| *** traveltissues has joined #buildstream | 06:06 | |
| *** rdale has joined #buildstream | 09:22 | |
| *** tme5 has joined #buildstream | 09:37 | |
| coldtom | does anyone fancy reviewing https://gitlab.com/BuildStream/bst-plugins-experimental/merge_requests/61 ? | 09:46 |
|---|---|---|
| *** tiagogomes has joined #buildstream | 09:56 | |
| benschubert | coldtom: just did, you a test be feasible by any chance? :) | 10:06 |
| *** tpollard has joined #buildstream | 10:09 | |
| *** jonathanmaw has joined #buildstream | 10:10 | |
| *** bochecha has joined #buildstream | 10:11 | |
| coldtom | benschubert: I could add a simple test that it works, but i'm not sure how I'd go about adding support for the standardised ones yet | 10:14 |
| benschubert | ah I meant any kind of test, not necessarily the standardised ones :) | 10:14 |
| *** traveltissues has joined #buildstream | 10:33 | |
| *** lachlan has joined #buildstream | 10:39 | |
| *** lachlan has quit IRC | 10:51 | |
| *** mohan43u has quit IRC | 10:52 | |
| *** mohan43u has joined #buildstream | 10:57 | |
| *** lachlan has joined #buildstream | 11:04 | |
| *** lachlan has quit IRC | 11:20 | |
| *** traveltissues has quit IRC | 11:22 | |
| tme5 | can anyone summarise the political status of the external plugins repo? I don't understand what will happen to it after the plugin migration. | 11:39 |
| tme5 | it seems that you don't want to officially maintain any plugins - in that case who will be the maintainers of bst-external? | 11:40 |
| tme5 | i'm curious about how the criteria for inclusion in bst-external will be any different to that of how it currently is in BuildStream | 11:43 |
| tlater[m] | tme5: In practice, the same people maintaining BuildStream. We don't want the plugins to be part of the main project, but bst-external is tested in sync with the main repository. | 11:43 |
| tme5 | but why? | 11:44 |
| tme5 | if you've moved all plugins from core to somewhere else, with the same maintainers, you haven't reduced any amount of work for yourself, in fact you've increased it substantially | 11:44 |
| tlater[m] | It was weird to essentially say "we think these tools are good" by having plugins for them, but not others. | 11:44 |
| tlater[m] | We didn't want to have plugins for any and all build tools in the main repository | 11:45 |
| tlater[m] | So bst-external was created, and things were moved to there to make it obvious that we're not exactly endorsing the tools themselves. | 11:45 |
| *** lachlan has joined #buildstream | 11:46 | |
| tlater[m] | In other words, this isn't about work reduction in the slightest | 11:47 |
| tme5 | perhaps it is weird to you, personally I don't see how it's weird to make a piece of software useful for common cases. But ignoring that, are you now going to have plugins for any and all build tools in the external repo? | 11:47 |
| tme5 | what do you mean by endorsing tools? BuildStream is a tool, that is designed to build software. If Make is a very common build tool, why is it bad to say "Make is a common build tool" ? | 11:50 |
| *** santi has joined #buildstream | 11:51 | |
| tlater[m] | tme5: When we last discussed this, we wanted a few select ones that are considered stable and important will be in bst-plugins (without the -experimental). bst-external will remain the bst 1.0 plugins repository. There will be a other plugin repositories that will contain what bst-external is today, as well as more specific ones for things like "this is docker related". | 11:51 |
| tlater[m] | Saying that "make is a common build tool" isn't bad, but it's bad to say "make is a common build tool, but npm is not" | 11:52 |
| tlater[m] | We don't want our (non-) inclusion of plugins to be perceived as such | 11:52 |
| tme5 | hmm, sounds like you are still endorsing tools to me | 11:52 |
| tlater[m] | Which ones? | 11:52 |
| tlater[m] | You mean the ones in bst-plugins? We explicitly don't. | 11:53 |
| tme5 | "a few select ones that are considered stable and important will be in bst-plugins" | 11:53 |
| tlater[m] | The split is purely there to make that more obvious. | 11:53 |
| tme5 | what? | 11:53 |
| tme5 | " There will be a other plugin repositories that will contain what bst-external is today" | 11:53 |
| *** lachlan has quit IRC | 11:53 | |
| coldtom | the stability is in the plugin, not the build tool itself | 11:54 |
| tlater[m] | bst-external today is a set of plugins that aren't considered good enough for inclusion in a few select plugins that we think are stable. | 11:54 |
| tme5 | coldtom, ok, well "important" is the key word | 11:55 |
| tlater[m] | There will have to be other repositories to take on that in the future. | 11:55 |
| tlater[m] | tme5: That said, what's your question behind all of this? | 11:55 |
| coldtom | aiui there still isn't actually a well defined policy for plugins in 2.0 | 11:55 |
| tlater[m] | Are you concerned we'll stop supporting your favorite plugin? | 11:55 |
| tme5 | tlater[m], my question is twofold, a) how does the plugin migration actually stop you "endorsing" particular tools, and b) how will the criteria for inclusion, support, maintenance, etc. functionally change after the migration? | 12:00 |
| tlater[m] | The answer to the latter is essentially not at all, except that some file paths will change, although this is all coming from discussions at a meetup in November last year. | 12:03 |
| tlater[m] | The former is essentially just that there will be a clear barrier between our code and our plugins. That makes it more obvious than when the plugins are simply part of BuildStream. | 12:03 |
| *** lachlan has joined #buildstream | 12:06 | |
| tme5 | so instead of buildstream endorin | 12:07 |
| tme5 | so instead of buildstream 'endorsing' plugins, bst-plugins will instead? | 12:08 |
| tlater[m] | Basically, yes | 12:08 |
| tlater[m] | And because that's called "plugins" it's a fair bit more obvious that it's not that strong of a suggestion to use those tools. | 12:09 |
| tlater[m] | This all came up when we didn't want to include docker plugins in the main repository, but someone pointed at the dpkg ones and asked why that's there | 12:09 |
| tlater[m] | Which was a good point | 12:10 |
| tme5 | "we didn't want to include docker plugins in the main repository" do you remember roughly why this was the case? | 12:10 |
| *** lachlan has quit IRC | 12:10 | |
| tlater[m] | We're now only keeping plugins that are necessary for BuildStream to function in there, i.e., junction import and such things. | 12:10 |
| tme5 | what was unsatisfactory about docker plugin that dpkg met? | 12:10 |
| tme5 | only if you remember, i'll dig later if needed | 12:11 |
| tlater[m] | Nothing. I think the dpkg plugin was even less stable than the docker plugin at the time. | 12:11 |
| tlater[m] | If anything the fact that it was made a year after the dpkg plugin | 12:11 |
| tlater[m] | And we suddenly had the infrastructure to consider that question. | 12:11 |
| tme5 | interesting, so in the future, docker will be accepted into bst-plugins? | 12:12 |
| tlater[m] | No, we decided to keep docker in a separate repository entirely. We didn't want bst-plugins to just be a dumping ground, because you don't want all plugins for all projects. | 12:12 |
| tlater[m] | We also decided to keep dpkg in a separate repository | 12:13 |
| tlater[m] | We didn't decide on details then, but the former exists | 12:13 |
| tlater[m] | And the latter will probably become some kind of "bst-distros" or "bst-package-managers" | 12:13 |
| tme5 | "because you don't want all plugins for all projects" why not? | 12:14 |
| tlater[m] | tme5: In a perfect world we'd have plugins for all build systems in existence | 12:15 |
| tme5 | sorry if it's unfair to quiz you so much, I'm just so confused | 12:15 |
| tlater[m] | That'd be one big repository | 12:15 |
| coldtom | tme5: this goliath mailing list thread was an attempt to clarify things, but it's a lot https://mail.gnome.org/archives/buildstream-list/2019-April/msg00022.html | 12:15 |
| tlater[m] | If we want to move towards that world, we'd like to have a bit more structure than a dumping ground... | 12:15 |
| tme5 | ok, well I can't claim to understand either the premise or the reasoning very well but thanks for helping me out nonetheless | 12:22 |
| tme5 | don't think it's worth me getting involved with this decision | 12:24 |
| tme5 | incidentally, this came about because of cs-shadow's comment that "whether or not ['git_tag'] lives in the same repository [as 'git'] shouldn't really affect its functionality. It's entirely possible for two git plugins to co-exist." I didn't understand why this would be desirable so I wanted to understand the philosophy of the migration | 12:29 |
| benschubert | tme5: will come back on that once I'm back from lunch, it's a very long discussion :'D | 12:30 |
| *** lachlan has joined #buildstream | 12:33 | |
| *** lachlan has quit IRC | 12:38 | |
| *** lachlan has joined #buildstream | 12:55 | |
| *** lachlan has quit IRC | 12:58 | |
| *** santi has quit IRC | 13:31 | |
| *** santi has joined #buildstream | 13:31 | |
| *** santix has joined #buildstream | 13:36 | |
| *** santi has quit IRC | 13:37 | |
| *** santix has quit IRC | 13:38 | |
| *** santix has joined #buildstream | 13:38 | |
| *** santix has quit IRC | 13:43 | |
| *** lachlan has joined #buildstream | 14:23 | |
| *** lachlan has quit IRC | 14:35 | |
| *** lachlan has joined #buildstream | 14:43 | |
| *** lachlan has quit IRC | 14:47 | |
| *** lachlan has joined #buildstream | 14:48 | |
| *** lachlan has quit IRC | 14:51 | |
| *** lachlan has joined #buildstream | 14:51 | |
| *** delli3 has quit IRC | 15:04 | |
| *** delli3 has joined #buildstream | 15:04 | |
| *** lachlan has quit IRC | 15:22 | |
| *** lachlan has joined #buildstream | 15:46 | |
| *** lachlan has quit IRC | 15:49 | |
| *** narispo has quit IRC | 16:22 | |
| *** narispo has joined #buildstream | 16:24 | |
| *** lachlan has joined #buildstream | 16:41 | |
| *** lachlan has quit IRC | 16:44 | |
| *** narispo has quit IRC | 16:47 | |
| *** narispo has joined #buildstream | 16:49 | |
| *** bochecha has quit IRC | 16:49 | |
| *** narispo has quit IRC | 17:08 | |
| *** narispo has joined #buildstream | 17:08 | |
| *** lachlan has joined #buildstream | 17:13 | |
| *** lachlan has quit IRC | 17:16 | |
| *** tme5 has quit IRC | 17:21 | |
| *** lachlan has joined #buildstream | 17:28 | |
| *** lachlan has quit IRC | 17:31 | |
| *** lachlan has joined #buildstream | 17:59 | |
| *** lachlan has quit IRC | 18:02 | |
| *** lachlan has joined #buildstream | 18:33 | |
| *** jonathanmaw has quit IRC | 18:33 | |
| *** lachlan has quit IRC | 18:36 | |
| *** tiagogomes has quit IRC | 18:44 | |
| *** lachlan has joined #buildstream | 18:56 | |
| *** lachlan has quit IRC | 18:59 | |
| *** lachlan has joined #buildstream | 19:05 | |
| *** lachlan has quit IRC | 19:09 | |
| *** lachlan has joined #buildstream | 19:09 | |
| *** lachlan has quit IRC | 19:14 | |
| *** lachlan has joined #buildstream | 19:45 | |
| *** lachlan has quit IRC | 19:51 | |
| *** lachlan has joined #buildstream | 20:02 | |
| *** lachlan has quit IRC | 20:06 | |
| *** lachlan has joined #buildstream | 20:51 | |
| *** lachlan has quit IRC | 20:57 | |
| *** rdale has quit IRC | 22:02 | |
| *** Trevinho has quit IRC | 23:24 | |
| *** Trevinho has joined #buildstream | 23:25 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!