| *** phoenix has quit IRC | 00:02 | |
| *** phoenix has joined #buildstream | 00:10 | |
| *** phoenix has quit IRC | 00:57 | |
| *** phoenix has joined #buildstream | 01:12 | |
| *** phoenix has quit IRC | 01:26 | |
| *** tristan has quit IRC | 04:43 | |
| *** tristan has joined #buildstream | 04:59 | |
| *** hasebastian has joined #buildstream | 05:13 | |
| gitlab-br-bot | tristanvb approved MR !1851 (valentindavid/include-options-from-junction->master: Process options in includes files with the options of their junction) on buildstream https://gitlab.com/BuildStream/buildstream/-/merge_requests/1851 | 05:52 |
|---|---|---|
| *** ChanServ sets mode: +o tristan | 05:52 | |
| tristan | Shall we go ahead and merge !1851 ? | 05:52 |
| *** benschubert has joined #buildstream | 07:06 | |
| *** tristan has quit IRC | 07:49 | |
| *** tpollard has joined #buildstream | 08:18 | |
| *** hasebastian has quit IRC | 08:23 | |
| *** santi has joined #buildstream | 08:34 | |
| gitlab-br-bot | willsalmon opened issue #1287 (New version of bst can not use repo that has had workspaces) on buildstream https://gitlab.com/BuildStream/buildstream/-/issues/1287 | 08:36 |
| gitlab-br-bot | BenjaminSchubert closed issue #1287 (New version of bst can not use repo that has had workspaces) on buildstream https://gitlab.com/BuildStream/buildstream/-/issues/1287 | 08:39 |
| *** rdale has joined #buildstream | 08:40 | |
| gitlab-br-bot | traveltissues reopened issue #1287 (New version of bst can not use repo that has had workspaces) on buildstream https://gitlab.com/BuildStream/buildstream/-/issues/1287 | 08:52 |
| WSalmon | anyone else had anything like this before, ill try and get a fix for this soon, it looks simple enough, if anyone has any implimentation details or other similar things i should bear in mind please put something on the issue https://gitlab.com/BuildStream/buildstream/-/issues/1287 | 08:52 |
| WSalmon | benschubert, you have mentioned my ticket as the duplicate | 08:53 |
| benschubert | yep, | 08:53 |
| traveltissues | i was a bit confused | 08:54 |
| benschubert | ah sorry | 08:54 |
| benschubert | meant #1266 | 08:54 |
| gitlab-br-bot | Issue #1266: Can't load a project with an older workspace to a junction https://gitlab.com/BuildStream/buildstream/-/issues/1266 | 08:54 |
| benschubert | problem is the same | 08:54 |
| benschubert | though your's actually a bit broader | 08:54 |
| traveltissues | ah, apologies | 08:55 |
| benschubert | I'd suggest closing yours WSalmon and adding that this would happen even after closing the workspace on the other ticket? | 08:56 |
| *** cs-shadow has joined #buildstream | 09:18 | |
| *** lachlan has joined #buildstream | 09:25 | |
| *** phoenix has joined #buildstream | 09:27 | |
| *** narispo has quit IRC | 09:44 | |
| *** tpollard has quit IRC | 09:50 | |
| *** tpollard has joined #buildstream | 09:50 | |
| gitlab-br-bot | willsalmon closed issue #1287 (New version of bst can not use repo that has had workspaces) on buildstream https://gitlab.com/BuildStream/buildstream/-/issues/1287 | 09:57 |
| *** tristan has joined #buildstream | 10:14 | |
| *** lachlan has quit IRC | 10:38 | |
| *** lachlan has joined #buildstream | 10:42 | |
| *** lachlan has quit IRC | 11:06 | |
| *** lachlan has joined #buildstream | 11:06 | |
| *** lachlan has quit IRC | 11:15 | |
| tpollard | is there a final goal for buildtrees for bst2? | 11:15 |
| tpollard | or is the current functionality the final cut? | 11:17 |
| *** lachlan has joined #buildstream | 11:31 | |
| *** lachlan has quit IRC | 11:32 | |
| *** lachlan has joined #buildstream | 11:37 | |
| juergbi | tpollard: there isn't anything in the current plan. is there a particular aspect that you think needs work? | 11:54 |
| *** ChanServ sets mode: +o tristan | 12:11 | |
| tristan | my memory is foggy and I've missed a great deal; but I suspect at least with buildbox and fuse that the performance of this use case is much improved since the original addition of build trees | 12:11 |
| tpollard | just wondering if they were being utilised as first envisioned that's all | 12:16 |
| tpollard | i.e, are they being utilised with workspaces & the underlying vcs, if appropriate | 12:22 |
| juergbi | tpollard: buildtrees are used for incremental workspace builds. however, it's not currently possible to do an incremental build using a non-workspace build as starting point | 12:28 |
| juergbi | we don't capture mtimes for non-workspace buildtrees | 12:28 |
| tristan | I think that `bst artifact list-contents` should prefix the paths with / | 12:33 |
| tpollard | so, the only use for buildtrees in standard build artifacts is for bst shell? | 12:33 |
| tristan | ... would be more consistent with split-rules settings and overlap whitelists | 12:33 |
| tpollard | if of course the config is enabled to cache/push/pull them | 12:33 |
| *** phoenix has quit IRC | 12:45 | |
| *** tpollard has quit IRC | 12:52 | |
| *** tpollard has joined #buildstream | 12:52 | |
| *** persia has quit IRC | 12:54 | |
| *** persia has joined #buildstream | 12:55 | |
| juergbi | tpollard: for me the main use case is to debug failed builds (and incremental workspace builds) and buidltrees are enabled by default for the failure case | 12:58 |
| WSalmon | juergbi, how do you inspect the buildtree in such a case without a workspace? can you use a checkout? | 12:59 |
| tpollard | shell | 12:59 |
| juergbi | yep, shell | 12:59 |
| juergbi | that said, allowing export/checkout would probably make sense as well | 13:00 |
| tpollard | Could be nice, say if you wanted to send tarball elsewhere for inspection (without using a common cache) | 13:05 |
| tristan | Any objections to merging !1851 now ? | 13:13 |
| gitlab-br-bot | MR !1851: Process options in includes files with the options of their junction https://gitlab.com/BuildStream/buildstream/-/merge_requests/1851 | 13:13 |
| * tristan now has an ever growing documentation MR to back it up | 13:14 | |
| *** lachlan has quit IRC | 13:32 | |
| juergbi | tristan: I haven't looked into the details with regards to correct processing order. however, I don't have any objections | 13:35 |
| tristan | juergbi, yeah I believe it's well tested | 13:37 |
| juergbi | tristan: btw: have you had a chance to take a look at !1865 ? | 13:37 |
| gitlab-br-bot | MR !1865: element.py: Fix strong cache key calculation in non-strict mode https://gitlab.com/BuildStream/buildstream/-/merge_requests/1865 | 13:37 |
| tristan | That's an MR now ! | 13:38 |
| tristan | I'll rebase my branch on it for a test | 13:38 |
| tristan | What the hell... | 13:44 |
| tristan | coldtom, valentind ... is this something that needs to be fixed in the infra: https://gitlab.com/BuildStream/buildstream/-/jobs/512056691 ? | 13:44 |
| tristan | Looks like docs built successfully but failed to coordinate their artifacts with the uploader, or smth | 13:44 |
| * tristan heads out for sushi... | 13:45 | |
| *** lachlan has joined #buildstream | 13:46 | |
| valentind | tristan, It looks like an issue from gitlab. | 13:53 |
| valentind | We do not have control on where the CI artifacts are stored I think. Only CI cache. | 13:53 |
| *** lachlan has quit IRC | 14:06 | |
| *** lachlan has joined #buildstream | 14:55 | |
| tristan | dang | 15:10 |
| * tristan hits retry with futility | 15:10 | |
| *** rdale_ct has joined #buildstream | 15:16 | |
| *** rdale has quit IRC | 15:19 | |
| valentind | This is configured on the gitlab instance. It is not on the runner side. | 15:28 |
| valentind | https://gitlab.com/gitlab-com/gl-infra/production/-/issues/1944 | 15:28 |
| *** lachlan has quit IRC | 15:34 | |
| *** lachlan has joined #buildstream | 15:36 | |
| *** lachlan has quit IRC | 16:10 | |
| philn | i'm trying to debug some non-whitelisted overlaps in my SDK... is there any tips&tricks besides trial and error poking? | 16:14 |
| philn | i'd need to know which element pulls a specific one | 16:15 |
| tristan | juergbi, anyway just to say, the branch definitely fixes the side effect, I've just been waiting for gitlab to get it's story straight so I can reply with links | 16:15 |
| juergbi | great, thanks for confirming | 16:16 |
| tristan | https://gitlab.com/BuildStream/buildstream/-/jobs/512252895 | 16:20 |
| tristan | valentind, ^^^^^^^^^^^^^ this one has been saying 3 seconds for a while longer than three seconds ... | 16:21 |
| tristan | :-S | 16:21 |
| tristan | sorry for blind ping, I wish this stuff "just worked" | 16:21 |
| tristan | I guess that having so many parallel test jobs is a contributing factor to things being randomly broken | 16:22 |
| juergbi | tristan: we again have the maximum amount of 100 droplets | 16:23 |
| juergbi | I'll drop older ones | 16:23 |
| juergbi | so we still have the issue of droplets not getting shut down :-/ | 16:23 |
| juergbi | tristan: new jobs should work now but automatic retry can take a very long time. manual cancel+retry stuck jobs speeds it up | 16:24 |
| juergbi | I'm wondering what change introduced this breakage. iirc, it was more reliable 6 months ago or so | 16:25 |
| tpollard | philn: I think fdsdk might have some tooling around working through overlaps | 16:26 |
| *** lachlan has joined #buildstream | 16:37 | |
| *** lachlan has quit IRC | 16:50 | |
| valentind | juergbi, I know there was a change introduce in some script that stopped removing droplets. | 16:54 |
| valentind | I reverted it. | 16:54 |
| valentind | So I expected it would work again. | 16:54 |
| valentind | But there was also an issue with gitlab runner creating way too many machines sometimes. | 16:56 |
| *** lachlan has joined #buildstream | 16:57 | |
| *** rdale_ct has quit IRC | 17:01 | |
| valentind | So, there was a typo in the systemd configuration file, that was stopping the docker machine to be cleaned. | 17:02 |
| valentind | I fixed that. | 17:02 |
| *** lachlan has quit IRC | 17:08 | |
| valentind | The docker-machine cache is cleaned. And made sure the script works. | 17:10 |
| *** santi has quit IRC | 17:12 | |
| juergbi | ta | 17:17 |
| *** tpollard has quit IRC | 17:19 | |
| *** lachlan has joined #buildstream | 17:28 | |
| *** lachlan has quit IRC | 17:35 | |
| douglaswinship | When you use bst shell to test out an element, can the sandbox reach an internet connection? | 17:43 |
| douglaswinship | I was told that it could, but I keep getting "could not resolve host" when I try to run curl as a test. | 17:43 |
| juergbi | douglaswinship: if you don't specify --isolate, bst shell does not restrict network access | 17:53 |
| juergbi | however, you typically need a suitable /etc/resolv.conf in the sandbox | 17:53 |
| juergbi | you can use project.conf host-files to bind-mount the host /etc/resolv.conf | 17:54 |
| douglaswinship | juergbi: thanks! I shall read up on what /etc/resolv.conf does! | 17:54 |
| juergbi | douglaswinship: see https://buildstream.gitlab.io/buildstream/format_project.html#host-files for an example | 17:54 |
| douglaswinship | thanks! | 17:55 |
| juergbi | resolv.conf contains the IP address of nameservers | 17:55 |
| juergbi | normally managed by your DHCP client but the sandbox normally doesn't know about this, of course | 17:55 |
| *** phoenix has joined #buildstream | 18:10 | |
| *** benschubert has quit IRC | 19:29 | |
| *** phoenix has quit IRC | 20:02 | |
| *** phoenix has joined #buildstream | 20:04 | |
| *** cs-shadow has quit IRC | 21:25 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!