| *** brlogger has joined #buildstream | 09:34 | |
| *** lachlan has joined #buildstream | 09:42 | |
| *** lachlan has quit IRC | 09:49 | |
| gitlab-br-bot | marge-bot123 merged MR !1416 (jennis/do_not_leak_project_specific_remotes->master: Do not leak subproject remotes) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1416 | 09:52 |
|---|---|---|
| *** lachlan has joined #buildstream | 09:56 | |
| *** tristan has joined #buildstream | 10:01 | |
| *** lachlan has quit IRC | 10:29 | |
| *** lachlan has joined #buildstream | 10:30 | |
| alexandrufazakas | If I use something like this https://pastebin.com/U64jkV8u and I want a test to trigger this, how should I check that it fails successfully? | 10:40 |
| alexandrufazakas | Not sure what to assert in the test | 10:40 |
| juergbi | alexandrufazakas: result.assert_main_error(ErrorDomain.APP, 'directory-not-empty') | 10:47 |
| alexandrufazakas | juergbi: oh, alright. Thank you :) | 10:47 |
| benschubert | juergbi: hey, with the new auto pull of junctions, is it normal that I don't get the fancy new UI while the junctions are pulled? | 10:56 |
| juergbi | hm, no, you should see the scheduler UI | 10:57 |
| juergbi | at least it seemed to work fine in my manual tests (that UI is not covered by CI) | 10:57 |
| benschubert | on WSL at least I am not | 10:57 |
| juergbi | but the fetching works fine, just no scheduler UI? | 10:59 |
| *** lachlan has quit IRC | 10:59 | |
| benschubert | juergbi: yep! | 10:59 |
| juergbi | there might certainly be rough edges. maybe UI only shows up for some bst commands. not sure why that would be the case, though | 11:00 |
| benschubert | I just get stuck at 'Loading Element' for 30 minutes, then I have my usual crash :) | 11:00 |
| tlater[m] | WSalmon: After some exploration, it looks like hardlinks in python are actually created as if they were the original file when the original file does not exist :) | 11:00 |
| juergbi | haha | 11:00 |
| benschubert | ah right I'm doing a 'show' | 11:00 |
| benschubert | that might be it | 11:00 |
| juergbi | I actually did test with bst show, though, iirc | 11:00 |
| WSalmon | tlater[m], so what you were worried about is a non issue? | 11:01 |
| WSalmon | with the python tar lib | 11:01 |
| juergbi | benschubert: is it possible that it's not about the fetching but about staging the subproject into the .bst temp directory? | 11:01 |
| tlater[m] | WSalmon: We can still hard link to files outside the staging directory | 11:01 |
| juergbi | benschubert: only fetching is done via scheduler | 11:02 |
| tlater[m] | It's just simpler to fix that (just check if we do that) | 11:02 |
| benschubert | juergbi: no, it's definitely fetching | 11:02 |
| *** lachlan has joined #buildstream | 11:04 | |
| gitlab-br-bot | aevri opened (was WIP) MR !1419 (aevri/set_resource_limits->master: platform: re-scope set_resource_limits) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1419 | 11:17 |
| *** tpollard has quit IRC | 12:02 | |
| *** tpollard has joined #buildstream | 12:03 | |
| *** lachlan has quit IRC | 12:15 | |
| gitlab-br-bot | cs-shadow opened MR !1422 (chandan/fix-readme-badges->master: README.rst: Fix path for badge images) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1422 | 12:51 |
| gitlab-br-bot | marge-bot123 merged MR !1420 (chandan/fix-egg-info-gitignore->master: .gitignore: Fix path of .egg-info) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1420 | 12:54 |
| *** lachlan has joined #buildstream | 13:07 | |
| benschubert | tristan: if you have time today for !1401 it would be much appreciated, I have 4-5 MRs waiting on this one | 13:17 |
| gitlab-br-bot | MR !1401: introduce a Node api, and remove node_get https://gitlab.com/BuildStream/buildstream/merge_requests/1401 | 13:17 |
| *** raoul_ is now known as raoul | 13:19 | |
| *** ChanServ sets mode: +o tristan | 13:26 | |
| tristan | benschubert, scanning it now... | 13:26 |
| tristan | benschubert, sorry for the delay... | 13:26 |
| benschubert | tristan: no worries, I know it's a huge one, it was hard to split :/ thanks a lot! | 13:27 |
| WSalmon | please could juergbi or tristan etc add Becky as a build stream developer, her gitlab username is beckyella16 | 13:33 |
| Becky | test | 13:38 |
| tlater[m] | Becky: test succeeded! | 13:39 |
| juergbi | WSalmon, Becky: done | 13:39 |
| tristan | juergbi, mid-air collision :) | 13:39 |
| Becky | juergbi thankyou | 13:39 |
| tristan | You get to doubly be a developer, as it appears to have succeeded twice ! | 13:40 |
| juergbi | hehe | 13:40 |
| tristan | benschubert, I think I'm done with comments... in general it's looking great, but I think incomplete | 13:49 |
| WSalmon | thanks juergbi tristan | 13:50 |
| benschubert | tristan: thanks a lot! It might have slipped through but the target branch is 'bschubert/new-node-api' not Master , I plan into making more PRs to this branch and merge the new-node-api to master with the whole new API | 13:50 |
| benschubert | so it's expected that it doesn't fix everything yet :) | 13:50 |
| tristan | We had our discussion about fetching special strings, like "project paths", that is kind of a separate discussion because the node needs to have knowledge of the project in order to resolve it - but in general I don't think we should be having APIs outside of `Node` which take nodes as arguments if we can avoid them | 13:50 |
| benschubert | I just wanted to have more manageable MRs and then a last huge round of review | 13:50 |
| tristan | Especially fetching the provenance seems to be a no-brainer that it should be part of Node | 13:51 |
| benschubert | agreed, I want to get rid of all of those when it make sense (basically everything in _yaml) | 13:51 |
| tristan | Yes, ok we are on the same page :) | 13:51 |
| *** lachlan has quit IRC | 13:51 | |
| benschubert | great! I'll change the default=, merge to new-node-api and continue then, thanks a lot for the thorough review! | 13:52 |
| tristan | yeah I missed that the target branch was not master | 13:52 |
| tristan | but I think that doesnt change my scan :) | 13:52 |
| benschubert | I know, it surprised everybody :D | 13:52 |
| juergbi | tristan: bst source checkout currently requires --fetch to fetch uncached sources. should we drop this as well (always fetching as needed), in line with the change for subprojects/junctions? | 13:52 |
| benschubert | we don't do that often in the project, but I think for such a thing it's easier | 13:52 |
| tristan | benschubert, yes it makes sense, I hope that things aren't churning too much for your rebases though :) | 13:53 |
| tristan | juergbi, good question | 13:53 |
| tristan | juergbi, I think it makes sense to drop it | 13:53 |
| benschubert | It's been surprisingly smooth, but that's one of the reasons I'm hunting for reviews, to not have to do it for too long ;) | 13:54 |
| tristan | juergbi, granted of course we should have the UI kick in for fetches which happen and everything... I think it will be more intuitive that way yes | 13:54 |
| juergbi | tristan: sounds good, should be better for consistency | 13:54 |
| tristan | juergbi, But it presents a more difficult question | 13:54 |
| juergbi | yes, should obviously use the usual UI | 13:55 |
| tristan | heh now that I think of it | 13:55 |
| tristan | Well, we could at least do this step for source fetches in a uniform way | 13:55 |
| tristan | juergbi, like... what about `bst artifact checkout` ? | 13:55 |
| tristan | haha | 13:55 |
| juergbi | I was thinking about this as well | 13:55 |
| tristan | "Should the pull be implied ?" | 13:55 |
| juergbi | ah, pull | 13:55 |
| tristan | Lets not think about this for now I think | 13:55 |
| juergbi | I was only thiking about building | 13:55 |
| benschubert | tristan: actually, we often used 'default' in plugins already (https://gitlab.com/BuildStream/buildstream/merge_requests/1401/diffs#diff-content-ae21ae20fa7c1172671e2b7cd0c6ba979c572eaf), do you want to get rid of all of them? | 13:56 |
| juergbi | implicit building is easier to argue against | 13:56 |
| tristan | juergbi, yes it is hehe :) | 13:56 |
| tristan | benschubert, I think it would be more consistent but honestly - I don't feel really very strongly on that point | 13:56 |
| benschubert | ok, I'll keep it for now for simplicity and we can re-review later on | 13:57 |
| benschubert | thanks! | 13:57 |
| juergbi | tristan: it's actually somewhat related to the second point here https://mail.gnome.org/archives/buildstream-list/2019-April/msg00027.html | 13:57 |
| tristan | benschubert, I'm happy that it is not required to be a keyword argument though | 13:57 |
| juergbi | completing a partial artifact is not that different from pulling one that is not yet local at all | 13:57 |
| tristan | juergbi, Good point | 13:57 |
| benschubert | tristan: if we want cython access, we need to allow positional arguments anyways ;) | 13:57 |
| juergbi | however, in any case, I think it's fine to consistently do implicit source fetching and think about pulling separately | 13:58 |
| tristan | benschubert, Does cython prevent us from our `*` signatures ? | 13:58 |
| benschubert | tristan: sadly yes if we want to be able to access it with C-speed | 13:58 |
| tristan | :-/ | 13:59 |
| benschubert | if we _really_ want to keep it, I can however have 2 versions, one for C access (the core) and one public | 13:59 |
| tristan | benschubert, my issue with that is mostly about API extensibility | 13:59 |
| benschubert | we can always have public wrappers for plugins | 13:59 |
| benschubert | its slightly less nice | 13:59 |
| benschubert | but works | 13:59 |
| tristan | it's really nice to be able to put that `*` and know that anything beyond the `*` *must* be keyword arg, it lets us add things without breaking API more easily | 13:59 |
| tristan | anyway, not such a huge hitch, I don't think we need to care much about it right now | 14:00 |
| tristan | benschubert, then again this depends how pervasive the cython API surface will be | 14:01 |
| gitlab-br-bot | marge-bot123 closed issue #1056 (Badge images in README are broken) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1056 | 14:01 |
| gitlab-br-bot | marge-bot123 merged MR !1422 (chandan/fix-readme-badges->master: README.rst: Fix path for badge images) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1422 | 14:01 |
| tristan | right now for Node APIs I think it's not a big deal | 14:01 |
| tristan | for Element API on the other hand, it's really good to have this control | 14:01 |
| benschubert | Sure, I don't want to bring API extensibility problems with Cython, if we have to do something we'll make sure we have a wrapper :) | 14:02 |
| tristan | yeah lets see when we get there | 14:03 |
| benschubert | agreed | 14:03 |
| tristan | at least there are avenues | 14:03 |
| gitlab-br-bot | BenjaminSchubert merged MR !1401 (bschubert/node-api-noget->bschubert/new-node-api: introduce a Node api, and remove node_get) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1401 | 14:11 |
| gitlab-br-bot | BenjaminSchubert opened MR !1424 (bschubert/node-api-nodel->bschubert/new-node-api: _yaml: Remove 'node_del' and support `del mapping[key]`) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1424 | 14:23 |
| gitlab-br-bot | BenjaminSchubert opened MR !1425 (bschubert/node-api-noset->bschubert/new-node-api: Remove 'node_set'. Now use MappingNode[key] = value) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1425 | 14:29 |
| *** tristan has quit IRC | 14:29 | |
| gitlab-br-bot | BenjaminSchubert opened MR !1426 (bschubert/node-api-keys->bschubert/new-node-api: Cleanup `MappingNode` key/values/items accessor) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1426 | 14:30 |
| gitlab-br-bot | juergbi opened MR !1427 (juerg/source-checkout->master: Fetch sources as needed for bst source checkout) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1427 | 14:35 |
| *** lachlan has joined #buildstream | 14:36 | |
| tlater[m] | alexandrufazakas: I'll give you developer permissions, would you mind reopening the MR on a branch on the main repo? It looks like CI doesn't want to run on your fork. | 14:37 |
| alexandrufazakas | tlater: sure thing | 14:37 |
| *** tristan has joined #buildstream | 14:38 | |
| tlater[m] | It's just annoying to have to pull your branch to see your tests ;) | 14:38 |
| alexandrufazakas | Heh, yeah, no worries | 14:38 |
| *** lachlan has quit IRC | 14:40 | |
| *** phildawson_ has joined #buildstream | 14:42 | |
| *** phildawson has quit IRC | 14:42 | |
| *** lachlan has joined #buildstream | 14:44 | |
| *** lachlan has quit IRC | 14:48 | |
| gitlab-br-bot | willsalmon opened MR !1429 (willsalmon/platformRefactor->master: Refactor Platform and Snadboxes) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1429 | 14:49 |
| raoul | With anyone with some grpc/protobuf knowledge care to look at !1410? It's had some review but it'd be good if someone familiar with those things had a look | 14:56 |
| gitlab-br-bot | MR !1410: Capabilities service https://gitlab.com/BuildStream/buildstream/merge_requests/1410 | 14:56 |
| raoul | maybe you juergbi? | 14:56 |
| WSalmon | hi benschubert i seem to remember but i could be wrong that tests-fedora-missing-deps are meant as a proxy to WSL i was wondering why the test are run as root https://gitlab.com/BuildStream/buildstream/blob/master/.gitlab-ci.yml#L124 is root helpful for the proxy? | 15:05 |
| juergbi | raoul: added a couple of comments | 15:08 |
| *** bochecha has quit IRC | 15:09 | |
| gitlab-br-bot | AlexFazakas opened MR !1430 (AlexFazakas/add-bst-init-argument->master: Add bst init argument) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1430 | 15:10 |
| *** lachlan has joined #buildstream | 15:19 | |
| tlater[m] | alexandrufazakas: Sorry, I'm a little distracted, haha. I'll take a look once your CI run is through. | 15:25 |
| alexandrufazakas | tlater: No hurry :) | 15:26 |
| benschubert | WSalmon: they are not a proxy, just to ensure we are marking our tests correctly :) | 15:50 |
| benschubert | running as root should not be needed | 15:50 |
| WSalmon | that test confused me by making a user and chown'ing everthing and then running as root | 15:52 |
| WSalmon | thanks benschubert | 15:52 |
| gitlab-br-bot | raoul.hidalgocharman opened issue #1057 (Use capabalities service in artifact cache) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1057 | 15:53 |
| *** lachlan has quit IRC | 15:53 | |
| alexandrufazakas | Is there a way I can see print's I use inside tests? -s doesn't seem to help | 15:53 |
| WSalmon | alexandrufazakas, the stdout and stderr get printed seperatly | 15:54 |
| WSalmon | look at the top of the stderr section to see the stdout | 15:54 |
| WSalmon | or ask print to print to stderr instead of stdout | 15:55 |
| alexandrufazakas | WSalmon: oh, you're right, it wasn't very obvious. Thanks :D | 15:55 |
| raoul | ta juergbi, responded to the discussions so I'll pass over to marge once the pipelines passed unless you've got any other complaints. There's one point on !1402 that still needs a response to if you can have a look at that too | 15:56 |
| gitlab-br-bot | MR !1402: Configuration option for disabling blob fetching with RE https://gitlab.com/BuildStream/buildstream/merge_requests/1402 | 15:56 |
| WSalmon | we care about what goes to stdout and what gose to stderr so its helpfull to have them seperated in some regards even if it is a bit of a trap for new people | 15:57 |
| WSalmon | we never want print() functions to end up in master but for debuging it can be handy, Becky alexandrufazakas -> https://stackoverflow.com/a/15808105 | 15:58 |
| juergbi | raoul: yes, I took a look earlier but I'm not sure yet about the 2 config options | 15:59 |
| juergbi | it doesn't seem to be ideal but I don't have a better suggestion otoh | 15:59 |
| raoul | yeah it's a bit more cumbersome than I'd like but seemed better imo than extended the dependency scope | 16:00 |
| raoul | If anyone else can think of a better way to do it, it'd be appreciated! | 16:01 |
| Becky | WSalmon ah thanks | 16:01 |
| *** phildawson_ has quit IRC | 16:08 | |
| *** phildawson_ has joined #buildstream | 16:08 | |
| benschubert | tristan: I have https://gitlab.com/BuildStream/buildstream/merge_requests/1424, https://gitlab.com/BuildStream/buildstream/merge_requests/1425, and https://gitlab.com/BuildStream/buildstream/merge_requests/1426 for followups already (much shorter :D) | 16:09 |
| *** phildawson_ has quit IRC | 16:10 | |
| *** phildawson has joined #buildstream | 16:11 | |
| WSalmon | juergbi, after way to much flaffing i have https://gitlab.com/BuildStream/buildstream/merge_requests/1429 | 16:15 |
| gitlab-br-bot | raoul.hidalgocharman approved MR !1427 (juerg/source-checkout->master: Fetch sources as needed for bst source checkout) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1427 | 16:15 |
| *** lachlan has joined #buildstream | 16:16 | |
| gitlab-br-bot | tpollard approved MR !1427 (juerg/source-checkout->master: Fetch sources as needed for bst source checkout) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1427 | 16:18 |
| *** lachlan has quit IRC | 16:20 | |
| gitlab-br-bot | marge-bot123 closed issue #915 (Artifact As A Proto: Capabilities Service) on buildstream https://gitlab.com/BuildStream/buildstream/issues/915 | 16:32 |
| *** raoul has quit IRC | 16:46 | |
| *** lachlan has joined #buildstream | 16:48 | |
| *** alexandrufazakas has quit IRC | 16:49 | |
| *** lachlan has quit IRC | 16:53 | |
| *** lachlan has joined #buildstream | 16:55 | |
| *** slaf has quit IRC | 17:00 | |
| *** slaf has joined #buildstream | 17:01 | |
| *** slaf has joined #buildstream | 17:01 | |
| *** slaf has joined #buildstream | 17:01 | |
| *** slaf has joined #buildstream | 17:01 | |
| *** tpollard has quit IRC | 17:05 | |
| *** tpollard has joined #buildstream | 17:05 | |
| *** slaf has quit IRC | 17:13 | |
| *** phildawson has quit IRC | 17:22 | |
| gitlab-br-bot | tlater opened MR !1431 (tar-target-renaming->master: tar.py: Make link target renaming work between base-dirs) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1431 | 17:26 |
| *** johnward has quit IRC | 17:47 | |
| *** johnward has joined #buildstream | 17:48 | |
| *** benbrown has quit IRC | 17:50 | |
| *** paulsherwood has quit IRC | 17:51 | |
| *** Becky has quit IRC | 17:51 | |
| *** ikerperez has quit IRC | 17:51 | |
| *** adds68 has quit IRC | 17:51 | |
| *** samkirkham has quit IRC | 17:51 | |
| *** samkirkham has joined #buildstream | 17:51 | |
| *** ikerperez has joined #buildstream | 17:51 | |
| *** adds68 has joined #buildstream | 17:51 | |
| *** Becky has joined #buildstream | 17:51 | |
| *** benbrown has joined #buildstream | 17:52 | |
| *** paulsherwood has joined #buildstream | 17:52 | |
| *** jonathanmaw has quit IRC | 18:09 | |
| *** lachlan has quit IRC | 18:19 | |
| *** ironfoot has quit IRC | 18:21 | |
| *** ironfoot has joined #buildstream | 18:22 | |
| *** ChanServ sets mode: +o ironfoot | 18:22 | |
| *** lachlan has joined #buildstream | 18:42 | |
| *** lachlan has quit IRC | 18:47 | |
| *** lachlan has joined #buildstream | 19:11 | |
| *** lachlan has quit IRC | 19:23 | |
| *** slaf has joined #buildstream | 20:39 | |
| *** slaf has joined #buildstream | 20:39 | |
| *** slaf has joined #buildstream | 20:39 | |
| *** slaf has joined #buildstream | 20:40 | |
| *** slaf has joined #buildstream | 20:40 | |
| *** slaf has joined #buildstream | 20:40 | |
| *** slaf has joined #buildstream | 20:40 | |
| *** slaf has joined #buildstream | 20:41 | |
| *** slaf has joined #buildstream | 20:41 | |
| *** slaf has joined #buildstream | 20:41 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!