| *** narispo has quit IRC | 01:33 | |
| *** narispo has joined #buildstream | 01:34 | |
| *** bochecha_ has joined #buildstream | 01:57 | |
| *** benschubert has quit IRC | 01:57 | |
| *** bochecha has quit IRC | 01:58 | |
| *** bochecha_ is now known as bochecha | 01:58 | |
| *** narispo has quit IRC | 02:04 | |
| *** narispo has joined #buildstream | 02:04 | |
| *** narispo has quit IRC | 02:34 | |
| *** narispo has joined #buildstream | 02:35 | |
| *** narispo has quit IRC | 03:31 | |
| *** narispo has joined #buildstream | 03:31 | |
| *** narispo has quit IRC | 04:02 | |
| *** narispo has joined #buildstream | 04:02 | |
| *** narispo has quit IRC | 05:03 | |
| *** narispo has joined #buildstream | 05:03 | |
| *** mohan43u has quit IRC | 05:39 | |
| *** narispo has quit IRC | 06:01 | |
| *** narispo has joined #buildstream | 06:01 | |
| *** bochecha has quit IRC | 06:09 | |
| *** slaf has quit IRC | 07:00 | |
| *** slaf has joined #buildstream | 07:02 | |
| *** phildawson_ has joined #buildstream | 07:48 | |
| *** benschubert has joined #buildstream | 08:20 | |
| *** tiagogomes has joined #buildstream | 08:20 | |
| *** narispo has quit IRC | 08:30 | |
| *** narispo has joined #buildstream | 08:30 | |
| *** tme5 has joined #buildstream | 08:46 | |
| tme5 | does anyone have any suggestions for a better help text for `bst shell --use-buildtree`? At the moment it says "Defaults to ask but if set to always the function will fail if a build tree is not available" | 09:02 |
|---|---|---|
| tme5 | it also accepts 'try' and 'never' | 09:03 |
| tpollard | tme5: I know what you mean, it definitely could be clearer | 09:07 |
| tme5 | it's just I was looking at #1110 and that help text only mentions the defaults, so it's a weird case haha, it doesn't really explain what the option does at all | 09:09 |
| gitlab-br-bot | Issue #1110: use `show_default` for Click.options https://gitlab.com/BuildStream/buildstream/issues/1110 | 09:10 |
| tme5 | does it stage the build dependencies of the element in the sandbox? | 09:10 |
| tme5 | if available? or something? | 09:10 |
| *** tristan has joined #buildstream | 09:12 | |
| tpollard | tme5: do you know what the buildtree is in buildstream? That's probably a good place to start | 09:15 |
| *** lachlan has joined #buildstream | 09:30 | |
| *** jonathanmaw has joined #buildstream | 09:34 | |
| *** lachlan has quit IRC | 09:47 | |
| gitlab-br-bot | marge-bot123 merged MR !1560 (jennis/bst_artifact_show->master: Introduce `bst artifact show`) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1560 | 09:49 |
| *** lachlan has joined #buildstream | 09:50 | |
| *** traveltissues has joined #buildstream | 09:52 | |
| *** traveltissues has quit IRC | 09:55 | |
| *** lachlan has quit IRC | 10:01 | |
| *** traveltissues has joined #buildstream | 10:05 | |
| *** lachlan has joined #buildstream | 10:09 | |
| *** lachlan has quit IRC | 10:14 | |
| *** lachlan has joined #buildstream | 10:20 | |
| *** tristan has quit IRC | 10:33 | |
| *** lachlan has quit IRC | 10:47 | |
| *** lachlan has joined #buildstream | 10:48 | |
| *** lachlan has quit IRC | 11:17 | |
| tlater[m] | tme5: Do you have any tips regarding accessing errors or suchlike from the artifact server process? | 11:28 |
| *** lachlan has joined #buildstream | 11:28 | |
| * tlater[m] has a hanging test that likely breaks on the server end, but can't see anything on account of it being on a subprocess. | 11:29 | |
| tlater[m] | I wish I could run this server in a terminal and point the test suite at it... | 11:30 |
| * tlater[m] checks if he can hardcode create_server to do that | 11:31 | |
| jennis | what server? | 11:31 |
| tme5 | on master, all the subprocess tests put the string form of the error on the queue back to the parent | 11:32 |
| tme5 | but other than that bit of info i have no idea, never attempted what you're doing | 11:32 |
| tlater[m] | tme5: This is during actual invocation | 11:33 |
| tlater[m] | RPC invocation, that is | 11:33 |
| tme5 | ahh, i have no knowledge about that sorry | 11:34 |
| tlater[m] | So the server's already running - the test doesn't report init errors (but I fixed that), but it can't asynchronously query the subprocess to ask it "hey, you still there" | 11:34 |
| tlater[m] | But ta, tme5 | 11:34 |
| tme5 | but I will resolve everything my MR right now! hopefully it will be merged soon | 11:34 |
| tlater[m] | jennis: artifact cache server | 11:34 |
| tlater[m] | I suspect it just crashes because I messed with it | 11:34 |
| tlater[m] | But the test suite just hangs instead (thanks pytest-timeout for giving me at least something) | 11:35 |
| tpollard | sounds like a lot of fun that raoul had when trying to debug | 11:35 |
| *** lachlan has quit IRC | 12:02 | |
| *** tristan has joined #buildstream | 12:13 | |
| tpollard | This looks nice https://github.com/pseudomuto/protoc-gen-doc | 12:41 |
| tpollard | if we wanted docs for out protos | 12:42 |
| *** jonathanmaw has quit IRC | 13:02 | |
| gitlab-br-bot | marge-bot123 closed issue #1078 (Replace bst build --all with bst build --deps all) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1078 | 13:02 |
| gitlab-br-bot | marge-bot123 merged MR !1559 (tmewett/build-deps-cli->master: Remove build --all flag in favour of --deps all/plan) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1559 | 13:02 |
| tme5 | everything has been resolved on !1557 besides one thread just mentioning test correctness - if everyone's happy with the MR i will resolve it and it will be ready for merge | 13:09 |
| gitlab-br-bot | MR !1557: Add in_subprocess pytest mark and modify tests which run in another process to use it https://gitlab.com/BuildStream/buildstream/merge_requests/1557 | 13:09 |
| *** lachlan has joined #buildstream | 13:13 | |
| *** lachlan has quit IRC | 13:22 | |
| *** jonathanmaw has joined #buildstream | 13:33 | |
| *** tiagogomes has quit IRC | 13:37 | |
| *** lachlan has joined #buildstream | 13:43 | |
| *** tiagogomes has joined #buildstream | 13:47 | |
| *** tiagogomes_ has joined #buildstream | 13:51 | |
| *** tiagogomes has quit IRC | 13:52 | |
| *** narispo has quit IRC | 13:55 | |
| *** narispo has joined #buildstream | 13:55 | |
| tme5 | hopefully this isn't a stupid question, but is there any documentation on what a build tree is? | 13:56 |
| tlater[m] | tme5: Not sure there is, but a build tree is, well, the source directory after you ran make | 14:02 |
| tlater[m] | Or whatever the appropriate build tool is for the specific project. | 14:02 |
| tlater[m] | It's not buildstream specific. | 14:03 |
| WSalmon | tlater[m], tpollard, raoul https://docs.pytest.org/en/latest/flaky.html the random test order seems like a thing we should try out! | 14:03 |
| tlater[m] | WSalmon: Oh, that'd be pretty cool | 14:04 |
| tlater[m] | It'd likely horribly break our tests | 14:04 |
| tlater[m] | But I think we want to :) | 14:05 |
| benschubert | our tests are supporting it quite well, I've been running this at home for quite some time | 14:05 |
| * tlater[m] is positively surprised | 14:06 | |
| *** lachlan has quit IRC | 14:06 | |
| WSalmon | dose `quite well` mean some dont? | 14:07 |
| tlater[m] | To think that our test suite can run out of order! | 14:07 |
| tlater[m] | We should try and run that in CI though, to keep it that way. | 14:07 |
| WSalmon | i found pleanty of integration tests that didnt like being run by them selves so i am surprised by this. | 14:08 |
| tlater[m] | I imagine the "integration" tests will struggle a bit with that. | 14:08 |
| tlater[m] | Yep :) | 14:08 |
| benschubert | WSalmon: not by themselves is another problem that this wouldn't solve | 14:08 |
| benschubert | but it's definitely a problem | 14:09 |
| WSalmon | oh i assumed that the ones that could not run by them selves would fail if they ended up going first? do we have so many that it dosent happen very often? | 14:11 |
| benschubert | yep | 14:12 |
| WSalmon | maybe a over night test that ran it 5 times.. but then it would get ignored like all our other over night tests.... ummm | 14:13 |
| tlater[m] | It's simple - benschubert fixes all the tetsts ;) | 14:15 |
| tlater[m] | * It's simple - @_gimpnet_benschubert:gnome.org fixes all the tests ;) | 14:15 |
| benschubert | WSalmon: if you want to see all the tests that don't run alone: pytest --forked --integration | 14:16 |
| benschubert | tlater[m]: I wish I could be everywhere at the same time, but that's not possible, so that won't happen | 14:16 |
| tlater[m] | Ah, that's the problem we need to solve first | 14:17 |
| tlater[m] | Let's clone benschubert | 14:17 |
| WSalmon | ohh so --forked is the magic, we should add this to the tests on master | 14:17 |
| WSalmon | and manual the rest of the time | 14:17 |
| benschubert | WSalmon: no we should not, that slows down things _a lot_ | 14:17 |
| WSalmon | hence only for master | 14:17 |
| benschubert | fair enough, yes that could be doable | 14:18 |
| benschubert | but the bleeding needs to stop first | 14:18 |
| WSalmon | is there a issue for this? | 14:18 |
| WSalmon | i cant see one | 14:19 |
| WSalmon | ill make one | 14:19 |
| tlater[m] | WSalmon: I'm convinced I've written a plethora of issues about this. But thanks for bringing it back to the light :) | 14:21 |
| tme5 | what are your thoughts on this help text? bst shell --use-buildtree [ask|try|always|never]: Stage the contents of the element's build-time sandbox | 14:27 |
| tme5 | want to make sure I get it right | 14:27 |
| tlater[m] | s/built-time/post-build/ would make more sense IMO | 14:28 |
| tlater[m] | But I don't like referring to the sandbox in our help text, because the user rarely ever interacts with it. | 14:29 |
| tlater[m] | So it ends up confusing | 14:29 |
| gitlab-br-bot | willsalmon opened issue #1111 (Some tests can not run by them selves) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1111 | 14:30 |
| *** lachlan has joined #buildstream | 14:35 | |
| *** tiagogomes_ has quit IRC | 14:35 | |
| tme5 | tlater[m], ty, will change that. the user does know that elements are built in a sandbox though right? from writing element files | 14:35 |
| tlater[m] | tme5: Yeah, but I'm thinking about a user with absolutely no clue here. I just know how to write my java project, what is a sanbdox? | 14:37 |
| tlater[m] | As a new user, I don't want to dig into the depths of buildstream's architecture documentation to figure out what the sandbox is. Hell, even as someone who knows about sandboxes, I don't know what specifically buildstream's does, let alone what a checkout of it would give me. | 14:37 |
| tlater[m] | So I feel like referring to the sandbox in any of our help entries is not great if we can avoid it at all. | 14:39 |
| tlater[m] | There must be a better way to describe "all the debris that is left lying around by your build tool" | 14:39 |
| traveltissues | 'side effects'? | 14:39 |
| tlater[m] | traveltissues: Yeah :D | 14:40 |
| * tlater[m] can't think of a good way to describe that this command gives you a shell with the side effects of a build though | 14:40 | |
| tme5 | hmm | 14:41 |
| benschubert | "files generated during your build step" | 14:43 |
| tme5 | you're right, the docs refer to the 'output artifact' instead of the 'sandbox' | 14:43 |
| tlater[m] | In general, I want the help texts be simple enough to understand without a full understanding of the tool | 14:44 |
| tlater[m] | Otherwise you inhibit explorative learning | 14:44 |
| tlater[m] | So even if the docs explained the sandbox in detail, I wouldn't like referring to it here. | 14:44 |
| tme5 | "Stage the element's post-build environment" | 14:45 |
| tme5 | "(including sources and dependencies)" ? | 14:45 |
| tlater[m] | Yeah, that sounds a lot better. I'm not sure I like "stage", but we use that elsewhere, so changing that would need to be a bigger refactor. | 14:46 |
| tme5 | great, and yeah, stage is used everywhere | 14:47 |
| tlater[m] | tme5: "Stage the element including the files generated during its build step" to take benschubert's suggestion | 14:48 |
| tme5 | (actually looking at it now, there are several other references in `bst shell --help` to the sandbox. But I won't introduce another haha) | 14:48 |
| tme5 | to me that sounds like *only* files generated by the build step will be staged, and not the whole environment in a post-build state | 14:50 |
| tme5 | unless that's actually what happens? | 14:50 |
| tlater[m] | I'm not 100% sure, frankly. I think the two are synonymous in practice, because nothing after the build step should be able to modify the sandbox anymore. | 14:52 |
| tlater[m] | Caveat that integration-commands might, but I think that's applied to other artifacts. That's where I'm unsure :) | 14:54 |
| tme5 | ah, i'm treating 'build step' as excluding the staging of dependencies. I think if you include it then they are basically the same as you say | 14:55 |
| tme5 | gah, documentation is hard | 14:55 |
| tlater[m] | Hehe, that's why we don't get it right most of the time. | 14:55 |
| tme5 | by my understanding of the behaviour I think "Stage the element's post-build environment, which includes its sources and dependencies" is pretty clear | 14:56 |
| * tlater[m] saw a talk recently where specialzied "docs teams" were suggested as good practice | 14:56 | |
| tlater[m] | Because a dev won't get docs right while also thinking about code. | 14:56 |
| tlater[m] | tme5: *Now* you can start thinking about semi-internationalization, where you want to use simpler nouns because not everybody's vocabulary is going to be amazing. | 14:57 |
| tlater[m] | Which is why I prefer benschubert's suggestion - it's simpler. | 14:57 |
| tlater[m] | But that decision is up to you, I don't have a very strong one here. | 14:57 |
| *** tiagogomes_ has joined #buildstream | 15:03 | |
| *** lachlan has quit IRC | 15:16 | |
| *** lachlan has joined #buildstream | 15:34 | |
| *** phil has joined #buildstream | 15:36 | |
| *** phildawson_ has quit IRC | 15:37 | |
| *** lachlan has quit IRC | 15:38 | |
| *** tiagogomes_ has quit IRC | 15:52 | |
| gitlab-br-bot | jennis opened MR !1565 (jennis/update_checkout->master: Update artifact checkout to handle ArtifactRefs) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1565 | 15:56 |
| *** tiagogomes has joined #buildstream | 15:58 | |
| gitlab-br-bot | jennis closed issue #1103 (Implement `bst artifact show`) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1103 | 16:06 |
| *** phil has quit IRC | 16:09 | |
| *** lachlan has joined #buildstream | 16:35 | |
| *** tme5 has quit IRC | 16:39 | |
| *** slaf has quit IRC | 16:45 | |
| *** slaf has joined #buildstream | 16:45 | |
| *** slaf has joined #buildstream | 16:45 | |
| *** slaf has joined #buildstream | 16:46 | |
| *** slaf has joined #buildstream | 16:46 | |
| *** slaf has joined #buildstream | 16:46 | |
| *** lachlan has quit IRC | 17:07 | |
| *** lachlan has joined #buildstream | 17:12 | |
| *** laurence has quit IRC | 17:16 | |
| *** bethw has quit IRC | 17:16 | |
| *** WSalmon has quit IRC | 17:16 | |
| *** johnward has quit IRC | 17:16 | |
| *** WSalmon has joined #buildstream | 17:17 | |
| *** Becky has quit IRC | 17:18 | |
| *** qinusty has quit IRC | 17:18 | |
| *** ikerperez has quit IRC | 17:18 | |
| *** samkirkham has quit IRC | 17:18 | |
| *** adds68 has quit IRC | 17:18 | |
| *** laurence has joined #buildstream | 17:20 | |
| *** johnward has joined #buildstream | 17:21 | |
| *** bethw has joined #buildstream | 17:21 | |
| *** lachlan has quit IRC | 17:24 | |
| *** lachlan has joined #buildstream | 17:27 | |
| *** lachlan has quit IRC | 17:39 | |
| *** traveltissues has quit IRC | 18:10 | |
| *** jonathanmaw has quit IRC | 18:15 | |
| *** adds68 has joined #buildstream | 18:23 | |
| *** lachlan has joined #buildstream | 18:29 | |
| *** narispo has quit IRC | 18:46 | |
| *** narispo has joined #buildstream | 18:46 | |
| *** qinusty has joined #buildstream | 19:03 | |
| *** narispo has quit IRC | 19:14 | |
| *** lachlan has quit IRC | 19:29 | |
| *** bochecha has joined #buildstream | 19:39 | |
| *** bochecha has joined #buildstream | 19:41 | |
| *** lachlan has joined #buildstream | 19:59 | |
| *** lachlan has quit IRC | 20:06 | |
| *** tristan has quit IRC | 20:44 | |
| *** narispo has joined #buildstream | 21:08 | |
| *** bochecha has quit IRC | 21:15 | |
| *** rdale has quit IRC | 21:17 | |
| *** bochecha has joined #buildstream | 21:23 | |
| *** bochecha_ has joined #buildstream | 21:26 | |
| *** bochecha has quit IRC | 21:27 | |
| *** bochecha_ is now known as bochecha | 21:27 | |
| *** bochecha has quit IRC | 21:46 | |
| *** slaf has quit IRC | 23:10 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!