| === JanC is now known as Guest86694 | ||
| === JanC_ is now known as JanC | ||
| Son_Goku | alan_g, hey | 09:00 |
|---|---|---|
| Son_Goku | I saw your blog posts about Mir earlier today | 09:00 |
| Son_Goku | I'm a bit confused by it though | 09:01 |
| Son_Goku | are you talking about Mir acting like a Wayland compositor or just offering support for the Wayland protocols so applications that are clients to Wayland compositors work on Mir, too? | 09:01 |
| === chihchun_afk is now known as chihchun | ||
| alan_g | Son_Goku: this is an area prone to misunderstandings. If *hypothetically* Mir offered support to clients using Wayland why wouldn't you consider it a Wayland compositor. | 09:19 |
| Son_Goku | I really don't see a problem with that, and in fact, I'd love that | 09:20 |
| Son_Goku | my *only* problem right now is that I *really* can't handle dealing with libinput having non-upstream APIs in order for Mir to build | 09:20 |
| Son_Goku | several months ago, I did build Mir for Fedora, but it became quickly unrealistic to maintain | 09:21 |
| Son_Goku | alan_g: Ubuntu moves too slowly :) | 09:21 |
| Son_Goku | the Mesa and libinput patches not being upstream makes adding Mir to Fedora *really* hard | 09:22 |
| Son_Goku | not to mention, libinput effectively has a different ABI when the patch is added | 09:23 |
| Son_Goku | as I've explained to tvoss, it would really be nice and helpful if I don't need to do terrible things like this in a copr repo that will break as Fedora moves forward | 09:24 |
| alan_g | Son_Goku: I've not tracked the upstreaming closely. I know what the plans were and the guys that were working on it, but with recent events there was no opportunity for a handover. | 09:27 |
| Son_Goku | well, the libinput patch was written by anpok_ | 09:28 |
| alan_g | yes | 09:28 |
| Son_Goku | he submitted it once long ago in 2015 | 09:28 |
| Son_Goku | but apparently he never followed up after the initial review | 09:29 |
| Son_Goku | so now we're stuck in this bad situation where no one can actually add mir without breaking libinput ABI | 09:29 |
| alan_g | I agree its a bad situation. I don't know any detail of how Canonical plans to proceed yet. I hope to get some decisions next week. | 09:33 |
| Son_Goku | I've gotten Mir to work on Fedora, for about a week :) | 09:33 |
| alan_g | Terrific! Have you blogged about the issues you encountered? | 09:35 |
| Son_Goku | I have not | 09:36 |
| Son_Goku | I didn't think anyone would be interested | 09:36 |
| Son_Goku | tvoss actually helped me get off the ground | 09:36 |
| Son_Goku | the big issue that I've heard from people about even thinking about Mir vs Wayland is that no one can reasonably use Mir due to non-portability | 09:37 |
| Son_Goku | the Big Idea(TM) was to make it so that Mir worked on Fedora without screwing with the core of the system so that people could reasonably consider it | 09:37 |
| Son_Goku | it helps that because Fedora is a very upstream-centric distribution (as opposed to Debian and Ubuntu), it's much easier to trust that others can pull from us and repackage for their own distros, and that opens the door to make Mir a very credible option | 09:38 |
| alan_g | Sure. That's also why there was work in progress to upstream to mesa and libinput. | 09:39 |
| === chihchun is now known as chihchun_afk | ||
| Son_Goku | alan_g: I even found a problem while packaging Mir which someone fixed once I told tvoss about it ;) | 09:42 |
| alan_g | Son_Goku: :) | 09:43 |
| alan_g | Really, we've tried to address problems but the Canonical re-org was a surprise. It's still in progress and I don't know what the future holds. | 09:46 |
| alan_g | Having a write-up of the issues you found could help the replanning. | 09:47 |
| Son_Goku | alan_g: I guess I can give it a shot | 10:17 |
| alan_g | Son_Goku: thanks. | 10:19 |
| zyga | Son_Goku: wow, I'm surprised to see you here :-) | 11:28 |
| Son_Goku | zyga: I've been here for months | 11:28 |
| * zyga imagines mir having a brigher future after being abandoned by canonical | 11:28 | |
| Son_Goku | the irony is that if Mir is mostly abandoned by Canonical, it could be more flexible | 11:29 |
| Son_Goku | certain things that might have been disallowed in the past could be done and merged in | 11:29 |
| zyga | yes, that's true | 11:29 |
| Son_Goku | really, I'd love to have Mir packaged in Fedora, but the libinput and mesa patches make it hard | 11:30 |
| Son_Goku | Peter Hutterer (libinput developer and Fedora package maintainer) and Adam Jackson (the chief maintainer for mesa in Fedora) aren't willing to accept patches like these that aren't backports from already merged upstream things | 11:31 |
| alan_g | Mir was always eager to consider pull requests | 11:32 |
| Son_Goku | but you're on Launchpad :( | 11:33 |
| zyga | Son_Goku: which is free software and supports git and bzr | 11:34 |
| Son_Goku | I maintain a private copy on GitLab because the GitLab CI is so nice | 11:34 |
| zyga | Son_Goku: pro tip: if you want to contribute, make it easy to get the contribution | 11:35 |
| Son_Goku | yep | 11:35 |
| Son_Goku | I know this | 11:35 |
| Son_Goku | launchpad doesn't make it "difficult" per se | 11:35 |
| Son_Goku | but afaik, it uses bzr and its MR mechanism is quite weird compared to what most people are used to these days | 11:35 |
| zyga | Son_Goku: so use that to make the contribution (not sure if mir is git or bzr) | 11:36 |
| Son_Goku | mir is bzr, I believe | 11:36 |
| Son_Goku | I did the conversion for gitlab | 11:36 |
| Son_Goku | yep | 11:36 |
| Son_Goku | still is bzr | 11:36 |
| zyga | one battle at a time | 11:36 |
| Son_Goku | eh, I'm not too concerned about mir itself | 11:37 |
| Son_Goku | mir is not the problem child yet | 11:37 |
| alan_g | Canonical's Mir is bzr. But... coming soon... https://github.com/unity8-team | 11:37 |
| Son_Goku | the problem is the infrastructure required for mir to build | 11:37 |
| Son_Goku | alan_g: that's unattached from Canonical entirely? | 11:37 |
| Son_Goku | and if I might make a suggestion, GitLab is a much nicer system | 11:38 |
| Son_Goku | I've been using it for my own projects and the much richer CI system built into it makes it so much more powerful | 11:38 |
| ogra_ | but they trash their servers :P | 11:38 |
| alan_g | Some of the ex-employees are copying the U8 ecosystem there | 11:38 |
| Son_Goku | ogra_: one of the things I've been doing is actually running builds against multiple distros in parallel and executing tests | 11:39 |
| Son_Goku | not with mir specifically, but with some of my own things | 11:39 |
| zyga | alan_g: nice! | 11:39 |
| Son_Goku | my original motivation for mir was because unity | 11:40 |
| Son_Goku | and I was told to not try unity 7 | 11:40 |
| Son_Goku | so I moved on to unity 8 | 11:40 |
| alan_g | U7 is based on compiz+gnome | 11:40 |
| Son_Goku | ah, that would have been a problem then | 11:41 |
| Son_Goku | iirc, ubuntu uses compiz 0.9.x | 11:41 |
| Son_Goku | but Fedora reverted back to 0.8.x after much breakages | 11:41 |
| Son_Goku | and now there's a group actively working on compiz 0.8.x development | 11:41 |
| alf_ | Great discussion with Jono Bacon: https://www.youtube.com/watch?v=4mOPmM4Y15o | 11:53 |
| zyga | odd, that test no longer fails | 13:36 |
| zyga | hrmm hrmm | 13:36 |
| zyga | let's retest | 13:36 |
| === lan3y is now known as Laney | ||
| === ChrisTownsend1 is now known as ChrisTownsend | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!