| *** alatiera has joined #buildstream | 01:19 | |
| *** tristan has quit IRC | 02:48 | |
| *** tristan has joined #buildstream | 03:21 | |
| *** nimish2711 has joined #buildstream | 03:23 | |
| *** nimish2711 has quit IRC | 03:38 | |
| *** nimish2711 has joined #buildstream | 03:38 | |
| *** nimish2711 has quit IRC | 04:03 | |
| *** nimish2711 has joined #buildstream | 04:08 | |
| *** nimish2711 has quit IRC | 04:19 | |
| *** nimish2711 has joined #buildstream | 04:20 | |
| *** alatiera has quit IRC | 04:25 | |
| *** cgmcintyre[m] has joined #buildstream | 04:27 | |
| *** connorshea[m] has joined #buildstream | 04:27 | |
| *** asingh_[m] has joined #buildstream | 04:27 | |
| *** awacheux[m] has joined #buildstream | 04:27 | |
| *** abderrahim[m] has joined #buildstream | 04:27 | |
| *** albfan[m] has joined #buildstream | 04:27 | |
| *** Demos[m] has joined #buildstream | 04:27 | |
| *** doras[m] has joined #buildstream | 04:27 | |
| *** dbuch has joined #buildstream | 04:27 | |
| *** Trevinho[m] has joined #buildstream | 04:27 | |
| *** dineshdb[m] has joined #buildstream | 04:27 | |
| *** jjardon[m] has joined #buildstream | 04:27 | |
| *** kailueke[m] has joined #buildstream | 04:27 | |
| *** pro[m] has joined #buildstream | 04:28 | |
| *** inigomartinez has joined #buildstream | 04:28 | |
| *** m_22[m] has joined #buildstream | 04:28 | |
| *** rafaelff[m] has joined #buildstream | 04:28 | |
| *** theawless[m] has joined #buildstream | 04:28 | |
| *** ssssam[m] has joined #buildstream | 04:28 | |
| *** waltervargas[m] has joined #buildstream | 04:28 | |
| *** segfault3[m] has joined #buildstream | 04:28 | |
| *** mattiasb has joined #buildstream | 04:28 | |
| *** krichter[m] has joined #buildstream | 04:28 | |
| *** tlater[m] has joined #buildstream | 04:28 | |
| *** nielsdg has joined #buildstream | 04:28 | |
| *** skullone[m] has joined #buildstream | 04:28 | |
| *** mrmcq2u[m] has joined #buildstream | 04:28 | |
| *** tchaik[m] has joined #buildstream | 04:28 | |
| *** reuben640[m] has joined #buildstream | 04:28 | |
| *** nimish2711 has quit IRC | 04:35 | |
| *** nimish2711 has joined #buildstream | 04:35 | |
| *** nimish2711 has quit IRC | 04:40 | |
| *** nimish2711 has joined #buildstream | 04:40 | |
| *** nimish2711 has quit IRC | 04:45 | |
| *** nimish2711 has joined #buildstream | 04:46 | |
| *** alatiera has joined #buildstream | 04:55 | |
| *** nimish2711 has quit IRC | 06:01 | |
| *** nimish2711 has joined #buildstream | 06:01 | |
| *** nimish2711 has quit IRC | 06:16 | |
| *** nimish2711 has joined #buildstream | 06:27 | |
| *** nimish2711 has quit IRC | 06:37 | |
| *** nimish2711 has joined #buildstream | 06:37 | |
| *** nimish2711 has quit IRC | 06:57 | |
| *** nimish2711 has joined #buildstream | 06:57 | |
| *** nimish2711 has quit IRC | 07:02 | |
| *** nimish2711 has joined #buildstream | 07:03 | |
| *** nimish2711 has quit IRC | 07:18 | |
| *** nimish2711 has joined #buildstream | 07:18 | |
| *** nimish2711 has joined #buildstream | 07:19 | |
| *** ikerperez has joined #buildstream | 07:23 | |
| *** rdale has joined #buildstream | 07:26 | |
| *** nimish2711 has quit IRC | 07:29 | |
| *** nimish2711 has joined #buildstream | 07:29 | |
| *** nimish2711 has joined #buildstream | 07:29 | |
| *** nimish2711 has joined #buildstream | 07:30 | |
| *** nielsdg_ has joined #buildstream | 07:32 | |
| *** nimish2711 has quit IRC | 07:40 | |
| *** nimish2711 has joined #buildstream | 07:40 | |
| *** nimish2711 has quit IRC | 07:45 | |
| *** nimish2711 has joined #buildstream | 07:45 | |
| * Kinnison takes a small box of cats and sets them amongst you pigeons :D | 07:46 | |
| *** jennis has joined #buildstream | 07:54 | |
| *** nimish2711 has quit IRC | 07:55 | |
| *** nimish2711 has joined #buildstream | 07:56 | |
| *** nimish2711 has quit IRC | 08:01 | |
| *** nimish2711 has joined #buildstream | 08:01 | |
| *** toscalix has joined #buildstream | 08:05 | |
| *** toscalix has quit IRC | 08:07 | |
| *** toscalix has joined #buildstream | 08:07 | |
| *** nimish2711 has quit IRC | 08:21 | |
| *** nimish2711 has joined #buildstream | 08:22 | |
| *** nimish2711 has quit IRC | 08:37 | |
| *** nimish2711 has joined #buildstream | 08:37 | |
| *** nimish2711 has quit IRC | 08:52 | |
| *** nimish2711 has joined #buildstream | 08:52 | |
| *** jonathanmaw has joined #buildstream | 08:57 | |
| *** nimish2711 has quit IRC | 09:02 | |
| *** nimish2711 has joined #buildstream | 09:03 | |
| *** phil has joined #buildstream | 09:03 | |
| *** tpollard has joined #buildstream | 09:04 | |
| *** nimish2711 has quit IRC | 09:13 | |
| *** nimish2711 has joined #buildstream | 09:13 | |
| *** lachlan has joined #buildstream | 09:46 | |
| *** nimish2711 has quit IRC | 09:48 | |
| *** nimish2711 has joined #buildstream | 09:48 | |
| phil | It looks like the CI in bst-experimental is still using a docker image with bst 1.2. I believe it needs to be updated to one which contains BuildStream master, but I'm not sure which image tag to use. I'm stuck between nightly, and nightly-master-56618785. My gut feeling is the later, as it would make for a more predictable CI, but given the rate of change in buildstream at the moment there's potential that this won't be updated as often as it | 09:51 |
|---|---|---|
| phil | should be. | 09:51 |
| *** nimish2711 has quit IRC | 09:53 | |
| Kinnison | experimental should surely use nightly? | 09:53 |
| Kinnison | it is, after all, experimental | 09:53 |
| Kinnison | At least until bst2 drops | 09:53 |
| *** nimish2711 has joined #buildstream | 09:54 | |
| jennis | I agree, nightly should be the one to use | 09:55 |
| *** lachlan has quit IRC | 10:04 | |
| *** nimish2711 has quit IRC | 10:04 | |
| *** nimish2711 has joined #buildstream | 10:04 | |
| phil | Nightly it is then | 10:05 |
| *** lachlan has joined #buildstream | 10:08 | |
| *** nimish2711 has quit IRC | 10:09 | |
| *** nimish2711 has joined #buildstream | 10:09 | |
| *** tristan has quit IRC | 10:12 | |
| *** tristan has joined #buildstream | 10:12 | |
| *** ChanServ sets mode: +o tristan | 10:12 | |
| tristan | phil, jennis, Kinnison ... I kind of disagree | 10:12 |
| tristan | Before you go ahead and change it, please just hear me out :) | 10:12 |
| tristan | I mean, obviously it should be using master | 10:12 |
| tristan | But... We have carefully setup CI before to always use explicit refs | 10:12 |
| tristan | This is because we want to be sure that the only thing that ever changes from on CI run to the next, is the commit which is under CI | 10:13 |
| tristan | So, the idea is that explicitly updating the ref to an external counterpart, *also* explicitly triggers CI | 10:14 |
| tristan | I believe I had a conversation with juergbi at some point where we discussed the possibility of a nightly automated MR, which would *attempt* to update the ref in a .gitlab-ci.yml, and merge if that passes | 10:15 |
| tristan | jennis, phil Does that make sense ^^^ ? | 10:15 |
| jennis | So you're proposing to have a testsuite-image container, and then clone + install BuildStream at the very latest master? Rather than whatever master was from last night? | 10:15 |
| phil | tristan, I could see arguments for both, hence my asking :) That was also my initial leaning, though as I said, I do worry slightly that a fixed image won't be updated often enough. | 10:15 |
| Kinnison | All I care is that at least one of the repos attempts to marry master-of-buildstream and master-of-bst-experimental | 10:15 |
| Kinnison | I don't mind how that happens | 10:16 |
| tristan | jennis, I am proposing that we *always* use an exact ref for any merge request | 10:16 |
| tristan | jennis, not a moving target, ever | 10:16 |
| tristan | jennis, and that *moving* that target equally needs to pass through CI | 10:17 |
| tristan | just trying to explain to jennis why this is important, to always know whether it was a change in BuildStream, or a change in a merge request, which causes a test to break | 10:18 |
| tristan | (I think Kinnison and phil already understood :)) | 10:18 |
| jennis | Yeah, I still don't think I'm following | 10:18 |
| *** nimish2711 has quit IRC | 10:19 | |
| tristan | jennis, Ok so this is the same reason that (A) we used to have all dependencies in a docker image preinstalled for BuildStream CI ... and that (B) When we changed that, we started pinning *all* dependencies to exact versions in a requirements.txt file (to allow easier docker maintenance) | 10:19 |
| *** lachlan has quit IRC | 10:20 | |
| *** nimish2711 has joined #buildstream | 10:20 | |
| tristan | jennis, Basically when someone submits a merge request... we don't want to introduce any changing noise, we want to know that it is the *commit* that we are testing, and not some external debris | 10:20 |
| jennis | Thanks tristan, also just had a quick chat with phil, I understand it now | 10:21 |
| tristan | jennis, Likewise, any update to the requirements.txt files also goes through CI, so when we update those, we know if an external dependency is what caused CI to break | 10:21 |
| tristan | So it would be possible for example, on some host... to setup a node.js script which automatically submits a merge request to bst-experimental every night | 10:22 |
| tristan | That would be the really awesome solution | 10:22 |
| phil | I do really like the idea of the CI periodically auto updating the image it uses | 10:22 |
| tristan | jennis, i.e. imagine that every night there is an automatically submitted merge request which updates .gitlab-ci.yml with the latest ref of BuildStream master | 10:23 |
| *** lachlan has joined #buildstream | 10:23 | |
| tristan | it's actually not all that hard to do | 10:23 |
| jennis | Yes I agree that something like that needs to happen otherwise we run the risk of our plugins becoming incompatible, right? | 10:23 |
| tristan | once you've got some sample of a node.js script and a host to run it on a cron job :) | 10:23 |
| tristan | jennis, Indeed, or it needs to be done periodically | 10:24 |
| tristan | automating it would be great though | 10:24 |
| jennis | I agree | 10:24 |
| phil | I imagine it could probably be done via a periodic gitlab CI job, which could commit directly if CI passes, rather than needing human intervention on a MR every day. | 10:24 |
| phil | And perhaps provide some sort of notification (e.g raise an issue) if the auto update fails. | 10:25 |
| phil | Any way, we're getting a bit sidetracked. For now I just want to update the image currently being used. | 10:26 |
| *** lachlan has quit IRC | 10:33 | |
| * phil goes for nightly-master-56618785 for now. | 10:34 | |
| phil | Crossing my fingers that CI passes | 10:34 |
| tristan | phil, I'm not sure that it is possible to do *on gitlab*, but it might even be ! | 10:36 |
| tristan | phil, I suspect we'd still need to use a node.js script from inside the job in order to trigger the merge request | 10:36 |
| tristan | as there is no self-merge-request-generation kind of feature that I know of | 10:36 |
| tristan | But automatically generating merge requests with the gitlab API is something I've done few years ago... | 10:36 |
| tristan | Could probably find that code, would have to search a bit | 10:37 |
| * phil wonders if the gitlab API has a python client | 10:37 | |
| phil | https://python-gitlab.readthedocs.io/en/stable/ | 10:37 |
| phil | It does :) | 10:37 |
| phil | I guess it was too much to hope that updating the image would just work | 10:39 |
| tristan | Oh that's cute, but in any case it's mostly just a simple restful API :) | 10:41 |
| gitlab-br-bot | marge-bot123 closed issue #995 (Cannot choose what tests to run) on buildstream https://gitlab.com/BuildStream/buildstream/issues/995 | 10:43 |
| gitlab-br-bot | marge-bot123 merged MR !1297 (phil/fixup-templated-test-collection->master: Ensure templated source tests aren't collected unconditionally) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1297 | 10:43 |
| *** lachlan has joined #buildstream | 10:52 | |
| * phil smiles at pytest --sw | 10:56 | |
| gitlab-br-bot | tpollard opened issue #1003 (Simplify metadata within Artifact class) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1003 | 11:09 |
| *** nimish2711 has quit IRC | 11:10 | |
| *** nimish2711 has joined #buildstream | 11:10 | |
| *** tristan has quit IRC | 11:11 | |
| *** rdale has quit IRC | 11:12 | |
| *** nimish2711 has quit IRC | 11:20 | |
| *** nimish2711 has joined #buildstream | 11:20 | |
| *** nimish2711 has quit IRC | 11:30 | |
| *** nimish2711 has joined #buildstream | 11:31 | |
| laurence | ah we reached issue #1000 | 11:33 |
| *** tristan has joined #buildstream | 11:36 | |
| *** nielsdg_ has quit IRC | 11:38 | |
| *** nimish2711 has quit IRC | 11:40 | |
| *** nimish2711 has joined #buildstream | 11:42 | |
| *** nielsdg_ has joined #buildstream | 12:01 | |
| *** kapil___ has joined #buildstream | 12:02 | |
| tpollard | nice summary email on the plugins tristan | 12:02 |
| *** alatiera_ has joined #buildstream | 12:35 | |
| *** alatiera has quit IRC | 12:35 | |
| *** alatiera_ is now known as alatiera | 12:36 | |
| *** rdale has joined #buildstream | 12:50 | |
| gitlab-br-bot | cs-shadow opened (was WIP) MR !1293 (chandan/pseudo-junction->master: Add support for defining target for junction elements) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1293 | 13:31 |
| *** alatiera_ has joined #buildstream | 13:36 | |
| *** alatiera has quit IRC | 13:37 | |
| *** alatiera_ is now known as alatiera | 13:37 | |
| *** nimish2711 has quit IRC | 13:43 | |
| *** phildawson_ has joined #buildstream | 14:03 | |
| *** phil has quit IRC | 14:03 | |
| *** nimish2711 has joined #buildstream | 14:09 | |
| *** nimish2711 has quit IRC | 14:10 | |
| *** nimish2711 has joined #buildstream | 14:11 | |
| *** phildawson_ has quit IRC | 14:14 | |
| *** phildawson_ has joined #buildstream | 14:14 | |
| *** nimish2711 has quit IRC | 14:15 | |
| *** nimish2711 has joined #buildstream | 14:16 | |
| tpollard | juergbi: I'll get around to replying to comments on !1292 just getting it rebased against master first | 14:27 |
| gitlab-br-bot | MR !1292: Artifact as a Proto: rework https://gitlab.com/BuildStream/buildstream/merge_requests/1292 | 14:27 |
| juergbi | sure, makes sense | 14:28 |
| coldtom | any interest for a hg plugin in bst-external/bst-plugins-experimental? i've got one which *seems* to work here https://gitlab.com/coldtom/bst-in-bst/blob/master/plugins/sources/hg.py | 14:31 |
| *** lachlan has quit IRC | 14:57 | |
| *** toscalix has quit IRC | 15:00 | |
| *** lachlan has joined #buildstream | 15:06 | |
| *** lachlan has quit IRC | 15:29 | |
| jonathanmaw | tristan: jennis and I have sorted out the changes after !1271, ready for you to have a look at | 15:42 |
| gitlab-br-bot | MR !1271: Follow-up from "Artifact 'abstraction' class" https://gitlab.com/BuildStream/buildstream/merge_requests/1271 | 15:42 |
| gitlab-br-bot | jennis approved MR !1293 (chandan/pseudo-junction->master: Add support for defining target for junction elements) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1293 | 15:54 |
| *** lachlan has joined #buildstream | 16:07 | |
| *** alatiera has quit IRC | 16:14 | |
| *** alatiera_ has joined #buildstream | 16:14 | |
| *** alatiera_ is now known as alatiera | 16:15 | |
| *** lachlan has quit IRC | 16:16 | |
| cs-shadow | juergbi: is !1293 good for you now? I was about to marge it but was wondering if you wanted to have a final look | 16:40 |
| *** lachlan has joined #buildstream | 16:42 | |
| juergbi | cs-shadow: it generally looks good to me. I'm wondering about 'Note that when targeting another junction, the names of the junction element must not be the same as the name of the target.' | 16:43 |
| juergbi | I understand why that's the case but we should probably lift this limitation | 16:43 |
| juergbi | and I'm wondering whether we should try to lift it now or defer this | 16:43 |
| cs-shadow | juergbi: I was thinking the same thing. I think it will be useful to not have this limitation but that may require some rework around the loader. Would you be okay with creating an issue to fix this separately? | 16:45 |
| cs-shadow | I do think we should do that but I don't want to block this change because of that | 16:45 |
| cs-shadow | (I mean I can create the issue if that's okay with you) | 16:46 |
| juergbi | cs-shadow: ok, let's open an issue for this. we might end up tweaking the semantics a bit but I think the approach in the MR is simple enough for this to be low risk | 16:46 |
| cs-shadow | juergbi: cool, thanks! I am bit keen to get this landed this soon as this will unblock a few things on our side | 16:47 |
| *** tpollard has quit IRC | 16:51 | |
| gitlab-br-bot | cs-shadow opened issue #1005 (Allow junction elements to be named same as junction targets) on buildstream https://gitlab.com/BuildStream/buildstream/issues/1005 | 17:07 |
| *** lachlan has quit IRC | 17:09 | |
| gitlab-br-bot | marge-bot123 merged MR !1293 (chandan/pseudo-junction->master: Add support for defining target for junction elements) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/1293 | 17:15 |
| *** jonathanmaw has quit IRC | 17:36 | |
| *** nielsdg_ has quit IRC | 18:33 | |
| *** nimish2711 has quit IRC | 19:33 | |
| *** kapil___ has quit IRC | 20:30 | |
| *** cs-shadow has quit IRC | 20:38 | |
| *** alatiera has quit IRC | 21:35 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!