| *** alatiera_ has quit IRC | 00:02 | |
| *** tiagogomes has quit IRC | 01:33 | |
| *** catonano has joined #buildstream | 03:24 | |
| *** catonano has quit IRC | 05:38 | |
| *** catonano has joined #buildstream | 05:44 | |
| *** catonano has quit IRC | 06:55 | |
| *** WSalmon has joined #buildstream | 07:02 | |
| *** iker has joined #buildstream | 07:10 | |
| *** bochecha has joined #buildstream | 07:22 | |
| *** WSalmon has quit IRC | 07:43 | |
| *** toscalix has joined #buildstream | 07:44 | |
| *** WSalmon has joined #buildstream | 07:56 | |
| *** finn has joined #buildstream | 08:09 | |
| *** finn has quit IRC | 08:31 | |
| Kinnison | juergbi: Did you get a chance to split out the work you did on mac_fixes ? | 08:31 |
|---|---|---|
| *** phildawson has joined #buildstream | 08:35 | |
| juergbi | Kinnison: yes, I opened !829 | 08:35 |
| juergbi | review appreciated | 08:36 |
| * Kinnison shall give it a proper going over this morning | 08:36 | |
| juergbi | thanks | 08:36 |
| *** tiagogomes has joined #buildstream | 08:36 | |
| * Kinnison would appreciate it if someone else would too | 08:36 | |
| juergbi | maybe jmac could take a look as well | 08:36 |
| juergbi | the branch has quite a few commits but it's really just refactoring, the individual commits are all small | 08:37 |
| jmac | I can do, I'm yet to start on my sprint's tasks yet though... | 08:37 |
| jmac | Not that sprints mean much on this channel | 08:38 |
| Kinnison | :-) | 08:38 |
| juergbi | thanks | 08:38 |
| juergbi | the most interesting commit review wise is likely "Move artifact cache creation to Context". also small but I'd be interested in feedback whether I'm moving things to the right place | 08:39 |
| * jmac wonders why https://gitlab.com/BuildStream/buildstream/merge_requests/825 hasn't merged | 08:39 | |
| juergbi | master moved on before CI finished, i.e., rebase required? | 08:42 |
| jmac | So it won't merge even if there's fast-forwardable changes? | 08:44 |
| juergbi | I think so, yes. gitlab never implicitly rebases | 08:44 |
| juergbi | freedesktop-sdk has a merge bot for this, afaict | 08:44 |
| juergbi | might be useful for us as well | 08:44 |
| Kinnison | Yeah, it may be worth looking into a merge bot | 08:44 |
| Kinnison | Rust uses one which can roll up multiple PRs into a single test-run+push-to-master which is nice for speeding things up | 08:45 |
| Kinnison | Unfortunately I think bors is github specific | 08:45 |
| juergbi | (to be precise, !825 is no longer fast forwardable because master moved on in the meantime) | 08:45 |
| juergbi | the fdo-sdk one (marge bot) should be good enough | 08:45 |
| juergbi | it serializes things, so immediately rebases the next MR in the queue after the merging of the previous one | 08:46 |
| Kinnison | What does it do for testing? Does it issue CI runs after each rebase, or does it rebase a stack of MRs together and test once? | 08:46 |
| juergbi | for each MR individually | 08:46 |
| gitlab-br-bot | buildstream: merge request (mablanch/668-remote-build-failure->master: Better handle remote build failures) #825 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/825 | 08:46 |
| juergbi | i.e., it really does the same thing as we can do as gitlab users, but organized by a single bot entity | 08:47 |
| Kinnison | I see | 08:49 |
| Kinnison | So it automates what currently happens | 08:49 |
| Kinnison | I guess that's a good step forward, saves human cognitive load for sure | 08:49 |
| Kinnison | juergbi: The commits on !829 look sane to me. I'm not really in a position to speak to the correctness of the approach as a whole, though I will say that it makes sense to me, I'm not fully au-fait with everything in the lower levels of bst. I raised a single discussion but otherwise I think 829 is good | 08:51 |
| *** finn has joined #buildstream | 08:53 | |
| *** WSlamon has joined #buildstream | 08:56 | |
| *** WSlamon has quit IRC | 08:57 | |
| *** jonathanmaw has joined #buildstream | 09:00 | |
| juergbi | thanks | 09:01 |
| gitlab-br-bot | buildstream: merge request (juerg/platform->master: Make platform independent of context) #829 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/829 | 09:05 |
| juergbi | added preflight method to artifact cache for basic sanity check | 09:06 |
| *** tpollard has joined #buildstream | 09:07 | |
| jmac | Yep, similarly, it looks sensible to me, but it'll take a while to review fully | 09:07 |
| jmac | I'm fine with the changes to the remote execution parts | 09:07 |
| Kinnison | juergbi: I'm satisfied with the preflight() change and have resolved the discussion | 09:08 |
| juergbi | jmac: ok, thanks. are you planning to do a full review later, i.e., shall I wait with the merge or just go ahead? | 09:14 |
| jmac | If you're happy to merge it then merge it | 09:15 |
| juergbi | ok | 09:15 |
| gitlab-br-bot | buildstream: merge request (mablanch/668-remote-build-failure->master: Better handle remote build failures) #825 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/825 | 09:16 |
| gitlab-br-bot | buildstream: merge request (juerg/platform->master: Make platform independent of context) #829 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/829 | 09:20 |
| *** raoul has joined #buildstream | 09:22 | |
| Nexus | I've not had to do it in a while, how do i rebuild the docs again? | 09:29 |
| juergbi | Nexus: if you have the dependencies, it should just be: make -C doc | 09:31 |
| *** catonano has joined #buildstream | 09:33 | |
| Nexus | ta :) | 09:35 |
| *** finn has quit IRC | 09:36 | |
| tiagogomes | jjardon https://docs.buildstream.build should work now without certificate related warnings | 09:37 |
| jjardon | Cool! | 09:39 |
| *** lachlan has joined #buildstream | 09:41 | |
| tiagogomes | we could apply some tweaks from http://docs.flatpak.org/ to https://docs.buildstream.build/ | 09:41 |
| jjardon | tiagogomes: thanks for reviewing the patches about the main website page later yesterday btw :) | 09:46 |
| gitlab-br-bot | buildstream: merge request (juerg/platform->master: Make platform independent of context) #829 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/829 | 09:48 |
| toscalix | tiagogomes: it works, thanks | 09:48 |
| toscalix | tiagogomes: maybe you want also to change the menu link | 09:49 |
| toscalix | in the front page of the site | 09:49 |
| gitlab-br-bot | buildstream: merge request (mac_fixes->master: Implement compatibility fixes for MacOSX and WSL Blocks #411 and #412") #726 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/726 | 09:49 |
| tiagogomes | toscalix which menu link? | 09:51 |
| toscalix | the top menu on the front page. One option is documentation | 09:52 |
| toscalix | the night builds of freedesktop-sdk failed | 09:52 |
| toscalix | hummm.... mails are moderated | 09:53 |
| toscalix | nice, we have notifications about the nightly builds on the mailing list | 09:55 |
| toscalix | notificatoins mailing list | 09:55 |
| gitlab-br-bot | buildstream: merge request (juerg/git-describe->master: git.py: Make `ref` human readable) #291 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/291 | 09:56 |
| toscalix | adds68: jjardon coldtom valentind you probably are interested in this notifications | 09:57 |
| toscalix | you can easily firter them among the rest of the notifications we send to the notifications-mailing -list | 09:58 |
| toscalix | https://mail.gnome.org/mailman/listinfo/buildstream-notifications-list | 09:58 |
| tiagogomes | I think that list is useless as it is. It is too noisy to track | 10:00 |
| juergbi | tiagogomes: for CI you've updated the fd sdk ref but kept the comment "18.08.12", which is now outdated, afaict | 10:01 |
| toscalix | tiagogomes: depends on how you manage firlters | 10:01 |
| toscalix | filters | 10:02 |
| toscalix | it is noisy because the activity is very high | 10:02 |
| tiagogomes | juergbi, right, I didn't bother to be accurate as that is just an indication. But I should have used the output of git describe | 10:02 |
| *** CTtpollard has joined #buildstream | 10:02 | |
| adds68 | toscalix, i already receive notifications from gitlab when pipelines fail? Have i missed some context here? | 10:03 |
| *** tpollard has quit IRC | 10:03 | |
| toscalix | adds68: this one is from buildstream repo | 10:04 |
| toscalix | from the new pipelines created by tiagogomes | 10:04 |
| toscalix | adds68: https://gitlab.com/BuildStream/buildstream/pipelines/31303716 | 10:05 |
| tiagogomes | If those pipelines fail, it is probably a problem with buildstream or a server being down, so it doesn't directly concern the freedesktop-sdk team | 10:05 |
| toscalix | https://gitlab.com/BuildStream/buildstream/pipelines/31303715 | 10:05 |
| toscalix | tiagogomes: which is why I pinged valentind | 10:06 |
| toscalix | but maybe users ping the freedesktop-sdk guys if something goes wrong | 10:06 |
| valentind | Are not those gitlab notifications? | 10:07 |
| tiagogomes | Hmm, in this case two jobs failed due lack of disk space | 10:07 |
| toscalix | yes, sent to the notifications list | 10:08 |
| tiagogomes | This one is more strange https://gitlab.com/BuildStream/buildstream/-/jobs/101978374 | 10:09 |
| toscalix | this is a better link https://gitlab.com/BuildStream/buildstream/pipeline_schedules | 10:09 |
| adds68 | There seems to be emails from every pipeline in that inbox, i think it's best if we get pinged for any freedesktop-sdk related issues | 10:10 |
| tiagogomes | Not easily possible | 10:11 |
| toscalix | adds68: we just got notifications from pipelines. It is an improvement. You can filter the ones you are interested on | 10:11 |
| *** finn has joined #buildstream | 10:12 | |
| valentind | OK, so I added notifications for all build fails on BuildStream/buildstream for me. | 10:13 |
| adds68 | toscalix, from what i'm aware if you run a pipeline Gitlab automatically notifies the person who started it/any MR/issue associated with that pipeline? | 10:15 |
| valentind | adds68, In this case this is a scheduled pipeline. | 10:15 |
| juergbi | tiagogomes: imo, the comment could also just indicate the branch (18.08) or have something like 18.08.12+ to indicate it's been updated after that tag | 10:16 |
| juergbi | just want to avoid confusion due to mismatch | 10:16 |
| adds68 | valentind, ah ok | 10:19 |
| gitlab-br-bot | buildstream: merge request (juerg/git-describe->master: git.py: Make `ref` human readable) #291 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/291 | 10:19 |
| toscalix | tiagogomes: correct me if I am wrong. The pipeline ID is fixed. | 10:20 |
| * tiagogomes doesn't know | 10:20 | |
| toscalix | the ID is part of the subject of the notification mail | 10:20 |
| valentind | toscalix, I do not think it is. | 10:20 |
| valentind | Otherwise you cannot go see the history of builds. | 10:21 |
| valentind | There seems to be an ID however, when you hover the "play" button, you can see the url contains a number. I think this is the number of the schedule. | 10:22 |
| valentind | Like: "https://gitlab.com/BuildStream/buildstream/pipeline_schedules/12238/play" | 10:22 |
| valentind | So this is 12238 | 10:23 |
| toscalix | https://docs.gitlab.com/ee/ci/variables/#using-the-ci-variables-in-your-job-scripts I see here several IDs, some internal | 10:23 |
| jjardon | toscalix: tiagogomes thanks | 10:23 |
| toscalix | valentind: sadly the default mail notification does not include such number | 10:24 |
| toscalix | so we would need to work with the API so include the tag on the notification so we can filter them in mail clients | 10:27 |
| jjardon | toscalix: this should not affect our users though (we used a fixed bst 1.2 version), but nice you care to not break fdsk with buildstream master :) | 10:31 |
| *** lachlan has quit IRC | 10:34 | |
| toscalix | jjardon: there is a pipeline for bst-1.2 | 10:48 |
| jjardon | juergbi: is the description in https://gitlab.com/BuildStream/buildstream/merge_requests/291 still valid? | 10:48 |
| jjardon | toscalix: cool | 10:48 |
| juergbi | jjardon: no, that was the initial proposal | 10:48 |
| jjardon | ah, ok | 10:49 |
| * jjardon keeps reading | 10:49 | |
| juergbi | don't really want to completely remove that but maybe i should still amend it | 10:49 |
| juergbi | jjardon: https://gitlab.com/BuildStream/buildstream/merge_requests/291#note_61261970 | 10:49 |
| jjardon | yeah, maybe an update after what is already there would help in case we have to look to that later | 10:49 |
| toscalix | jjardon: https://gitlab.com/BuildStream/buildstream/pipelines/31303716/builds | 10:49 |
| jjardon | cool, I will try and remember to update https://gitlab.com/BuildStream/buildstream/blob/master/.gitlab-ci.yml#L166 when a new freedesktop-sdk release is out | 10:52 |
| gitlab-br-bot | buildstream: merge request (juerg/git-describe->master: git.py: Make `ref` human readable) #291 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/291 | 10:54 |
| juergbi | description updated | 10:54 |
| *** lachlan has joined #buildstream | 10:55 | |
| toscalix | jjardon: is there a way to avoid in that line a hardcoded ref so you do not have to update it manually? | 10:55 |
| tiagogomes | I don't want to use the tip of 18.06, but rather a fixed ref | 10:57 |
| jjardon | toscalix: sure, you can build wherever is latest, but that's a question for tiagogomes | 10:57 |
| toscalix | tiagogomes: ok, your call | 10:57 |
| tiagogomes | fixed ref | 10:57 |
| tiagogomes | Otherwise we need to workout if what is causing the failure is buildstream or fd-sdk | 10:58 |
| tiagogomes | fd-dsk could have a release document with the steps involved in doing a release (tag, announcement, release party…). That could be one of the steps so that is not forgotten | 10:59 |
| toscalix | I wonder if we need to document that or it is obvious from reading .gitlab-ci.yml | 11:00 |
| raoul | How would I go about creating a flatpak binary with buildstream? Have managed to build doom with freedesktop following the flatpak-autotools example, but can't seem to find how I'd output a binary I can then use (maybe this should be in the docs somewhere?) | 11:00 |
| tiagogomes | From reading .gitlab-ci.yml it is obviously what needs to be done, but we still need to remember to do that | 11:01 |
| jjardon | tiagogomes: It can not be fdsk; nothing goes to master without passing the CI. But I understand your argument | 11:05 |
| *** abderrahim has quit IRC | 11:06 | |
| *** abderrahim has joined #buildstream | 11:07 | |
| gitlab-br-bot | buildstream: merge request (coldtom/strip-rules->master: WIP: Upstream freedesktop-sdk strip rules) #750 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/750 | 11:07 |
| *** finn has quit IRC | 11:07 | |
| valentind | raoul, we have made flatpaks of tuxracer with buildstream. That still involves a lot of boiler plate. See https://gitlab.com/valentindavid/org.gnu.Hello/tree/valentindavid/supertuxkart | 11:08 |
| valentind | Sorry super tux kart. Not tuxracer. | 11:08 |
| raoul | I'll have a gander at that, thanks valentind | 11:10 |
| valentind | raoul, note it is using a branch of freedesktop sdk. | 11:10 |
| valentind | I hope to merge that soon. But I have not had time. | 11:10 |
| jjardon | juergbi: I like https://gitlab.com/BuildStream/buildstream/merge_requests/291, but maybe we should make it optional? | 11:11 |
| juergbi | jjardon: might make sense. via project-wide defaults this should be convenient enough | 11:14 |
| jjardon | yeah, mostly for users doesnt get confused by their refs suddenly not being the standard git SHA1; probably better to have an option so people can opt in | 11:15 |
| juergbi | track-describe as option? | 11:16 |
| jjardon | I'd say "ref-format: sha1/git-describe"? | 11:18 |
| jjardon | and default to sha1? | 11:18 |
| *** finn has joined #buildstream | 11:33 | |
| gitlab-br-bot | buildstream: issue #682 ("Handle rogue plugin exceptions.") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/682 | 11:52 |
| *** finn has quit IRC | 12:03 | |
| gitlab-br-bot | buildstream: merge request (juerg/git-describe->master: git.py: Make `ref` human readable) #291 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/291 | 12:08 |
| gitlab-br-bot | buildstream: merge request (juerg/git-describe->master: git.py: Make `ref` human readable) #291 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/291 | 12:12 |
| gitlab-br-bot | buildstream: merge request (juerg/git-describe->master: git.py: Make `ref` human readable) #291 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/291 | 12:17 |
| gitlab-br-bot | buildstream: merge request (juerg/git-describe->master: git.py: Make `ref` human readable) #291 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/291 | 12:18 |
| gitlab-br-bot | buildstream: merge request (willsalmon/outOfSourecBuild->master: Out of source builds) #776 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/776 | 12:19 |
| paulsher1ood | raising this here, in case anyone has helpful suggestions https://lists.trustable.io/pipermail/trustable-software/2018-September/000457.html | 12:27 |
| coldtom | how can i stop pytest skipping tests when i'm testing locally? | 12:28 |
| paulsher1ood | how can we clean this up? it's frustrating that we are still relying on binaries and huge configure files in 2018 | 12:28 |
| juergbi | coldtom: to not skip integration tests: ./setup.py test --addopts --integration | 12:31 |
| juergbi | tests with missing dependencies are also skipped (e.g., in case you don't have bzr) | 12:31 |
| coldtom | thanks juergbi :) | 12:31 |
| gitlab-br-bot | buildstream: merge request (tpollard/494->master: WIP: Don't pull artifact buildtrees by default) #786 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/786 | 12:31 |
| juergbi | paulsher1ood: I can't think of a generic way how to solve this. if pre-generated configure files and similar are to be avoided (which would be good, I agree), either all the extra dependencies/steps have to be added or upstream has to be convinced to improve things | 12:49 |
| juergbi | both require extra work, the former is likely less time consuming in many cases, but the latter would be better long term | 12:50 |
| juergbi | (the former less time consuming and actually feasible for the most part; the latter might often result in a dead end, depending on the project/community) | 12:51 |
| juergbi | potential third approach would be to patch up the build systems even if the patch is not accepted upstream | 12:51 |
| paulsher1ood | flatmush: ^^ ? | 12:52 |
| flatmush | paulsher1ood: I don't disagree, but it requires someone with the correct understanding of autotools and the time to do it, for me it seems like a very large task | 12:53 |
| juergbi | I agree. especially gcc config is very complex | 12:54 |
| juergbi | although it is possible that it's mostly compatible with latest autoconf | 12:54 |
| *** gitlab-br-bot has quit IRC | 12:54 | |
| *** gitlab-br-bot has joined #buildstream | 12:54 | |
| *** gitlab-br-bot has joined #buildstream | 12:54 | |
| *** gitlab-br-bot has quit IRC | 12:55 | |
| *** gitlab-br-bot has joined #buildstream | 12:55 | |
| juergbi | at least it's not autoconf 2.13 which firefox still appears to require... (they have a mix of old autoconf project and some new stuff) | 12:55 |
| *** ctgriffiths has joined #buildstream | 13:05 | |
| gitlab-br-bot | buildstream: merge request (tpollard/494->master: WIP: Don't pull artifact buildtrees by default) #786 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/786 | 13:06 |
| gitlab-br-bot | buildstream: merge request (tpollard/494->master: WIP: Don't pull artifact buildtrees by default) #786 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/786 | 13:09 |
| *** lachlan has quit IRC | 13:12 | |
| gitlab-br-bot | buildstream: merge request (juerg/git-describe->master: git.py: Make `ref` human readable) #291 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/291 | 13:13 |
| juergbi | jjardon: added ref-format config key to !291 | 13:17 |
| juergbi | jonathanmaw: does this addition make sense to you as well? | 13:17 |
| jonathanmaw | juergbi: yep | 13:17 |
| juergbi | ta | 13:17 |
| *** CTtpollard is now known as tpollard | 13:20 | |
| tpollard | juergbi: do I have to any anything special for my artifact server to use batch? I updated the checkout of buildstream to the head of master, but it's still catching the exceptions batchreadblobs not implemented | 13:20 |
| juergbi | tpollard: hm, your artifact server is running from master and you see these exceptions? or have you only updated the client? | 13:21 |
| tpollard | juergbi: the artifact server is checked out against master. and my client pushing to it is on a branch based off of master | 13:22 |
| juergbi | BatchReadBlobs is definitely implemented on master. just to be sure, you've restarted the artifact server after updating the checkout? | 13:24 |
| tpollard | I believe so | 13:25 |
| tpollard | will try again | 13:25 |
| tpollard | yep still happening | 13:26 |
| tpollard | bst-version on the server seems to be correct too | 13:27 |
| *** lachlan has joined #buildstream | 13:29 | |
| *** finn has joined #buildstream | 13:30 | |
| jmac | tpollard: Is it practical to wipe the caches on both client and server? | 13:30 |
| tiagogomes | juergbi with your recent changes on moving the creating of the artifact server to a different place, any chance the tests are more reliable in parallel? | 13:30 |
| tpollard | jmac: I could overnight, the cache that I'm pushing up to it came from a pre batch artifact server | 13:31 |
| tpollard | if that could be part of the issue | 13:31 |
| tiagogomes | The current runners have two vCPUs, so we should take benefit from that if possible | 13:31 |
| jmac | tpollard: I'm just wondering if you have a cached failure | 13:31 |
| juergbi | tiagogomes: they might. it fixes parallel test issue with the mac_fixes branch, but it might also fix other existing issues | 13:32 |
| juergbi | tiagogomes: last I checked the integration tests didn't work in parallel, though, and I don't expect my changes to help there | 13:33 |
| juergbi | however, it might not be too difficult to fix them up | 13:33 |
| *** finn has quit IRC | 13:33 | |
| tiagogomes | ok, I will do a local run to see how it goes | 13:34 |
| juergbi | tiagogomes, tpollard: the exception is on the server side, as I understand it. cached failures can't trigger that | 13:34 |
| juergbi | eh, tpollard, jmac | 13:34 |
| *** ahmed89 has quit IRC | 13:36 | |
| *** finn has joined #buildstream | 13:42 | |
| *** catonano has quit IRC | 13:45 | |
| gitlab-br-bot | buildstream: merge request (juerg/git-describe->master: git.py: Make `ref` human readable) #291 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/291 | 13:46 |
| gitlab-br-bot | buildstream: issue #683 ("Let's Encrypt certificate needs to be renewed") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/683 | 13:47 |
| gitlab-br-bot | buildstream: merge request (valentindavid/silverblue-path->master: Fix outside-of-project check when project path is not canonical.) #832 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/832 | 13:50 |
| tiagogomes | I had that ^^ on my radar valentind :) | 13:51 |
| valentind | tiagogomes, oh sorry. | 13:52 |
| tiagogomes | no worries, I didn't start any work on it besides reproducing it | 13:52 |
| valentind | tiagogomes, then you can review it. | 13:52 |
| valentind | But wait for the tests to pass first. | 13:52 |
| tiagogomes | yup, as I wrote the original code | 13:53 |
| valentind | I did that quickly before I go play MTG. | 13:53 |
| * tiagogomes used to have a good starving for blood deck of Zombies :) | 13:54 | |
| tpollard | cheers juergbi, looks like I didn't have -e on the server bst install | 13:56 |
| *** catonano has joined #buildstream | 14:01 | |
| * Kinnison is having difficulty getting gitlab to load MR pages, is anyone else? | 14:02 | |
| tpollard | wfm | 14:04 |
| * Kinnison blames his notwork then :-) | 14:05 | |
| gitlab-br-bot | buildstream: merge request (tpollard/494->master: WIP: Don't pull artifact buildtrees by default) #786 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/786 | 14:11 |
| gitlab-br-bot | buildstream: merge request (mac_fixes->master: Implement compatibility fixes for MacOSX and WSL Blocks #411 and #412") #726 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/726 | 14:16 |
| *** alatiera_ has joined #buildstream | 14:19 | |
| gitlab-br-bot | buildstream: issue #665 ("Buildstream tries to use message without a message handler if linux platform and no user namespace support") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/665 | 14:19 |
| *** catonano has quit IRC | 14:23 | |
| *** catonano has joined #buildstream | 14:24 | |
| WSalmon | if i get a failure mid build then bst tells me what make said to make it unhappy but if ./configure etc fail then i have not message just the command so i have to drop in to a shell and run the command my self, is this know? a bug or a feature? its supper anoying | 14:24 |
| juergbi | hm, it should normally show the last few log lines independent of which command failed | 14:25 |
| juergbi | and with (l) you can take a look at the whole log | 14:25 |
| gitlab-br-bot | buildstream: merge request (mac_fixes->master: Implement compatibility fixes for MacOSX and WSL Blocks #411 and #412") #726 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/726 | 14:30 |
| gitlab-br-bot | buildstream: merge request (valentindavid/silverblue-path->master: Fix outside-of-project check when project path is not canonical.) #832 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/832 | 14:30 |
| tpollard | can anyone on ubuntu 18.04 currently run the tests? | 14:49 |
| tpollard | 18.04.1 actually | 14:49 |
| jmac | I'm 18.04.1, I'll try | 14:50 |
| tpollard | adding pytest == 3.6.3 to install requires used to work for me the last time I had to run them | 14:50 |
| tpollard | but that conflicts now | 14:51 |
| jmac | They appear to be running and from what I can tell I have pytest 3.8.0 | 14:51 |
| * tpollard kills .eggs and tries again | 14:52 | |
| tpollard | that looks to have done the trick, jumped to 3.8.1 | 14:52 |
| gitlab-br-bot | buildstream: merge request (mac_fixes->master: Implement compatibility fixes for MacOSX and WSL Blocks #411 and #412") #726 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/726 | 14:54 |
| WSalmon | juergbi, you are right, however the last thing in the log is a echo of the last command run but it was so long it was shaddowing the error, the (l) option is just what i was after! thanks | 14:57 |
| gitlab-br-bot | buildstream: merge request (jonathan/pickle-yaml->master: WIP: Add a cache of parsed and provenanced yaml) #787 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/787 | 15:02 |
| jmac | tpollard: I get a huge number of RemovedInPytest4Warning messages, which is annoying, but the tests pass | 15:05 |
| gitlab-br-bot | buildstream: issue #411 ("Native bst support on OS X, remote execution only") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/411 | 15:09 |
| gitlab-br-bot | buildstream: issue #412 ("Native bst support on Windows w/ WSL, remote execution only") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/412 | 15:10 |
| gitlab-br-bot | buildstream: issue #575 ("os.sched_getaffinity() not supported on MacOSX Blocks #411") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/575 | 15:10 |
| gitlab-br-bot | buildstream: issue #269 ("buildstream hangs forever if fuse device is not present") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/269 | 15:16 |
| gitlab-br-bot | buildstream: merge request (richardmaw/push-failed-build-regression->master: tests: Add regression test for pushing failed builds) #824 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/824 | 15:17 |
| *** finn has quit IRC | 15:17 | |
| jjardon | juergbi: cool! thanks a lot | 15:18 |
| *** finn has joined #buildstream | 15:18 | |
| gitlab-br-bot | buildstream: merge request (tpollard/494->master: WIP: Don't pull artifact buildtrees by default) #786 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/786 | 15:19 |
| *** finn has quit IRC | 15:23 | |
| tlater[m] | jmac: I've recently started seeing those messages from flycheck... I'll probably want to write a patch to fix these soon. | 15:27 |
| tlater[m] | Interestingly, we don't seem to support python 3.8 at all, because we use features that are going to be deprecated. | 15:27 |
| juergbi | tlater[m]: python 3.8 or pytest 3.8? | 15:28 |
| tlater[m] | juergbi: python 3.8 | 15:30 |
| tlater[m] | Pylint handily informed me that some of the @abc stuff will be deprecated, for example | 15:30 |
| juergbi | tlater[m]: ok, but if it's just deprecated, it should still work | 15:31 |
| tlater[m] | juergbi: The warning suggests that it will be moved entirely and not work anymore | 15:31 |
| juergbi | we should still port things over to the corresponding replacements, of course, as long as that doesn't break Python 3.5 | 15:32 |
| jmac | Could someone with GNOME wiki permissions allow the user 'JMacArthur' to edit the BuildStream wiki? | 15:32 |
| juergbi | tlater[m]: oh, that would be Python breaking backward compatibility? I thought that shouldn't happen... | 15:32 |
| juergbi | jmac: sure, just a sec | 15:32 |
| tlater[m] | I agree, I haven't looked into exactly how migration is supposed to happen yet. | 15:33 |
| juergbi | jmac: there is already a JimMacArthur in the list. two accounts? shall I still add it? | 15:33 |
| * tlater[m] will take a look when fixing these annoying new pylint errors | 15:33 | |
| jmac | juergbi: JimMacArthur is me, but I don't have the password or email address for that account | 15:33 |
| juergbi | ok | 15:33 |
| jmac | juergbi: Hence JMacArthur | 15:33 |
| juergbi | jmac: done | 15:34 |
| jmac | Thanks | 15:34 |
| *** finn has joined #buildstream | 15:35 | |
| *** finn has quit IRC | 15:38 | |
| gitlab-br-bot | buildstream: issue #429 ("BuildStream refuses to cache artifacts with files that have very strict permissions") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/429 | 15:38 |
| gitlab-br-bot | buildstream: issue #534 ("Failed builds not pushing artifacts on quit") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/534 | 15:42 |
| gitlab-br-bot | buildstream: merge request (richardmaw/push-failed-build-regression->master: tests: Add regression test for pushing failed builds) #824 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/824 | 15:42 |
| *** lachlan has quit IRC | 15:56 | |
| juergbi | valentind: commented on !830 but overall looks good to me | 16:00 |
| *** iker has quit IRC | 16:06 | |
| * tiagogomes doesn't understand why if `path` is given, we don't always link | 16:17 | |
| tiagogomes | Is this because the original object could reside in a different filesystem? | 16:18 |
| tiagogomes | Anyway, link_file is a confusing parameter, as there will be always be a link operation occurring | 16:19 |
| jmac | I've written up my reverse engineering of the file import rules at https://wiki.gnome.org/Projects/BuildStream/ImportRules. This is only likely to be of niche interest, but I think it should be part of our developer documentation, if we create such a thing. | 16:21 |
| juergbi | tiagogomes: it's mainly a precaution against potential CAS repo corruption. if the input file was not created by CAS, the one that created it might modify it later on, which would corrupt the cache | 16:23 |
| juergbi | it's possible that we could link in more cases, maybe even all, but we'd have to be very carefully check where the files are coming from | 16:24 |
| juergbi | and make sure we don't accidentally introduce issues at the call sites in the future | 16:24 |
| tiagogomes | That makes sense | 16:27 |
| gitlab-br-bot | buildstream: merge request (valentindavid/silverblue-path->master: Fix outside-of-project check when project path is not canonical.) #832 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/832 | 16:33 |
| *** toscalix has quit IRC | 16:33 | |
| *** catonano has quit IRC | 16:41 | |
| *** lachlan has joined #buildstream | 16:42 | |
| gitlab-br-bot | buildstream: issue #22 ("failed to upgrade buildstream at 640a734ec1db923b5") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/22 | 16:42 |
| gitlab-br-bot | buildstream: issue #429 ("BuildStream refuses to cache artifacts with files that have very strict permissions") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/429 | 16:53 |
| gitlab-br-bot | buildstream: issue #673 ("Silverblue: Specified path 'elements' must not lead outside of the project directory") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/673 | 16:57 |
| gitlab-br-bot | buildstream: merge request (valentindavid/silverblue-path->master: Fix outside-of-project check when project path is not canonical.) #832 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/832 | 16:57 |
| *** jonathanmaw has quit IRC | 17:00 | |
| gitlab-br-bot | buildstream: issue #673 ("Silverblue: Specified path 'elements' must not lead outside of the project directory") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/673 | 17:00 |
| gitlab-br-bot | buildstream: issue #429 ("BuildStream refuses to cache artifacts with files that have very strict permissions") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/429 | 17:02 |
| *** catonano has joined #buildstream | 17:09 | |
| *** lachlan has quit IRC | 17:36 | |
| bochecha | how do I set the cache.quota option in buildstream.conf ? | 18:57 |
| bochecha | the doc isn't very clear on this :( | 18:57 |
| bochecha | can I set something like `quota: 10G` ? | 18:57 |
| bochecha | or do I have to use an integer in bytes? | 18:58 |
| juergbi | bochecha: K/M/G/T suffices are supported | 19:03 |
| juergbi | *suffixes | 19:04 |
| *** toscalix has joined #buildstream | 19:14 | |
| bochecha | juergbi: thanks | 19:29 |
| bochecha | juergbi: would a MR explaining that be welcome? :) | 19:29 |
| *** catonano has quit IRC | 19:54 | |
| *** catonano has joined #buildstream | 19:54 | |
| *** xjuan has joined #buildstream | 22:10 | |
| jjardon | bochecha: yup | 22:45 |
| *** xjuan has quit IRC | 23:08 | |
| *** WSalmon has quit IRC | 23:14 | |
| *** bochecha has quit IRC | 23:15 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!