| *** traveltissues has quit IRC | 05:39 | |
| *** tristan has quit IRC | 05:41 | |
| *** tristan has joined #buildstream | 06:54 | |
| *** ChanServ sets mode: +o tristan | 06:54 | |
| juergbi | pointswaves: at quick glance, in the junction you set the bootstrap_build_arch option but in the CI config of the other job you only set target_arch | 07:29 |
|---|---|---|
| juergbi | if I didn't misread anything | 07:29 |
| *** benschubert has joined #buildstream | 07:33 | |
| pointswaves | juergbi: i was trying to match https://gitlab.com/freedesktop-sdk/freedesktop-sdk/-/blob/willsalmon/bst2-buildbox-fix/Makefile#L24 i did another run including the snap_grade with no change | 07:45 |
| pointswaves | thanks for looking | 07:45 |
| juergbi | ah, right, only checked -o in .gitlab-ci.yml | 07:49 |
| tristan | benschubert, I'm in the middle of writing this up... and I think that the "overrides" table (currently on junctions in !1901) should be moved to project.conf inside a "junctions" configuration group, beside the new "duplacate" list | 07:49 |
| tristan | This is particularly relevant considering that we have to support full paths of element names, whereas it made more sense at the "junction" level as we were saying "While declaring this junction, override this junction's junction foo.bst with my junction bar.bst" | 07:50 |
| tristan | Having full paths at the junction level is less bound to the declaration of the junction itself, and I think it makes sense to put these both at the same place | 07:51 |
| tristan | juergbi, I know that it was a long conversation and don't know if you followed; feel free to comment now if you want or wait for the mail to take it in all at once :) | 07:52 |
| benschubert | tristan: I am for it, a single place to see all of it! :) | 07:54 |
| tristan | fwiw the relevant thing right now is: https://gitlab.com/BuildStream/buildstream/-/blob/tristan/junction-jungle/src/buildstream/plugins/elements/junction.py#L132 | 07:54 |
| tristan | Yeah, I liked having everything in junctions but the need to move the whitelisting feature to project.conf (and the need for deep pathing), means it's better to have it all in project.conf | 07:54 |
| tristan | I will do the writeup that way in that case | 07:54 |
| juergbi | I haven't been fully following the conversation. will try to catch up either here or via ML | 07:56 |
| *** traveltissues has joined #buildstream | 08:08 | |
| *** tristan has quit IRC | 08:24 | |
| *** phildawson has joined #buildstream | 08:30 | |
| *** santi has joined #buildstream | 08:48 | |
| *** tristan has joined #buildstream | 09:15 | |
| *** ChanServ sets mode: +o tristan | 09:15 | |
| *** phildawson has quit IRC | 09:33 | |
| tristan | whew | 10:27 |
| tristan | benschubert, juergbi; sent. | 10:27 |
| juergbi | ta, will take a look | 10:28 |
| *** tristan has quit IRC | 10:40 | |
| douglaswinship | Has anyone encountered the following error before? | 10:45 |
| douglaswinship | "Unable to mount /home/myusername/.cache/buildstream/build/jupyter-notebook-rwxv7xo0/scratch/_/mount: fusermount: unknown option 'nonempty'" | 10:46 |
| douglaswinship | I wonder if i'm using the wrong version of fuse/fusermount for bst 1.93.1? | 10:46 |
| *** tristan has joined #buildstream | 11:01 | |
| *** ChanServ sets mode: +o tristan | 11:01 | |
| juergbi | douglaswinship: this can happen on some distros with 1.93.1. fixed in master and bst-1, see https://gitlab.com/BuildStream/buildstream/-/merge_requests/1834 | 11:15 |
| douglaswinship | aha! thanks! | 11:20 |
| coldtom | i suppose https://gitlab.com/BuildStream/buildstream/-/issues/860 can be closed then :P | 11:20 |
| juergbi | ta | 11:32 |
| *** hasebastian has joined #buildstream | 12:57 | |
| scottclarke | if anyone has a minute, I would appreciate someone having a look at this https://gitlab.com/BuildStream/buildstream/-/merge_requests/1955 | 14:47 |
| WSalmon | I have been thinking similar scottclarke thanks for raising these! | 15:43 |
| scottclarke | WSalmon, thanks for your comment, regarding the second point I think my change also causes the hash to appear in the second box anyway | 15:45 |
| WSalmon | oh nice, hence the before and after comment :D | 15:51 |
| scottclarke | yeah makes sense | 15:56 |
| WSalmon | I need to have a look at that code section to see if there is any other useful bits of info we can brake out as i reagually have to dig in other places as bst dose not help in letting me know what is happening | 15:58 |
| *** santi has quit IRC | 17:44 | |
| *** hasebastian has quit IRC | 18:14 | |
| *** xjuan has joined #buildstream | 18:43 | |
| *** benschubert has quit IRC | 18:56 | |
| pointswaves | juergbi: turns out i had got the `sandbox: build-arch: %{arch}` in my project but it needed to be in the junction too | 19:30 |
| pointswaves | juergbi: i also found https://gitlab.com/BuildStream/buildstream/-/issues/1338 would it be better as a issue for one of the buildboxes? | 20:08 |
| *** jjardon has quit IRC | 22:35 | |
| *** jjardon has joined #buildstream | 22:36 | |
| *** ChanServ sets mode: +o jjardon | 22:36 | |
| *** xjuan has quit IRC | 22:44 | |
| *** xjuan has joined #buildstream | 22:47 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!