| *** noisecell has joined #buildstream | 00:15 | |
| *** noisecell has quit IRC | 00:18 | |
| *** mcatanzaro has quit IRC | 00:29 | |
| *** Prince781 has quit IRC | 00:31 | |
| *** Prince781 has joined #buildstream | 01:49 | |
| *** Prince781 has quit IRC | 02:21 | |
| *** Prince781 has joined #buildstream | 03:17 | |
| *** Prince781 has quit IRC | 05:24 | |
| *** tristan has joined #buildstream | 07:18 | |
| *** toscalix has joined #buildstream | 09:04 | |
| *** noisecell has joined #buildstream | 09:20 | |
| *** tm has joined #buildstream | 09:22 | |
| *** jonathanmaw has joined #buildstream | 09:23 | |
| *** ssam2 has joined #buildstream | 09:26 | |
| *** valentind has quit IRC | 09:33 | |
| *** aday has joined #buildstream | 09:56 | |
| *** dominic has joined #buildstream | 10:15 | |
| tristan | Note, this is pretty severe and relatively easy to work on: https://gitlab.com/BuildStream/buildstream/issues/216 | 10:34 |
|---|---|---|
| gitlab-br-bot | buildstream: merge request (sam/integration-tests-alpine->master: Use a custom base sysroot for the integration tests) #243 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/243 | 10:49 |
| *** ssam3 has joined #buildstream | 11:00 | |
| *** ssam2 has quit IRC | 11:01 | |
| tristan | gah, any reason why `fail-on-overlaps` is documented in the "External Plugins" section of the project.conf docs http://buildstream.gitlab.io/buildstream/projectconf.html#fail-on-overlaps ? | 11:03 |
| gitlab-br-bot | buildstream: merge request (sam/integration-tests-alpine->master: Use a custom base sysroot for the integration tests) #243 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/243 | 11:05 |
| * tristan fixes that in his branch | 11:05 | |
| *** aday has quit IRC | 11:19 | |
| *** noisecell has quit IRC | 11:32 | |
| *** noisecell has joined #buildstream | 11:33 | |
| *** noisecell has quit IRC | 11:40 | |
| *** noisecell has joined #buildstream | 11:42 | |
| gitlab-br-bot | buildstream: merge request (239-use-pylint-for-linting->master: WIP: Resolve "Use pylint for linting") #283 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/283 | 11:49 |
| *** noisecell has quit IRC | 11:50 | |
| *** noisecell has joined #buildstream | 11:52 | |
| *** noisecell has quit IRC | 12:02 | |
| *** adds68 has quit IRC | 12:05 | |
| *** adds68 has joined #buildstream | 12:05 | |
| *** adds68 has quit IRC | 12:06 | |
| *** adds68 has joined #buildstream | 12:06 | |
| gitlab-br-bot | buildstream: merge request (239-use-pylint-for-linting->master: WIP: Resolve "Use pylint for linting") #283 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/283 | 12:31 |
| *** xjuan has joined #buildstream | 13:03 | |
| tristan | mkay, so tracking into project.refs, works across junctions, as does the overrides if your depending project has them | 13:18 |
| tristan | more optionality than I like here; code needs cleanup, and still need to add something explicit for cross-junction tracking | 13:18 |
| tristan | warnings in place for ignored refs in existing element.bst files (in the case you chose to enable project.refs style recoding of refs) | 13:19 |
| tristan | and round tripping works for manually organizing and commenting your project.refs file | 13:20 |
| *** tlater has quit IRC | 13:22 | |
| *** ltu has quit IRC | 13:22 | |
| *** aiden has quit IRC | 13:22 | |
| *** aiden has joined #buildstream | 13:23 | |
| *** jonathanmaw has quit IRC | 13:27 | |
| *** jonathanmaw has joined #buildstream | 13:27 | |
| *** tiago has quit IRC | 13:31 | |
| *** tiago has joined #buildstream | 13:31 | |
| *** laurenceurhegyi has joined #buildstream | 13:36 | |
| *** tlater has joined #buildstream | 13:37 | |
| *** aday has joined #buildstream | 13:37 | |
| *** aday has quit IRC | 13:39 | |
| *** aday has joined #buildstream | 13:40 | |
| *** tristan has quit IRC | 13:40 | |
| *** tristan has joined #buildstream | 13:44 | |
| gitlab-br-bot | buildstream: merge request (214-we-need-a-way-to-make-elements-depend-on-a-subset-of-an-element-s-output->master: WIP: Resolve "We need a way to make elements depend on a subset of an element's output") #284 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/284 | 13:48 |
| *** adds68 has quit IRC | 13:58 | |
| *** dominic has quit IRC | 14:00 | |
| *** dominic has joined #buildstream | 14:00 | |
| *** noisecell has joined #buildstream | 14:00 | |
| *** adds68 has joined #buildstream | 14:04 | |
| *** aday has quit IRC | 14:26 | |
| *** aday has joined #buildstream | 14:30 | |
| gitlab-br-bot | buildstream: merge request (239-use-pylint-for-linting->master: WIP: Resolve "Use pylint for linting") #283 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/283 | 14:41 |
| tlater | I like the branch name of #214 | 14:49 |
| tristan | the longer the better ? | 14:51 |
| tlater | I pity anyone without bash-completions trying to check it out ;) | 14:52 |
| tristan | copy/paste works too :) | 14:52 |
| tlater | Btw, does anyone know why this MRs pipeline doesn't start: https://gitlab.com/BuildStream/buildstream/merge_requests/283/commits ? | 14:52 |
| tlater | I suspect gitlab tries to run it in the other repo's runners, but that isn't configured correctly or something | 14:53 |
| tristan | tlater, because it's in jennis's repo | 14:53 |
| tristan | right | 14:53 |
| tristan | I think it needs to be configured / enabled first time around | 14:54 |
| tlater | Seems like a bit of a burden to new contributors :/ | 14:54 |
| tristan | Well, I still prefer it to patches sent on mailing lists and attached to bugzilla | 14:54 |
| tristan | but then, the old school ways have some advantages too | 14:56 |
| tlater | Like? | 14:56 |
| tlater | I'd be very surprised | 14:56 |
| * tristan had to add "Submitted branches must not contain a history of the work done in the feature branch. Please use git's interactive rebase feature in order to compose a clean patch series suitable for submission." to the HACKING.rst after quite a few submissions | 14:57 | |
| *** adds68 has quit IRC | 14:57 | |
| tristan | which include dirt proposed to apply against master history | 14:57 |
| jmac | How do you run a subset of tests with setup.py test? | 14:57 |
| tristan | this just doesnt happen when the way to submit a patch, is to just submit a patch | 14:57 |
| tlater | jmac: ./setup.py test --addopts '-k <expression>' iirc | 14:58 |
| tristan | beat me to it | 14:58 |
| * tlater sees that point, but then you don't really have a history either | 14:58 | |
| tristan | where <expression> is either the exact test, or a prefix (not sure if it does globs) | 14:59 |
| tristan | tlater, of course we do; we've been using git with send-patch-to-bugzilla for many years | 14:59 |
| tristan | before that, svn, before that, cvs | 15:00 |
| *** adds68 has joined #buildstream | 15:00 | |
| jmac | tlater: That's got it, thanks | 15:00 |
| jmac | -k frontend appears to match anything with frontend in the path | 15:00 |
| tristan | ah, I think without -k does prefix-or-only-exact-match | 15:01 |
| tlater | I'd have said more about the expression if I understood it better, it's black pytest voodoo | 15:01 |
| * tlater generally just uses it to run a specific test | 15:01 | |
| tristan | yeah I havent dug too deep, but we've collected some useful things in the HACKING file | 15:01 |
| tristan | bad tlater :) | 15:02 |
| tlater | ;P | 15:02 |
| tristan | or you mean, the -k option | 15:02 |
| *** adds68 has quit IRC | 15:11 | |
| *** adds68 has joined #buildstream | 15:21 | |
| *** adds68 has joined #buildstream | 15:26 | |
| *** laurenceurhegyi is now known as ltu | 15:29 | |
| *** mcatanzaro has joined #buildstream | 15:38 | |
| jennis | Dealing with this bare-exception: https://gitlab.com/BuildStream/buildstream/blob/master/setup.py#L81 | 16:01 |
| jennis | Seems ok to me to use an ImportError there | 16:01 |
| jennis | thoughts? | 16:01 |
| gitlab-br-bot | buildstream: merge request (jmac/configurable-logging->master: WIP: Configurable log line formatting) #282 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/282 | 16:02 |
| * tlater thinks that import can sometimes cause other exceptions because of gi.require_version() and because gi sometimes acts weirdly when it's not installed correctly | 16:02 | |
| jennis | ok so maybe just log it...? | 16:04 |
| jennis | (+ leave bare) | 16:04 |
| *** noisecell has quit IRC | 16:04 | |
| tlater | jennis: Depending on how much time you have for this, it would be nice if you could figure out what exceptions those two statements can actually throw | 16:05 |
| tlater | There's documentation out there somewhere | 16:05 |
| jennis | tlater, will look into it | 16:05 |
| tlater | Note that you're looking at gobject-introspection | 16:06 |
| tlater | Sometimes a bit hard to get package names from imports ;) | 16:06 |
| jennis | tlater, looks like the gi.require_version('OSTree', '1.0') can throw ValueError and AttributeError and the `from gi.repository import OSTree` can throw ImportError and potentially RuntimeError | 16:19 |
| tlater | Don't catch RuntimErrors | 16:19 |
| jennis | ok | 16:19 |
| tlater | But with the other three you can make that code a bit more robust :) | 16:19 |
| tlater | jennis: A RuntimeError is specifically an error designed to tell you that something is seriously wrong with the underlying system or python. Part of why bare `except` is bad is because it also catches those, afaik. | 16:20 |
| *** noisecell has joined #buildstream | 16:20 | |
| *** aday has quit IRC | 16:27 | |
| *** aday has joined #buildstream | 16:47 | |
| jonathanmaw | what's a good metric for deciding whether a change in buildstream-docker-images is good to be merged? The CI passed, but when I try to replicate the docker build locally I end up with my own error messages | 17:10 |
| jonathanmaw | I suppose I'll see if the cache.zip file that's uploaded is illuminating. | 17:12 |
| ssam3 | if the CI passes, it should be good | 17:14 |
| ssam3 | also worth checking that the image size isn't increased dramatically | 17:14 |
| ssam3 | the CI runs `docker images` (which lists the sizes) for that purpose | 17:14 |
| * tlater wondered why it did that | 17:15 | |
| ssam3 | did you follow the newly documented process to build ? it's at https://gitlab.com/BuildStream/buildstream-docker-images/merge_requests/23/diffs | 17:15 |
| jonathanmaw | ssam3: yep, it worked right up until "Error unpacking rpm package iputils-20161105-7.fc27.x86_64 | 17:16 |
| jonathanmaw | Error unpacking rpm package iputils-20161105-7.fc27.x86_64 | 17:16 |
| jonathanmaw | iputils-20161105-7.fc27.x86_64 was supposed to be installed but is not!" | 17:16 |
| jonathanmaw | oh, I've actually had a bunch of other errors | 17:16 |
| jonathanmaw | "Failed to connect to bus: No such file or directory" is a pretty odd thing to get | 17:17 |
| jonathanmaw | I think in the scriptlet for lzip | 17:17 |
| jonathanmaw | ah, seems to be benign, as that shows up in the CI as well | 17:18 |
| jonathanmaw | ssam3: it's gone from 470MB to 500MB from adding quilt | 17:19 |
| jonathanmaw | quilt (with its dependencies) seems pretty heavy, but the image has been bigger before, so I'm not concerned | 17:20 |
| * jonathanmaw mashes the approve button | 17:20 | |
| nexus | oh ,i know that error | 17:22 |
| nexus | its the one i had! | 17:22 |
| nexus | it's a docker.service thing i think | 17:22 |
| jonathanmaw | nexus: the iputils one? | 17:22 |
| * nexus flails wildly at skullman | 17:23 | |
| nexus | yes | 17:23 |
| skullman | that error is because aufs doesn't support file capabilities | 17:24 |
| *** tm has quit IRC | 17:26 | |
| nexus | ^ | 17:26 |
| skullman | we fixed this locally by changing the docker container tree backend to devicemapper | 17:28 |
| skullman | I don't think this will be possible for our runners, which is going to be a problem. | 17:28 |
| skullman | I don't know if we can tell dnf to not care that we can't add those filesystem attributes. | 17:30 |
| ssam3 | hmm, so installing quilt in a container fails? | 17:30 |
| ssam3 | but if it worked in the CI, then it's not an issue | 17:30 |
| ssam3 | we run the build pipeline for candidate branches, so it should already have been tested | 17:30 |
| ssam3 | i mean, it would still be nice if it didn't fail | 17:31 |
| ssam3 | locally | 17:31 |
| ssam3 | but whatever | 17:31 |
| skullman | I think the issue is that we got it to work locally with aconfiguration change we can't make on the CI | 17:31 |
| ssam3 | but it must work in the CI already | 17:31 |
| ssam3 | otherwise the pipeline for that branch would be failing | 17:31 |
| ssam3 | or is this a different branch ? | 17:31 |
| skullman | ah, I'm getting confused with something else | 17:32 |
| skullman | or am I, I'm not sure | 17:32 |
| skullman | ah, tangentially related which is how I was getting confused | 17:33 |
| skullman | ssam3: should be ok then | 17:33 |
| persia | It is slightly worrisome that something that works in CI cannot be replicated locally, but yes, if it actually works in CI (rather than only appearing to work), that is a good indication that the latest change didn't break it. | 17:33 |
| nexus | it can be replicates locally,it just requires a config change | 17:35 |
| nexus | *or have we changed subject?* | 17:35 |
| *** ssam3 has quit IRC | 17:36 | |
| *** toscalix has quit IRC | 17:45 | |
| persia | nexus: If it requires a change, that's not replication. | 17:56 |
| persia | That's doing something subtly different that gets the same result. | 17:56 |
| * persia is picky about these sorts of things | 17:56 | |
| skullman | perhaps, but it's a docker bug that causes this not ours | 17:56 |
| persia | (mostly as a result of many years of listening to "it worked on my machine") | 17:56 |
| persia | skullman: Yep. We should probably try to fix that, assuming it isn't huge. Or at least understand and report it. | 17:57 |
| skullman | persia: from what I understand it's well understood and why there was a push to get away from AUFS as a first-class Docker tree backend, but that was a couple of years ago and I didn't keep up with the discussion. | 17:58 |
| persia | Ah, excellent. Bugs well-understood upstream with a plan for resolution are my favorite kind. | 17:58 |
| *** jonathanmaw has quit IRC | 18:03 | |
| *** tm has joined #buildstream | 18:17 | |
| *** dominic has quit IRC | 18:26 | |
| *** valentind has joined #buildstream | 18:34 | |
| *** Prince781 has joined #buildstream | 20:13 | |
| *** Prince781 has quit IRC | 20:16 | |
| *** aday has quit IRC | 21:00 | |
| *** xjuan has quit IRC | 21:01 | |
| *** valentind has quit IRC | 22:09 | |
| *** jennis has quit IRC | 23:06 | |
| *** jennis has joined #buildstream | 23:07 | |
| *** tristan has quit IRC | 23:09 | |
| *** tm has quit IRC | 23:50 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!