| *** Prince781 has quit IRC | 00:43 | |
| *** mcatanzaro has joined #buildstream | 01:23 | |
| *** ruben has quit IRC | 01:47 | |
| *** classtime has joined #buildstream | 01:47 | |
| classtime | СНRОПО IS TEACHING A CLASS @ iяс.sцреяиетs.ояg сни sцреявоwl | 01:48 |
|---|---|---|
| classtime | mcatanzaro tristan nexus rafaelff[m] | 01:48 |
| *** classtime has left #buildstream | 01:48 | |
| *** shitface has joined #buildstream | 03:00 | |
| *** shitface has quit IRC | 03:03 | |
| *** mcatanzaro has quit IRC | 06:20 | |
| gitlab-br-bot | buildstream: issue #240 ("Make a release of BuildStream") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/240 | 06:27 |
| *** Prince781 has joined #buildstream | 07:44 | |
| *** Prince781 has quit IRC | 07:49 | |
| *** Prince781 has joined #buildstream | 07:52 | |
| *** jonathanmaw has joined #buildstream | 09:39 | |
| *** toscalix has joined #buildstream | 09:40 | |
| *** tristan has quit IRC | 09:40 | |
| *** jonathanmaw has quit IRC | 09:51 | |
| *** jonathanmaw has joined #buildstream | 09:51 | |
| *** ssam2 has joined #buildstream | 09:54 | |
| *** aday has joined #buildstream | 10:28 | |
| *** tristan has joined #buildstream | 10:31 | |
| *** toscalix_ has joined #buildstream | 11:21 | |
| *** toscalix has quit IRC | 11:22 | |
| tlater | jonathanmaw: Could you have another look at the docker image MR? https://gitlab.com/BuildStream/bst-external/merge_requests/9 | 11:58 |
| * tlater fixed the issue you were having | 11:58 | |
| jonathanmaw | tlater: cool | 11:58 |
| *** zalupik has quit IRC | 12:17 | |
| tlater | Anyone mind if I create a buildstream-examples repo? jmac, I think you were interested in seeing one of these as well. | 12:32 |
| jmac | I am, please create it | 12:32 |
| * tlater notices that the bst-external repo doesn't have a license | 12:45 | |
| ssam2 | oops | 12:45 |
| ssam2 | I'm sure it should be the same as buildstream.git | 12:45 |
| tlater | I'll license both that and the buildstream-examples as lgpl v2.1 | 12:46 |
| jmac | I now seem to be in a rabbit hole, finding dependencies for a python module, making .bst files for those, then repeating with the dependencies for those. Is there an easier way to do this? | 12:48 |
| tlater | (There's no way this affects anything, I'll just push to master) | 12:48 |
| tlater | Hrm, I guess we could have a nice pip-to-bst conversion tool that throws out a bunch of .bst files | 12:48 |
| * tlater doesn't think there's a way currently, jmac | 12:48 | |
| jmac | I'm thinking of making such a tool just for this task | 12:49 |
| ssam2 | there's no such tool for buildstream yet | 12:49 |
| ssam2 | https://gitlab.com/baserock/import has a way of doing this | 12:49 |
| ssam2 | https://gitlab.com/baserock/import/blob/master/baserockimport/exts/python.to_lorry and https://gitlab.com/baserock/import/blob/master/baserockimport/exts/python.find_deps are the relevant bits | 12:50 |
| ssam2 | the first one finds the source code, the second one finds the dependencies | 12:51 |
| persia | Seems like deriving a tool from that for buildstream-extras would likely make folk happy | 12:51 |
| * tlater should think about doing this for npm as well, given his delving into that over Christmas | 12:52 | |
| persia | I believe baserock has examples for npm and gems, although both might have aged and not match current upstream practice | 12:52 |
| ssam2 | yes, all in here: https://gitlab.com/baserock/import/tree/master/baserockimport/exts | 12:54 |
| juergbi | tlater: for examples we actually might want a very permissive license, if we want people to use the examples as starting point for their own projects | 12:59 |
| tlater | juergbi: I'm not sure what license would be appropriate then, MIT? | 13:00 |
| * tlater knows far too little about licenses | 13:00 | |
| tlater | Frankly, it might be best to leave licensing up to individual projects in here | 13:01 |
| tlater | But I'm not sure how to do that - careful review? | 13:01 |
| juergbi | MIT would be fine with me. don't have strong opinions for this particular case | 13:01 |
| persia | That makes sense, although a first pass might be very broad (something like MIT or ISC) | 13:01 |
| persia | tlater: Put a LICENSES file in root, explaiing that each subdirectory has it's own license. | 13:02 |
| persia | Then make sure that happens. | 13:02 |
| tlater | persia: ta, seems most sensible - would the root directory need a separate license? | 13:02 |
| gitlab-br-bot | buildstream: merge request (juerg/recursive-pipelines->master: Junctions and subprojects) #160 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/160 | 13:06 |
| juergbi | jonathanmaw: can you please review this commit? https://gitlab.com/BuildStream/buildstream/merge_requests/160/diffs?commit_id=a83c94b74e0f65aa3cee328bc37be57b9ae7a5c3 | 13:06 |
| jonathanmaw | juergbi: will do after I've reviewed ssam2/tlater's docker source MR | 13:07 |
| juergbi | ta | 13:07 |
| persia | tlater: No, because it wouldn't contain anything other than directories and the instructions saying where to find licenses: nothing that needs licensing. | 13:21 |
| *** adds68 has quit IRC | 13:22 | |
| * tlater created the examples repository, will wait for his other branches to merge before he adds the first example | 13:22 | |
| *** adds68 has joined #buildstream | 13:22 | |
| tlater | jmac: If you had an example project you wanted to push, this is the repo: https://gitlab.com/BuildStream/buildstream-examples | 13:22 |
| jmac | OK. I don't have any examples right now but might be able to make something | 13:23 |
| * tlater thinks jonathanmaw may have had one for dpkg stuff, but isn't sure anymore | 13:24 | |
| ssam2 | we talked about putting most of the benchmark script in a separate repo | 13:57 |
| ssam2 | so, shall I create https://gitlab.com/BuildStream/benchmarks ? | 13:57 |
| * ssam2 will do so after lunch if nobody objects | 13:58 | |
| *** toscalix_ is now known as toscalix | 14:11 | |
| * jonathanmaw finishes reviewing the docker merge request | 14:18 | |
| jonathanmaw | tlater, ssam2 ^ | 14:18 |
| tlater | ta jonathanmaw | 14:18 |
| * tlater saw a few of the comments already | 14:18 | |
| *** adds68 has quit IRC | 14:40 | |
| *** adds68 has joined #buildstream | 14:40 | |
| jonathanmaw | juergbi: reviewed that commit | 14:41 |
| juergbi | ta | 14:42 |
| tlater | Heh, not a single one of our files currently passes pylint's code quality checks | 14:49 |
| persia | Fixing that might be a reasonable thing for someone to do whilst learning the source code, if anyone is starting fresh. | 14:50 |
| tlater | persia: I don't think that's possible. We simply don't adhere to normal python standards | 14:51 |
| persia | On purpose? Or because we didn't try? | 14:51 |
| * tlater thinks we didn't try | 14:52 | |
| tlater | But at this point it would be a massive refactor, including rewriting some of our class infrastructure | 14:52 |
| tlater | It would be a lot easier to just change pylint's rules to adhere to our standards | 14:52 |
| tlater | Maybe start enforcing some of the easier stuff (like ordering of imports) | 14:53 |
| juergbi | i guess we could enable additional checks one by one when we want to? | 14:55 |
| * tlater will look at disabling most things that seem hard to fix and be happy to pass on refactoring everything else, after that we can do what juergbi suggests | 14:57 | |
| persia | For the OpenStack project, when they decided they were large enough to need coding standards, they set up a test framework that let them opt out of things, and then each repo opted in when they had time. There was some social pressure on repos with lots of opt-outs, but not so much everyone had to rewrite everything from scratch. | 14:57 |
| tlater | Do we have a line-length limit? I think I recall the Hacking.rst file saying it was 121, but it doesn't specify a number anymore | 15:01 |
| jmac | ISTR hacking.rst saying we ignore the line length bits of PEP8 | 15:02 |
| ssam2 | from setup.cfg | 15:02 |
| ssam2 | pep8maxlinelength = 119 | 15:02 |
| *** mcatanzaro has joined #buildstream | 15:02 | |
| tlater | ta ssam2 | 15:02 |
| ssam2 | tlater, which of us should do the fixups on bst-external!9 | 15:10 |
| ssam2 | ? | 15:10 |
| tlater | Ah, sorry ssam2 | 15:10 |
| tlater | I could probably do a few of them, but I lack context for others | 15:10 |
| tlater | It might be easier if you did them | 15:10 |
| ssam2 | ok | 15:10 |
| * ssam2 creates gitlab.com/buildstream/benchmarks | 15:27 | |
| tlater | So, looking at slightly more refined pylint output, we have a lot of issues such as too-many-x and too-few-x (where x is variables, arguments, methods, ...) | 15:31 |
| tlater | Those would be *nice* to sort out, since they decrease perceived complexity - but I doubt this is work we'd get to. | 15:32 |
| ssam2 | i don't find those warnings very helpful to be honest | 15:32 |
| ssam2 | you can't automate taste :-) | 15:32 |
| persia | +1 on not automating taste. | 15:32 |
| persia | That said, if one has very complex objects, it is sometimes useful to reconsider them to see if they sensibly decompose. | 15:33 |
| * tlater suspects that having these warnings enabled in pylint would help in pointing it out | 15:35 | |
| tlater | But I think our codebase is just too large to do this at this point | 15:35 |
| gitlab-br-bot | buildstream: merge request (146-use-minimum-python-version-3-4-for-integration-tests->master: Resolve "Use minimum python version (3.4) for integration tests") #266 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/266 | 15:41 |
| gitlab-br-bot | buildstream: merge request (juerg/recursive-pipelines->master: Junctions and subprojects) #160 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/160 | 15:41 |
| *** ruben has joined #buildstream | 15:57 | |
| gitlab-br-bot | buildstream: merge request (juerg/recursive-pipelines->master: Junctions and subprojects) #160 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/160 | 16:16 |
| juergbi | tlater: !256 i assume this is leftover from debugging? | 16:24 |
| juergbi | # pprint(sys.modules) | 16:24 |
| * tlater thought he got rid of that | 16:25 | |
| tlater | But yes | 16:25 |
| juergbi | i've rebased the branch locally and removed that | 16:25 |
| tlater | Also gitlab appears to be down | 16:25 |
| juergbi | ok, will update the branch | 16:25 |
| juergbi | oh? | 16:25 |
| tlater | tyvm juergbi | 16:25 |
| tlater | I get a 500 | 16:25 |
| juergbi | here as well now | 16:25 |
| *** tristan has quit IRC | 16:26 | |
| tlater | Hm, do we want to follow pep8 import order? | 16:29 |
| juergbi | tlater: shall i also add a Fixes #171? | 16:29 |
| juergbi | tlater: that would be alphabetical? not sure how this would look like with from imports | 16:29 |
| tlater | It also groups by internal-external and stdlib imports | 16:30 |
| juergbi | we tend to currently group internal vs external imports | 16:30 |
| tlater | Yeah, but we don't do so consistently | 16:30 |
| juergbi | sounds sensible to me but i don't know the details | 16:30 |
| * tlater enables it for now and will show the fallout later | 16:30 | |
| tlater | Ah, yeah, I think #171 can be closed with this | 16:30 |
| tlater | Hm | 16:30 |
| juergbi | ok, i'll add it to the commit message as reference | 16:30 |
| juergbi | or not? | 16:30 |
| tlater | it's still slow | 16:31 |
| juergbi | ah, gitlab | 16:31 |
| * tlater just doubts we'll be able to improve it much | 16:31 | |
| juergbi | or completions? | 16:31 |
| tlater | No, the bash completions | 16:31 |
| tlater | Maybe leave the issue open, someone might come with a bright idea | 16:31 |
| juergbi | ok | 16:31 |
| gitlab-br-bot | buildstream: merge request (completion-optimizations->master: Completion optimizations) #256 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/256 | 16:32 |
| juergbi | jonathanmaw: !259 should no longer be marked WIP, right? | 16:36 |
| * paulsherwood wonders if there's any way at all to get a link to tristan's talk today | 16:39 | |
| gitlab-br-bot | buildstream: issue #242 ("Allow building as non-root inside the sandbox") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/242 | 16:41 |
| persia | paulsherwood: Not that I see yet, other than pestering volunteers (which often results in the opposite of one's desire). I see both the immediately preceeding and immediately following talks posted though, so they might be getting through the day. | 16:42 |
| juergbi | state is still waiting for reviewer but so are a lot of others: https://review.video.fosdem.org/overview | 16:43 |
| gitlab-br-bot | buildstream: merge request (completion-optimizations->master: Completion optimizations) #256 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/256 | 16:44 |
| jonathanmaw | juergbi: that's right | 16:44 |
| * jonathanmaw un-WIPs it | 16:45 | |
| gitlab-br-bot | buildstream: merge request (212-git-source-needs-a-way-to-disable-checking-out-submodules->master: Resolve "Git source needs a way to disable checking out submodules") #259 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/259 | 16:45 |
| ssam2 | could someone approve https://gitlab.com/BuildStream/buildstream-docker-images/merge_requests/19/diffs ? | 16:47 |
| ssam2 | i've just spent 10 minutes trying to figure out why it wasn't in builds, but turned out I never clicked the 'merge' button | 16:47 |
| ssam2 | and now I had to rebase, which means it needs approval all over again | 16:47 |
| juergbi | done | 16:48 |
| ssam2 | thanks | 16:48 |
| tlater | ta for merging 256 juergbi :) | 16:51 |
| juergbi | yw | 16:52 |
| gitlab-br-bot | buildstream: merge request (completion-optimizations->master: Completion optimizations) #256 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/256 | 17:07 |
| gitlab-br-bot | buildstream: merge request (juerg/dbus->master: WIP: Inherit user id and group id for bst shell) #265 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/265 | 17:10 |
| gitlab-br-bot | buildstream: merge request (sam/document-profiling-and-benchmarks->master: WIP: HACKING.rst: Mention benchmarking and profiling tools) #267 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/267 | 17:15 |
| *** jonathanmaw has quit IRC | 17:15 | |
| gitlab-br-bot | buildstream: merge request (juerg/dbus->master: WIP: Inherit user id and group id for bst shell) #265 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/265 | 17:15 |
| gitlab-br-bot | buildstream: issue #243 ("dpkg import source plugin") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/243 | 17:19 |
| *** ruben has quit IRC | 17:20 | |
| * tlater wasn't aware how frequently we use `global` | 17:20 | |
| tlater | I'm already finding small bugs in error messages and such though :) | 17:22 |
| *** tristan has joined #buildstream | 17:24 | |
| * tlater wonders if pep8 testing should be a separate task, considering we're currently duplicating the work for the linux and unix platforms | 17:29 | |
| *** aday has quit IRC | 17:30 | |
| *** aday has joined #buildstream | 17:31 | |
| persia | tlater: Yes, I think code quality tests should probably only be run on one (selected) platform. Running the same series of code analysis and static analysis tooling on multiple platforms provides little developer benefit. | 17:31 |
| juergbi | it doesn't take that much time, though. not sure whether spinning up another worker is worth it | 17:33 |
| tlater | Hrm, that would be a separate issue, but I think I'll just add it to this MR | 17:33 |
| juergbi | my comment was based on current pep8 time, with pylint it may be significant, no idea | 17:34 |
| tlater | juergbi: Ah, sorry, skipped that line somehow | 17:35 |
| tlater | With pylint it does become more significant | 17:35 |
| tlater | I also think with more platforms it's a waste of resources, tests are slow enough as it is | 17:35 |
| persia | In that case, if we're considering the feedback loop from push to CI results, it may be interesting to split. | 17:36 |
| tlater | I'll draft it out as part of this MR, see what people say at merge time - tristan will probably want to be involved in this discussion eventually | 17:37 |
| jmac | Is sdk.gnome.org down? | 17:54 |
| persia | jmac: Does not respond to ICMP or HTTP for me. | 17:55 |
| jmac | Boo. | 17:56 |
| jmac | I'll do something else then: Do we have any elements which create bootable images as artifacts? | 17:57 |
| * tlater has one | 17:59 | |
| persia | freedesktop-sdk merged a change from ssam2 today, including sdk/elements/vm/desktop-vm-image-x86_64.bst | 17:59 |
| tlater | About to add it to the buildstream-examples repo, just waiting for ssam2 to finish merging his docker plugin | 17:59 |
| tlater | jmac: This is the most barebones image we could come up with: https://gitlab.com/BuildStream/buildstream-examples/tree/x86image | 18:00 |
| jmac | Great, just what I was after! | 18:01 |
| *** ssam2 has quit IRC | 18:03 | |
| juergbi | persia: i don't think the feedback loop is a big concern, developers are expected to run pytest locally before pushing | 18:23 |
| * tlater would also add a recommendation to run pylint in your IDE to the hacking file | 18:24 | |
| tlater | juergbi: Linting currently takes about 51 seconds, so it's probably not necessary to split. | 18:25 |
| tlater | I think the output is neater if it isn't interwoven with everything else, though. | 18:25 |
| tlater | Then again, hopefully future branches won't have quite as much to nag about | 18:25 |
| persia | Adding linters does tend to reduce nagging over time | 18:26 |
| *** toscalix has quit IRC | 18:37 | |
| *** valentind has joined #buildstream | 19:41 | |
| *** aday has quit IRC | 19:44 | |
| *** bethw has joined #buildstream | 21:38 | |
| *** valentind has quit IRC | 21:51 | |
| *** ruben has joined #buildstream | 23:02 | |
| *** kektrain has joined #buildstream | 23:08 | |
| *** kektrain has quit IRC | 23:10 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!