| *** dftxbs3e has quit IRC | 04:41 | |
| *** dftxbs3e has joined #buildstream | 05:05 | |
| *** alexandrufazakas has joined #buildstream | 08:09 | |
| *** traveltissues has joined #buildstream | 09:12 | |
| *** bochecha has joined #buildstream | 09:14 | |
| gitlab-br-bot | LaurenceUrhegyi closed MR !1415 (laurence/update-contributing-guide->master: Adds a Committer Policy section to CONTRIBUTING.rst) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1415 | 09:14 |
|---|---|---|
| *** jonathanmaw has joined #buildstream | 09:17 | |
| *** lachlan has joined #buildstream | 09:24 | |
| juergbi | tristan: any thoughts on this? https://gitlab.com/BuildStream/buildstream/merge_requests/1402#note_187384193 | 09:26 |
| laurence | juergbi, would be good to raise as a follow up issue and not block that patch landing, if possible | 09:29 |
| juergbi | I think I'll rather move the corresponding commit to a separate branch/MR as it's mostly orthogonal to the rest | 09:29 |
| *** ChanServ sets mode: +o tristan | 09:44 | |
| tristan | juergbi, The merge request says "configuration for disabling blob fetching with RE", but only adds the "whether to build all dependencies" configuration option ? | 09:44 |
| juergbi | tristan: no, that's just one commit | 09:44 |
| tristan | juergbi, I agree that it would make sense to split that into two parts (first add the "all" build option and then do the other one ?) | 09:45 |
| juergbi | that extra option came up in our discussion about disabling blob fetching | 09:45 |
| tristan | oh strange | 09:45 |
| tristan | https://gitlab.com/BuildStream/buildstream/merge_requests/1402/diffs#note_187384193 | 09:45 |
| tristan | When I show the whole file on gitlab it shows me only one change to userconfig.yaml | 09:46 |
| juergbi | the other is in remote-execution section which is not in userconfig.yaml as it's handled more dynamically | 09:46 |
| juergbi | (similar to artifact servers) | 09:46 |
| tristan | Eh ? | 09:46 |
| juergbi | however, maybe we could actually put the new option into userconfig.yaml, that shouldn't be an issue | 09:46 |
| juergbi | that's already the case in master. I'll try whether I can fix this up | 09:47 |
| juergbi | tristan: any opinion on new userconfig group `build` vs. top-level option `build-all`? | 09:48 |
| tristan | Yeah not sure why we'd want to start fragmenting the config... is it because we're giving the user too many knobs and buttons ? | 09:48 |
| juergbi | no, I think it wasn't completely intentional. originally, the remote execution config was incorrectly only supported in project.conf | 09:49 |
| tristan | juergbi, I agree with you on that | 09:49 |
| * tristan commented such on the issue | 09:50 | |
| juergbi | and it was handled similarly to artifact servers, and artifact server nodes are handled dynamically because they are in a list | 09:50 |
| juergbi | ok, ta. might make sense to call the key inside the new group `all-dependencies` instead of just `all` | 09:51 |
| tristan | juergbi, lets fix it to be consistent then | 09:51 |
| tristan | if we can open an issue for that right away that'd be great | 09:51 |
| tristan | juergbi, I'd go another step | 09:51 |
| juergbi | ok, will either open an issue or directly fix it | 09:51 |
| tristan | juergbi, If you want to ask raoul to split it into a separate MR it might make more sense for redesign though | 09:52 |
| juergbi | yes, makes sense. raoul might not have time for this but I'll take care of it | 09:52 |
| tristan | juergbi, what about something like a value which can be 'all' 'run' or 'build' | 09:52 |
| tristan | like a --deps thing, expressing what you expect to be building | 09:53 |
| tristan | then we could remove `bst build --all` and have `bst build --deps all` | 09:53 |
| juergbi | hm, does deps 'build' even make sense for 'bst build'? | 09:53 |
| tristan | yeah it does | 09:53 |
| tristan | juergbi, For instance I wanna build something new | 09:54 |
| tristan | Or... maybe it doesnt | 09:54 |
| tristan | I *has* in the past, it might be irrelevant | 09:54 |
| tristan | juergbi, point is people want to be in a state where they can try to build something which is either new or failed | 09:54 |
| tristan | but current releases dont give them the ability to delete an artifact | 09:54 |
| juergbi | hm, I don't see the connection to deps option | 09:55 |
| tristan | juergbi, I think the goal posts certainly change in BuildStream 2 regarding that question | 09:55 |
| juergbi | the default is to build targets and their (transitive) runtime dependencies - and build dependencies only as needed | 09:55 |
| juergbi | --all builds all dependencies | 09:56 |
| tristan | Yeah it's hard to say... I think it may also be from a BuildStream developer perspective we wanted that | 09:56 |
| juergbi | not sure what --deps build would do. skip unneeded runtime-only dependencies? | 09:56 |
| tristan | I.e. as long as "BuildStream works perfectly", you don't need to rebuild the same thing over and over and debug what went wrong | 09:56 |
| juergbi | I'm still not seeing the connection between the deps option and the rebuild use case | 09:57 |
| tristan | Ok... in bst 1 - there is no such thing as cached failed builds | 09:57 |
| juergbi | the latter should be covered by bst artifact delete, but it seems orthogonal | 09:57 |
| tristan | You have a failure in CI | 09:57 |
| tristan | I mean I guess you could just build the target which failed, that's true | 09:57 |
| tristan | maybe you just want to run `bst shell --build` later on and know that you wont need network connectivity though | 09:58 |
| tristan | I.e. you "Wanna get your cache in a state where you have everything ready to build 'foo'" | 09:58 |
| juergbi | ah, not building target itself either | 09:58 |
| tristan | You know the other stuff is not gonna fail, cause it succeeded in CI... so you run `bst build --deps build foo.bst` and go grab lunch | 09:58 |
| juergbi | that might make sense in some cases | 09:59 |
| tristan | yeah Scope.RUN includes the target | 09:59 |
| tristan | Scope.BUILD does not | 09:59 |
| juergbi | probably rather niche use case but it's not unreasonable | 09:59 |
| juergbi | right, I forgot about that | 09:59 |
| benschubert | hey tristan, juergbi I documented the new 'MappingNode' class on https://gitlab.com/BuildStream/buildstream/merge_requests/1467/diffs, I'll tackle the other classes ASAP, but I'd appreciate early reviews on the documentation style if you have time! Might be better to check it out and view the built documentation | 09:59 |
| juergbi | will try to take a look later today | 10:00 |
| benschubert | thanks! | 10:00 |
| juergbi | tristan: ok, I'm happy to aim towards run/all/build. bst build --deps run would be slightly different from other commands as it would still include build dependencies that are needed, of course. but that shouldn't be an issue | 10:01 |
| juergbi | tristan: we could add a userconfig option with just `run` (default) and `all` first, and tackle `build` and the CLI syntax switch separately. could also do all at once, I suppose | 10:02 |
| tristan | juergbi, right, it could be documented as the build/run scope only relative to the target (build-of-build-of-build dependencies are always on demand unless `all` is specified) | 10:02 |
| tristan | bla | 10:02 |
| tristan | :) | 10:02 |
| juergbi | yes, that makes sense | 10:03 |
| juergbi | mostly matching the other commands except the on demand indirect build dependencies | 10:03 |
| juergbi | I don't think anyone will ever want to put deps `build` in userconfig, but we should still support it for consistency with CLI | 10:05 |
| *** tristan has quit IRC | 10:07 | |
| *** tristan has joined #buildstream | 10:14 | |
| *** ChanServ sets mode: +o tristan | 10:15 | |
| gitlab-br-bot | traveltissues opened (was WIP) MR !1466 (traveltissues/cache-key-changes->master: element.py: changes to __cache_key_dict) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1466 | 10:18 |
| *** lachlan has quit IRC | 10:33 | |
| *** lachlan has joined #buildstream | 10:33 | |
| *** lachlan has quit IRC | 10:45 | |
| *** lachlan has joined #buildstream | 10:46 | |
| *** lachlan has quit IRC | 10:55 | |
| *** lachlan has joined #buildstream | 10:57 | |
| *** lachlan has quit IRC | 11:21 | |
| *** lachlan has joined #buildstream | 11:31 | |
| *** bochecha has quit IRC | 11:38 | |
| *** dftxbs3e has quit IRC | 12:25 | |
| *** dftxbs3e has joined #buildstream | 12:25 | |
| *** bochecha has joined #buildstream | 12:55 | |
| *** tristan has quit IRC | 13:00 | |
| *** lachlan has quit IRC | 14:30 | |
| *** lachlan has joined #buildstream | 14:43 | |
| traveltissues | can i please get reviews on https://gitlab.com/BuildStream/buildstream/merge_requests/1468 and https://gitlab.com/BuildStream/buildstream/merge_requests/1466 | 15:05 |
| gitlab-br-bot | jennis opened (was WIP) MR !1449 (jennis/update_ready_for_runtime->master: Remove __update_state_recursively()) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1449 | 15:09 |
| *** bochecha has quit IRC | 15:16 | |
| *** bochecha has joined #buildstream | 15:17 | |
| gitlab-br-bot | shashwatdalal opened issue #1075 (Stack elements permit `build-depends`) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1075 | 15:40 |
| gitlab-br-bot | juergbi opened issue #1076 (Add remote-execution section to userconfig.yaml) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1076 | 15:48 |
| *** bochecha has quit IRC | 15:51 | |
| *** lachlan has quit IRC | 16:03 | |
| *** lachlan has joined #buildstream | 16:10 | |
| gitlab-br-bot | traveltissues approved MR !1452 (raoul/994-further-re-testing->master: Further RE testing) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1452 | 16:11 |
| gitlab-br-bot | shashwatdalal opened issue #1077 (Build skipped when element has duplicate build dependencies.) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1077 | 16:11 |
| gitlab-br-bot | juergbi opened issue #1078 (Replace bst build --all with bst build --deps all) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1078 | 16:13 |
| *** alexandrufazakas has quit IRC | 16:22 | |
| Kinnison | I can look in the morning | 16:51 |
| *** jonathanmaw has quit IRC | 17:03 | |
| gitlab-br-bot | marge-bot123 closed issue #1043 (Configuration option for disabling blob fetching with RE) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1043 | 17:04 |
| gitlab-br-bot | marge-bot123 merged MR !1402 (raoul/1043-disabling-blob-fetching->master: Configuration option for disabling blob fetching with RE) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1402 | 17:04 |
| *** traveltissues has quit IRC | 17:20 | |
| benschubert | tristan: when you have time: https://gitlab.com/BuildStream/buildstream/merge_requests/1467, the public API is done, still need the private one | 17:23 |
| *** lachlan has quit IRC | 17:56 | |
| *** tristan has joined #buildstream | 18:01 | |
| *** tristan_ has joined #buildstream | 18:01 | |
| *** ChanServ sets mode: +o tristan | 18:05 | |
| tristan | tristan_, ping | 18:05 |
| tristan | tristan_, I don't see your xchat pid on my computer... would you please pop up a notification on my desktop so I can kill the duplicate instance of hexchat ? | 18:07 |
| * tristan only sees one | 18:07 | |
| *** tristan_ has quit IRC | 18:09 | |
| *** tristan has quit IRC | 18:09 | |
| *** tristan has joined #buildstream | 18:09 | |
| *** ChanServ sets mode: +o tristan | 18:10 | |
| tristan | whoa that was weird, some GNOME 3.30 bugs I guess | 18:10 |
| gitlab-br-bot | valentindavid opened (was WIP) MR !1146 (valentindavid/local-cache-exec-leak->master: Deduplicate files in local cache with or without exec rights) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1146 | 18:14 |
| *** bochecha has joined #buildstream | 18:26 | |
| *** admin has left #buildstream | 18:26 | |
| *** admin has joined #buildstream | 18:26 | |
| *** dtf has joined #buildstream | 18:37 | |
| gitlab-br-bot | valentindavid opened MR !1471 (valentindavid/chroot_fixes->master: Some small fixes in chroot sandbox) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1471 | 18:39 |
| *** bochecha has quit IRC | 21:26 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!