| *** leopi has quit IRC | 00:02 | |
| *** coldtom has joined #buildstream | 04:59 | |
| *** tristan has joined #buildstream | 06:01 | |
| *** ChanServ sets mode: +o tristan | 06:12 | |
| *** leopi has joined #buildstream | 06:22 | |
| gitlab-br-bot | buildstream: merge request (tristan/docs-version-badge->master: WIP: Release badges) #742 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/742 | 06:40 |
|---|---|---|
| *** noisecell has joined #buildstream | 07:08 | |
| *** tristan has quit IRC | 07:17 | |
| *** adds68 has joined #buildstream | 07:33 | |
| *** tristan has joined #buildstream | 07:46 | |
| *** ChanServ sets mode: +o tristan | 07:47 | |
| *** toscalix has joined #buildstream | 07:53 | |
| *** finn has joined #buildstream | 08:03 | |
| *** qinusty has joined #buildstream | 08:04 | |
| gitlab-br-bot | buildstream: issue #516 ("BuildStream can not build from non-numeric tags") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/516 | 08:09 |
| gitlab-br-bot | buildstream: merge request (537-mirror-fallback-does-not-work-for-git->master: WIP: Resolve "Mirror fallback does not work for git") #618 changed state ("closed"): https://gitlab.com/BuildStream/buildstream/merge_requests/618 | 08:10 |
| gitlab-br-bot | buildstream: issue #537 ("Mirror fallback does not work for git") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/537 | 08:10 |
| gitlab-br-bot | buildstream: issue #584 ("Slow cache, 25 seconds before anything is printed") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/584 | 08:12 |
| *** rdale has joined #buildstream | 08:16 | |
| gitlab-br-bot | buildstream: issue #603 ("CI takes a very long time: ~1h , ideally should complete in 15min") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/603 | 08:32 |
| *** tiagogomes has joined #buildstream | 08:49 | |
| *** bochecha has joined #buildstream | 08:51 | |
| *** jonathanmaw has joined #buildstream | 09:08 | |
| *** tristan has quit IRC | 09:13 | |
| *** tiagogomes has quit IRC | 09:14 | |
| *** tiagogomes has joined #buildstream | 09:16 | |
| valentind | jjardon, cas.prune() does not seem to be called on the server. So the refs are removed, but the objects are probably never removed. | 09:21 |
| valentind | No I was wrong. | 09:25 |
| *** tristan has joined #buildstream | 09:26 | |
| *** ChanServ sets mode: +o tristan | 09:27 | |
| gitlab-br-bot | buildstream: merge request (tristan/docs-version-badge->master: WIP: Release badges) #742 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/742 | 09:36 |
| gitlab-br-bot | buildstream: merge request (tristan/docs-version-badge->master: WIP: Release badges) #742 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/742 | 09:36 |
| gitlab-br-bot | buildstream: merge request (Qinusty/597-non-alias-url-fix->master: Fix ValueError traceback for URL without alias or URI scheme) #700 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/700 | 09:43 |
| *** dtf has quit IRC | 09:46 | |
| *** dtf has joined #buildstream | 09:46 | |
| *** tlater[m] has joined #buildstream | 09:47 | |
| gitlab-br-bot | buildstream: merge request (Qinusty/597-non-alias-url-fix->master: Fix ValueError traceback for URL without alias or URI scheme) #700 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/700 | 09:50 |
| gitlab-br-bot | buildstream: merge request (tpollard/591->master: buildstream/_project.py: Report if project.conf is missing name) #680 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/680 | 09:55 |
| *** finn_ has joined #buildstream | 09:57 | |
| *** finn has quit IRC | 09:59 | |
| gitlab-br-bot | buildstream: merge request (jonathan/expose-downloadablefilesource->master: Make DownloadableFileSource publically accessible) #743 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/743 | 09:59 |
| gitlab-br-bot | buildstream: merge request (tpollard/591->master: buildstream/_project.py: Report if project.conf is missing name) #680 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/680 | 09:59 |
| gitlab-br-bot | buildstream: merge request (tpollard/591->master: buildstream/_project.py: Report if project.conf is missing name) #680 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/680 | 10:05 |
| qinusty | Does anyone have experience with pytest.mark.timeout? | 10:08 |
| tristan | qinusty, good idea | 10:14 |
| tpollard | jjardon: could you spare any time to review or comment on this please? https://gitlab.com/BuildStream/buildstream/merge_requests/564 I know part of your usecase would be having it backported | 10:15 |
| gitlab-br-bot | buildstream: merge request (Qinusty/597-non-alias-url-fix->master: Fix ValueError traceback for URL without alias or URI scheme) #700 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/700 | 10:17 |
| gitlab-br-bot | buildstream: merge request (tristan/docs-version-badge->master: WIP: Release badges) #742 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/742 | 10:20 |
| gitlab-br-bot | buildstream: merge request (tpollard/483->master: plugins/git.py: Warn if ref is not in given track) #564 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/564 | 10:21 |
| gitlab-br-bot | buildstream: merge request (tristan/docs-version-badge->master: WIP: Release badges) #742 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/742 | 10:21 |
| jjardon | tpollard: sure, sorry I missed that ping in the MR | 10:22 |
| gitlab-br-bot | buildstream: merge request (tristan/docs-version-badge->master: WIP: Release badges) #742 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/742 | 10:24 |
| gitlab-br-bot | buildstream: merge request (tristan/docs-version-badge->master: Release badges) #742 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/742 | 10:24 |
| gitlab-br-bot | buildstream: merge request (mac_fixes->master: WIP: Resolve "os.sched_getaffinity() not supported on MacOSX Blocks #411") #726 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/726 | 10:26 |
| qinusty | I quickly reviewed it tpollard :D | 10:27 |
| qinusty | tristan, with pytest-timeout. It kills the test session rather than failing the test... | 10:27 |
| qinusty | Which seems pretty extreme for what we're after | 10:28 |
| tpollard | qinusty: cheers | 10:29 |
| tristan | qinusty, I thought it was for the test of recursive variable resolution, to avoid a hanging pipeline and fail | 10:30 |
| qinusty | It is, It'd bail the entire test sequence | 10:30 |
| tristan | Ahh, it kills the *whole session* | 10:30 |
| tristan | yikes | 10:30 |
| qinusty | just tests/format/variables.py::test_cyclic_variables<NEWLINE> | 10:30 |
| qinusty | exit | 10:30 |
| tristan | well, I'm not fussed even if the failing test hangs | 10:30 |
| qinusty | But it has SOME configuration | 10:30 |
| qinusty | I'll take a look into it a bit deeper | 10:30 |
| tristan | As long as we do have a test, hanging is an undesirable failure mode for the test suite but it is certainly much better than not having a test | 10:31 |
| tristan | qinusty, arguably, just because it failed by hanging in the past, doesnt mean it will fail the same way in the future | 10:32 |
| qinusty | Agreed | 10:32 |
| tristan | qinusty, and the same goes for any other test we do have, maybe in the future they might fail with a hang | 10:32 |
| * tristan is going to merge the version badges now | 10:33 | |
| qinusty | nice! | 10:33 |
| tristan | I've tinkered too much with them already, there are drawbacks I dont like, but meh | 10:33 |
| gitlab-br-bot | buildstream: merge request (mac_fixes->master: WIP: Resolve "os.sched_getaffinity() not supported on MacOSX Blocks #411") #726 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/726 | 10:33 |
| qinusty | Also, there was a patch a while back which overwrote and undid a change from another MR. We should be a little careful about what we're committing when dealing with merge conflicts | 10:34 |
| tristan | qinusty, indeed, did that one get fixed ? | 10:34 |
| qinusty | Yeah I reviewed and merged the refix | 10:34 |
| tristan | Good, I think that kind of thing is quite unlikely to happen - developers should rebase their branches frequently to minimize the diffs | 10:36 |
| Nexus | i'm having a problem with pytest==3.7 if gives me a "AttributeError: 'Pep8Item' object has no attribute '_collectfile'" error | 10:36 |
| qinusty | 3.7.? | 10:37 |
| tristan | Nexus, the solution to that locally, is to `pip upgrade` your pep8 package, the solution at a larger scale would be to pin pep8 to a specific version in dev-requirements.txt | 10:37 |
| tristan | jjardon, is currently working on migrating from the pep8 checker to it's renamed package pycodestyle, too | 10:38 |
| tristan | so it would be good to make sure it's pinned with that change as well | 10:38 |
| Nexus | qinusty: 3.7.0 | 10:38 |
| tristan | pep8 (or pycodestyle) falls under the same category of pylint: It should have an exact version, because it's unacceptable that CI starts to fail because of external factors (linters growing new checks) | 10:39 |
| Nexus | tristan: upgrading my pep8 to 1.7.1 didn't fix it :/ | 10:39 |
| gitlab-br-bot | buildstream: merge request (valentindavid/post_tracking_errors->master: Report processing errors from tracking) #747 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/747 | 10:41 |
| tristan | Nexus, and maybe pytest-pep8 also | 10:41 |
| tristan | I suppose it would be pytest-pep8 which claims to have intimate knowledge of a Pep8Item | 10:41 |
| tristan | and having those out of sync would be bad | 10:41 |
| Nexus | still didnt fix it | 10:42 |
| *** solid_black has quit IRC | 10:44 | |
| tristan | Nexus, please let us know when you find out what is causing this... in the meantime, a dump of the versions of things in dev-requirements.txt you are using in an issue would be a good start | 10:44 |
| gitlab-br-bot | buildstream: merge request (jjardon/pycodestyle_fixes->master: Some pycodestyle (PEP8) fixes) #746 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/746 | 10:45 |
| gitlab-br-bot | buildstream: merge request (tristan/docs-version-badge->master: Release badges) #742 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/742 | 10:46 |
| tristan | Nexus, sorry I forgot to ask first... I canceled a pipeline that you had running, because it was not the latest pipeline running for your branch | 10:48 |
| Nexus | np | 10:48 |
| tristan | unless you had intentionally been testing things in subsequent commits, all should be fine :) | 10:49 |
| Nexus | https://github.com/pytest-dev/pytest/issues/3742 | 10:49 |
| Nexus | seems to be a known problem with 3.7 | 10:49 |
| Nexus | hasn't been fixed in the latest version yet | 10:49 |
| tristan | Nice | 10:50 |
| tristan | Hmmm, strange though, how come it doesnt happen in CI ? | 10:50 |
| tristan | We use pytest >= 3.7 there | 10:50 |
| Nexus | odd | 10:51 |
| Nexus | fixed it locally with a workaround, but am confused about the error | 10:55 |
| gitlab-br-bot | buildstream: merge request (Qinusty/600-recursive-variables->master: Add cyclic check within variable resolution) #712 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/712 | 10:55 |
| gitlab-br-bot | buildstream: merge request (mac_fixes->master: WIP: Resolve "os.sched_getaffinity() not supported on MacOSX Blocks #411") #726 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/726 | 11:01 |
| tristan | Nexus, how did you fix it ? | 11:01 |
| tristan | pin something ? | 11:02 |
| Nexus | sure 1 min | 11:02 |
| tristan | I wonder if the best is to just wait for pytest 3.7.1 ultimately, but would be good to know for now | 11:02 |
| tristan | (I expect they release often enough, but am not sure) | 11:02 |
| Nexus | it's currently on 3.7.3 and not fixed yet | 11:03 |
| tristan | ewww | 11:03 |
| Nexus | yup | 11:03 |
| *** cs-shadow has joined #buildstream | 11:03 | |
| gitlab-br-bot | buildstream: merge request (jjardon/pycodestyle->master: WIP: Use pycodestyle instead pep8 python module) #638 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/638 | 11:04 |
| Nexus | i'm not sure how to make this nicer | 11:07 |
| Nexus | https://hastebin.com/jafexenisu.rb | 11:07 |
| Nexus | i added " # <---- Replaced `root` with `self`" | 11:07 |
| Nexus | at each point where it needed to be done | 11:07 |
| Nexus | the workaround on the thread wasn't enough | 11:07 |
| gitlab-br-bot | buildstream: merge request (chandan/pypi-badge->master: README.rst: Add status badges for PyPI release and Python versions) #719 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/719 | 11:08 |
| tristan | Eeek, so you patched the code | 11:09 |
| tristan | Nexus, rather than that; we should repeat the process we did to fix the pylint issues; I think that if you push a test branch to CI which prints the versions of things installed there, you can try pinning them locally in dev-requirements.txt, and one of the combinations is sure to work | 11:11 |
| tristan | Nexus, in fact, after confirming with bochecha, it will be fine to pin everything in dev-requirements.txt to an exact version and that will not interfere with the packaging story | 11:11 |
| gitlab-br-bot | buildstream: merge request (jjardon/pycodestyle->master: WIP: Use pycodestyle instead pep8 python module) #638 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/638 | 11:13 |
| gitlab-br-bot | buildstream: merge request (tristan/docs-version-badge->master: Release badges) #742 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/742 | 11:14 |
| tristan | cs-shadow, any idea why the PyPI badges are broken ? maybe we're missing some package metadata for it to work ? | 11:14 |
| tristan | i.e. https://gitlab.com/BuildStream/buildstream/tree/chandan/pypi-badge shows "invalid json response", although I seem to recall it worked last week | 11:15 |
| cs-shadow | tristan: No idea why. Looking into that at the moment. | 11:15 |
| cs-shadow | I could've sworn i saw the first badge working, the python version one isn't supposed to work until we add "py_requires" | 11:15 |
| tristan | yeah I saw the PyPI badge work for sure | 11:15 |
| cs-shadow | funnily enough it works at a different badge provider: https://badge.fury.io/for/py/BuildStream | 11:16 |
| tristan | An didn't see the python one, I wonder if showing the minimal python dependency in a badge (is that what it does ?) is really of any value, though | 11:16 |
| tristan | I mean, we could add a badge for every dependency then, but that would just be clutter | 11:17 |
| gitlab-br-bot | buildstream: merge request (valentindavid/roundtripping_only_when_modified->master: Disable round-tripping when element is not modified) #748 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/748 | 11:17 |
| cs-shadow | tristan: the version badge will add something like https://img.shields.io/pypi/pyversions/Django.svg I included it in the MR just because i was experimenting with badges. | 11:18 |
| tristan | cs-shadow, eww, that badge provider does something crazy haha :) | 11:19 |
| tristan | cs-shadow, did you look at the content ? it's an svg with javascript | 11:19 |
| tristan | cs-shadow, wait for the pipeline on master (https://gitlab.com/BuildStream/buildstream/pipelines/28943003) to complete, and see new custom badges :D | 11:20 |
| valentind | qinusty, I added !748 to deal with not round tripping unmodified files. Maybe you can have a look. It is a very small patch. | 11:22 |
| tristan | cs-shadow, looking at the django example, I think that it requires that BuildStream be precompiled for each python version (are those wheels ?)... I think we opted out of that and wanted to ensure that setup.py runs on each host and asserts the deps, isn't it ? | 11:23 |
| tristan | cs-shadow, another random question; all that text from the README on the PyPI site; is it automatically extracted from the latest dist tarball's README ? or is it redundant ? | 11:24 |
| cs-shadow | tristan: reg. wheels, yeah we are not publishing them but I'm not sure if the python versions is supposed to reflect that. I thought it was about supported versions but I'll need to check | 11:25 |
| cs-shadow | The text is generated from the long_description field of setup.py, which in turn is read from README: https://gitlab.com/BuildStream/buildstream/blob/master/setup.py#L232 | 11:26 |
| cs-shadow | I'll look a bit more into why the badges are coming back with invalid json response and report back | 11:26 |
| gitlab-br-bot | buildstream: merge request (chandan/setup.py-metadata->master: WIP: setup.py: Add more metadata useful for PyPI) #718 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/718 | 11:31 |
| tristan | cs-shadow, good to know about the README, was worried we had another thing to maintain :D | 11:35 |
| cs-shadow | tristan: yeah, that would've been a pain | 11:36 |
| gitlab-br-bot | buildstream: merge request (Qinusty/600-recursive-variables->master: Add cyclic check within variable resolution) #712 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/712 | 11:38 |
| gitlab-br-bot | buildstream: merge request (chandan/setup.py-metadata->master: WIP: setup.py: Add more metadata useful for PyPI) #718 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/718 | 11:40 |
| gitlab-br-bot | buildstream: merge request (tpollard/483->master: plugins/git.py: Warn if ref is not in given track) #564 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/564 | 11:48 |
| gitlab-br-bot | buildstream: merge request (chandan/setup.py-metadata->master: setup.py: Add more metadata useful for PyPI) #718 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/718 | 11:48 |
| gitlab-br-bot | buildstream: merge request (tpollard/483->master: plugins/git.py: Warn if ref is not in given track) #564 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/564 | 11:49 |
| gitlab-br-bot | buildstream: merge request (tpollard/483->master: plugins/git.py: Warn if ref is not in given track) #564 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/564 | 11:49 |
| gitlab-br-bot | buildstream: merge request (tpollard/591->master: buildstream/_project.py: Report if project.conf is missing name) #680 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/680 | 11:52 |
| gitlab-br-bot | buildstream: merge request (tpollard/591->master: buildstream/_project.py: Report if project.conf is missing name) #680 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/680 | 11:53 |
| gitlab-br-bot | buildstream: merge request (chandan/setup.py-metadata->master: setup.py: Add more metadata useful for PyPI) #718 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/718 | 12:13 |
| gitlab-br-bot | buildstream: merge request (tristan/setup.py-metadata-1.2->bst-1.2: setup.py: Add more metadata useful for PyPI) #749 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/749 | 12:24 |
| tristan | cs-shadow, !749 is backporting your !718 btw | 12:25 |
| tristan | cs-shadow, anything else we might be missing in bst-1.2 ? | 12:25 |
| tristan | regarding PyPI I mean :) | 12:25 |
| cs-shadow | tristan: thanks! i was about to do that myself but you beat me to it :D | 12:26 |
| cs-shadow | the only other thing that I think is not backported are the docs. But I guess that's not needed? | 12:27 |
| tristan | Nah we've been leaving them behind | 12:28 |
| tristan | We could do a post release effort but it's rather unimportant; the docs are currently only published from master | 12:28 |
| tristan | I might do it one weekend, the result will be that the docs packages delivered from bst-1.2 releases on distros will be more up to date | 12:29 |
| persia | For folk following master, that's probably fine. As 1.3 diverges from 1.2, it might make sense to put together 1.2 docs, but I suspect users will provide guidance on whether that is necessary (if only by being confused). | 12:29 |
| tristan | persia, essentially I agree, but we are blocked by https://gitlab.com/BuildStream/buildstream/issues/178 | 12:30 |
| cs-shadow | slightly tangential, have we considered hosting our docs on https://readthedocs.org/ instead of GitLab that supports multiple versions? | 12:31 |
| tristan | cs-shadow, yeah | 12:31 |
| tristan | cs-shadow, I remember that the process of considering that was quite a headache, too | 12:31 |
| cs-shadow | tristan: I see, I also missed Sam's notes from earlier: https://gitlab.com/BuildStream/buildstream/issues/178#note_58726525 | 12:32 |
| tristan | cs-shadow, yeah they have "their own way" anyway | 12:32 |
| tristan | as I read that comment, I recall looking at the upstream issue - its basically another case of "You should all do it our way" | 12:33 |
| cs-shadow | fair enough | 12:33 |
| cs-shadow | I'll try https://robpol86.github.io/sphinxcontrib-versioning/v2.2.1/index.html and see if that works | 12:33 |
| tristan | oh yeah I think jjardon provided a link to that before | 12:34 |
| * cs-shadow should finish reading issues :) | 12:35 | |
| tristan | cs-shadow, Not sure what it'll do though, I think we do want to keep using our gitlab pages automated approach, and I suppose in any case it means checking out the latest of each release branch in a loop | 12:35 |
| tristan | We could do that directly in our .gitlab-ci.yml, I guess | 12:35 |
| tristan | Or maybe we'd want "Latest release tag for every release line... plus master" | 12:36 |
| tristan | Meh | 12:36 |
| gitlab-br-bot | buildstream: merge request (chandan/setup.py-metadata->master: setup.py: Add more metadata useful for PyPI) #718 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/718 | 12:37 |
| gitlab-br-bot | buildstream: merge request (Qinusty/597-non-alias-url-fix->master: Fix ValueError traceback for URL without alias or URI scheme) #700 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/700 | 12:38 |
| cs-shadow | looping over release branches doesn't sound too bad | 12:39 |
| *** alatiera_ has joined #buildstream | 12:42 | |
| gitlab-br-bot | buildstream: merge request (tristan/setup.py-metadata-1.2->bst-1.2: setup.py: Add more metadata useful for PyPI) #749 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/749 | 12:44 |
| gitlab-br-bot | buildstream: merge request (mac_fixes->master: WIP: Resolve "os.sched_getaffinity() not supported on MacOSX Blocks #411") #726 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/726 | 12:45 |
| gitlab-br-bot | buildstream: merge request (coldtom/strip-rules->master: Upstream freedesktop-sdk strip rules) #750 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/750 | 12:49 |
| tristan | tpollard, that wont work :) | 12:52 |
| tristan | tpollard, strip-fail-on-error is a variable, not a project option | 12:52 |
| tristan | tpollard, Anyway, lets consider this for master but I think we wont do it this week | 12:53 |
| tristan | To be honest, we're probably leaning towards removing the strip-binaries completely in a later release, due to it being too linux specific | 12:54 |
| tpollard | wrong tom I presume heh | 12:55 |
| tpollard | coldtom ^ | 12:55 |
| tristan | Ohhhh sorry coldtom ^^^ | 12:58 |
| coldtom | tristan: ah, it should work with %{} around it in the (?) directive though right? | 13:00 |
| tristan | coldtom, nope | 13:01 |
| *** noisecell has quit IRC | 13:01 | |
| gitlab-br-bot | buildstream: merge request (Qinusty/597-backport-ValueError-fix->bst-1.2: Backport !700: Prevent ValueError on URLs without an alias) #751 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/751 | 13:01 |
| tristan | coldtom, It will work only by checking it expanded in the shell, using shell conditionals | 13:01 |
| gitlab-br-bot | buildstream: merge request (Qinusty/597-backport-ValueError-fix->bst-1.2: Backport !700: Prevent ValueError on URLs without an alias) #751 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/751 | 13:02 |
| tristan | coldtom, That said, it will also be better I think to not do one single switch and duplicate the huge script, but rather only check it at the exit points | 13:02 |
| tristan | coldtom, Here is a nice little hack... The variable is called "non-fatal-strip", and it's default is "true" (note that it's default is not "True", it is "true") | 13:03 |
| tristan | coldtom, then, *during the script*, the commands which can be fatal, you just append "|| %{non-fatal-strip}" to those | 13:03 |
| gitlab-br-bot | buildstream: issue #597 ("Git url without alias causes ValueError") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/597 | 13:03 |
| gitlab-br-bot | buildstream: merge request (Qinusty/597-non-alias-url-fix->master: Fix ValueError traceback for URL without alias or URI scheme) #700 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/700 | 13:04 |
| coldtom | tristan: thanks, i'll work on doing that instead :) | 13:04 |
| tristan | coldtom, that will result in "possiblyfatalthing || true" if non-fatal, and "possiblyfatalthing || false" if it would be fatal | 13:04 |
| qinusty | tristan, re: Adding a new item to devrequirements.txt. CI is failing because --index-url is invalid://uri. How do I get the CI to have all of the dev requirements installed? | 13:09 |
| tpollard | tiagogomes: are you happy for this to be merged for your issue? https://gitlab.com/BuildStream/buildstream/merge_requests/680 it has an approval | 13:12 |
| tpollard | Nexus raises a good point on the tests not being specific enough however, as it didn't catch the issue in the first place. Although that's more directed towards provenance | 13:15 |
| Nexus | Does anyone recognise this error? https://hastebin.com/sadadaxiko.cs | 13:19 |
| tristan | qinusty, that is intentional... Why do you want to pin a version to something that is not already in the CI docker images ? | 13:20 |
| tristan | qinusty, or, it's a completely new dependency ? | 13:20 |
| qinusty | Not pin, to add pytest-timeout | 13:20 |
| tristan | Ahhh, right... for the moment that requires participation from cs-shadow | 13:20 |
| qinusty | I can sort the docker images if needed tristan | 13:21 |
| qinusty | Via gitlab, but not via Dockerhub | 13:21 |
| qinusty | Unless gitlab CI pushes to dockerhub | 13:21 |
| tristan | Alternatively, we could go the other way, pin *everything* in dev-requirements.txt, and allow the CI to download them | 13:21 |
| tristan | but that is a bit more risky, because of inadvertent side effects from implicit dependencies | 13:22 |
| qinusty | Leaving it to the docker images seems fine, as long as we can modify the images | 13:22 |
| tristan | qinusty, I think it does push to dockerhub automatically, after that BuildStream's explicit refs to the docker images need an update | 13:22 |
| qinusty | I'll give it a go | 13:22 |
| tristan | i.e. in BuildStream's .gitlab-ci.yml | 13:23 |
| qinusty | yup | 13:23 |
| tristan | juergbi, I'm going to roll out a 1.3.1 this weekend, along with 1.2.0 | 13:24 |
| tristan | juergbi, I don't really expect it to be ready for testing, mostly a matter of unbreaking the links to the latest dev snapshot | 13:25 |
| juergbi | ok | 13:25 |
| gitlab-br-bot | buildstream: issue #597 ("Git url without alias causes ValueError") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/597 | 13:27 |
| qinusty | https://asciinema.org/a/DNd8LkfJySps518w84zgwkHk5 needs to make its mind up | 13:30 |
| * tpollard had forgotten about asciinema | 13:33 | |
| qinusty | Can't push to buildstream docker images :( Perhaps jjardon can help? | 13:52 |
| tristan | cs-shadow, can you give qinusty a hand adding a new dep to the CI images please ? | 13:53 |
| cs-shadow | tristan: sure, just got back to my desk | 13:55 |
| tristan | cs-shadow, thanks :D | 13:55 |
| cs-shadow | qinusty: the CI should push it automatically | 13:55 |
| cs-shadow | let me add you as a collaborator | 13:55 |
| qinusty | Cheers | 13:55 |
| tristan | coldtom, so we can close 604 ! | 13:57 |
| tristan | I will close now | 13:57 |
| cs-shadow | qinusty: added you now, please try again :) | 13:57 |
| gitlab-br-bot | buildstream: issue #604 ("Cache not storing objects, but storing refs") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/604 | 13:57 |
| qinusty | Pushed :D cs-shadow, what's the review process with the docker images repo? I've only added a dependency so I'm sure it's fine :) | 13:58 |
| * paulsher1ood wonders when and why we started allowing application of .patch files in bst | 14:00 | |
| cs-shadow | qinusty: It's more free-form than the main repo since the number of changes were relatively small. For now, someone with maintainer rights should take a look in a reasonable time. | 14:01 |
| paulsher1ood | tristan: ^^ | 14:01 |
| cs-shadow | for this one, I've marked it to land when the CI passes | 14:01 |
| qinusty | sounds good, cheers cs-shadow | 14:02 |
| cs-shadow | qinusty: Once the MR lands and the CI passes on master, you'll need to submit another MR to buildstream repo to update the tags of the testsuite images in .gitlab-ci.yml | 14:02 |
| qinusty | Yup, I assume the tag will be generated once the CI passes? | 14:02 |
| * paulsher1ood is looking at things like https://gitlab.com/freedesktop-sdk/freedesktop-sdk/commit/31d4cd88b48d0e771f6e609b0906bf6250670312#8bdfc19d592cd0e84a58eea04244544c4e04dafe | 14:03 | |
| cs-shadow | qinusty: yeah, it's not exported anywhere so we need to look into the pipeline to find it. For example, https://gitlab.com/BuildStream/buildstream-docker-images/-/jobs/93014671 has something like "Pushing image buildstream/testsuite-fedora:28-master-116-5d508779" | 14:04 |
| tristan | paulsher1ood, since forever, there is nothing about BuildStream which forces you to work with or without patches; we don't dictate terms | 14:05 |
| paulsher1ood | let me try this in the other direction then... | 14:07 |
| paulsher1ood | valentind: why would we choose to apply a huge patch, rather than starting with a desired state/branch from git? | 14:07 |
| gitlab-br-bot | buildstream: issue #597 ("Git url without alias causes ValueError") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/597 | 14:08 |
| gitlab-br-bot | buildstream: merge request (Qinusty/597-backport-ValueError-fix->bst-1.2: Backport !700: Prevent ValueError on URLs without an alias) #751 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/751 | 14:09 |
| gitlab-br-bot | buildstream: merge request (valentindavid/roundtripping_only_when_modified->master: Disable round-tripping when element is not modified) #748 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/748 | 14:11 |
| valentind | paulsher1ood, this is not upstreamed. | 14:17 |
| valentind | It is based on the patch that Debian uses. But all distribution have their own patch to version symbols of openssl 1.0 | 14:17 |
| valentind | Other distributions also have clear patch files for what is not upstreamed. | 14:20 |
| gitlab-br-bot | buildstream: issue #612 ("Warn users when using a URL without an alias") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/612 | 14:22 |
| toscalix | tiagogomes: do you have what you need already? | 14:24 |
| tiagogomes | For the moment yes | 14:25 |
| gitlab-br-bot | buildstream: issue #613 ("Buildstream should exit early if an unknown alias is used within a URL") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/613 | 14:30 |
| qinusty | toscalix I have raised https://gitlab.com/BuildStream/buildstream/issues/612, https://gitlab.com/BuildStream/buildstream/issues/613 to follow up on issues I noticed during my work on https://gitlab.com/BuildStream/buildstream/issues/597 | 14:30 |
| toscalix | qinusty: thanks | 14:36 |
| *** noisecell has joined #buildstream | 14:39 | |
| qinusty | Okay tristan, I'm gonna do something I've wanted to do for a while which is throw together a page about buildstream UI | 14:43 |
| gitlab-br-bot | buildstream: merge request (valentindavid/roundtripping_only_when_modified->master: Disable round-tripping when element is not modified) #748 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/748 | 14:43 |
| qinusty | Mainly for my benefit, but it benefits the release too I guess. I don't know what everything even means | 14:44 |
| tristan | qinusty, what do you have in mind ? | 14:44 |
| qinusty | An image of a build in progress perhaps, with some of the UI elements highlighted and described in a sort of glossary below the image | 14:45 |
| toscalix | juergbi: thanks for paying attention as requested in the bug. Just saw it has been closed (won't fix). | 14:45 |
| qinusty | e.g. The Fetched: x y z at the bottom | 14:45 |
| qinusty | Built: x y z | 14:45 |
| tristan | qinusty, note that http://buildstream.gitlab.io/buildstream/using_commands.html is ready to add custom explanations to each command, that might help | 14:45 |
| tristan | qinusty, right, I guess that could be interesting, would be good to have it auto generated like the examples (but would be tricky to get a snapshot of the statusbar) | 14:47 |
| tristan | A glossary is a popular request, but I think that is separate | 14:47 |
| qinusty | Well, by glossary I mean | 14:47 |
| qinusty | If I label parts of the UI | 14:47 |
| qinusty | I'll describe them below | 14:47 |
| qinusty | by label | 14:47 |
| qinusty | In https://i.imgur.com/gjrXaOp.png, we can see a fairly straight forward view of the bst build output | 14:48 |
| tristan | qinusty, yeah; maintainability is a bit worrisome though, how do you regenerate the stuff when we tweak/improve the UI ? | 14:49 |
| qinusty | The tricky bits for me coming in, were understanding what the message output is [duration][hash key][job:element??] ACTION message? | 14:49 |
| tristan | I guess maybe that doesnt happen too often at this point at least | 14:49 |
| qinusty | Well I imagine we would tweak and not exactly change massively | 14:49 |
| tristan | indeed; but it means any tweak has a consequence in docs now | 14:50 |
| qinusty | But I do think regardless of what we do going forward, we want something out there for release which just gives people a glimpse before they install/run buildstream | 14:50 |
| tristan | qinusty, I think it's just something to keep in mind, maybe an autogenerated output with some text below explaining the above, so that text is the only part which needs updating | 14:50 |
| tristan | rather than having to open up photoshop | 14:51 |
| qinusty | I agree, But we can still label the individual elements, Even if we describe them differently | 14:51 |
| toscalix | qinusty: I like it | 14:54 |
| qinusty | And maybe a slightly contentious point but... Perhaps we want an image of buildstream in some shape or form in our README. | 14:54 |
| toscalix | qinusty: definetly on the website | 14:54 |
| toscalix | showing key features | 14:55 |
| toscalix | or something that clearly differentiate us from Yocto | 14:55 |
| jjardon | qinusty: are you sorted? | 14:56 |
| qinusty | To me, when I open a project. I feel slightly more comfortable after seeing what I'm getting before I install. The information in the text is obviously the deciding factor. But if Buildstream is shared around on release and people open the page for a quick glimpse | 14:56 |
| toscalix | an output with colours, a self explanatory command.... | 14:56 |
| qinusty | Yeah I'm good jjardon :D | 14:56 |
| toscalix | qinusty: you are right | 14:56 |
| toscalix | we have a page in the web called feature page | 14:57 |
| toscalix | explaining what we are releasing | 14:57 |
| toscalix | that page should have a couple of those | 14:57 |
| jjardon | qinusty: so about the logs of the cache server in gnome-build-meta; what do you need exactly? | 14:58 |
| * qinusty isn't certain they're logged to file | 14:59 | |
| toscalix | and by adding them to than page, which is release version specific, the maintenance point brought up by tristan does not apply, which was an important one | 14:59 |
| toscalix | so qinustythink about them and get a couple, with a one or two lines explanation of what we are seing, and we add them to the feature page | 14:59 |
| qinusty | https://gitlab.com/BuildStream/buildstream/blob/master/buildstream/_artifactcache/casserver.py#L368 logs messages regarding the outcome of the clean up code. This is what I'm after jjardon. | 15:01 |
| jjardon | qinusty: abderrahim[m] is waiting to purge the server so we can keep building; any idea if something useful would be in the journald logs? | 15:02 |
| * jjardon checks | 15:02 | |
| jjardon | qinusty: would that be logged to the journal? | 15:02 |
| qinusty | I'm not sure, I can't even see where the logging is initialised within the project. | 15:03 |
| jjardon | juergbi: do you know? ^ | 15:04 |
| juergbi | jjardon: when run via systemd it might indeed end up in the journal, however, I haven't verified this | 15:05 |
| jjardon | abderrahim[m]: do you remember the IP of the cache server? I will try to take a look around | 15:06 |
| jjardon | ah wait I can look for it in our CI | 15:06 |
| qinusty | https://stackoverflow.com/questions/34588421/how-to-log-to-journald-systemd-via-python By the looks of it, you need to configure journald logging | 15:06 |
| *** tristan has quit IRC | 15:06 | |
| qinusty | I'd like to recreate this issue anyway but had trouble creating a docker contain with a restricted storage size | 15:06 |
| qinusty | s/contain/container | 15:07 |
| *** leopi has quit IRC | 15:08 | |
| *** leopi has joined #buildstream | 15:08 | |
| *** tristan has joined #buildstream | 15:15 | |
| *** ChanServ sets mode: +o tristan | 15:15 | |
| jjardon | qinusty: something is logged, but none of the messages you linked | 15:15 |
| jjardon | https://www.irccloud.com/pastebin/vDGXnfqo/ | 15:16 |
| jjardon | this is the backtrace of the first time it failed: | 15:17 |
| jjardon | https://www.irccloud.com/pastebin/ZTYRLUkm/ | 15:17 |
| gitlab-br-bot | buildstream: merge request (coldtom/strip-rules->master: Upstream freedesktop-sdk strip rules) #750 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/750 | 15:20 |
| qinusty | cheers jjardon, guess we'll have to try and reproduce this issue | 15:29 |
| jjardon | abderrahim[m]: feel free to empty the server, please: I think stop cas-pullpush.service and remove /home/artifacts/artifacts/* would do it? | 15:31 |
| gitlab-br-bot | buildstream: merge request (Qinusty/600-recursive-variables->master: Add cyclic check within variable resolution) #712 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/712 | 15:36 |
| *** jjardon sets mode: -M | 15:39 | |
| *** xjuan has joined #buildstream | 15:44 | |
| gitlab-br-bot | buildstream: merge request (bschubert/log-missed-cache->master: Log not-found objects in the cache as SKIPPED) #729 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/729 | 15:51 |
| gitlab-br-bot | buildstream: merge request (bschubert/log-missed-cache->master: Log not-found objects in the cache as SKIPPED) #729 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/729 | 15:51 |
| *** dtf has joined #buildstream | 15:55 | |
| qinusty | benschubert, reviewed your MR for logging skipped PULL jobs. If we can get that patched and backported before Friday that'd be good :D | 15:55 |
| benschubert | qinusty: Sure, let me split the commit and push again! | 15:56 |
| qinusty | No worries :D Remember to rebase to origin/master | 15:56 |
| gitlab-br-bot | buildstream: merge request (bschubert/log-missed-cache->master: Log not-found objects in the cache as SKIPPED) #729 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/729 | 16:01 |
| benschubert | qinusty: here you go! Is that good for you? | 16:01 |
| qinusty | Yeah looks good :) Are you able to merge on pipeline success? | 16:02 |
| benschubert | No, I don't have that level of permission :( | 16:03 |
| qinusty | :( maybe tristan or jjardon can add you to the list since this is your second patch going into Builstream. | 16:04 |
| qinusty | I've flagged it to merge though | 16:05 |
| benschubert | thanks! | 16:05 |
| gitlab-br-bot | buildstream: merge request (Qinusty/600-recursive-variables->master: Add cyclic check within variable resolution) #712 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/712 | 16:13 |
| gitlab-br-bot | buildstream: merge request (Qinusty/600-recursive-variables->master: Add cyclic check within variable resolution) #712 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/712 | 16:13 |
| gitlab-br-bot | buildstream: merge request (bschubert/log-missed-cache->master: Log not-found objects in the cache as SKIPPED) #729 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/729 | 16:25 |
| *** leopi has quit IRC | 16:29 | |
| gitlab-br-bot | buildstream: merge request (Qinusty/600-recursive-variables->master: Add cyclic check within variable resolution) #712 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/712 | 16:34 |
| *** leopi has joined #buildstream | 16:37 | |
| benschubert | qinusty: should I open a cherry-pick MR for this to bst-1.2? Is that what you were looking for? | 16:43 |
| qinusty | When I backport 1-2 commit patches, I usually just branch from origin/bst-1.2 and cherry pick the commits | 16:44 |
| qinusty | Then target your merge request onto bst-1.2 | 16:44 |
| benschubert | sure, 1.2 was the branch you wanted the backport in right? | 16:45 |
| qinusty | Yup, backport it to 1.2. I'll get tristan to have a final say, but I'm pretty sure it's good to go in there. It's just extra information for the user | 16:46 |
| qinusty | s/I'll/I'd | 16:46 |
| qinusty | Does anyone know where the entry point for bst-artifact-server is? | 16:53 |
| * qinusty actually looks and finds it *facepalm* | 16:54 | |
| toscalix | website repo moved. I explained the reasons in the mailing list | 16:55 |
| toscalix | good progress today on this front | 16:55 |
| *** toscalix has quit IRC | 16:57 | |
| gitlab-br-bot | buildstream: merge request (Qinusty/600-recursive-variables->master: Add cyclic check within variable resolution) #712 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/712 | 16:58 |
| gitlab-br-bot | buildstream: merge request (bschubert/backport-log-missed-cache->bst-1.2: Bschubert/backport log missed cache) #752 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/752 | 17:00 |
| gitlab-br-bot | buildstream: issue #614 ("Follow-up from "Bschubert/backport log missed cache"") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/614 | 17:22 |
| gitlab-br-bot | buildstream: merge request (bschubert/backport-log-missed-cache->bst-1.2: Bschubert/backport log missed cache) #752 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/752 | 17:24 |
| *** jonathanmaw has quit IRC | 17:45 | |
| *** leopi has quit IRC | 18:54 | |
| *** xjuan has quit IRC | 20:32 | |
| gitlab-br-bot | buildstream: issue #615 ("bst build seems to be stuck forever, not downloading anything") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/615 | 20:43 |
| *** alatiera_ has quit IRC | 21:32 | |
| *** tristan has quit IRC | 21:50 | |
| bochecha | bst 1.1.7 is so much faster to start! big thanks to whoever made that happen :) | 22:49 |
| *** rdale has quit IRC | 23:07 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
