+/840566 | 04:40 | |
| @d-j-j:matrix.org | Hi, in my review for the fileset spec https://review.opendev.org/c/zuul/zuul/+/839550 the topic of switching to https://pypi.org/project/re2/ was brought up by corvus. I'm wondering since it's a drop-in replacement if you considered before to switch the whole repo to re2 in one change? Would it be sufficient to rely on the successful regression or are there some additional concerns / need for performance measurements, etc? I think at least the performance comparison on the re2 page look very promising and since I'm also very interested in any performance-related improvements I'd be happy to work on this topic :) | 07:05 |
|---|---|---|
| -@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: [zuul/zuul] 840543: Cleanup actiongeneral usage https://review.opendev.org/c/zuul/zuul/+/840543 | 10:21 | |
| @mhuin:matrix.org | Hello zuul-maint, could we land https://review.opendev.org/c/zuul/zuul-client/+/837908 to make it easier to use zuul-client with existing zuul.conf files? | 12:52 |
| @jim:acmegating.com | d-j-j: we don't want to switch for performance reasons, but rather safety from user input. it's not a drop in replacement; it has significant functional shortcomings compared to the standard re library, including features people currently use with file matchers (especially around negative lookaheads). that's why it needs careful thought as to how it interacts with the proposal. | 14:09 |
| @jim:acmegating.com | mhu: qq on that | 14:19 |
| -@gerrit:opendev.org- daniel.pawlik https://matrix.to/#/@dpawlik:matrix.org proposed: [zuul/zuul] 840679: Add passlib to Ansible venvs https://review.opendev.org/c/zuul/zuul/+/840679 | 14:32 | |
| @jpew:matrix.org | Do child job`dependencies` override (not combine) with parent `dependencies`? e.g | 14:35 |
| ``` | ||
| - job: | ||
| name: parent | ||
| dependencies: | ||
| - foo | ||
| - job: | ||
| name: child | ||
| parent: parent | ||
| dependencies: | ||
| - bar | ||
| ``` | ||
| it appears that `child` only has a dependency on `bar` | ||
| @mhuin:matrix.org | corvus, thks, will check | 14:48 |
| @jim:acmegating.com | jpew: that's correct (necessary so that a child job can override the dependency in order to depend on an overridden dependency). | 14:49 |
| @jpew:matrix.org | @corvus: Thanks | 14:50 |
| @jpew:matrix.org | * corvus: thanks | 14:50 |
| @blaisep-sureify:matrix.org | TIL about https://softwarefactory-project.io/docs/index.html | 16:08 |
| I didn't know it was a thing. | ||
| Looks like a cool disto | ||
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 840709: Add support for GHE repository cache https://review.opendev.org/c/zuul/zuul/+/840709 | 16:49 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 840381: Update docs to clarify inheritance and global repo state https://review.opendev.org/c/zuul/zuul/+/840381 | 16:50 | |
| @jakobmeng:fedora.im | jpew, fungi, Clark: Just wanted to report back that once i removed all duplicate jobs from our branches (e.g. master jobs now only in master branch), it now works as expected ☺️ Thank you again for helping out! https://review.opendev.org/c/openstack/ansible-collections-openstack/+/840437 | 18:50 |
| @jpew:matrix.org | I was able to get my use-case working also: I had to make a "branch-control-job" that most all other jobs in the project `depend` on, and then that job adjusts it's `zuul_return` `child_jobs` based on some pattern matching on `zuul.branch`. It's not pretty (better suggestions are welcome) but if fixes the problem where we can say "don't run these jobs on a branch made from master unless it matches a pattern" | 19:03 |
| @clarkb:matrix.org | I've been looking into why the auto scroll to line doesn't work in zuul web on log files. In my firefox debugger using breakpoints I can see that the useEffect callback set here https://opendev.org/zuul/zuul/src/branch/master/web/src/containers/logfile/LogFile.jsx#L75 is called after the file is done isFetching (and so the flag isFetching is set to True). But if I set a breakpoint at the scroll location: https://opendev.org/zuul/zuul/src/branch/master/web/src/containers/logfile/LogFile.jsx#L94 it is not called after isFetching toggles to true and the file is present. Any idea why this might be happening? Part of me wonders if it is due to some exception being thrown by the first pass because the file isn't loaded yet but my console doesn't show any related exceptions from what I can tell | 19:34 |
| @clarkb:matrix.org | it is calling the scrollto function twice but both occurences before the file seems to exist which means the line tags aren't present to find and scroll to | 19:36 |
| @clarkb:matrix.org | Chrome's debugger seems to eb better about showing me variable values which is nice | 19:47 |
| @clarkb:matrix.org | I managed to do something that turned my little desktop into a space heater. I think I managed to get it to run the hook too often and that caused it to spin a lot? I still can't get it to run the hook after loading though | 20:22 |
| @clarkb:matrix.org | anyway lunch now. And maybe someone who groks this better than me will see something | 20:22 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 840709: Add support for GHE repository cache https://review.opendev.org/c/zuul/zuul/+/840709 | 20:39 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 840758: Add USER to Dockerfile https://review.opendev.org/c/zuul/zuul/+/840758 | 20:56 | |
| -@gerrit:opendev.org- Joshua Watt proposed: [zuul/zuul] 838924: merger: Handle merges with cherry-pick merge-mode https://review.opendev.org/c/zuul/zuul/+/838924 | 21:09 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 840760: Add a script to find deprecated queue syntax https://review.opendev.org/c/zuul/zuul/+/840760 | 21:13 | |
| -@gerrit:opendev.org- Joshua Watt proposed: [zuul/nodepool] 839226: Do not reset quota cache timestamp when invalid https://review.opendev.org/c/zuul/nodepool/+/839226 | 21:19 | |
| @clarkb:matrix.org | I think the js issue is that we're called initially with isFetchingLogfile set to false which causes us to update the highlight lines values. Changing the highlight values triggers the next event but it has nothing to do because no lines are loaded. Then we we're called again with isFetching set to true while actually fetching which skips the highlight lines updates. Then finally called again with it set to false when fetching is complete. But this last time doesn't trigger the scroll event because it doesn't change the values for the highlights. The values remain the same | 21:36 |
| -@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul] 840769: Set logfile isFetching initial state to true https://review.opendev.org/c/zuul/zuul/+/840769 | 22:04 | |
| @clarkb:matrix.org | I've begun to wonder if chrome is properly reloading my local jsx edits so am just pushing that. I can't get things to work locally yet, but I'm fairly certain the comment in that change captures what is broken at least | 22:05 |
| @clarkb:matrix.org | Basically when logfile is created isFetching is false. This allows the first useEffect callback to trigger and call setHighlightStart(). Calling setHighlightStart() updates highlightStart which allows the second useEffect to trigger, but it is doing so before the logfile is fetched so we don't do any scrolling. Then later the first useEffect callback is called again and it calls setHighlightStart() again but with the same value. Since the highlightStart value doesn't change we don't trigger the second useEffect callback and never scroll | 22:07 |
| @clarkb:matrix.org | I think if we convince the build that logfile starts out with isFetching to true then we won't set highlightStart until the page has loaded. Then we can highlight the line and scroll to it | 22:07 |
| @clarkb:matrix.org | but I'm bashing my head against my tools (chrome and firefox) and not entirely sure if they are executing modified code or not (because I think it may be executing minified code its just that the plain text is served along side it and I'm modifying the plain text and going nowhere) | 22:08 |
| @clarkb:matrix.org | hrm you can't really test my change beacuse if you reload the site preview url to try and have it scroll for you you get Not Found instead | 22:26 |
| @clarkb:matrix.org | but also I put an if (false && oldcondition) modification in chrome and saved it and stepped through and it continued to enter that block so I'm 99% sure that those edits aren't being used either. I guess we need to deploy a zuul web at / so that we can test the reload and school? thats fun | 22:29 |
| @jim:acmegating.com | Clark: ^ | 22:45 |
| @jim:acmegating.com | * Clark: wfm locally; i tried to make a gif | 22:47 |
| @jim:acmegating.com | i tested it by checking out the change; cd web; yarn start:opendev | 22:48 |
| @clarkb:matrix.org | TIL we haev an opendev shortcut | 22:51 |
| @clarkb:matrix.org | corvus: it is hard to tell from the gif, but I guess after you selected the line and you refreshed it jumped to the line? | 22:52 |
| @jim:acmegating.com | Clark: not quite -- the only interaction was pasting the url in a new tab | 22:52 |
| @clarkb:matrix.org | ah that works too | 22:52 |
| @jim:acmegating.com | (i intentionally chose a giant log file so we can see all the delays) | 22:53 |
| @clarkb:matrix.org | so ya there were two problems. First I was editing things and not getting the edited results back as expected (would be nice if chrome made this more apparent "I'm not running your edited code despite you thinking I am") and the actual issue is that we start with isFetching set to false so we set the highlightStart value which doesn't get a new value later allowing the scroll to trigger after the page loads. If we set isFetching to True then we don't set highlightStart until after we finish fetching the document and we trigger the scroll after the page is loaded | 22:54 |
| @clarkb:matrix.org | In more positive news I know a lot more about react now :) The other thing I was looking at is preserving log tree expansion/navigation via the url bar but that seems a bit more comlpicated | 22:55 |
| @jim:acmegating.com | yeah, that one has stumped me for some time | 22:56 |
| @clarkb:matrix.org | I think to fix that we need to add on expansion handlers to update the url as we click and expand things and then when we load the url have to expand as appropriate. What makes this super tricky is that you can expand multiple divergent trees so you can't just do a 1:1 in the url bar | 22:56 |
| @jim:acmegating.com | Clark: or maybe for that, just some local storage of the expansion? | 22:57 |
| @jim:acmegating.com | i don't think it's as important to be able to share a url with the tree expansion? or is it? | 22:57 |
| @clarkb:matrix.org | its probably less important. For me the main use case is viewing a log file, then hitting back to go check another file but I have to dig back down into the tree each time | 22:58 |
| @clarkb:matrix.org | then when I share links its of specific files:lines typically | 22:58 |
| @iwienand:matrix.org | that seems a little bit like state that should be stored in redux? | 23:51 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!