| *** tristan has quit IRC | 05:32 | |
| *** tristan has joined #buildstream | 05:59 | |
| gitlab-br-bot | push on buildstream@sam/compose-list-exception (by Tristan Van Berkom): 11 commits (last: _pipeline.py: Make fetching of remote refs optional) https://gitlab.com/BuildStream/buildstream/commit/5c2ef6d076921bc0121e61efaa7a719c34ea1912 | 06:59 |
|---|---|---|
| gitlab-br-bot | buildstream: merge request (sam/compose-list-exception->master: Catch attempts to compose a list) #129 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/129 | 06:59 |
| gitlab-br-bot | buildstream: issue #142 ("Potentially printing provenance more than once in LoadErrors") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/142 | 07:07 |
| *** semanticdesign has joined #buildstream | 08:12 | |
| *** valentind has joined #buildstream | 08:36 | |
| gitlab-br-bot | buildstream: merge request (sam/compose-list-exception->master: Catch attempts to compose a list) #129 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/129 | 08:42 |
| gitlab-br-bot | push on buildstream@master (by Tristan Van Berkom): 1 commit (last: Catch attempts to compose a list) https://gitlab.com/BuildStream/buildstream/commit/8c4d8ca2b7c886690fb933428c6768c745667db7 | 08:42 |
| gitlab-br-bot | buildstream: Tristan Van Berkom deleted branch sam/compose-list-exception | 08:42 |
| gitlab-br-bot | push on buildstream@options (by Tristan Van Berkom): 9 commits (last: tests/frontend/track.py: Added tests for recursive tracking and --except) https://gitlab.com/BuildStream/buildstream/commit/e472f5abd7f9aeff994a14e581c7a89199bd255f | 08:44 |
| gitlab-br-bot | buildstream: merge request (options->master: doc: Explain how to configure project options) #130 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/130 | 08:44 |
| *** givascu has joined #buildstream | 08:54 | |
| *** tlater has joined #buildstream | 09:24 | |
| gitlab-br-bot | push on buildstream@user-namespace-fallback (by Tristan Van Berkom): 4 commits (last: _frontend/main.py: Hold on to unconditional startup messages) https://gitlab.com/BuildStream/buildstream/commit/fc5213d03c012a5ec2213214d36f0097ede3f791 | 09:58 |
| gitlab-br-bot | buildstream: merge request (user-namespace-fallback->master: User namespace fallback) #132 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/132 | 09:59 |
| tristan | jjardon[m], ping | 09:59 |
| tristan | jjardon[m], can you please try building GNOME with user-namespace-fallback branch ? i.e. the above MR: https://gitlab.com/BuildStream/buildstream/merge_requests/132 | 10:00 |
| tristan | jjardon[m], if that passes integration tests on not arch linux, and it "worksforyou" on Arch linux, I'll merge | 10:01 |
| *** ssam2 has joined #buildstream | 10:05 | |
| gitlab-br-bot | push on buildstream@master (by Tristan Van Berkom): 2 commits (last: _frontend/main.py: Hold on to unconditional startup messages) https://gitlab.com/BuildStream/buildstream/commit/e8309af42d6e9ce554961503638b4925c1a8a53d | 10:06 |
| gitlab-br-bot | buildstream: Tristan Van Berkom deleted branch options | 10:06 |
| gitlab-br-bot | buildstream: merge request (options->master: doc: Explain how to configure project options) #130 changed state ("closed"): https://gitlab.com/BuildStream/buildstream/merge_requests/130 | 10:06 |
| gitlab-br-bot | push on buildstream@user-namespace-fallback (by Tristan Van Berkom): 6 commits (last: Catch attempts to compose a list) https://gitlab.com/BuildStream/buildstream/commit/8c4d8ca2b7c886690fb933428c6768c745667db7 | 10:07 |
| gitlab-br-bot | buildstream: merge request (user-namespace-fallback->master: User namespace fallback) #132 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/132 | 10:08 |
| jjardon[m] | tristan: https://paste.gnome.org/pxnkzmsah how can I force a rebuild in bst? | 10:20 |
| ssam2 | delete your artifact cache ? | 10:24 |
| *** valentind has quit IRC | 10:25 | |
| gitlab-br-bot | push on buildstream@tracking-changes (by Tristan Maat): 5 commits (last: WIP: Add support for multiple arguments (*very* WIP)) https://gitlab.com/BuildStream/buildstream/commit/65dbf24543c4eafdb066558e0560f9b433fba91f | 10:28 |
| gitlab-br-bot | buildstream: merge request (tracking-changes->master: WIP: Tracking changes) #119 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/119 | 10:28 |
| tristan | Ummm good question | 10:38 |
| tristan | jjardon[m], maybe build something a bit higher up ? | 10:38 |
| tristan | like core-deps/glib.bst ? | 10:38 |
| tristan | Because... downloading and constructing the base system is... ugh, long time | 10:38 |
| tristan | jjardon[m], another approach might be to first run `bst track base/ninja.bst` and rerun that with a newer ninja | 10:39 |
| tristan | then you only rebuild ninja | 10:39 |
| jjardon[m] | tristan: mmm, not sure this is what we expected: https://paste.gnome.org/p5u3xdc6a | 10:45 |
| tristan | jjardon[m], tracking didnt pull a new version of ninja | 10:48 |
| tristan | :-/ | 10:48 |
| tristan | It looks like stderr is not caught by my simple minded call to subprocess.check_output(), and instead it gets rudely sent to the terminal | 10:48 |
| tristan | But, that's not a huge deal I think | 10:48 |
| tristan | jjardon[m], just try `bst track core-deps/glib.bst` and try building that, certainly that will have a new commit to build | 10:49 |
| tristan | or maybe gettext | 10:49 |
| tristan | those are pretty low level in the stack | 10:49 |
| jjardon[m] | tristan: same happened for glib, let me try gettext | 10:50 |
| jjardon[m] | same for gettext | 10:50 |
| tristan | jjardon[m], no new versions ? | 10:51 |
| tristan | jjardon[m], I wonder, how did you get all of those built in the first place ? | 10:51 |
| tristan | here, I can produce some mumbo jumbo ostree command for you... | 10:52 |
| tristan | one sec | 10:52 |
| jjardon[m] | tristan: I followed the steps on the wiki | 10:52 |
| tristan | jjardon[m], yeah, but you have Arch, which means... you couldnt have built it :D | 10:53 |
| tristan | jjardon[m], you did it with commented lines in _sandboxbwrap.py I guess ? | 10:54 |
| tristan | jjardon[m], here: ostree refs --repo ~/.cache/buildstream/artifacts/ostree --delete `ostree refs --repo ~/.cache/buildstream/artifacts/ostree | grep gnome/core-deps-gettext` | 10:54 |
| tristan | jjardon[m], that'll delete your local cache of gettext | 10:54 |
| jjardon[m] | tristan: yes, I commented the s.py line | 10:59 |
| tristan | :) | 10:59 |
| tristan | jjardon[m], so this is most probably going to work, at least we've seen in your output that we're detecting the incapability to do user namespaces properly there | 11:00 |
| jjardon[m] | tristan: same thing after trying that command :/ | 11:00 |
| tristan | what ?! | 11:00 |
| tristan | jjardon[m], the ostree thing didnt delete your artifact ? | 11:00 |
| tristan | what happened, is it on an artifact share ? | 11:00 |
| tristan | did it download the artifact instead of building it ? | 11:00 |
| tristan | did you run `bst build core-deps/gettext.bst` after that ? | 11:01 |
| tristan | jjardon[m], gettext.bst is not a part of base.bst, so; you'd have to try to build it explicitly after deleting it | 11:01 |
| jjardon[m] | I did bst build base.bst, let me try that | 11:01 |
| jjardon[m] | ah! | 11:01 |
| tristan | unfortunately, pipelines are damn slow on gitlab today | 11:02 |
| jjardon[m] | tristan: something interesting: https://paste.gnome.org/pkejdhda7 | 11:02 |
| tristan | hrmmmm | 11:03 |
| tristan | jjardon[m], mkayyyyy... | 11:03 |
| * tristan looks into that | 11:04 | |
| tristan | ohhhhh | 11:05 |
| tristan | tlater, looks like your Platform.get_platform() is actually instantiating a new platform in the child tasks, is that intentional for any reason ? | 11:06 |
| tlater | tristan: Where does it do that? It's possible that that's used only to access a specific function, probably a bit heavy handed if so. | 11:07 |
| tristan | tlater, in jjardon[m]'s log: https://paste.gnome.org/pkejdhda7#line-62 | 11:10 |
| tristan | line 62 | 11:10 |
| tristan | indicates that in the child task, we're instantiating a platform... via Element.__sandbox() | 11:10 |
| * tristan also found another bug thanks to that stack trace | 11:10 | |
| tristan | looks like at times we can be logging integration commands to the wrong log file | 11:11 |
| * tristan quickly fixes this, which should fix jjardon[m]'s issue | 11:11 | |
| tlater | Hmm... Yeah, it's definitely not supposed to do that. | 11:11 |
| tlater | I think for a while I had the platform as a singleton | 11:11 |
| tristan | tlater, if you could keep your eye on that and fix it, I'd really like that | 11:12 |
| tristan | I bet it's an easy fix | 11:12 |
| tristan | maybe you just need a toplevel variable around in platform.py, and refer to it as `global` from within that function | 11:12 |
| * tlater will, do you prefer making it a singleton or trying to fiddle the pipeline into places it probably doesn't belong? | 11:12 | |
| tristan | get_platform() | 11:12 |
| tristan | I dont know what you mean by fiddling the pipeline, ensuring it's singleton is better I think | 11:13 |
| tristan | I guess you might mean, passing it through a whole lot of internal apis | 11:13 |
| tlater | Yeah, I think I prefer that too. Technically the pipeline currently owns the platform | 11:13 |
| tristan | yeah dont do that | 11:13 |
| tristan | just singleton, maybe you are just missing global | 11:13 |
| tristan | tlater, note that a variable assignment in python triggers a new variable to be declared | 11:13 |
| tristan | which will not have any warning, but shadow existing variables | 11:14 |
| tristan | which is, totally horrible | 11:14 |
| tristan | and is a probable cause of a bug like this :D | 11:14 |
| * tlater is aware, my linter warns me :) | 11:14 | |
| tristan | Hmmm, we should add another linter to our test suite too | 11:14 |
| tristan | anyway | 11:14 |
| gitlab-br-bot | push on buildstream@master (by Tristan Van Berkom): 1 commit (last: _scheduler/job.py: Ensure messages are logged to the correct log file) https://gitlab.com/BuildStream/buildstream/commit/40f5fe7789d4085183717cfc158297959d8b380b | 11:16 |
| gitlab-br-bot | push on buildstream@user-namespace-fallback (by Tristan Van Berkom): 4 commits (last: _scheduler/job.py: Ensure messages are logged to the correct log file) https://gitlab.com/BuildStream/buildstream/commit/40f5fe7789d4085183717cfc158297959d8b380b | 11:16 |
| tristan | jjardon[m], wanna try it one more time with a fresh `user-namespace-fallback` branch ? | 11:17 |
| jjardon[m] | sure, give me a second | 11:17 |
| jjardon[m] | tristan: mmm, seems its downloading the artifact: https://paste.gnome.org/ptnfbvsx9 Is there any way to disable any network connection? (like -n in jhbuild) | 11:20 |
| gitlab-br-bot | buildstream: merge request (user-namespace-fallback->master: User namespace fallback) #132 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/132 | 11:21 |
| tristan | jjardon[m], ummmm, modify project.conf :) | 11:22 |
| tristan | jjardon[m], comment out the artifact share | 11:22 |
| tristan | jjardon[m], and of course, rerun that ostree mumbo jumbo | 11:23 |
| jjardon[m] | tristan: Its alive! https://paste.gnome.org/pzpj2xtzp | 11:28 |
| tristan | w00t \o/ | 11:29 |
| tristan | jjardon[m], great ! | 11:29 |
| tristan | jjardon[m], that status line (after start, before the warning) is because of the thing tlater needs to fix | 11:30 |
| * tlater will be done in a second, is stumped by failing tests | 11:30 | |
| * jjardon[m] proceed to build the whole GNOME now | 11:30 | |
| tristan | jjardon[m], I'll set this one to merge, and that messy thing will go away by itself after Platform becomes a proper singleton | 11:30 |
| * tristan is heading out "early" today | 11:31 | |
| tristan | have dinner plans | 11:31 |
| jjardon[m] | tristan: I gues is normal that "bst build base.bst" doesnt do anything as the artifact cache of gettext is the same? | 11:31 |
| tristan | jjardon[m], gettext.bst is not a part of base.bst (it's a part of core-deps) | 11:32 |
| tristan | jjardon[m], base.bst is only the debootstrap runtime | 11:32 |
| *** adds68_ has quit IRC | 11:32 | |
| jjardon[m] | tristan: rigth, you said that before :) sorry | 11:32 |
| *** adds68 has joined #buildstream | 11:32 | |
| tlater | tristan: Would you have a bit of time to go over what I wrote for Planner.plan? It's a lot larger of a change than I'd like it to be :/ | 11:33 |
| tlater | tristan: It's here atm: https://gitlab.com/BuildStream/buildstream/merge_requests/119/diffs#029d46ba7e1b2197fd4337e31e3d8422a390dad5_84_94 | 11:33 |
| tristan | I have to go now though :-/ | 11:33 |
| tlater | Ah, that's alright | 11:34 |
| tlater | Sorry, I should have asked sooner | 11:34 |
| tristan | tlater, I was busy trying to nail the Arch thing anyway | 11:34 |
| tristan | tlater, I think anyway, my hair brained ideas of reducing that code with Element.dependencies() are wrong | 11:35 |
| tristan | because of runtime dependencies anyway needing to be sorted differently | 11:35 |
| tlater | tristan: Well, sure, but there's a lot that changes about the actual algorithm because the original doesn't create a proper parallel sort for multiple root nodes. | 11:36 |
| tristan | tlater, not looking too deeply at this, maybe some alternative which takes the toplevel elements you want to build and uses some virtual toplevel to depend on them might help reduce that | 11:36 |
| * tlater is quite happy to wait until you can have a proper look :) | 11:37 | |
| tristan | tlater, also, maybe it doesnt really have to be an "Element" per se, just some simple spoofy thing which "quacks like an element" | 11:37 |
| tristan | just an idea | 11:37 |
| * tristan out | 11:37 | |
| tlater | o/ | 11:37 |
| tristan | tlater, if you happen to be needing something smaller to take care of, you might look at this: https://gitlab.com/BuildStream/buildstream/issues/142 | 11:38 |
| tristan | but that is really "just in case" :D | 11:38 |
| * tlater has enough to do :) | 11:38 | |
| *** tristan has quit IRC | 11:41 | |
| *** tristan has joined #buildstream | 11:48 | |
| tristan | ok friend is a bit late... gives me some minutes... | 11:52 |
| tristan | tlater, ok so, it looks like your algo grew mostly in terms of comments | 11:52 |
| tristan | tlater, are you able to create a link to that which is not a diff, but a revision ? | 11:52 |
| * tristan tries | 11:53 | |
| tlater | tristan: https://gitlab.com/BuildStream/buildstream/blob/486df70383c51a455f9d50963e1043ad471fee65/buildstream/_pipeline.py#L94 | 11:53 |
| tristan | https://gitlab.com/BuildStream/buildstream/blob/tracking-changes/buildstream/_pipeline.py#L61 | 11:53 |
| tristan | yep, approximately the same | 11:54 |
| tristan | so, to be honest, I dont want to review this part | 11:54 |
| tristan | juergbi, are you able to take some time to ensure that https://gitlab.com/BuildStream/buildstream/blob/tracking-changes/buildstream/_pipeline.py#L61 is correct ? | 11:54 |
| tristan | please ? | 11:54 |
| juergbi | will take a look | 11:55 |
| tristan | Not necessarily ensure I guess | 11:55 |
| * tlater will add tests to help with 'correctness' anyway | 11:55 | |
| tristan | tlater, I was going to say that :) | 11:55 |
| tlater | I'd mostly like to know if this is a sensible change :) | 11:55 |
| tristan | tlater, but you are on the ball anyway :D | 11:55 |
| tristan | I dont mind about the code size fwiw, not obsessive about LOC counts, I'm more obsessive about general complexity and not having duplication of logic | 11:56 |
| ssam2 | I can see having one (?) per toplevel dict entry getting a bit restrictive | 12:14 |
| ssam2 | e.g., if I have %{gcc-arch-flags} and %{gcc-libdir} variables in my element, and i have complex but separate logic for configuring each one ... i have to put both sets of conditionals under one (?) | 12:15 |
| ssam2 | which is fine for now | 12:15 |
| ssam2 | but maybe for 2.0 we should allow conditional entries like (?.1), (?.2), (?.foo) so that different concerns can be logically separated | 12:16 |
| ssam2 | or maybe it's good that this discourages people from going too crazy with conditionals. i could certainly rewrite this whole bootstrap more clearly, if I had months spare to retest everything | 12:17 |
| tristan | ssam2, I'm not sure this is needed; each toplevel is a list anyway right ? | 12:32 |
| tristan | And it's an if, if, if, if list | 12:32 |
| tristan | not an if, elif, elif, list | 12:32 |
| tristan | ssam2, anyway I'm on my way out the door literally, maybe if you showed me an example I could maybe point out how to do it with the current thing | 12:33 |
| tristan | or, understand what it is that is impossible or unwieldy | 12:33 |
| ssam2 | no rush | 12:34 |
| ssam2 | once you see the baserock gnu-toolchain you'll be able to see what i mean, i expect | 12:34 |
| ssam2 | once i'm done converting to conditionals, i mean | 12:34 |
| *** adds68 has quit IRC | 12:39 | |
| *** adds68 has joined #buildstream | 12:39 | |
| gitlab-br-bot | push on buildstream@platform_singleton (by Tristan Maat): 1 commit (last: Make the platform object a singleton) https://gitlab.com/BuildStream/buildstream/commit/dda93aa347d34204e1f22f65d41c0b778e8bf889 | 13:26 |
| gitlab-br-bot | buildstream: merge request (platform_singleton->master: Make the platform object a singleton) #133 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/133 | 13:26 |
| gitlab-br-bot | buildstream: merge request (platform_singleton->master: Make the platform object a singleton) #133 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/133 | 13:27 |
| gitlab-br-bot | push on buildstream@platform_singleton (by Tristan Maat): 1 commit (last: Make the platform object a singleton) https://gitlab.com/BuildStream/buildstream/commit/5e28b4a3de1f42a1d4240d429c1c85641da5f40c | 13:27 |
| gitlab-br-bot | push on buildstream@platform_singleton (by Tristan Maat): 1 commit (last: Clean old Platform.get_platform references) https://gitlab.com/BuildStream/buildstream/commit/e268c297788f8175ae82e8978e44222da8acad64 | 13:31 |
| gitlab-br-bot | buildstream: merge request (platform_singleton->master: Make the platform object a singleton) #133 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/133 | 13:31 |
| gitlab-br-bot | push on buildstream@tracking-changes (by Tristan Maat): 14 commits (last: Catch attempts to compose a list) https://gitlab.com/BuildStream/buildstream/commit/8c4d8ca2b7c886690fb933428c6768c745667db7 | 13:39 |
| gitlab-br-bot | buildstream: merge request (tracking-changes->master: WIP: Tracking changes) #119 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/119 | 13:39 |
| gitlab-br-bot | buildstream: issue #92 ("Handle kernels which lack CONFIG_USER_NS") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/92 | 13:40 |
| gitlab-br-bot | buildstream: merge request (user-namespace-fallback->master: User namespace fallback) #132 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/132 | 13:40 |
| gitlab-br-bot | buildstream: Tristan Van Berkom deleted branch user-namespace-fallback | 13:40 |
| gitlab-br-bot | push on buildstream@master (by Tristan Van Berkom): 3 commits (last: _platform/linux.py: Add preflight check to detect user namespaces) https://gitlab.com/BuildStream/buildstream/commit/a9a2bcece3b90092d9260f5c46675a0e17df381d | 13:40 |
| *** tiagogomes has quit IRC | 14:33 | |
| tlater | Hm, I'm getting a lot of test failures due to the changes in the internal API | 15:18 |
| tlater | Loader.load() and various other commands now expect lists | 15:18 |
| tlater | Should I make those functions implicitly change single arguments to lists? | 15:18 |
| * tlater feels like that might be a bad idea, but it would fix ~450 failing test cases | 15:19 | |
| tlater | Hmm, nah, it makes no sense to change API to accomodate tests... | 15:20 |
| * tlater stops being lazy | 15:20 | |
| *** semanticdesign has quit IRC | 15:31 | |
| gitlab-br-bot | buildstream: merge request (provenance_fmt->master: WIP: De-dup brackets around provenance in error messages) #134 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/134 | 16:08 |
| *** jude has quit IRC | 16:46 | |
| gitlab-br-bot | buildstream: merge request (incremental-build->master: WIP: Incremental build) #126 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/126 | 17:06 |
| gitlab-br-bot | buildstream: merge request (incremental-build->master: WIP: Incremental build) #126 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/126 | 17:06 |
| gitlab-br-bot | push on buildstream@multiple_targets (by Tristan Maat): 13 commits (last: _pipeline.py: Implement Planner.plan using Kahn's algorithm) https://gitlab.com/BuildStream/buildstream/commit/5ce8d1904e0e19434e584283433065452f4d1fee | 17:20 |
| *** valentind has joined #buildstream | 17:28 | |
| gitlab-br-bot | buildstream: merge request (incremental-build->master: WIP: Incremental build) #126 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/126 | 17:30 |
| gitlab-br-bot | push on buildstream@multiple_targets (by Tristan Maat): 3 commits (last: Adjust `bst workspace` to multiple targets) https://gitlab.com/BuildStream/buildstream/commit/3e525e2506eabb2c947195b4b8e23585c7c86b5b | 17:34 |
| gitlab-br-bot | push on buildstream@multiple_targets (by Tristan Maat): 4 commits (last: main.py: Adjust `bst track` to multiple targets) https://gitlab.com/BuildStream/buildstream/commit/bec2330c50c7d52db9dba516f7cf59569ef4e381 | 18:03 |
| *** tlater has quit IRC | 18:04 | |
| gitlab-br-bot | buildstream: merge request (incremental-build->master: WIP: Incremental build) #126 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/126 | 18:16 |
| *** ssam2 has quit IRC | 18:40 | |
| *** semanticdesign has joined #buildstream | 19:03 | |
| *** givascu has quit IRC | 20:27 | |
| *** adds68 has quit IRC | 21:34 | |
| *** valentind has quit IRC | 22:32 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!