| *** tristan has quit IRC | 00:02 | |
| *** dftxbs3e has quit IRC | 00:21 | |
| *** dftxbs3e has joined #buildstream | 00:24 | |
| *** tristan has joined #buildstream | 02:03 | |
| *** tristan has quit IRC | 03:05 | |
| *** tristan has joined #buildstream | 04:02 | |
| *** ironfoot has quit IRC | 04:21 | |
| *** ironfoot has joined #buildstream | 04:22 | |
| *** ChanServ sets mode: +o ironfoot | 04:22 | |
| *** tristan has quit IRC | 06:05 | |
| *** tristan has joined #buildstream | 06:28 | |
| *** bochecha has joined #buildstream | 07:47 | |
| *** tiagogomes has quit IRC | 08:29 | |
| *** tiagogomes has joined #buildstream | 08:30 | |
| *** jonathanmaw has joined #buildstream | 08:34 | |
| *** tpollard has joined #buildstream | 09:02 | |
| *** bochecha_ has joined #buildstream | 09:15 | |
| *** bochecha has quit IRC | 09:17 | |
| *** bochecha_ is now known as bochecha | 09:17 | |
| *** alexandrufazakas is now known as alexfazakas | 09:42 | |
| tpollard | is tab completion broken on master? | 09:57 |
|---|---|---|
| tpollard | wait, all commands are broke for me :S | 09:58 |
| * tpollard reinstalls | 09:58 | |
| benschubert | tpollard: 'all commands are broken', what do you mean by that? | 09:58 |
| * Kinnison has found that sometimes pip3 install -e . doesn't quite DTRT | 09:59 | |
| tpollard | everything was producing a filenotfounderror | 09:59 |
| tpollard | I've reinstalled, and it's now fine... | 09:59 |
| benschubert | weird | 10:00 |
| tpollard | although I think the flatpak-autotools example needs updating for the node validate changes | 10:00 |
| tpollard | Or I need to update my bst experimental | 10:00 |
| * tpollard sighs into monday morning | 10:00 | |
| benschubert | tpollard: that's for bst-plugins-experimental correct? Yep, they all need updating, I have a WIP MR for that, but this repo is in such a bad state that I wasn't able of getting any CI green | 10:01 |
| benschubert | if someone wants to take ownership of fixing it, I'll happily leave it, as it's clearly not a priority for me :) Otherwise, I'll fix it when I have time | 10:01 |
| *** phoenix has joined #buildstream | 10:03 | |
| *** kapip has joined #buildstream | 10:32 | |
| gitlab-br-bot | marge-bot123 merged MR !1487 (bschubert/node-enum->master: Add a 'as_enum' on Scalar nodes to help with constraining inputs) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1487 | 10:36 |
| *** cs-shadow has joined #buildstream | 11:13 | |
| *** phoenix has quit IRC | 11:32 | |
| laurence | Only 19 open MRs on the project, nice | 12:07 |
| laurence | our table on the README about the availability in distros is looking a little ..... red | 12:12 |
| laurence | Is it best to remove that until 2.0? | 12:12 |
| laurence | https://gitlab.com/BuildStream/buildstream | 12:12 |
| tlater[m] | I think we have a couple too many distros in there atm, at the very least | 12:14 |
| tlater[m] | Do we really need to note down every recent version of every distro under the sun? Haha | 12:15 |
| tlater[m] | Not that useful of a table until we get some people maintaining packages | 12:15 |
| jjardon | laurence: removing the table will not fix reality; better to try to fix the status of the packages on distros if you want to see less red | 12:18 |
| jjardon | seems there is a bug there though; 1.2.8 should be green, not red | 12:19 |
| jjardon | I have reported that; It will get fixed in a couple of days | 12:21 |
| jjardon | https://repology.org/project/buildstream/report | 12:22 |
| gitlab-br-bot | jjardon opened MR !1506 (jjardon/ruamel->master: Depend on ruamel >= 0.16) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1506 | 12:29 |
| *** laurence has left #buildstream | 12:39 | |
| *** laurence has joined #buildstream | 12:39 | |
| gitlab-br-bot | danielsilverstone-ct closed MR !1506 (jjardon/ruamel->master: Depend on ruamel >= 0.16) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1506 | 12:59 |
| gitlab-br-bot | danielsilverstone-ct reopened MR !1506 (jjardon/ruamel->master: Depend on ruamel >= 0.16) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1506 | 12:59 |
| Kinnison | jjardon: I apologise for the close/open noise on your ruamel MR, I was pressing a button on a page which reloaded. V. confused :D | 13:01 |
| *** tpollard has quit IRC | 13:02 | |
| tlater[m] | Argh | 13:04 |
| tlater[m] | Turns out clicking "commit" in gitlab's UI commits straight to master | 13:05 |
| tlater[m] | Sorry! | 13:05 |
| *** tpollard has joined #buildstream | 13:07 | |
| Kinnison | oops | 13:08 |
| Kinnison | Many misclicks on gitlab today :D | 13:08 |
| tlater[m] | I hope that doesn't break anything marge related? | 13:08 |
| benschubert | tlater[m]: at least it was for a typo :) | 13:23 |
| *** narispo has quit IRC | 13:36 | |
| * tlater[m] is annoyed that buildstream only says "connect failed" when it doesn't manage to connect to an artifact server | 13:42 | |
| tlater[m] | Why? 404? 500 | 13:42 |
| tlater[m] | Unauthorized for some reason? :| | 13:43 |
| tlater[m] | This is the culprit: https://gitlab.com/BuildStream/buildstream/blob/master/src/buildstream/_cas/casremote.py#L202 | 13:51 |
| tlater[m] | e.details() is too vague | 13:51 |
| benschubert | tlater[m]: when running with '--debug' it would be good to have the `str()` version of it then :) | 13:55 |
| benschubert | (I would actually like getting stack traces when running with '--debug'. Would that be something that would be accepted? | 13:55 |
| *** traveltissues has joined #buildstream | 13:57 | |
| *** traveltissues has quit IRC | 13:57 | |
| tlater[m] | benschubert: Agreed, but I think we need a better default than e.details(), too | 13:57 |
| tlater[m] | The actual error turned out to be "Connection refused (111)" | 13:57 |
| tlater[m] | Maybe an upstream thing, tbf | 13:58 |
| tlater[m] | I'm also definitely pro-stacktraces. No reason not to show them with --debug on. | 14:00 |
| benschubert | I'll make a PR with that then | 14:01 |
| *** rdale has quit IRC | 14:01 | |
| *** rdale has joined #buildstream | 14:02 | |
| tlater[m] | ta benschubert :) | 14:06 |
| tlater[m] | Hm, is https://buildstream.gitlab.io/buildstream/using_configuring_cache_server.html#artifact-command-reference up-to-date? | 14:19 |
| tlater[m] | I get an error claiming it doesn't have the "BuildStream capabilities service": https://pastebin.com/raw/tp0G1enw | 14:20 |
| *** narispo has joined #buildstream | 14:28 | |
| tpollard | is your local artifact server install from current master? | 14:37 |
| tpollard | 'old' servers can't speak new artifacts | 14:37 |
| tpollard | but 'new' servers should speak old artifact storage for now at least | 14:37 |
| tlater[m] | tpollard: I thought it was, on account of it being a version published by us to dockerhub 7 days ago | 14:42 |
| tlater[m] | But I don't actually think it is, since manually building the image on master doesn't work. | 14:42 |
| * tlater[m] wonders if bst-artifact-server should display a version message on startup | 14:48 | |
| tlater[m] | Technically gpl demands it, if nothing else | 14:48 |
| Kinnison | I don't really hold with services writing things like that on startup, mostly because services shouldn't be seen | 14:50 |
| tlater[m] | Yeah, I'm not too convinced either | 14:52 |
| tlater[m] | It would make answering the question "is my server outdated" a lot easier though. | 14:52 |
| tlater[m] | After getting my docker image to build, yup, looks like the server was indeed outdated | 15:01 |
| * tlater[m] will try and remember to patch the Dockerfile so it can build newer versions too. | 15:01 | |
| *** tristan has quit IRC | 15:05 | |
| *** tristan has joined #buildstream | 15:32 | |
| tlater[m] | Re https://gitlab.com/BuildStream/buildstream/issues/1041 - I'm starting to grok the code around this, but now I'm a little lost on what the difference between the CAS and artifact endpoints really is. | 15:46 |
| tlater[m] | They seem to be described by the same proto, looking at _artifactcache.py:542 | 15:47 |
| tlater[m] | Is it "the place we upload blobs to" vs "the place we upload the "artifact proto" to"? | 15:47 |
| *** alexfazakas has quit IRC | 15:56 | |
| *** alexandrufazakas has joined #buildstream | 15:56 | |
| alexandrufazakas | how can I lint one file only? | 15:56 |
| alexandrufazakas | is it not `tox -e lint $file`? | 15:57 |
| *** alexandrufazakas is now known as alexfazakas | 15:57 | |
| tpollard | tlater[m]: that's what I believe is the goal | 15:57 |
| tlater[m] | Right, so I'll basically need two remotes, and the code in _artifactcache needs to learn how to use the right remote for the right thing | 15:58 |
| tlater[m] | ta tpollard | 15:58 |
| tlater[m] | alexfazakas: I think you'll need a -- in there | 16:00 |
| tlater[m] | i.e. `tox -e lint -- $file` | 16:00 |
| tlater[m] | But don't quote me on that :) | 16:00 |
| cs-shadow | tlater[m]: unfortunately that won't work as we don't handle that case in our tox config | 16:26 |
| cs-shadow | alexfazakas: without changing the tox config, this is what I usually do: | 16:26 |
| cs-shadow | ``` | 16:26 |
| cs-shadow | ``` | 16:27 |
| cs-shadow | tox -e lint --notest | 16:27 |
| cs-shadow | . .tox/lint/bin/activate | 16:27 |
| cs-shadow | # Now you can run linters directly | 16:27 |
| cs-shadow | pylint FILENAME | 16:27 |
| cs-shadow | pycodesytle FILENAME | 16:27 |
| cs-shadow | ``` | 16:27 |
| tlater[m] | Hm, maybe worth changing the tox config, if you can figure out how to, alexfazakas | 16:31 |
| *** tristan has quit IRC | 16:34 | |
| benschubert | alexfazakas: https://tox.readthedocs.io/en/latest/example/general.html for how to do that (posargs with defaults) :) | 16:34 |
| * tlater[m] wonders what the "instance name" on remotes means | 16:50 | |
| tlater[m] | It seems to be passed to requests, is there somewhere that defines what the protocol looks like? | 16:51 |
| *** jonathanmaw has quit IRC | 16:57 | |
| tlater[m] | benschubert: are you still about? | 17:12 |
| benschubert | tlater[m]: yep | 17:13 |
| *** tristan has joined #buildstream | 17:16 | |
| tlater[m] | Ah, sorry, notifications aren't working :| | 17:19 |
| tlater[m] | Re the note on type hints | 17:19 |
| tlater[m] | Do you really not want them there? | 17:19 |
| tlater[m] | I mean, sure, they're fairly low value, but at worst they make understanding the API we interact with a bit easier. And it's not particularly much effort to add them. | 17:20 |
| tlater[m] | The only real downside is that the style would be slightly different from that in the main codebase, and I think eventually we do want a future in which we have type hints everywhere. | 17:20 |
| benschubert | tlater[m]: I'm not against it, just not sure it's the best place to start :) | 17:22 |
| benschubert | and they are fairly low value on small scripts | 17:22 |
| benschubert | but removing it is not worth it either :) | 17:23 |
| tlater[m] | We do need to start somewhere ;) | 17:23 |
| benschubert | Sure, I was just saying that the discussion on the ML wasn't about adding types everywhere but specifically on the public API | 17:24 |
| benschubert | so I personally don't have any problem, but I don't know if others might | 17:24 |
| tlater[m] | Right, gotcha | 17:26 |
| tlater[m] | I read it as "we want them everywhere, but that'll be difficult, so let's add them to where we already have a convention" | 17:27 |
| tlater[m] | I'm happy to risk someone furiously removing type hints from a contrib script in a week or two ;p | 17:27 |
| benschubert | fair | 17:28 |
| *** tiagogomes has quit IRC | 17:32 | |
| *** tiagogomes has joined #buildstream | 18:26 | |
| *** flatmush has joined #buildstream | 18:42 | |
| *** tiagogomes has quit IRC | 19:21 | |
| juergbi | tlater[m]: regarding #1041, that sounds correct. however, please note that the CAS code is changing in !1499 | 19:32 |
| gitlab-br-bot | Issue #1041: Support separate end points for artifact caches https://gitlab.com/BuildStream/buildstream/issues/1041 | 19:32 |
| gitlab-br-bot | MR !1499: WIP: Use buildbox-casd for CAS access https://gitlab.com/BuildStream/buildstream/merge_requests/1499 | 19:33 |
| *** flatmush has quit IRC | 20:32 | |
| *** kapip has quit IRC | 20:51 | |
| *** tristan has quit IRC | 22:23 | |
| *** tristan has joined #buildstream | 22:23 | |
| *** Drakola has joined #buildstream | 22:56 | |
| *** Drakola has quit IRC | 22:58 | |
| *** Drakola has joined #buildstream | 22:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!