| *** Prince781 has quit IRC | 01:18 | |
| *** Prince781 has joined #buildstream | 01:25 | |
| *** wsgalaxy has joined #buildstream | 04:35 | |
| *** wsgalaxy has left #buildstream | 05:01 | |
| *** Prince781 has quit IRC | 05:10 | |
| *** dominic has joined #buildstream | 07:15 | |
| *** Prince781 has joined #buildstream | 07:28 | |
| dominic | Bwrap PR for #286 passed all checks, just need to change a couple of things and it should be good to merge | 07:30 |
|---|---|---|
| *** Prince781 has quit IRC | 07:42 | |
| *** bochecha_ has joined #buildstream | 07:57 | |
| *** toscalix has joined #buildstream | 08:16 | |
| *** jonathanmaw has joined #buildstream | 08:18 | |
| *** slaf has quit IRC | 08:31 | |
| *** slaf has joined #buildstream | 08:33 | |
| *** slaf has joined #buildstream | 08:33 | |
| *** slaf has joined #buildstream | 08:34 | |
| *** slaf has joined #buildstream | 08:34 | |
| *** slaf has joined #buildstream | 08:34 | |
| *** slaf has joined #buildstream | 08:34 | |
| *** slaf has joined #buildstream | 08:35 | |
| *** slaf has joined #buildstream | 08:35 | |
| *** slaf has joined #buildstream | 08:35 | |
| *** slaf has joined #buildstream | 08:35 | |
| Nexus | Does anyone know why the Element.__extract() method is private? juergbi tlater ? I need to use it (or re-write the functionality) in another class | 08:40 |
| *** jennis has joined #buildstream | 09:03 | |
| *** tiago has joined #buildstream | 09:05 | |
| skullman | Nexus: I think it's probably because it's meant to be an implementation detail since it's meant to be behind the scenes that it gets extracted during other functions | 09:09 |
| juergbi | Nexus: we typically make everything private that is not used from the outside | 09:14 |
| juergbi | I'd have to take a closer look but it could well be that you can make it public (or rather, make it API-private, i.e., one instead of two underscores) | 09:14 |
| tlater | Nexus: Isn't Element.stage enough for what you'd like to do? | 09:14 |
| Nexus | juergbi: can i A) Re-write the functionality or B) make it less private? | 09:15 |
| juergbi | I would definitely prefer less private to code duplication | 09:15 |
| juergbi | but it depends on what you want to do | 09:15 |
| Nexus | juergbi: i want to do the extract within a function in pstreecache.py | 09:17 |
| Nexus | ostreecache.py* | 09:17 |
| juergbi | what's the context? | 09:17 |
| juergbi | will it also be needed by other artifact caches? | 09:18 |
| juergbi | if so, shouldn't it be outside the artifact cache impl.? | 09:18 |
| tlater | Nexus: It would help if you could try and explain what you're trying to do - possible there's a better way that doesn't require private API (: | 09:19 |
| tlater | juergbi: It looks like pycoverage creates an empty coverage file specifically on this branch with this docker container, and breaks when trying to combine it |: | 09:24 |
| * tlater wonders if it would be acceptable to just remove empty .coverage.* files before trying to combine them | 09:25 | |
| Nexus | Ok so in "caching buildtrees" i have to check that the artifact exists, and if it does, extract the build-tree directory, before this was done in element, however that didn't account for tarcache. So now tarcache handles it itself and passes back a string/path, so i want ostreecache to do the same, meaning i need to extract it inside that class | 09:25 |
| tlater | Nexus: So you want to extract *just* the build tree? | 09:27 |
| Nexus | i'm extracting the whole thing afaik, but just returning the path for the build tree dir | 09:28 |
| tlater | In either case, you'll find that Element.__extract just calls ArtifactCache.extract | 09:29 |
| * tlater suggests using that function instead | 09:29 | |
| Nexus | cool :) | 09:42 |
| *** bochecha_ has quit IRC | 09:45 | |
| * tlater wonders if he's seeing this pycoverage bug because we run it concurrently... | 10:05 | |
| *** lantw44 has quit IRC | 10:05 | |
| *** lantw44 has joined #buildstream | 10:07 | |
| *** bochecha_ has joined #buildstream | 10:12 | |
| gitlab-br-bot | buildstream: merge request (tlater/context-trailing-slash->master: tests/context/context.py: Don't break with trailing slashes) #470 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/470 | 10:19 |
| noisecell | juergbi, /usr/sbin is not an option either, with /usr/bin seems to work, but I would like to check that the minimal base image for definitions boot using init in that path. | 10:25 |
| noisecell | boots* | 10:28 |
| gitlab-br-bot | buildstream: merge request (379-_sandboxbwrap-py-post-bwrap-cleanup-shouldn-t-attempt-to-remove-preexisting-folders->master: WIP: Resolve "_sandboxbwrap.py: Post-bwrap cleanup shouldn't attempt to remove preexisting folders") #478 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/478 | 10:30 |
| gitlab-br-bot | buildstream: merge request (379-_sandboxbwrap-py-post-bwrap-cleanup-shouldn-t-attempt-to-remove-preexisting-folders->master: WIP: Resolve "_sandboxbwrap.py: Post-bwrap cleanup shouldn't attempt to remove preexisting folders") #478 changed state ("closed"): https://gitlab.com/BuildStream/buildstream/merge_requests/478 | 10:33 |
| gitlab-br-bot | buildstream: merge request (valentindavid/331_include->master: WIP: Add support for include in project.conf) #471 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/471 | 10:39 |
| gitlab-br-bot | buildstream: merge request (caching_build_trees->master: WIP: Caching buildtrees) #474 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/474 | 10:46 |
| gitlab-br-bot | buildstream: merge request (gokcen_sandbox_cleanup_test->master: WIP: _sandboxbwrap.py: Fix post-bwrap cleanup behaviour) #479 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/479 | 11:05 |
| gitlab-br-bot | buildstream: merge request (gokcen_sandbox_cleanup_test->master: WIP: _sandboxbwrap.py: Fix post-bwrap cleanup behaviour) #479 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/479 | 11:15 |
| jjardon | Hi, I'd like to have a Fedora28 image of bst, what is the policy to update https://gitlab.com/BuildStream/buildstream-docker-images/tree/master/fedora ? | 11:26 |
| gitlab-br-bot | buildstream: issue #10 ("Implement dpkg Source") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/10 | 11:54 |
| gitlab-br-bot | buildstream: merge request (caching_build_trees->master: WIP: Caching buildtrees) #474 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/474 | 11:58 |
| tlater | jjardon: Create a new image under testsuite | 12:02 |
| tlater | We'll keep the old image around | 12:02 |
| tlater | Or do you want a fedora 28 image to run buildstream for actual build work? | 12:02 |
| jjardon | tlater: both really | 12:04 |
| tlater | jjardon: Hm, alright | 12:04 |
| tlater | Well, we would like to separate images used for the test suite from images used for running buildstream | 12:04 |
| jjardon | any special reason for doing that? Is it worth the extra maintenance? | 12:05 |
| tlater | If you'd like to create a runtime image, I suggest adding a "buildstream" directory, and adding a fedora-28 subdirectory to that | 12:05 |
| tlater | jjardon: It's currently quite confusing to figure out which image to use | 12:05 |
| tlater | Also, if we have buildstream installed already, it's possible that the test suite gets confused | 12:05 |
| jjardon | oh, I actually do not need buildstream installed | 12:06 |
| tlater | Ah, in that case just create a new testsuite image :) | 12:06 |
| jjardon | great, ok | 12:06 |
| * tlater will remember to ask "Do you need buildstream installed" next time - or add it to CONTRIBUTING.md | 12:07 | |
| gitlab-br-bot | buildstream: merge request (caching_build_trees->master: WIP: Caching buildtrees) #474 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/474 | 12:42 |
| gitlab-br-bot | buildstream: merge request (gokcen_sandbox_cleanup_test->master: WIP: _sandboxbwrap.py: Fix post-bwrap cleanup behaviour) #479 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/479 | 12:44 |
| Nexus | i'm getting "buildstream._exceptions.PlatformError: Root privileges are required to run without bubblewrap." | 12:48 |
| Nexus | last tim i ran bst as root, things broke | 12:49 |
| Nexus | options? :) | 12:49 |
| tlater | Nexus: Use a docker container | 12:49 |
| tlater | If you run with the unix platform, you require root | 12:50 |
| tlater | No way around it - unix doesn't support sandboxing without root access, at least atm | 12:50 |
| tlater | We looked into using usrchroot at some point but that turned out not to be exactly feasible | 12:50 |
| Nexus | can you lionk your bst-here script again pls? | 12:51 |
| Nexus | link* | 12:51 |
| tlater | It's in a snippet in the main repository | 12:51 |
| tlater | No need to link, really | 12:51 |
| gitlab-br-bot | buildstream: merge request (gokcen_sandbox_cleanup_test->master: WIP: _sandboxbwrap.py: Fix post-bwrap cleanup behaviour) #479 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/479 | 13:05 |
| *** xjuan has joined #buildstream | 13:09 | |
| Nexus | tlater: i'm getting the "//" issue again, it seems to just be when i use this docker img | 13:13 |
| tlater | Nexus: Yes, as I've explained before that's because the docker image sets XDG_CACHE_HOME | 13:13 |
| tlater | Buildstream doesn't normalize its paths correctly, which casues a few test failures, but is otherwise harmless | 13:14 |
| tlater | Also see this MR: https://gitlab.com/BuildStream/buildstream/merge_requests/470 | 13:14 |
| Nexus | i'm suddenly getting `py.error.EACCES: [Permission denied]: rmtree('/home/phillipsmyth/projects/buildstream/tmp',` | 13:38 |
| Nexus | no idea why | 13:38 |
| paulsherwood | Nexus: did something run there as root? | 13:38 |
| Nexus | not under my instructions | 13:39 |
| paulsherwood | do ls -la on that dir to see whether it has special permissions | 13:39 |
| Nexus | drwxr-xr-x 688 root root 69632 Jun 1 14:12 tmp | 13:39 |
| paulsherwood | there ya go | 13:39 |
| Nexus | eugh | 13:39 |
| Nexus | ty | 13:39 |
| paulsherwood | np | 13:39 |
| gitlab-br-bot | buildstream: merge request (gokcen_sandbox_cleanup_test->master: WIP: _sandboxbwrap.py: Fix post-bwrap cleanup behaviour) #479 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/479 | 13:41 |
| tlater | Nexus: That is a temporary directory created by the test suite | 13:41 |
| noisecell | is it possible to know which artifact is build latest when changing one element? https://paste.baserock.org/kawekojuga | 13:41 |
| tlater | Since you ran the test suite as root, it created the temporary directory as root | 13:41 |
| Nexus | tlater: but i didnt | 13:41 |
| tlater | Then you tried running the test suite as non-root | 13:41 |
| tlater | You did, inside the docker container | 13:41 |
| Nexus | i only ran as root in the docker | 13:41 |
| Nexus | yeah | 13:41 |
| tlater | Yep | 13:41 |
| Nexus | im out of it now | 13:41 |
| tlater | That created a root-owned directory | 13:41 |
| Nexus | outside the container??? | 13:41 |
| tlater | Yes, because the directory was mounted | 13:41 |
| tlater | That *specific* directory can be affected | 13:42 |
| paulsherwood | noisecell: apparently not... that looks like a bug to me | 13:42 |
| tlater | Otherwise you'd have to keep rebuilding the image every time you wanted to test buildstream | 13:42 |
| paulsherwood | is bst forcing all creation times to 1970? | 13:42 |
| tlater | Which would be a PITA to say the least | 13:42 |
| tlater | paulsherwood: Yes | 13:42 |
| paulsherwood | tlater: why? | 13:42 |
| tlater | We want builds to be perfectly reproducible | 13:43 |
| paulsherwood | i understand that... i don't see why the artifact needs to be force-timed to achieve it | 13:43 |
| paulsherwood | (i can understand force-timing of the contents of the artifact) | 13:43 |
| tlater | Ah, that in turn is a side effect of using ostree, I believe | 13:44 |
| tlater | Not quite as sure about that | 13:44 |
| * paulsherwood increasingly wishes he'd never heard of ostree in this context. seems to contributoe to more problems than it solves | 13:44 | |
| * paulsherwood apologises to the developers of ostree, who are no doubt lovely people | 13:45 | |
| tlater | Well, we'll be using a different cache soon enough :) I suspect ostree just wasn't created with this use case in mind at all | 13:45 |
| * paulsherwood concurs | 13:46 | |
| noisecell | tlater, will this issue be solved in the new cache? The use case would be: 1 developer changing 1/multiple elements and want to see the meta/files/logs from them, at the moment to be sure which one is which I would have to either log the build, save the cache (in both cases if I would like to compare outputs) or remove these artifacts | 13:49 |
| tlater | noisecell: I *believe* it will, but I haven't looked into the details of the new cache yet | 13:49 |
| noisecell | tlater, should I create a bug for this and you can market as review when new cache arrives? | 13:50 |
| tlater | Yes, that's fair. We can also think about a different solution for this specific use case, because you shouldn't have to poke in buildstream internals. | 13:51 |
| gitlab-br-bot | buildstream: issue #169 ("Documentation improvements") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/169 | 13:52 |
| gitlab-br-bot | buildstream: merge request (caching_build_trees->master: WIP: Caching buildtrees) #474 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/474 | 14:00 |
| gitlab-br-bot | buildstream: issue #416 ("Imposible to determine creation time in local artifacts") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/416 | 14:02 |
| tlater | ta noisecell | 14:03 |
| noisecell | tlater, no probs | 14:03 |
| Nexus | fatal: reference is not a tree: c9ed485f7afd2e457f6a6027b9e4398cbefa02ae | 14:13 |
| Nexus | wut | 14:13 |
| tlater | Nexus: That happens when you force push too quickly | 14:14 |
| Nexus | how do i fix it? | 14:14 |
| tlater | Look at the latest Job, instead of one that started before your last force push | 14:14 |
| Nexus | ahh, i needed to refresh | 14:15 |
| gitlab-br-bot | buildstream: merge request (caching_build_trees->master: WIP: Caching buildtrees) #474 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/474 | 14:36 |
| gitlab-br-bot | buildstream: merge request (gokcen_sandbox_cleanup_test->master: WIP: _sandboxbwrap.py: Fix post-bwrap cleanup behaviour) #479 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/479 | 15:05 |
| *** Prince781 has joined #buildstream | 15:09 | |
| gitlab-br-bot | buildstream: merge request (richardmaw/cache-fail->master: WIP: Store failed builds in the cache) #475 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/475 | 15:43 |
| gitlab-br-bot | buildstream: merge request (richardmaw/cache-fail->master: WIP: Store failed builds in the cache) #475 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/475 | 15:45 |
| *** dominic has quit IRC | 15:51 | |
| gitlab-br-bot | buildstream: merge request (valentindavid/331_include->master: WIP: Add support for include in project.conf) #471 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/471 | 16:03 |
| gitlab-br-bot | buildstream: merge request (gokcen/sandbox_cleanup_fix->master: WIP: Fix post-bwrap cleanup behaviour) #480 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/480 | 16:10 |
| gitlab-br-bot | buildstream: merge request (gokcen_sandbox_cleanup_test->master: WIP: _sandboxbwrap.py: Fix post-bwrap cleanup behaviour) #479 changed state ("closed"): https://gitlab.com/BuildStream/buildstream/merge_requests/479 | 16:11 |
| *** jonathanmaw has quit IRC | 16:28 | |
| *** xjuan has joined #buildstream | 16:46 | |
| *** toscalix has quit IRC | 16:56 | |
| *** bochecha_ has quit IRC | 17:05 | |
| *** tiago has quit IRC | 17:07 | |
| gitlab-br-bot | buildstream: merge request (gokcen/sandbox_cleanup_fix->master: Fix post-bwrap cleanup behaviour) #480 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/480 | 17:15 |
| *** jennis has quit IRC | 17:40 | |
| *** Prince781 has quit IRC | 17:57 | |
| *** Prince781 has joined #buildstream | 18:52 | |
| *** xjuan has quit IRC | 19:22 | |
| *** Prince781 has quit IRC | 19:41 | |
| *** Prince781 has joined #buildstream | 20:34 | |
| *** Prince781 has quit IRC | 20:51 | |
| *** Prince781 has joined #buildstream | 20:55 | |
| *** Prince781 has quit IRC | 21:01 | |
| *** cs_shadow has quit IRC | 21:12 | |
| *** Prince781 has joined #buildstream | 22:24 | |
| *** Prince781 has quit IRC | 22:49 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!