| *** tristan_ has joined #buildstream | 04:11 | |
| *** tristan_ has quit IRC | 04:12 | |
| *** tristan_ has joined #buildstream | 04:12 | |
| *** tristan_ has quit IRC | 04:14 | |
| *** tristan_ has joined #buildstream | 04:14 | |
| *** tristan_ has quit IRC | 04:16 | |
| *** tristan_ has joined #buildstream | 04:16 | |
| *** tristan_ has quit IRC | 04:18 | |
| *** tristan_ has joined #buildstream | 04:21 | |
| *** tristan_ has quit IRC | 04:22 | |
| *** tristan_ has joined #buildstream | 04:22 | |
| *** tristan_ has quit IRC | 04:23 | |
| *** tristan_ has joined #buildstream | 04:24 | |
| *** tristan_ has quit IRC | 04:25 | |
| *** tristan_ has joined #buildstream | 04:26 | |
| *** tristan_ has quit IRC | 04:27 | |
| *** tristan_ has joined #buildstream | 04:31 | |
| *** tristan_ has quit IRC | 04:32 | |
| *** tristan_ has joined #buildstream | 04:32 | |
| *** tristan_ has quit IRC | 05:37 | |
| *** tristan_ has joined #buildstream | 05:38 | |
| *** tristan_ has quit IRC | 05:42 | |
| *** tristan_ has joined #buildstream | 05:43 | |
| *** tristan_ has quit IRC | 05:50 | |
| *** tristan_ has joined #buildstream | 05:50 | |
| *** tristan_ has quit IRC | 05:51 | |
| *** tristan_ has joined #buildstream | 05:52 | |
| *** tristan_ has quit IRC | 05:57 | |
| *** tristan_ has joined #buildstream | 05:59 | |
| *** tristan_ has quit IRC | 06:00 | |
| *** tristan_ has joined #buildstream | 06:01 | |
| *** tristan_ has quit IRC | 06:02 | |
| *** tristan_ has joined #buildstream | 06:03 | |
| *** tristan_ has quit IRC | 06:04 | |
| *** tristan_ has joined #buildstream | 06:05 | |
| *** tristan_ has quit IRC | 06:06 | |
| *** tristan_ has joined #buildstream | 06:07 | |
| *** tristan_ has quit IRC | 06:08 | |
| *** tristan_ has joined #buildstream | 06:09 | |
| *** tristan_ has quit IRC | 06:10 | |
| *** tristan_ has joined #buildstream | 06:11 | |
| *** tristan_ has quit IRC | 06:12 | |
| *** tristan_ has joined #buildstream | 06:13 | |
| *** tristan_ has quit IRC | 06:16 | |
| *** tristan_ has joined #buildstream | 06:17 | |
| *** tristan_ has quit IRC | 06:18 | |
| *** tristan_ has joined #buildstream | 06:19 | |
| *** tristan_ has quit IRC | 06:20 | |
| *** tristan_ has joined #buildstream | 06:21 | |
| *** tristan_ has quit IRC | 06:22 | |
| *** tristan_ has joined #buildstream | 06:23 | |
| *** tristan_ has quit IRC | 06:24 | |
| *** tristan_ has joined #buildstream | 06:27 | |
| *** tristan_ has quit IRC | 06:28 | |
| *** tristan_ has joined #buildstream | 06:36 | |
| *** tristan_ has quit IRC | 06:39 | |
| gitlab-br-bot | coldtom opened MR !1538 (coldtom/move-makemaker->master: Move makemaker plugin to bst-plugins-experimental) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1538 | 07:06 |
|---|---|---|
| *** bochecha has joined #buildstream | 07:47 | |
| gitlab-br-bot | marge-bot123 merged MR !1538 (coldtom/move-makemaker->master: Move makemaker plugin to bst-plugins-experimental) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1538 | 08:31 |
| *** tme5 has joined #buildstream | 08:35 | |
| *** tiagogomes has quit IRC | 08:38 | |
| gitlab-br-bot | coldtom opened MR !1539 (coldtom/move-modulebuild->master: Move modulebuild plugin to bst-plugins-experimental) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1539 | 08:48 |
| *** tiagogomes has joined #buildstream | 08:48 | |
| *** traveltissues has joined #buildstream | 08:58 | |
| *** tpollard has joined #buildstream | 08:58 | |
| *** jonathanmaw has joined #buildstream | 08:59 | |
| *** tpollard has quit IRC | 08:59 | |
| *** tpollard has joined #buildstream | 09:01 | |
| *** phildawson_ has joined #buildstream | 09:06 | |
| *** juergbi has quit IRC | 09:07 | |
| *** juergbi has joined #buildstream | 09:12 | |
| tme5 | is the wildcard expansion that bst does documented anywhere? eg bst show '*.bst' | 09:12 |
| qinusty | I'd assume it's just a standard glob? bst commands work on files within the root element directory which is defined in the project.conf | 09:22 |
| *** lachlan has joined #buildstream | 09:22 | |
| Kinnison | if you quote the glob, the glob is passed to buildstream | 09:23 |
| Kinnison | rather than being shell expanded | 09:23 |
| jennis | tme5, I know that it's documented in `artifact log` and `artifact delete` | 09:24 |
| jennis | But perhaps it should be implemented at a higher level, as it works for element names as well as artifact refs | 09:24 |
| tme5 | ahh, brilliant thanks | 09:25 |
| jennis | https://buildstream.gitlab.io/buildstream/using_commands.html#bst-artifact-log | 09:25 |
| jennis | tme5 ^ | 09:25 |
| tme5 | it is in fact implemented at a higher level | 09:25 |
| tme5 | and this bug is a result of that https://gitlab.com/BuildStream/buildstream/issues/959 | 09:25 |
| tme5 | well, i guess it's not technically a bug | 09:25 |
| tme5 | but thank you for pointing me to the docs. Do you think it would be a good idea to make it more clear where this expansion happens, and to document it more generally? | 09:26 |
| jennis | Yes definitely | 09:27 |
| jennis | And I think that issue is correct | 09:27 |
| jennis | bst show * shouldn't error out like that, because we do support element names | 09:28 |
| jennis | In short, I think `bst show *` == `bst show *.bst` behaviour-wise | 09:28 |
| jennis | Which is also equivalent to the currently supported `bst show` IIRC | 09:28 |
| tme5 | so there should be two types of wildcard expansion? | 09:29 |
| tme5 | most shells could do `bst show **.bst`, right? | 09:30 |
| tme5 | I feel like it's clearer behaviour to have bst do only artifact ref expansion, where applicable | 09:31 |
| jennis | It seems that the two types are alrady kind of supported | 09:32 |
| Kinnison | tme5: ** implies an understanding of a structure on the filesystem -- buildstream's elements are a flat namespace that happens to contain slashes as valid characters in a limited fashion | 09:32 |
| tme5 | so there are use cases where elements do not come from files? | 09:33 |
| *** lachlan has quit IRC | 09:56 | |
| *** lachlan has joined #buildstream | 10:00 | |
| *** phoenix has joined #buildstream | 10:07 | |
| *** phoenix has quit IRC | 10:13 | |
| *** phoenix has joined #buildstream | 10:14 | |
| *** phoenix has quit IRC | 10:18 | |
| *** phoenix has joined #buildstream | 10:20 | |
| *** phoenix has quit IRC | 10:22 | |
| *** phoenix_ has joined #buildstream | 10:23 | |
| *** phoenix_ is now known as phoenix | 10:23 | |
| jjardon | qinusty: hey, I've heard you were recollecting what is remaining from buildstream to supporting remote-execution API's; did you have a chance to write that somewhere? | 10:31 |
| qinusty | jjardon partially, there appears to be two main issues, caching which is a discussion you were involved in and support of sandboxed environments from RE server implementations | 10:32 |
| *** phoenix has quit IRC | 10:32 | |
| qinusty | I'll tidy up what I've got and let you know | 10:33 |
| adds68 | Does anyone know if there is an endpoint you can query in the artifact-cache that will give a reliable "health status"? | 10:33 |
| adds68 | i can curl 1101 and 1102, but they return an empty response, which is an error =/ | 10:33 |
| jjardon | qinusty: cool, I was thinking maybe we can use https://gitlab.com/BuildStream/buildstream/issues/1050 for the remote caching and then we can open a new one for the remote execution part, as probably is more complicated? | 10:34 |
| *** lachlan has quit IRC | 10:35 | |
| qinusty | Yes that issue encapsulates basically everything I know about the caching issue so it can primarily reside there. I think the issues with sandboxing can be addressed in a single issue detailing each client. I haven't looked into buildfarm as an option yet though | 10:37 |
| *** lachlan has joined #buildstream | 10:39 | |
| *** phoenix has joined #buildstream | 10:47 | |
| *** phoenix has quit IRC | 10:54 | |
| *** lachlan has quit IRC | 11:07 | |
| *** lachlan has joined #buildstream | 11:11 | |
| *** lachlan has quit IRC | 11:14 | |
| *** phoenix has joined #buildstream | 11:18 | |
| *** phoenix has quit IRC | 11:19 | |
| gitlab-br-bot | traveltissues approved 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 | 11:22 |
| *** jonathanmaw has quit IRC | 11:23 | |
| *** jonathanmaw has joined #buildstream | 11:23 | |
| *** phoenix has joined #buildstream | 11:26 | |
| *** lachlan has joined #buildstream | 11:31 | |
| jennis | juergbi, are there installation instructions for bb-casd? | 11:33 |
| tlater[m] | adds68: This is for bst 1.2, right? | 11:34 |
| juergbi | jennis: not yet but this may help: https://gitlab.com/BuildStream/buildstream-docker-images/commit/57ce3e339bca3839ab52598121fdcc2702c5cdb4#7c1688a0c480b7b1d571df7be349080b1122e041 | 11:35 |
| tlater[m] | It's a little complicated because it uses grpc. I'm not 100% sure what http calls those things turn into. | 11:35 |
| tlater[m] | juergbi: Any idea? | 11:35 |
| jennis | thanks juergbi | 11:36 |
| tlater[m] | There is StatusRequest, but well, it's grpc :) | 11:36 |
| tlater[m] | adds68: Looks like you'd need to do something along these lines: https://medium.com/google-cloud/grpc-with-curl-a65d879a18f7 | 11:40 |
| tlater[m] | Not nice, to say the least | 11:40 |
| tlater[m] | It's probably easiest to write a tiny python script that uses grpc to make a StatusRequest | 11:41 |
| * tlater[m] is a little annoyed at this | 11:41 | |
| *** phoenix has quit IRC | 11:41 | |
| *** lachlan has quit IRC | 11:42 | |
| *** phoenix has joined #buildstream | 11:48 | |
| *** phoenix has quit IRC | 11:50 | |
| *** lachlan has joined #buildstream | 11:51 | |
| jennis | juergbi, after your branch lands, we will no longer be supporting old style bst-artifact-server, correct? | 11:51 |
| jennis | If I've understood the changes to casremote.py correctly | 11:51 |
| juergbi | jennis: as a client? that's not supported in master either | 11:52 |
| jennis | Thought it currently was | 11:52 |
| jennis | juergbi, for the remotes ^ | 11:53 |
| juergbi | no, the server still supports older clients but the new bst client never checks non-proto artifacts on the remote | 11:53 |
| juergbi | cache keys are anyway different, so there isn't really a point | 11:53 |
| jennis | yeah sorry. Will the server still be supporting older clients? | 11:53 |
| jennis | After your MR lands | 11:53 |
| juergbi | yes, that should still be the case | 11:53 |
| juergbi | although, it might not be covered by CI | 11:54 |
| juergbi | do you see a reason why the MR would break that? | 11:54 |
| jennis | ok, so we still need all the "CASCache.contains(), CASCache.resolve_ref()" stuff in CASCache | 11:54 |
| *** lachlan has quit IRC | 11:54 | |
| gitlab-br-bot | traveltissues approved MR !1536 (jennis/fix_failed_workspaces->master: Don't reset a failed (but cached) workspaced Element) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1536 | 11:54 |
| jennis | No I don't | 11:54 |
| juergbi | yes | 11:54 |
| *** phoenix has joined #buildstream | 12:01 | |
| *** phoenix has quit IRC | 12:05 | |
| gitlab-br-bot | tlater opened MR !1540 (tlater/cache-endpoints->master: Tlater/cache endpoints) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1540 | 12:09 |
| gitlab-br-bot | marge-bot123 merged MR !1531 (tmewett/report-weird-return-codes->master: Report error when job process unexpectedly dies) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1531 | 12:18 |
| gitlab-br-bot | marge-bot123 closed issue #1089 (Reporting of unexpected deaths of worker processes could be better) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1089 | 12:18 |
| adds68 | tlater[m], i ended up working with ironfoot to make a little one liner shell command to make the empty response truthy | 12:45 |
| adds68 | tlater[m], should i open an issue for this? ;D | 12:45 |
| tlater[m] | adds68: Yeah, sure. It probably won't see much traction for a while, but it's stupid that we can't see if our server is alive. | 12:46 |
| adds68 | cool | 12:46 |
| * adds68 adds to list of TODOs | 12:46 | |
| tlater[m] | Once is enough | 12:46 |
| tlater[m] | Don't need 68 of them ;p | 12:46 |
| adds68 | tlater[m] haha! | 12:47 |
| ironfoot | this ideally can be added into the Dockerfile | 12:48 |
| ironfoot | I think it's possible to use curl to fake a grpc call, but won't look good | 12:48 |
| tlater[m] | ironfoot: I wouldn't like it if it were a docker-only feature | 12:53 |
| tlater[m] | IMO it should be added to bst-artifact-server, perhaps as a custom little http server thing that relays everything to GRPC except a specific curl request. | 12:54 |
| ironfoot | yes, that would be the ideal solution | 12:58 |
| tpollard | +1 | 12:58 |
| *** lachlan has joined #buildstream | 13:04 | |
| *** traveltissues has quit IRC | 13:05 | |
| *** traveltissues has joined #buildstream | 13:06 | |
| *** narispo has quit IRC | 13:24 | |
| *** narispo has joined #buildstream | 13:24 | |
| *** lachlan has quit IRC | 13:46 | |
| *** lachlan has joined #buildstream | 13:55 | |
| *** lachlan has quit IRC | 13:58 | |
| *** lachlan has joined #buildstream | 14:07 | |
| *** lachlan has quit IRC | 14:13 | |
| tlater[m] | Is there something like `bst source push`? | 14:33 |
| tlater[m] | I feel like I got rid of it accidentally somehow | 14:33 |
| *** lachlan has joined #buildstream | 14:34 | |
| tpollard | I dont think so tlater | 14:34 |
| tpollard | afaik only bst build makes a SourcePush queue | 14:35 |
| tlater[m] | That seems a little odd | 14:35 |
| *** narispo has quit IRC | 14:37 | |
| tpollard | I tend to agree | 14:37 |
| *** narispo has joined #buildstream | 14:37 | |
| tlater[m] | Well, I suppose I'm in the best position to figure out how to change that. | 14:38 |
| * tlater[m] digs into the frontend | 14:38 | |
| *** lachlan has quit IRC | 14:44 | |
| *** lachlan has joined #buildstream | 14:51 | |
| tme5 | i'm noticing, especially in bst artifact subcommand help text, that "artifact" and "element" are often used interchangably | 14:56 |
| *** lachlan has quit IRC | 14:57 | |
| tme5 | i'm working on making more of a distinction, is that useful? | 15:03 |
| tme5 | especially with the fact the commands take elements and artifact refs interchangably | 15:07 |
| *** lachlan has joined #buildstream | 15:10 | |
| *** tristan_ has joined #buildstream | 15:14 | |
| tpollard | tme5: I think it's useful | 15:15 |
| jennis | yeah tme5, that will be more useful | 15:15 |
| tpollard | especially as internally the distinction of element-artifact is now more abstracted in the classes | 15:16 |
| *** lachlan has quit IRC | 15:17 | |
| *** lachlan has joined #buildstream | 15:31 | |
| *** traveltissues has quit IRC | 15:36 | |
| *** phildawson_ has quit IRC | 15:36 | |
| *** phildawson_ has joined #buildstream | 15:37 | |
| *** tristan_ has quit IRC | 15:38 | |
| *** tristan_ has joined #buildstream | 15:39 | |
| *** lachlan has quit IRC | 15:44 | |
| *** lachlan has joined #buildstream | 15:48 | |
| tlater[m] | Hrm, I just realized that we have a source-servers and an artifacts configuration setting in project.conf | 15:51 |
| tlater[m] | I wonder if it'd make more sense to name that "thing" remotes and distinguish between source and artifact remotes under that key | 15:52 |
| tlater[m] | It would resolve the inconsistency in their naming and make it more obvious what they do | 15:53 |
| tlater[m] | I should say, remote-caches | 15:53 |
| *** lachlan has quit IRC | 16:03 | |
| jennis | and also aliases for remotes? :D | 16:09 |
| *** bochecha has quit IRC | 16:11 | |
| *** lachlan has joined #buildstream | 16:16 | |
| tlater[m] | jennis: Honestly, I think that'd be awesome | 16:20 |
| tlater[m] | Would make error messages more informative and --remote easier to use. | 16:21 |
| tlater[m] | Maybe... | 16:21 |
| * tlater[m] sent a long message: < https://gnome.modular.im/_matrix/media/v1/download/gnome.org/hRIIbjgmwVNcgPccmMlVPocp > | 16:23 | |
| tlater[m] | And then add cas/meta split to the individual caches :) | 16:23 |
| tlater[m] | Hell, you could define them at the root level and then tell them to be source/artifact caches with the alias. | 16:25 |
| *** lachlan has quit IRC | 16:28 | |
| *** lachlan has joined #buildstream | 16:32 | |
| *** tristan_ has quit IRC | 16:38 | |
| gitlab-br-bot | tmewett opened MR !1541 (tmewett/artifact-help->master: Clarify bst artifact subcommand help text) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1541 | 16:41 |
| *** lachlan has quit IRC | 16:42 | |
| *** tme5 has quit IRC | 16:44 | |
| *** tristan_ has joined #buildstream | 17:04 | |
| adds68 | I'm seeing "task: non-zero exit (1)" returned from the bst-artifact-cache docker images | 17:06 |
| adds68 | Would anyone have any idea what could cause that? | 17:06 |
| *** lachlan has joined #buildstream | 17:13 | |
| *** lachlan has quit IRC | 17:19 | |
| *** lachlan has joined #buildstream | 17:27 | |
| *** jonathanmaw has quit IRC | 17:34 | |
| *** tristan_ has quit IRC | 17:38 | |
| *** lachlan has quit IRC | 18:02 | |
| *** lachlan has joined #buildstream | 18:22 | |
| *** tristan_ has joined #buildstream | 18:36 | |
| *** tristan_ has quit IRC | 18:37 | |
| *** dftxbs3e has quit IRC | 19:23 | |
| *** tristan_ has joined #buildstream | 19:36 | |
| *** tristan_ has quit IRC | 19:38 | |
| *** lachlan has quit IRC | 19:59 | |
| *** tristan_ has joined #buildstream | 20:31 | |
| *** cs-shadow has quit IRC | 20:36 | |
| *** tristan_ has quit IRC | 20:38 | |
| *** tristan_ has joined #buildstream | 20:39 | |
| gitlab-br-bot | tristanvb opened MR !1542 (tristan/strict-rebuild->master: Support strict build dependencies) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1542 | 20:57 |
| *** ChanServ sets mode: +o tristan | 20:59 | |
| tristan | juergbi, I was thinking the above might be nice to have in 1.4, as we're using non-strict in CI now... any thoughts ? | 20:59 |
| * tristan thinks it feels silent today, like I didnt get the memo that it was actually a second sunday in a row :-/ | 20:59 | |
| *** tristan_ has quit IRC | 21:38 | |
| *** tristan_ has joined #buildstream | 21:38 | |
| *** tristan_ has quit IRC | 21:40 | |
| *** tristan_ has joined #buildstream | 21:40 | |
| *** tristan_ has quit IRC | 21:43 | |
| *** tristan_ has joined #buildstream | 21:44 | |
| *** tristan_ has quit IRC | 21:45 | |
| *** tristan_ has joined #buildstream | 21:45 | |
| *** tristan_ has quit IRC | 21:47 | |
| *** tristan_ has joined #buildstream | 22:36 | |
| *** tristan_ has quit IRC | 22:38 | |
| *** tristan_ has joined #buildstream | 23:36 | |
| *** tristan_ has quit IRC | 23:38 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!