| *** mohan43u has quit IRC | 03:11 | |
| *** nimish2711 has joined #buildstream | 03:42 | |
| *** nimish2711 has quit IRC | 05:09 | |
| *** mohan43u has joined #buildstream | 05:20 | |
| *** nimish2711 has joined #buildstream | 06:41 | |
| *** tristan has quit IRC | 07:27 | |
| *** tristan has joined #buildstream | 07:41 | |
| *** bochecha has joined #buildstream | 07:54 | |
| *** raoul has joined #buildstream | 08:55 | |
| *** jonathanmaw has joined #buildstream | 09:23 | |
| *** lachlan has joined #buildstream | 09:41 | |
| *** lachlan has quit IRC | 09:51 | |
| *** lachlan has joined #buildstream | 10:16 | |
| *** lachlan has quit IRC | 10:19 | |
| *** lachlan has joined #buildstream | 10:31 | |
| *** lachlan has quit IRC | 10:34 | |
| raoul | Just noticed that gitlab supposedly has interactive web terminals, which may be useful for trying to track down #1023, would whoever deals with the CI infrastructure be able to add that? | 10:36 |
|---|---|---|
| gitlab-br-bot | Issue #1023: CI test suite hanging https://gitlab.com/BuildStream/buildstream/issues/1023 | 10:36 |
| raoul | https://docs.gitlab.com/ee/ci/interactive_web_terminal/ | 10:36 |
| raoul | looks like it would just require configuring a [session_server] section | 10:36 |
| tpollard | ooo | 10:37 |
| raoul | Is it valentind that deals with infrastructure? | 10:42 |
| abderrahim[m] | jonathanmaw Hi, can I have a bst-external release with https://gitlab.com/BuildStream/bst-external/merge_requests/88? | 10:42 |
| *** lachlan has joined #buildstream | 10:45 | |
| *** lachlan has quit IRC | 10:49 | |
| *** tristan has quit IRC | 10:50 | |
| *** raoul_ has joined #buildstream | 10:56 | |
| *** jonathanmaw_ has joined #buildstream | 10:56 | |
| *** johnward- has joined #buildstream | 10:57 | |
| *** bochecha_ has joined #buildstream | 10:57 | |
| *** CTtpollard has joined #buildstream | 10:58 | |
| *** hergertme_ has joined #buildstream | 10:59 | |
| *** jonathanmaw has quit IRC | 10:59 | |
| *** raoul has quit IRC | 10:59 | |
| *** bochecha has quit IRC | 10:59 | |
| *** johnward has quit IRC | 10:59 | |
| *** tpollard has quit IRC | 10:59 | |
| *** hergertme has quit IRC | 10:59 | |
| *** bochecha_ is now known as bochecha | 10:59 | |
| *** johnward- is now known as johnward | 10:59 | |
| *** tristan has joined #buildstream | 11:11 | |
| *** jonathanmaw_ is now known as jonathanmaw | 11:13 | |
| jonathanmaw | abderrahim[m], okie doke | 11:13 |
| abderrahim[m] | ta | 11:15 |
| *** CTtpollard is now known as tpollard | 11:21 | |
| tpollard | I'm trying to freedesktop against bst master, I've installed bst-experimental but still get 'No Source type registered for kind 'ostree'' | 11:22 |
| tpollard | Is there something else I need to do? | 11:22 |
| *** ChanServ sets mode: +o tristan | 11:25 | |
| tristan | tpollard, you need to (A) get the freedesktop-sdk branch which has the addition commit which uses bst-experimental ... (B) edit the freedesktop-sdk project.conf to tell it where the ostree plugin comes from | 11:25 |
| tpollard | cheers tristan! sorted | 11:32 |
| *** lachlan has joined #buildstream | 11:47 | |
| *** lachlan has quit IRC | 11:51 | |
| *** lachlan has joined #buildstream | 11:54 | |
| *** lachlan has quit IRC | 11:55 | |
| jonathanmaw | abderrahim[m]: new tag pushed https://gitlab.com/BuildStream/bst-external/tree/0.13.0 | 12:12 |
| abderrahim[m] | jonathanmaw: thanks | 13:56 |
| *** lachlan has joined #buildstream | 14:00 | |
| *** lachlan has quit IRC | 14:25 | |
| *** lachlan has joined #buildstream | 14:31 | |
| gitlab-br-bot | raoul.hidalgocharman merged MR !1345 (raoul/440-fix->master: Fix init_workspace for multiple sources) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1345 | 14:31 |
| juergbi | jonathanmaw: now that update_state() is split into more readable/manageable parts, I'm not convinced that the introduction of the cache key classes, as they exist in the branch, improve things overall | 14:52 |
| juergbi | what's your take on this now? do you consider it a clear improvement compared to master? | 14:55 |
| jonathanmaw | hmm, not if the only cases it covers are strict, non-strict, and ArtifactElement | 14:56 |
| *** lachlan has quit IRC | 15:25 | |
| *** lachlan has joined #buildstream | 15:26 | |
| *** lachlan has quit IRC | 15:30 | |
| *** lachlan has joined #buildstream | 15:35 | |
| *** lachlan has quit IRC | 15:38 | |
| *** bochecha has quit IRC | 15:44 | |
| *** nimish2711 has quit IRC | 15:46 | |
| *** dtf has joined #buildstream | 15:52 | |
| *** lachlan has joined #buildstream | 15:53 | |
| *** lachlan has quit IRC | 15:58 | |
| jonathanmaw | juergbi: would you mind taking a look at !1325? I've reworked it so that it doesn't depend on !1323 | 16:07 |
| gitlab-br-bot | MR !1325: Move Element.__*cached variable to Artifact class https://gitlab.com/BuildStream/buildstream/merge_requests/1325 | 16:07 |
| gitlab-br-bot | MR !1323: Delegate cache key calculation to CacheKeyController objects https://gitlab.com/BuildStream/buildstream/merge_requests/1323 | 16:07 |
| *** bochecha has joined #buildstream | 16:10 | |
| *** bochecha has quit IRC | 16:16 | |
| *** lachlan has joined #buildstream | 16:17 | |
| gitlab-br-bot | danielsilverstone-ct opened (was WIP) MR !1346 (danielsilverstone-ct/classmethod-inits->master: Make various Element/Source init helpers @classmethod to clarify data flow) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1346 | 16:19 |
| *** lachlan has quit IRC | 16:21 | |
| *** lachlan has joined #buildstream | 16:22 | |
| *** tpollard has quit IRC | 16:25 | |
| *** lachlan has quit IRC | 16:25 | |
| *** jonathanmaw has quit IRC | 16:28 | |
| *** lachlan has joined #buildstream | 16:48 | |
| *** raoul_ has quit IRC | 16:51 | |
| *** lachlan has quit IRC | 16:53 | |
| *** oem has joined #buildstream | 17:12 | |
| *** lachlan has joined #buildstream | 17:28 | |
| *** oem has quit IRC | 17:32 | |
| *** lachlan has quit IRC | 17:33 | |
| *** lachlan has joined #buildstream | 17:48 | |
| *** lachlan has quit IRC | 17:55 | |
| *** lachlan has joined #buildstream | 18:11 | |
| *** nimish2711_ has joined #buildstream | 18:33 | |
| *** nimish2711_ is now known as nimish2711 | 18:35 | |
| *** lachlan has quit IRC | 18:43 | |
| *** nimish2711 has quit IRC | 18:50 | |
| *** nimish2711 has joined #buildstream | 18:58 | |
| *** lachlan has joined #buildstream | 19:00 | |
| *** lachlan has quit IRC | 19:12 | |
| dftxbs3e | hi, is there a way to resume a build from object files with buildstream? so that if a single source file has changed, then only the concerned object files get rebuilt and then the whole is linked again | 19:50 |
| dftxbs3e | FYI; I'm building LLVM then Clang, ninja, gn and then ungoogled-chromium | 19:51 |
| dftxbs3e | oh also, ffmpeg | 19:51 |
| coldtom | dftxbs3e: you mean if a dep changes don't rebuild things higher up the stack? because if so pass `--no-strict` to bst | 19:53 |
| dftxbs3e | coldtom: no not that, more like the make, "already up to date" effect | 19:54 |
| benschubert | dftxbs3e: you could open a workspace on your element when building it, that would leave the object files there. However, BuildStream would then kick off a build whenever you build something down your stack, since we have no way of knowing is something in the workspace changed | 19:58 |
| benschubert | (I wouldn't recommend doing a production build like that, but when hacking on the project/setting up the bst file, it's quite useful) | 19:59 |
| dftxbs3e | benschubert: hm, I'm wasting so much time rebuilding stuff that's already built and wont change | 20:05 |
| benschubert | dftxbs3e: if the build finishes, it should never rebuild it. Do you need to stop your builds in the middle? | 20:07 |
| dftxbs3e | benschubert: FYI, I don't use buildstream currently, I'm contemplating it, the big of the work is ungoogled-chromium, it goes from version to version, so if the source for that project gets updated, I'd like to avoid rebuilding unaffected object files | 20:09 |
| benschubert | Buildstream is sadly not able of doing it like that :/ | 20:11 |
| benschubert | If you can break the project in multiple chunks and have one element per chunk, you could alleviate that | 20:12 |
| benschubert | But that's not ideal | 20:12 |
| dftxbs3e | benschubert: I think it's one of the big flaws of CI systems currently | 20:13 |
| *** brlogger has joined #buildstream | 20:17 | |
| benschubert | https://gitlab.com/bloomberg/recc a caching compiler (kind of distcc + ccache, bit a bit easier) | 20:17 |
| dftxbs3e | hmm | 20:23 |
| benschubert | What would be your exact use case? | 20:31 |
| dftxbs3e | benschubert: I just can't stand having to spin up unnecessary resources for a single line of code changed | 20:43 |
| dftxbs3e | benschubert: this is the project: https://gitlab.com/lle-bout/ungoogled-chromium | 20:44 |
| benschubert | Ok, and you would like buildstream in the CI correct? | 20:45 |
| *** nimish2711 has quit IRC | 20:52 | |
| dftxbs3e | benschubert: I thought it could improve things in some way.. | 20:59 |
| benschubert | If you use a persistent local or remote cache, it will for elements that did not change. I remember us talking about making it more granular long ago, but I don't remember how the discussion staled | 21:06 |
| dftxbs3e | benschubert: well yeah.. chromium is a whole big element | 21:07 |
| benschubert | I know, I remember the pain in coming it :/ | 21:08 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!