| *** nimish2711 has joined #buildstream | 02:03 | |
| *** tristan has quit IRC | 04:56 | |
| *** tristan has joined #buildstream | 05:18 | |
| *** kapil___ has joined #buildstream | 05:39 | |
| *** alatiera has joined #buildstream | 06:20 | |
| *** bochecha has joined #buildstream | 06:41 | |
| *** ChanServ sets mode: +o tristan | 06:47 | |
| tristan | Hmmm, here is another INTERNALERROR> https://gitlab.com/BuildStream/buildstream/-/jobs/200600652 | 06:52 |
|---|---|---|
| tristan | Also, whats with this buildstream-bastion-overnight runner ? taking soo long to start: https://gitlab.com/BuildStream/buildstream/-/jobs/200600662 | 06:53 |
| tristan | Maybe it *really* only runs in the nighttime ? | 06:53 |
| tristan | Hmmm, any idea about this bastion ? | 07:21 |
| tristan | I'm seeing: ERROR: Preparation failed: exit status 1 | 07:21 |
| tristan | after not starting in a long time | 07:21 |
| *** tristan has quit IRC | 07:32 | |
| *** toscalix has joined #buildstream | 07:52 | |
| *** tristan has joined #buildstream | 07:53 | |
| *** ChanServ sets mode: +o tristan | 07:58 | |
| *** rdale has joined #buildstream | 08:26 | |
| *** tpollard has joined #buildstream | 09:03 | |
| *** jonathanmaw has joined #buildstream | 09:07 | |
| gitlab-br-bot | phildawson opened MR !1307 (phil/move-integration-cache-to-testing->master: Make Integration cache fixture available in testing package) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1307 | 09:18 |
| phildawson_ | Do the buildstream/buildstream:nightly docker images get automatically updated or need manually updating? | 09:45 |
| phildawson_ | Either way, I need there to be a new nightly image, does anyone know how I can go about making this happen? | 09:47 |
| benschubert | phildawson_: hey, do you remember why you added ïnclude_package_data"in setup.py ? | 10:05 |
| phildawson_ | Let me take a quick look to remind myself | 10:07 |
| phildawson_ | Okay, so I added when making the templated source tests available. I think I added it to quickly make the data files necessary for the templated source tests in buildstream.testing be installed along with buildstream. I meant to go back and do that properly before making an MR out of it, but clearly missed it :/ | 10:13 |
| *** cs-shadow has joined #buildstream | 10:14 | |
| *** alatiera has quit IRC | 10:15 | |
| *** alatiera_ has joined #buildstream | 10:15 | |
| *** alatiera_ is now known as alatiera | 10:16 | |
| *** tristan has quit IRC | 10:22 | |
| *** tristan has joined #buildstream | 10:24 | |
| benschubert | phildawson_: ok, because it's preventing me from packaging BuildStream using source distributions (wheels do work fine though??) and I don't really know what's the problem, just know that removing it will allow me to run buildstream x) | 10:31 |
| benschubert | phildawson_: what are the things you wanted to cleanup there? | 10:31 |
| phildawson_ | benschubert, I'll quickly tidy that up now. | 10:31 |
| benschubert | oh ok great, let me know when you have a PR :) | 10:31 |
| phildawson_ | Will do :) | 10:31 |
| phildawson_ | I wonder what's going on there though | 10:32 |
| benschubert | I wish I knew | 10:32 |
| cs-shadow | phildawson_: the nightly images do get rebuilt. They sometimes get hit by flaky CI, as they did today | 10:33 |
| cs-shadow | https://gitlab.com/BuildStream/buildstream/issues/1008 for the missing package data in case you hadn't noticed before | 10:34 |
| phildawson_ | Cheers cs-shadow. | 10:35 |
| * cs-shadow just hit the rebuild button | 10:35 | |
| phildawson_ | I wasn't sure, because it had been 10 days since the latest version. | 10:36 |
| cs-shadow | yeah, the CI failures happen more often than ideal :/ If i find time at some point, i need to look into that | 10:37 |
| * cs-shadow isn't sure what to do when `dnf install` seg faults : https://gitlab.com/BuildStream/buildstream-docker-images/-/jobs/200757345 | 10:38 | |
| phildawson_ | erm, that's worrying. | 10:39 |
| cs-shadow | seems to happen rather consistently, on the same fedora 30 image. that also explains no updates for the past week or so | 10:41 |
| * phildawson_ wonders if it could be to do with this: https://bit.ly/2W10fwH | 10:51 | |
| benschubert | phildawson_: last updated upstream: 19 days ago for the fedora docker image, seems consistent | 10:54 |
| phildawson_ | That was my thought. I don't really understand the issue, but the symptoms and timing seem to match up. | 10:58 |
| benschubert | basically, they introduced a new compression mechanism that ended up segfaulting and they had to disable it on the repos and update the client before reenabling, but if you are on a bad version now, you need to explicitely disable to be able to update -_-'' | 11:07 |
| cs-shadow | i'd have hoped that the they will publish a new base image with the fix | 11:23 |
| gitlab-br-bot | tristanvb opened MR !1308 (tristan/fdsdk-ci-refresh->master: .gitlab-ci.yml: Updating overnight tests to test ported freedesktop-sdk) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1308 | 11:25 |
| *** ChanServ sets mode: +o tristan | 11:26 | |
| tristan | phildawson_, So that mr above ^^ should make the overnight tests work again... unless there is a bug still in my port of the flatpak_image but I doubt it | 11:27 |
| tristan | phildawson_, I left you hanging on that issue/mr about tests - I am personally really keen on having something more explicit as I described (have an external plugin declare a test case and call an explicit API handing over a repo implementation and plugin kind, instead of some mysterious hooks triggering tests) | 11:28 |
| tristan | phildawson_, but I dont want to push more work onto you because of that, so I hope maybe I can help out and try to figure out a more nicely integrated API surface for source tests, if you'll allow me :) | 11:29 |
| *** tristan has quit IRC | 11:53 | |
| *** lachlan has joined #buildstream | 11:58 | |
| *** tristan has joined #buildstream | 11:58 | |
| *** lachlan has quit IRC | 12:02 | |
| laurence | jonathanmaw, jennis - I'm wondering if we want to trim the table in the weekly performance email a little | 12:32 |
| laurence | i usually just look at the last 4 or 5 rows | 12:33 |
| phildawson_ | tristan, it'll be nice to have the overnight test working again :) About the source tests, I freely accept that it's a little bit magic and I'm definitely happy to discuss improvements. As you say though, I'm not sure quite how much time I'll be able to spend on it. | 12:38 |
| jennis | laurence, yeah nice idea, it's getting quite long | 13:01 |
| *** lachlan has joined #buildstream | 13:03 | |
| phildawson_ | +1 it's easy enough to go looking for the historical data if you want to :) | 13:08 |
| *** lachlan has quit IRC | 13:11 | |
| gitlab-br-bot | phildawson opened MR !1309 (phil/1008->master: setup.py: Include buildstream.testing datafiles in package_data) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1309 | 13:15 |
| phildawson_ | ^benschubert | 13:16 |
| benschubert | phildawson_: great thanks ! | 13:24 |
| benschubert | let me test this | 13:25 |
| *** phil has joined #buildstream | 13:32 | |
| *** phildawson_ has quit IRC | 13:33 | |
| gitlab-br-bot | BenjaminSchubert approved MR !1309 (phil/1008->master: setup.py: Include buildstream.testing datafiles in package_data) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1309 | 13:46 |
| *** lachlan has joined #buildstream | 13:53 | |
| *** lachlan has quit IRC | 13:58 | |
| gitlab-br-bot | phildawson opened issue #1009 (CI doesn't test sdist packaging includes necessary data files) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1009 | 14:04 |
| *** lachlan has joined #buildstream | 14:04 | |
| *** lachlan has quit IRC | 14:09 | |
| laurence | jennis, also, and this may be a dumb question, but we still have 'show' and 'show with cache' columns | 14:12 |
| laurence | do we need both? | 14:12 |
| laurence | i guess it can't hurt, but i was a little surprised we still tested with a hot cache - isn't it redundant now? or doesn't it add an overhead? | 14:13 |
| benschubert | guess we can drop it in the next email :) | 14:15 |
| jonathanmaw | tristan: do you have https://gitlab.com/BuildStream/buildstream/merge_requests/1299 on your agenda at some point? I'd like a sanity check if you have time. | 14:33 |
| *** lachlan has joined #buildstream | 14:33 | |
| jennis | laurence, yeah it was just legacy stuff | 14:47 |
| jennis | Will also drop this in the next email | 14:48 |
| tpollard | is there any reason that we 'force' certain abstract methods to be implemented by certain subclasses, but don't use @abstractmethod to enforce it? | 14:48 |
| *** lachlan has quit IRC | 14:51 | |
| gitlab-br-bot | abderrahimk reopened issue #990 (BUG in push with non-strict mode) on buildstream https://gitlab.com/BuildStream/buildstream/issues/990 | 14:55 |
| *** phil is now known as phildawson | 15:02 | |
| *** toscalix has quit IRC | 15:19 | |
| *** alatiera has quit IRC | 15:22 | |
| *** toscalix has joined #buildstream | 15:27 | |
| *** toscalix has quit IRC | 15:27 | |
| *** alatiera has joined #buildstream | 15:28 | |
| gitlab-br-bot | marge-bot123 closed issue #1008 (Data files are missing from python package) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1008 | 15:55 |
| gitlab-br-bot | marge-bot123 merged MR !1309 (phil/1008->master: setup.py: Include buildstream.testing datafiles in package_data) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1309 | 15:55 |
| gitlab-br-bot | jennis opened MR !1310 (jennis/revert_gc_management->master: Revert !1164 - Manage GC during pipeline load) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1310 | 16:00 |
| tpollard | cs-shadow: around? | 16:51 |
| cs-shadow | tpollard: hey! yes | 16:51 |
| tpollard | cool :) | 16:51 |
| tpollard | cs-shadow: I'm currently trying to get !1292 over the line, after recently rebasing it against master it now has a single test failure. Working my way back and cherry picking the the patchset it occurs when https://gitlab.com/BuildStream/buildstream/commit/1cbc5e63dc10c7676237fa923273a3937d6977ca lands | 16:53 |
| gitlab-br-bot | MR !1292: Artifact as a Proto: rework https://gitlab.com/BuildStream/buildstream/merge_requests/1292 | 16:53 |
| tpollard | any ideas? | 16:53 |
| cs-shadow | interesting, is the failure in a CI job somewhere? | 16:54 |
| tpollard | https://gitlab.com/BuildStream/buildstream/-/jobs/200866731 is an instance of the failure | 16:54 |
| tpollard | but it occurs locally too | 16:54 |
| tpollard | the MR applied on top of the previous commit to that one linked it passes fine | 16:54 |
| cs-shadow | let me have a look. We did stop creating an empty `bst` directory so if anything was relying on that directory being there, it might be affected | 16:55 |
| tpollard | That's my current thought | 16:56 |
| tpollard | `git cherry-pick e2285e58472633b7cf836c9c950e32507231190a^..744ca6b9c7cd99af8439db761390f2dfd3ab9140` for the set | 16:56 |
| tpollard | also this isn't my MR, so makes it even more fun :) | 16:57 |
| *** lachlan has joined #buildstream | 17:03 | |
| *** jonathanmaw has quit IRC | 17:05 | |
| *** lachlan has quit IRC | 17:08 | |
| cs-shadow | tpollard: my testing setup is somewhat busted at the moment (different story) but my suspicion is that the tests needs to be updated. We are asserting that the size of our output is non-zero. This is fine for a stack element, which is what are seeing in the CI. `assert directory_hash and directory_size` fails because the size is 0 | 17:14 |
| cs-shadow | I think this is because of the changes in _artifact.py related to how the size is calculated. | 17:14 |
| tpollard | yes I think it's stemming from cache() in _artifact.py | 17:15 |
| tpollard | hmm | 17:17 |
| cs-shadow | I don't have a place to test it somewhere but I'd be tempted to see what the size is at the moment. Maybe it is taking into account the size of the parent directory or something like that | 17:20 |
| *** nimish2711 has quit IRC | 18:08 | |
| *** alatiera has quit IRC | 18:59 | |
| *** kapil___ has quit IRC | 19:02 | |
| *** bochecha has quit IRC | 22:06 | |
| *** rdale has quit IRC | 23:36 | |
| *** benschubert has quit IRC | 23:48 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!