| *** benschubert has quit IRC | 00:05 | |
| *** nimish2711 has joined #buildstream | 04:07 | |
| *** nimish2711 has quit IRC | 04:12 | |
| *** nimish2711 has joined #buildstream | 04:13 | |
| *** nimish2711 has quit IRC | 04:37 | |
| *** nimish2711 has joined #buildstream | 04:38 | |
| *** nimish2711 has quit IRC | 04:43 | |
| *** nimish2711 has joined #buildstream | 04:43 | |
| *** nimish2711 has quit IRC | 04:53 | |
| *** juergbi has quit IRC | 05:07 | |
| *** juergbi has joined #buildstream | 05:09 | |
| *** nimish2711 has joined #buildstream | 05:39 | |
| *** nimish2711 has quit IRC | 05:59 | |
| *** nimish2711 has joined #buildstream | 05:59 | |
| *** nimish2711 has quit IRC | 06:24 | |
| *** nimish2711 has joined #buildstream | 06:25 | |
| *** paulsherwood has joined #buildstream | 06:32 | |
| *** paulsherwood has joined #buildstream | 06:32 | |
| *** paulsherwood has joined #buildstream | 06:32 | |
| *** nimish2711 has quit IRC | 06:40 | |
| *** tristan has joined #buildstream | 06:41 | |
| *** nimish2711 has joined #buildstream | 06:55 | |
| *** nimish2711 has quit IRC | 07:31 | |
| *** nimish2711 has joined #buildstream | 07:34 | |
| *** ikerperez has joined #buildstream | 07:56 | |
| *** rdale has joined #buildstream | 07:58 | |
| *** benschubert has joined #buildstream | 08:07 | |
| *** bochecha has joined #buildstream | 08:18 | |
| *** raoul has joined #buildstream | 08:43 | |
| *** bochecha has quit IRC | 08:45 | |
| WSalmon | hi guys, i have been doing to work with the python tar lib, and i think i have some idea's for https://gitlab.com/BuildStream/buildstream/issues/912 could i get some feedback if this sulution seems good? I others like it then i think we can mark it as new comers but only if we are exact about what we want | 08:54 |
|---|---|---|
| WSalmon | tristan, juergbi ^ | 08:55 |
| *** ChanServ sets mode: +o tristan | 09:02 | |
| tristan | WSalmon, numeric_owner according to the docs, will attempt to set the ownership user/group bits according to the numbers recorded in the tarfile rather than attempting to lookup the named owners on the target system where things are being extracted, if I am reading the docs correctly I believe that is what it means | 09:02 |
| tristan | WSalmon, However for source extractions, I think we *always* want ownership of extracted files to be given to the effective uid/gid of the extracting process | 09:03 |
| tristan | I could be wrong about that, or we might want an option later down the road, for extracting sysroots where we want to preserve the ownership bits, but that must block on support for recording ownership bits in CAS fist | 09:04 |
| tristan | *first | 09:04 |
| *** phil has joined #buildstream | 09:04 | |
| *** bochecha has joined #buildstream | 09:04 | |
| tristan | However currently, extractions happen on a local filesystem without a privileged user, before being injected into a build | 09:04 |
| WSalmon | yep, thats why extractfile seems like such a good solution, for now, until the cas becomes more fancy. | 09:04 |
| WSalmon | yep, thats why `extractfile` seems like such a good solution, for now, until the cas becomes more fancy. | 09:05 |
| tristan | WSalmon, I'm a bit skeptical about dumbing down the surrounding code to only do what CAS *currently* supports | 09:06 |
| WSalmon | Not only have i been working with Tars, but something simmilar to this bug got me the other day for something i was trying to build | 09:06 |
| tristan | i.e. I would much prefer that the day that we add support to CAS to record richer FS metadata, that we don't have to go chasing down other code paths | 09:06 |
| WSalmon | once the cas can take more info we can just add exactly what we can handle | 09:06 |
| tristan | in order to leverage it | 09:06 |
| tristan | WSalmon, however I think these are rather separate topics | 09:07 |
| tristan | We are talking about source extractions | 09:07 |
| tristan | Not artifacts | 09:07 |
| tristan | So the barrier here is that the user will always have limited permissions of what they can do with a source extraction | 09:08 |
| tristan | Once it is in an artifact, permissions, ownership bits, abstract OS dependent filesystem metadata can start to be recorded | 09:08 |
| tristan | And CAS supporting that should just transparently/automatically support everything it can (not only CAS, I should say the sandboxing also plays a part in this) | 09:08 |
| *** phil has quit IRC | 09:09 | |
| *** phildawson has joined #buildstream | 09:10 | |
| tristan | WSalmon, I think that first off, extracted files need to belong to the active user in any case | 09:10 |
| *** benbrown has joined #buildstream | 09:11 | |
| tristan | The behavior of GNU Tar is ... IIRC, to automatically assign files to the active user if the user is not superuser, and to assign UID/GID recorded in the tarfile in the case that the extracting user is superuser | 09:12 |
| tristan | WSalmon, I dont recall if it is that, or if it is a try/fallback approach for each file, but the point is that it is anyway nondeterministic (you can get different results depending on your privileges), that's not good | 09:12 |
| tristan | So making sure that extracted files always belong to the user extracting them is certainly a good thing | 09:13 |
| *** jonathanmaw has joined #buildstream | 09:14 | |
| WSalmon | it seems like we can make the plugin sufficently reprojecable with `extractfile` but anything else would have some unknowns | 09:30 |
| tristan | It's quite possible that we should be looping through the tarball itself rather than extractall yes | 09:31 |
| tristan | It's probably not a newcommers issue, there are a few things that need careful consideration (for instance, the globbing of what to extract needs to be preserved) | 09:32 |
| tristan | Well, not sure if that's what newcommers means, we should probably not paint it as "dead easy to solve" | 09:33 |
| *** bochecha has quit IRC | 09:39 | |
| *** bochecha has joined #buildstream | 09:39 | |
| *** lachlan has joined #buildstream | 09:42 | |
| *** tristan has quit IRC | 09:49 | |
| *** lachlan has quit IRC | 09:50 | |
| *** lachlan has joined #buildstream | 09:56 | |
| *** lachlan has quit IRC | 10:01 | |
| Kinnison | jonathanmaw: are our WSL runners dead? | 10:02 |
| * jonathanmaw has a look | 10:02 | |
| jonathanmaw | WSL runner's log has a lot of "your browser sent an invalid request" messages, not sure if that's client error or server error though | 10:04 |
| *** tristan has joined #buildstream | 10:05 | |
| jonathanmaw | hmm, the CI settings say that they can't contact the runner | 10:05 |
| jonathanmaw | I'll give it a kick | 10:05 |
| jonathanmaw | It's receiving jobs now I restarted it | 10:07 |
| *** lachlan has joined #buildstream | 10:19 | |
| *** lachlan has quit IRC | 10:31 | |
| *** lachlan has joined #buildstream | 10:37 | |
| raoul | phildawson, is there a way of running a single plugins test suite now that it uses the generic tests? Would like to run those without having to run the whole test suite | 10:44 |
| phildawson | raoul, there is indeed. See https://gitlab.com/BuildStream/buildstream/blob/master/tox.ini#L134 for the details :) | 10:48 |
| raoul | Ah nice, thank you :) | 10:49 |
| raoul | can you not pass pytest arguments with this? i.e. not use -n 4 | 10:56 |
| *** lachlan has quit IRC | 11:06 | |
| *** lachlan has joined #buildstream | 11:08 | |
| *** lachlan has quit IRC | 11:12 | |
| *** nimish2711 has quit IRC | 11:39 | |
| *** nimish2711 has joined #buildstream | 11:48 | |
| *** tchaik[m] has quit IRC | 12:34 | |
| *** tristan has quit IRC | 13:14 | |
| jonathanmaw | jennis: are you satisfied by the explanation to https://gitlab.com/BuildStream/buildstream/merge_requests/1342#note_170862989 ? | 13:17 |
| jennis | jonathanmaw, yeah looks good, I must've missed the if statement in update state | 13:19 |
| *** lachlan has joined #buildstream | 13:40 | |
| *** tristan has joined #buildstream | 13:41 | |
| *** lachlan has quit IRC | 13:46 | |
| gitlab-br-bot | coldtom opened issue #1026 (Add meson subproject support) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1026 | 13:47 |
| *** lachlan has joined #buildstream | 13:53 | |
| *** lachlan has quit IRC | 14:04 | |
| gitlab-br-bot | raoul.hidalgocharman opened MR !1345 (raoul/440-fix->master: Fix init_workspace for multiple sources) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1345 | 14:16 |
| raoul | ^ Noticed a mistake I made in source cache so made a fix MR. Pretty small so should be easy to review and get in. | 14:17 |
| *** lachlan has joined #buildstream | 14:20 | |
| *** lachlan has quit IRC | 14:29 | |
| *** lachlan has joined #buildstream | 14:33 | |
| *** lachlan has quit IRC | 14:45 | |
| WSalmon | *Im asking for a friend, is there a easy way to allow all overlaps even if they seem dangerous? only to be used while hacking you understand ;) i think overlap-whitelist should be set to something like `*` can i do that in the project.conf vars? | 14:50 |
| *** lachlan has joined #buildstream | 15:01 | |
| *** juergbi has quit IRC | 15:03 | |
| *** juergbi has joined #buildstream | 15:15 | |
| *** lachlan has quit IRC | 15:53 | |
| *** bochecha has quit IRC | 15:57 | |
| *** lachlan has joined #buildstream | 16:01 | |
| *** nimish2711 has quit IRC | 16:18 | |
| *** raoul has quit IRC | 16:30 | |
| gitlab-br-bot | tpollard approved MR !1345 (raoul/440-fix->master: Fix init_workspace for multiple sources) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1345 | 17:03 |
| *** rdale has quit IRC | 17:06 | |
| *** lachlan has quit IRC | 17:14 | |
| *** jonathanmaw has quit IRC | 17:35 | |
| *** xjuan has joined #buildstream | 17:46 | |
| gitlab-br-bot | marge-bot123 merged MR !1342 (shared/split_out_update_state_2->master: Split Element._update_state into cache key and artifact state methods) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1342 | 18:19 |
| *** slaf has quit IRC | 18:22 | |
| *** slaf has joined #buildstream | 18:36 | |
| *** slaf has joined #buildstream | 18:37 | |
| *** slaf has joined #buildstream | 18:37 | |
| *** slaf has joined #buildstream | 18:37 | |
| *** slaf has joined #buildstream | 18:37 | |
| *** slaf has joined #buildstream | 18:38 | |
| *** slaf has joined #buildstream | 18:38 | |
| *** slaf has joined #buildstream | 18:38 | |
| *** nimish2711 has joined #buildstream | 18:39 | |
| *** xjuan has quit IRC | 19:42 | |
| *** nimish2711 has quit IRC | 19:57 | |
| *** em has joined #buildstream | 21:39 | |
| *** em has quit IRC | 21:40 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!