| *** tristan has quit IRC | 00:58 | |
| *** tristan has joined #buildstream | 01:02 | |
| *** tristan has quit IRC | 06:55 | |
| *** jude has joined #buildstream | 06:58 | |
| *** tristan has joined #buildstream | 07:25 | |
| *** valentind has joined #buildstream | 08:22 | |
| adds68 | Hi, would anybody know what option in buildstream would map to this option in the flatpak builder >> https://github.com/flatpak/freedesktop-sdk-images/blob/344108d97b33f95a39f0d8aae801714515d29362/org.freedesktop.Sdk.json.in#L2463 ? | 08:25 |
|---|---|---|
| *** jonathanmaw has joined #buildstream | 08:34 | |
| *** semanticdesign has joined #buildstream | 08:48 | |
| *** semanticdesign_ has joined #buildstream | 08:48 | |
| tristan | adds68, config-opts ? | 08:53 |
| tristan | is that what you mean ? | 08:53 |
| adds68 | tiagogomes, hey, yea | 08:53 |
| adds68 | tristan, ** | 08:53 |
| adds68 | tristan, i'm looking at https://buildstream.gitlab.io/buildstream/format.html#kind 'configure-commands' | 08:53 |
| adds68 | would that be correct? | 08:53 |
| tristan | adds68, I assume you've got a checkout of flatpak-builder, but to know the correct answer, it's around here: https://github.com/flatpak/flatpak-builder/blob/master/src/builder-module.c#L1301 | 08:55 |
| tristan | adds68, for an autotools element (i.e. https://buildstream.gitlab.io/buildstream/elements/autotools.html#module-elements.autotools) | 08:56 |
| tristan | (that page is reachable via the top page -> Builtin Plugins -> autotools) | 08:56 |
| tristan | adds68, you probably want to just convert to an autotools element which sets the %{conf-extra} variable | 08:57 |
| tristan | adds68, that way you just tell the regular/default configure-commands to add those extra options | 08:57 |
| tristan | adds68, of course, according to flatpak-builder source code, it might be slightly different depending on the type of build system being used (cmake, meson, etc) | 08:57 |
| adds68 | tristan, ahh that is where the confusion is for me, do i use 'conf-args' or 'configure-commands' | 08:57 |
| tristan | everything is a possibility | 08:58 |
| tristan | but conf-args are there for this sort of convenience | 08:58 |
| adds68 | tristan, yea sam suggested that, so i'm just assuming autotools for now | 08:58 |
| adds68 | tristan, awesome thanks, i shall try this for the first pass | 08:58 |
| tristan | builder-module.c should reveal which build system you need to care about for what you are parsing :) | 08:58 |
| adds68 | tristan, as i'm only parsing the json data i sadly don't seem to have that information atm | 08:59 |
| tristan | I'm not sure what "simple" means in there, looks like a fairly recent addition | 08:59 |
| tristan | adds68, of course you do, in flatpak json it's "buildsystem" | 09:00 |
| tristan | I think autotools must be the default so it's not listed for that particular one you linked | 09:00 |
| tristan | adds68, look at "openal" in that same json, it uses "buildsystem" : "cmake" | 09:00 |
| tristan | adds68, in builder-module.c, it does some conditional stuff based on the build system, which you would need to take into account to decide what element kind to generate and such | 09:01 |
| adds68 | tristan, ah! | 09:03 |
| adds68 | tristan, how did i miss that! Well i've designed the parser based on classes which map to buildstream elements | 09:04 |
| adds68 | so i can just create some more classes to do with these build systems | 09:04 |
| tristan | So now you know how to decide which class to use :) | 09:04 |
| adds68 | tristan, awesome, thanks :) | 09:04 |
| tristan | indeed, it's quite convertible stuff | 09:04 |
| adds68 | tristan, i must say buildstream definitions are so much easier to read/understand than the json/flatpak builder stuff | 09:05 |
| *** tlater has joined #buildstream | 09:10 | |
| *** ssam2 has joined #buildstream | 09:18 | |
| gitlab-br-bot | push on buildstream@tracking-changes (by Tristan Maat): 5 commits (last: Issue #113: Split tracking and saving in `bst build`) https://gitlab.com/BuildStream/buildstream/commit/9fa7a1a9557a13c37976b45716937a335c5d818d | 09:27 |
| gitlab-br-bot | buildstream: merge request (tracking-changes->master: WIP: Tracking changes) #119 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/119 | 09:27 |
| tlater | :O gitlab-br-bot is back | 09:27 |
| tlater | tristan: This may be a slightly stupid question, but I'm assuming we want to track Scope.ALL for recursive elements? | 09:33 |
| gitlab-br-bot | buildstream: merge request (sam/artifacts-url-fix->master: _pipeline.py: Give the correct artifact pull URL in remote ticker) #120 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/120 | 09:33 |
| gitlab-br-bot | push on buildstream@sam/artifacts-url-fix (by Tristan Van Berkom): 2 commits (last: Issue #124: Add test for staging to element build directory) https://gitlab.com/BuildStream/buildstream/commit/35d7051a07451e0647f60963d0485908b1d0139a | 09:33 |
| tristan | tlater, that's a good question, indeed I think so - which means that it would expand on the default pipeline for building | 09:34 |
| tristan | Also, please use your imagination, I think right now there are a *lot* of proposed new tracking options to `bst build`, I like the recursive/except semantic for users (more practical than listing individual elements)... | 09:35 |
| tristan | tlater, but what I mean by imagination is, we want to ideally support all the things which have been proposed, but with the most convenient CLI semantics (and least CLI options) as possible | 09:36 |
| tristan | not sure the current proposed options achieves that, maybe it does, dont have time to think about it myself right now though. | 09:36 |
| * tlater has not seen all proposals yet, but will have a read | 09:36 | |
| tristan | tlater, it may sound tempting to add a subcommand to `bst build` for tracking specifics, but I worry that it may make something else I want to do impossible | 09:37 |
| tristan | tlater, i.e., now that we've ditched variants, it will be possible for all of the `bst <options> command <options> ELEMENT` to instead become `bst <options> command <options> ELEMENT [ELEMENT2 [ELEMENT3...]]` | 09:38 |
| tlater | Hm, yeah, if we had a subcommand that won't work too well | 09:39 |
| tristan | And I wanna do that, transform every command that currently takes a single TARGET option, to make it an unlimited number of TARGETS on the same invocation | 09:39 |
| ssam2 | great idea | 09:39 |
| * tristan should file it so it's on the map | 09:41 | |
| gitlab-br-bot | buildstream: issue #129 ("Allow specifying multiple target elements in ever command") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/129 | 09:44 |
| tlater | tristan: Are the proposals for build tracking accumulated somewhere? I can't find any on the ML | 09:45 |
| tlater | Or is it just what's listed on #117? | 09:46 |
| tristan | tlater, just in the issue tracking afaik | 09:49 |
| tristan | gah ! | 09:50 |
| tristan | unacceptable TLS cert on gnome7 | 09:50 |
| tlater | tristan: I'm pretty sure that TLS cert caused me to run into #128, so not too bad ;) | 09:53 |
| tristan | any ideas how to enable http on a reverse proxied apache server ? | 09:54 |
| * tristan wants to solve this immediately, not on Dave Page's schedule | 09:54 | |
| ssam2 | do you control the reverse proxy ? | 09:54 |
| tristan | I do | 09:54 |
| tristan | I just have no clue, whatsoever, about anything like that | 09:54 |
| ssam2 | is port 80 forwarded ? | 09:54 |
| * tristan opens config | 09:54 | |
| ssam2 | i could have a look if you want | 09:54 |
| tristan | I have <VirtualHost *:80> for the forwarding of every sled | 09:55 |
| tristan | I might assume that to mean "yes" | 09:55 |
| tristan | but, I suppose it actually means "no" | 09:55 |
| tristan | (everything open *except* port 80 ?) | 09:56 |
| tristan | weird | 09:56 |
| tristan | ssam2, still have creds to log into jumpserv.colo.codethink.co.uk ? | 09:56 |
| ssam2 | seems I do | 09:57 |
| tristan | ssam2, I know about /etc/apache2/sites-available/gnome-proxy.conf | 09:57 |
| tristan | there might be more in this monster of configuration stuff that I dont know about | 09:57 |
| ssam2 | i'll have a dig | 09:57 |
| tlater | tristan: A final little annoyance from me before I try to finish up the tracking MR - would you like the implicit tracking option to be part of this or of the - future - tracking domains? I think it would make a lot of sense to just have 'bst build --fix-my-bloody-pipeline' as well as more fine grained control. | 10:00 |
| tristan | tlater, right now what is blocking is that with the current options, we're stuck with recursive tracking on --track, and that will be permanent if it slips passed 1.0 | 10:01 |
| tristan | But we do *need* a recursive option | 10:01 |
| tristan | tlater, and it's not really acceptable to chose another approach that does not consider the end game, and again paints us into an undesirable corner | 10:02 |
| tristan | So, to be clear; not everything needs to be implemented to close it | 10:02 |
| tristan | But, the design has to be perfect and satisfy all the cases we might think up, without adding *too many* options in the future | 10:02 |
| tristan | Understand ? | 10:02 |
| tlater | tristan: Yup, I do | 10:03 |
| tristan | tlater, just in case you missed it, there are 2 related issues which should be interlinked somehow | 10:05 |
| tristan | I think you need both issues to see the whole picture | 10:05 |
| tlater | #113 and #117? | 10:05 |
| tristan | maybe | 10:05 |
| tlater | --track-save and --track-recurse/--track? | 10:06 |
| tlater | x) | 10:06 |
| * tristan sees if he can fix the (<) replacing configure commands when specified in project.conf element overrides bug while Sam fixes http | 10:06 | |
| tristan | tlater, yep that's it | 10:06 |
| *** bochecha has quit IRC | 10:14 | |
| gitlab-br-bot | buildstream: merge request (sam/artifacts-url-fix->master: _pipeline.py: Give the correct artifact pull URL in remote ticker) #120 changed state ("merged"): https://gitlab.com/BuildStream/buildstream/merge_requests/120 | 10:17 |
| gitlab-br-bot | push on buildstream@master (by Tristan Van Berkom): 1 commit (last: _pipeline.py: Give the correct artifact pull URL in remote ticker) https://gitlab.com/BuildStream/buildstream/commit/9285f004162efeb09b4622ffe28edeeeb43f4262 | 10:17 |
| gitlab-br-bot | buildstream: Sam Thursfield deleted branch sam/artifacts-url-fix | 10:17 |
| jonathanmaw | tristan: How would you like the documentation for bst-external to be served? looking at the repo's Settings -> Pages page, I can define a domain, and then get CI runners to push to there | 10:19 |
| jonathanmaw | should I endeavour to have bst-external shared to the same domain as buildstream, or create my own domain and have buildstream's docs link to it? | 10:19 |
| tristan | jonathanmaw, I dont know what a "domain" is in this context, but I presume that if your project is called 'bst-external', and lives in the BuildStream gitlab group... and if you do "pretty much exactly the same thing"... | 10:24 |
| tristan | Then the docs will land at http://buildstream.gitlab.io/bst-external | 10:24 |
| tristan | beside buildstream main docs at http://buildstream.gitlab.io/buildstream - which sounds sensible to me, right ? | 10:25 |
| jonathanmaw | tristan: ok, I'll give that a go | 10:25 |
| *** sstriker has joined #buildstream | 10:27 | |
| *** valentind has quit IRC | 10:42 | |
| ssam2 | i can't figure out the https issue, sorry | 10:51 |
| ssam2 | dave says he will "get to it soon" | 10:51 |
| ssam2 | i suspect there's something outside our control going on, as the jumpserver doesn't contain that SSL cert as far as I can see | 10:52 |
| tristan | :-S | 10:54 |
| tristan | crap | 10:54 |
| *** sstriker has quit IRC | 11:02 | |
| ssam2 | tlater, there's an interesting exception in https://gitlab.com/baserock/definitions/-/jobs/37567134 | 11:04 |
| tlater | ssam2: We already have a branch that fixes that | 11:05 |
| ssam2 | ah, cool | 11:05 |
| * tlater would like to add a MR for it, but I'm not 100% sure it's unrelated to #128 | 11:05 | |
| gitlab-br-bot | push on buildstream@master (by Tristan Van Berkom): 2 commits (last: project.py: Fetch element type overrides before assertions) https://gitlab.com/BuildStream/buildstream/commit/b648bcab88340ea470098c9d2af60d3f556a51d6 | 11:06 |
| gitlab-br-bot | buildstream: issue #127 ("Error when trying to use list prepend to replace pre-* commands in project.conf") changed state ("closed") https://gitlab.com/BuildStream/buildstream/issues/127 | 11:07 |
| tlater | Mind you, it would be nice if someone could confirm #128. ssam2, do you think you could `bst build --track apps.bst` in gnome-modulesets and check if the second counters keep running? If they do, we could merge a fix for that exception :) | 11:08 |
| ssam2 | i'll give it a go | 11:09 |
| ssam2 | breaks because of CA cert errors pulling the debian ostree repo :-( | 11:11 |
| tristan | frak | 11:11 |
| ssam2 | i don't know if there's an ostree equivalent to GIT_SSL_NO_VERIFY | 11:11 |
| tlater | ssam2: That's actually useful in this case | 11:11 |
| tlater | Just 'continue' | 11:12 |
| tlater | I expect it's related to that | 11:12 |
| ssam2 | ah, ok | 11:12 |
| ssam2 | time does seem to be moving forwards | 11:12 |
| tlater | ssam2: In the normal 1s intervals or does it happen to update whenever a message appears? | 11:13 |
| ssam2 | ah | 11:13 |
| ssam2 | only when a message appears | 11:13 |
| ssam2 | yeah they have all stopped now | 11:13 |
| tlater | \o/ So yeah, that's a bug | 11:13 |
| gitlab-br-bot | buildstream: issue #130 ("Interactive prompts requiring full words is annoying") changed state ("opened") https://gitlab.com/BuildStream/buildstream/issues/130 | 11:14 |
| *** jude has quit IRC | 11:45 | |
| *** jude has joined #buildstream | 11:46 | |
| *** bochecha has joined #buildstream | 12:06 | |
| tlater | tristan: How should something like `bst --track-recursive top-level.bst --track-except dependency.bst --track-recursive dependency-of-dependency.bst` behave? | 12:12 |
| tlater | Atm it seems impossible to re-include a dependency of a dependency once it's excepted | 12:12 |
| tlater | That seems wrong, but it also seems overkill to add some way around that. | 12:13 |
| tristan | I would ignore anything to do with ordering on the command line and expect an error | 12:16 |
| tristan | telling me about it, I guess | 12:16 |
| tristan | "Explicitly asked for 'dependency-of-dependency.bst', which is already excepted by 'dependency.bst'" | 12:17 |
| tristan | tlater, OR - consider toplevel targets as orthogonal dependencies | 12:18 |
| tristan | tlater, i.e. the expected semantics of --except is already to trim the elements of a pipeline _only_ if they are not orthogonally depended on by other elements no --except'ed | 12:19 |
| tristan | tlater, so in that case, the best actually is probably to just to ahead and track top-level.bst and dependency-of-dependency.bst | 12:19 |
| tristan | but not dependency.bst | 12:20 |
| tristan | Seems the natural answer | 12:20 |
| tristan | (its the natural answer if you consider explicitly asking --track-recurse as a *separate* dependency on that element which cannot be ignored) | 12:20 |
| *** bochecha has quit IRC | 12:21 | |
| tristan | tlater, the same question arises for `bst track --except dependency.bst top-level.bst dependency-of-dependency.bst` | 12:21 |
| tristan | (I.e. when we allow multiple targets on all bst commands) | 12:21 |
| tlater | Gah, sorry, just missed that | 12:37 |
| tlater | Alright, that makes sense. | 12:39 |
| *** xjuan has joined #buildstream | 12:43 | |
| adds68 | In a buildstream source definition, what would happen if i left something blank? | 12:51 |
| adds68 | i.e leaving url: blank in this example? https://buildstream.gitlab.io/buildstream/format.html#sources | 12:51 |
| ssam2 | it depends on the field | 12:55 |
| ssam2 | url is probably required, because you can't pull something if you don't know where it is | 12:55 |
| ssam2 | but there are also optional fields in some contexts | 12:55 |
| adds68 | ssam2, ok, it's just that the json files don't seem to enforce many rules so i'm kind of having to just take it as fact | 12:57 |
| adds68 | ssam2, i'll just have to add some form of error checking before i write the file to avoid writing any blank dependencies and hope that the required sources are there for buildstream | 12:57 |
| ssam2 | bear in mind that the thing you're converting is built fairly regularly | 12:58 |
| ssam2 | so you can expect it to work, within reason | 12:58 |
| gitlab-br-bot | push on buildstream@jonathan/external-plugins (by Jonathan Maw): 3 commits (last: Remove dpkg and x86image elements) https://gitlab.com/BuildStream/buildstream/commit/d1e29e2015373d1558e595c8fc9ed138ddab7900 | 12:58 |
| ssam2 | I think http://sdkbuilder1.gnome.org/logs/ has build logs for the SDK, and build failures also get announced by a bot in #flatpak | 13:00 |
| adds68 | ssam2, yea javier mentioned that to me, i'm still working from the git json atm | 13:07 |
| adds68 | ssam2, it's pretty much ready now, it's just i need to then add further classes for other modules that don't use Autotools etc | 13:08 |
| adds68 | Although i'm still not sure what 'simple' means | 13:08 |
| ssam2 | i think it means, do whatever is defined in "build-commands" and nothing more | 13:08 |
| ssam2 | it's not documented that clearly ... but https://www.mankier.com/5/flatpak-manifest says of build-commands: "An array of commands to run during build (between make and make install if those are used). This is primarily useful when using the "simple" buildsystem." | 13:09 |
| adds68 | ssam2, that guide is very useful, i didn't think to look in there | 13:17 |
| gitlab-br-bot | buildstream: merge request (jonathan/external-plugins->master: Remove and refer to elements that have been moved to bst-external) #122 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/122 | 13:33 |
| *** tristan has quit IRC | 14:45 | |
| *** tristan has joined #buildstream | 15:02 | |
| tristan | tlater, I dont get this often, but this might be helpful: https://bpaste.net/show/8bb969637fdc | 15:21 |
| tlater | ta tristan, hopefully I can track that down soon | 15:21 |
| tlater | I'm assuming the final few lines were cut out by the status widget? | 15:23 |
| tristan | tlater, yeah :-/ | 15:25 |
| tristan | that doesnt happen in you `bst command ... | cat` | 15:26 |
| tristan | s/in/if | 15:26 |
| * tristan gotta run | 15:26 | |
| tlater | o/ | 15:26 |
| tristan | midnight video chat with family in canada | 15:26 |
| tristan | laterz \o | 15:26 |
| gitlab-br-bot | push on buildstream@tracking-changes (by Tristan Maat): 3 commits (last: Make `bst build --track-recurse` specify the element to recurse) https://gitlab.com/BuildStream/buildstream/commit/ea9fc2fc493d7f007e8a5474cd4f030e6c5165f2 | 15:32 |
| gitlab-br-bot | push on buildstream@tracking-changes (by Tristan Maat): 8 commits (last: _pipeline.py: Give the correct artifact pull URL in remote ticker) https://gitlab.com/BuildStream/buildstream/commit/9285f004162efeb09b4622ffe28edeeeb43f4262 | 15:40 |
| gitlab-br-bot | buildstream: merge request (tracking-changes->master: WIP: Tracking changes) #119 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/119 | 15:45 |
| gitlab-br-bot | buildstream: merge request (tracking-changes->master: WIP: Tracking changes) #119 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/119 | 15:47 |
| *** adds68 has quit IRC | 15:59 | |
| *** sstriker has joined #buildstream | 16:05 | |
| *** sstriker has quit IRC | 16:30 | |
| *** jonathanmaw has quit IRC | 16:31 | |
| gitlab-br-bot | push on buildstream@107-failing-child-processes-when-tracking-missing-git-branches (by Tristan Maat): 4 commits (last: _pipeline.py: Give the correct artifact pull URL in remote ticker) https://gitlab.com/BuildStream/buildstream/commit/9285f004162efeb09b4622ffe28edeeeb43f4262 | 16:34 |
| *** ssam2 has quit IRC | 16:38 | |
| gitlab-br-bot | buildstream: Tristan Maat created branch 128-status-ticker-fails-to-update-periodically-on-some-builds | 17:07 |
| gitlab-br-bot | buildstream: merge request (128-status-ticker-fails-to-update-periodically-on-some-builds->master: WIP: Resolve "Status ticker fails to update periodically on some builds") #123 changed state ("opened"): https://gitlab.com/BuildStream/buildstream/merge_requests/123 | 17:07 |
| *** jude has quit IRC | 17:20 | |
| *** tlater has quit IRC | 17:23 | |
| *** valentind has joined #buildstream | 17:50 | |
| *** xjuan has quit IRC | 19:52 | |
| *** semanticdesign_ has quit IRC | 22:09 | |
| *** semanticdesign has quit IRC | 22:09 | |
| *** adds68 has joined #buildstream | 22:34 | |
| *** adds68 has quit IRC | 23:00 | |
| *** adds68 has joined #buildstream | 23:10 | |
| *** adds68 has quit IRC | 23:15 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!