| *** rdale has quit IRC | 00:08 | |
| *** xjuan has quit IRC | 01:16 | |
| *** tristan has joined #buildstream | 03:08 | |
| *** tristan has quit IRC | 04:33 | |
| *** ctolentino has quit IRC | 05:58 | |
| *** cs-shadow has quit IRC | 05:58 | |
| *** lchlan has quit IRC | 05:58 | |
| *** ctolentino has joined #buildstream | 05:58 | |
| *** cs-shadow has joined #buildstream | 05:58 | |
| *** lchlan has joined #buildstream | 05:58 | |
| *** alatiera_ has joined #buildstream | 08:12 | |
| Kinnison | juergbi: It ought to be pretty simple to spot and eliminate workspace path prefixes from scanning | 08:58 |
|---|---|---|
| juergbi | that's true | 08:58 |
| juergbi | probably makes sense to do this | 08:59 |
| *** rdale has joined #buildstream | 09:16 | |
| *** toscalix has joined #buildstream | 09:24 | |
| gitlab-br-bot | valentindavid opened (was WIP) MR !862 (valentindavid/link_files_sort_resolved->master: Resolve paths before ordering them in link_files/copy_files) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/862 | 09:31 |
| *** jonathanmaw has joined #buildstream | 09:54 | |
| tiagogomes | tpollard wat | 09:59 |
| tpollard | tiagogomes: you were asking if bst could have the option to build without trying to use artifact servers | 10:10 |
| tiagogomes | aah, missed the MR, thanks for that | 10:11 |
| tpollard | np :) | 10:12 |
| *** raoul has joined #buildstream | 10:21 | |
| *** lachlan has joined #buildstream | 10:27 | |
| *** tristan has joined #buildstream | 10:37 | |
| jennis | tiagogomes are you still around? | 10:54 |
| tiagogomes | yes | 10:55 |
| jennis | You did a commit a while back: plugin: bake API to get and validate a project path | 10:55 |
| jennis | which basically ensures that any `path:` we've declared in a bst will need to be contained within the project | 10:56 |
| jennis | was the intent of this to also restrict local junctions to be hosted within the project? Because I don't think I agree with this | 10:56 |
| *** toscalix has quit IRC | 10:57 | |
| tiagogomes | It was. I have committed a change to freedesktop-sdk for that | 10:58 |
| jennis | For example, I have the freedesktop-junction.bst in my gnome-build-meta project, now instead of junctioning to the stable version on git, I want to junction locally to the version I have. And providing the absolute path to this project throws this exception | 10:58 |
| jennis | I also noticed the change preventing '../path/to/somewhere' | 10:58 |
| jennis | I don't see why we're restricting users to keep junctions (other BuildStream projects) inside of their current project? | 10:59 |
| *** alatiera__ has joined #buildstream | 11:00 | |
| tiagogomes | How can you ensure reproducibility if you are using the content of random directories in your host file system | 11:00 |
| *** alatiera_ has quit IRC | 11:01 | |
| *** toscalix has joined #buildstream | 11:01 | |
| Kinnison | You can't, but you may need to do the latter without care for the former | 11:01 |
| tiagogomes | jennis can't you still use a junction with a git source, but this file passing a filepath in "url:"? | 11:04 |
| tiagogomes | That should work | 11:05 |
| tiagogomes | You'd still need to have to commit the changes in your clone of freedesktop-sdk, but you don't need to push them | 11:06 |
| jennis | Heh, I could, but that's not the point | 11:07 |
| jennis | It's more convenient for me as a user to just use my local version, and make changes as and when without having to push to a branch | 11:08 |
| jennis | I understand both arguments and it'd be nice if more people would weigh in with their opinion | 11:09 |
| jennis | s/would/could | 11:09 |
| *** brlogger has joined #buildstream | 11:15 | |
| tiagogomes | jennis tbc, you don't need to push them to a branch. You just need create a commit with the changes | 11:16 |
| jennis | And then specify a file url to the junction'd git project? | 11:19 |
| jennis | This isn't the point I'm making though, the point is whether we should be this restrictive :p | 11:20 |
| tiagogomes | File an issue then :) | 11:21 |
| jennis | Will do | 11:25 |
| abderrahim[m] | jennis I think you want to use a workspace | 11:25 |
| jennis | abderrahim[m], yes this also would work, thanks | 11:31 |
| toscalix | jennis: maybe the ML help you to get input to define the use case | 11:34 |
| juergbi | tpollard: I'm a bit confused about the CLI option naming in !786. it's well possible that I missed a comment about this but in #494 I see --pull-buildtrees, not --pull-build-trees, and in !786 you comment that you're not completely happy with --pull-build-trees. what was the reason to use --pull-build-trees instead of --pull-buildtrees in the first place? | 12:15 |
| WSalmon | tlater[m], ctolentino has just posted https://mail.gnome.org/archives/buildstream-list/2018-November/msg00038.html to the ML, the last paragraph talks about he cools ideas we were discussing about dealing with test and reusing build trees etc | 12:16 |
| juergbi | tpollard: I'm starting to think --pull-buildtrees would make most sense. as treating buildtree as single word sounds sensible enough and we already use 'buildtree' e.g. as subdirectory name (that's not user-visible, though) | 12:16 |
| juergbi | (don't really want to start bikeshedding. main point is consistency, see my latest comment on the MR) | 12:16 |
| *** gitlab-br-bot has joined #buildstream | 12:20 | |
| tpollard | I got the impression that build-trees were preferred as separate words | 12:25 |
| tpollard | I can change it | 12:25 |
| tpollard | juergbi: to match https://gitlab.com/BuildStream/buildstream/merge_requests/786#note_113769559 | 12:27 |
| juergbi | tpollard: however, before that, tristan proposed or at least mentioned --pull-buildtrees, though. I think in the MR comment it was more about using _ as separator where we want separation | 12:30 |
| juergbi | tristan: if you happen to be around: quick question, --pull-buildtrees or --pull-build-trees as CLI option (and identical spelling for userconfig, of course)? | 12:31 |
| tpollard | my preference is --pull-buildtrees | 12:35 |
| *** alatiera__ has quit IRC | 12:41 | |
| juergbi | mine too. jonathanmaw also mentioned unhappiness with "pullbuildtrees" | 13:00 |
| tpollard | I've done a rework locally, will check everything still passes in the meantime | 13:01 |
| juergbi | great | 13:01 |
| tiagogomes | I also prefer --pull-buildtrees | 13:01 |
| tpollard | I left a reply to your comment re setting a default value of something parsed from userconf data | 13:01 |
| tpollard | oh, got the reply just now... | 13:02 |
| juergbi | I'll make a quick branch for the quota thing to avoid future confusion | 13:03 |
| tpollard | cool, I'll drop the default | 13:05 |
| *** lachlan has quit IRC | 13:13 | |
| *** ChanServ sets mode: +o tristan | 13:16 | |
| tristan | juergbi, tpollard --pull-buildtrees is fine with me... for user config I'm a bit concerned about grouping | 13:16 |
| juergbi | yes, I just commented about this | 13:16 |
| juergbi | of the existing groups, 'cache' might fit but not sure | 13:16 |
| juergbi | or we could create a new group but don't have a concrete idea | 13:16 |
| tristan | myeah, is it 'cache' or is it 'artifacts' ? | 13:17 |
| tristan | I guess it must be 'cache' | 13:17 |
| *** alatiera__ has joined #buildstream | 13:20 | |
| juergbi | tristan: yes, artifacts is a list, not a group | 13:20 |
| juergbi | I think cache is a reasonable fit | 13:21 |
| tristan | juergbi, sounds good to me | 13:21 |
| tpollard | ok | 13:23 |
| gitlab-br-bot | juergbi opened MR !953 (juerg/context-default-values->master: _context.py: Drop duplicated default values for user configuration) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/953 | 13:25 |
| *** alatiera_ has joined #buildstream | 13:34 | |
| *** jonathanmaw has quit IRC | 14:00 | |
| *** jonathanmaw has joined #buildstream | 14:00 | |
| *** finn_ has joined #buildstream | 14:00 | |
| *** phildawson_ has quit IRC | 14:01 | |
| *** phildawson_ has joined #buildstream | 14:01 | |
| *** finn has quit IRC | 14:02 | |
| *** tristan has quit IRC | 14:06 | |
| *** alatiera_ has quit IRC | 14:12 | |
| *** tristan has joined #buildstream | 14:27 | |
| * paulsherwood wonders what could lead to "pipeline: failed" on main badges at https://gitlab.com/BuildStream/buildstream | 14:34 | |
| Kinnison | OSError: [Errno 30] Read-only file system: 'debugedit-disable-md2.patch' | 14:37 |
| Kinnison | Looks like the runner had a snafu | 14:37 |
| Kinnison | I think valentind and possibly jjardon are aware and on it | 14:38 |
| juergbi | there is also the issue that we have two different CI pipelines for the same commit (normal and overnight) and the badge simply shows the latest pipeline | 14:39 |
| juergbi | the overnight one tests bst against fdo sdk and that's the one that failed | 14:40 |
| valentind | Something wrong with one builder. | 14:40 |
| valentind | Ah, it is the filesystem of the builder that gets back in readonly. | 14:40 |
| valentind | It happened already. | 14:40 |
| valentind | When operations fix crash usually it remounts read only., | 14:40 |
| valentind | I will ask Dave. | 14:42 |
| valentind | Rebooted the machine. Now it is mounted rw. We will see if it passes this night. | 14:44 |
| *** lachlan has joined #buildstream | 14:47 | |
| jjardon | valentind: FYI, you can trigger the schelude jobs manually at any point (simply press "play" in the scheludes page). I have triggered https://gitlab.com/BuildStream/buildstream/pipelines/36738166 and https://gitlab.com/BuildStream/buildstream/pipelines/36738158 to check if it's fixed now | 14:55 |
| valentind | jjardon, There is only one builder for the moment. So I suppose one of them will just timeout. | 14:56 |
| jjardon | valentind: I though buildstream used to have 2 for the overnight tests? | 14:57 |
| valentind | jjardon, Yes, but we are still waiting for operations to split the storage. So one is down because it does not have space. | 14:58 |
| jjardon | valentind: let's setup something in DO, so we do not need to wait? It would take 10 min maximum | 14:59 |
| valentind | Would freedesktop sdk need more resource than a the builders provided by the bastion? | 15:04 |
| jjardon | valentind: freedesktop-sdk doesnt share any infra with buildstream | 15:05 |
| valentind | jjardon, builstream has builders on DO. | 15:05 |
| valentind | I am talking about those. | 15:05 |
| jjardon | valentind: ah sorry | 15:11 |
| jjardon | misunderstood | 15:11 |
| *** alatiera_ has joined #buildstream | 15:11 | |
| jjardon | yeah, so we create a new bastion that will generate runners with the new spec | 15:11 |
| jjardon | valentind: I have already done: see buildstream-bastion-overnight at https://gitlab.com/BuildStream/buildstream/settings/ci_cd#js-runners-settings (It's paused at the moment) | 15:12 |
| *** alatiera_ has joined #buildstream | 15:12 | |
| jjardon | valentind: if you are ok, active it and give it a try | 15:12 |
| valentind | jjardon, let me look at the configuration. | 15:13 |
| jjardon | valentind: It will generate machines with 4 CPU, 8GB of RAM and 160 GB of hard drive: that should be enough to build freedesktop-sdk | 15:14 |
| gitlab-br-bot | coldtom opened issue #767 (Yaml Dumper removes blank lines seemingly arbitrarily) on buildstream https://gitlab.com/BuildStream/buildstream/issues/767 | 15:14 |
| valentind | jjardon, OK | 15:15 |
| valentind | We need to tag it | 15:15 |
| valentind | Oh it is. | 15:15 |
| valentind | Good | 15:15 |
| valentind | jjardon, You can start it. | 15:15 |
| jjardon | cool, sec | 15:16 |
| valentind | Oh gitlab-runner does not clean correctly docker container. There are also failures on the old builder due to that. | 15:17 |
| valentind | I have a fix for the docker issues on the arm builders. I will apply it tomorrow. | 15:24 |
| jjardon | valentind: new runner seems to work well: https://gitlab.com/BuildStream/buildstream/-/jobs/121057968 | 15:30 |
| jjardon | valentind: I have paused the codethink ones (codethink-x86_64-1 and codethink-x86_64-2), feel free to tell Codethink ops to remove them when you think new ones are tested enough | 15:31 |
| valentind | What do we do with them then? | 15:33 |
| jjardon | valentind: with codethink-x86_64-1 and codethink-x86_64-2? nothing; they will take control back of them and use them for another stuff | 15:36 |
| valentind | ok | 15:36 |
| toscalix | valentind: please recod the deprecation in the tickets | 15:49 |
| *** alatiera_ has quit IRC | 15:50 | |
| *** alatiera_ has joined #buildstream | 15:54 | |
| *** alatiera_ has quit IRC | 15:57 | |
| * jjardon updates https://wiki.gnome.org/Projects/BuildStream/Infrastructure | 15:57 | |
| jjardon | valentind: are codethink-x86_64-1 and codethink-x86_64-2 gone already or you prefer to keept them for a while? | 15:58 |
| gitlab-br-bot | jjardon closed issue #763 (Overnigth tests are failing) on buildstream https://gitlab.com/BuildStream/buildstream/issues/763 | 16:01 |
| jjardon | toscalix: what ticket number, please? | 16:05 |
| toscalix | checking | 16:05 |
| toscalix | jjardon: there is none | 16:09 |
| toscalix | I have on my todo to create the subgroup for infra and move this repo there: https://gitlab.com/BuildStream/sled-ansible-config | 16:10 |
| toscalix | we will add the tickets there | 16:10 |
| *** alatiera_ has joined #buildstream | 16:13 | |
| *** rdale has quit IRC | 16:19 | |
| *** rdale has joined #buildstream | 16:23 | |
| coldtom | are there plans in the works to upgrade the ruamel.yaml dependency? currently running `bst track` will "unfold" folded blocks (using >) | 16:31 |
| *** alatiera_ has quit IRC | 16:36 | |
| *** alatiera_ has joined #buildstream | 16:36 | |
| *** jonathanmaw has quit IRC | 16:53 | |
| *** lachlan has quit IRC | 16:54 | |
| *** lachlan has joined #buildstream | 16:55 | |
| lachlan | Hi, I have a WIP branch for benchmarking that I would like to have reviewed for the broad approach being taken. | 16:55 |
| gitlab-br-bot | coldtom opened MR !955 (coldtom/bst-fmt->master: bst fmt: Add basic functionality) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/955 | 16:55 |
| gitlab-br-bot | jennis opened MR !956 (jennis/less_restrictive_pathways->master: We shouldn't fail on an absolute path to our project) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/956 | 16:55 |
| jennis | A nice 'end of the day and I'm tired' patch up for review there ^^ | 17:01 |
| jennis | Would recommend commit-by-commit as the total changes looks messy xD | 17:02 |
| lachlan | MR !9 (lachlanmackenzie/AddGraphingOfResults->master: WIP: Add runner to get and correlate results for plotting ) on buildstream/benchmarking https://gitlab.com/BuildStream/benchmarks/merge_requests/9 | 17:08 |
| gitlab-br-bot | MR !9: Leeming/sandbox only https://gitlab.com/BuildStream/buildstream/merge_requests/9 | 17:08 |
| *** toscalix has quit IRC | 17:10 | |
| tpollard | juergbi: I should have everything pushed now bar not having the directory logic in the pullqueue | 17:10 |
| juergbi | great | 17:11 |
| *** MOLLET6719 has joined #buildstream | 17:11 | |
| *** MOLLET6719 has left #buildstream | 17:11 | |
| juergbi | tpollard: does the comment about the directory logic move make sense to you? | 17:11 |
| tpollard | juergbi: vaguely | 17:16 |
| tpollard | now that it's in global context and not an option passed to the queue from stream, I should be able to reproduce the same logic within element? | 17:17 |
| gitlab-br-bot | jjardon opened issue #768 (Overnigth tests are failing for bst-1.2 branch) on buildstream https://gitlab.com/BuildStream/buildstream/issues/768 | 17:18 |
| tpollard | I'm trying to think if it will have ramifications in the future (one being an attempted partial pushed spawning a pull queue for the buildtree) | 17:22 |
| juergbi | tpollard: yes. if the partial pull will go via pullqueue, Element should likely grow a method such as _pull_buildtree() | 17:26 |
| jennis | juergbi, thanks for the review, I've replied | 17:30 |
| *** abderrahim has joined #buildstream | 17:31 | |
| *** abderrahim4 has quit IRC | 17:32 | |
| *** finn_ has quit IRC | 17:53 | |
| *** alatiera_ has quit IRC | 18:00 | |
| juergbi | jennis: btw: doesn't this sound like you want a workspace for the junction? don't we support this? | 18:04 |
| *** lachlan has quit IRC | 18:37 | |
| gitlab-br-bot | juergbi approved MR !943 (aevri/unit_tests->master: contributing: add guidance on unit tests) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/943 | 19:48 |
| *** ctolentino has quit IRC | 20:05 | |
| gitlab-br-bot | juergbi merged MR !943 (aevri/unit_tests->master: contributing: add guidance on unit tests) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/943 | 20:25 |
| *** tristan has quit IRC | 20:30 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!