[01:51] <robert_ancell> duflu, hold off committing any changes to XMir - I'm about to rebase it
[01:51] <duflu> robert_ancell: Kay
[01:51] <duflu> robert_ancell: I removed the changes outside of hw/xmir/ so should be easier now
[01:51] <robert_ancell> duflu, yeah, nice work!
[01:51] <duflu> The code changes anyway
[01:52] <robert_ancell> I've kept the rootless as a separate change for now, but we probably can roll it into one patch going forward now those changes are gone.
[01:57] <robert_ancell> duflu, ok all done. Hopefully I haven't broken anything...
[01:57] <robert_ancell> Now to set up a CI train thingy for this
[02:05] <robert_ancell> oh yay. https://requests.ci-train.ubuntu.com/ makes me want to kill myself
[02:08] <duflu> Can we have traditional distros back please?
[02:10] <robert_ancell> duflu, I think the intentions were good but this is a million steps backwards in complexity
[02:11] <duflu> robert_ancell: Absolutely. You're entitled to create complexity, but not when you make it everyone else's problem
[02:16] <robert_ancell> duflu, do you want the LP bug numbers in the debian/changelog?
[02:17] <robert_ancell> It's kind of hard to track them post rebasing
[02:17] <duflu> robert_ancell: Can't hurt, there's only a few. But more interesting are the enhancements to list
[02:17] <robert_ancell> duflu, do you want to summarise the changes then?
[02:17] <duflu> robert_ancell: OK
[02:18] <duflu> robert_ancell: What git commit was the last release?
[02:18] <robert_ancell> duflu, on the pre-rebased branch?
[02:18] <duflu> Never mind. I'll just cover the past few weeks
[02:19] <robert_ancell> There were just those four commits over the 1.17 changes (xmir, rootless, desktop_file_hint and keeping stdout open)
[02:19] <robert_ancell> I don't know why the stdout change was there, but we haven't been using it in the xorg-server package so we can probably drop it.
[02:19] <duflu> Yeah stdout doesn't work still
[02:20] <duflu> robert_ancell: Gimme an hour or so (other tasks in the way)
[02:20] <robert_ancell> duflu, ok
[02:20] <robert_ancell> duflu, do you know what the stdout change is for?
[02:21] <duflu> robert_ancell: Not even looked at it. I just know printf still doesn't work
[02:21] <robert_ancell> But ErrorF works, so is there any reason to make stdout work?
[02:21] <robert_ancell> I'm guessing this is a design decision by upstream?
[02:22] <duflu> robert_ancell: It would be nice if we could use some InfoF instead. DebugF exists but logs nothing
[02:42] <duflu> Heh, Mir gained a new build-dep overnight that is actually good to see... libinput-dev
[02:49] <robert_ancell> duflu, nice!
[03:04] <duflu> robert_ancell: In that email it might be more correct to s/xmir:/Xmir:/
[03:05] <robert_ancell> duflu, which email?
[03:05] <duflu> robert_ancell: The one in your inbox
[03:07] <robert_ancell> ah
[03:07] <robert_ancell> thanks
[03:11] <duflu> robert_ancell: Actually let me rewrite that
[03:11] <robert_ancell> ok
[03:13] <duflu> Done
[03:18] <duflu> For a while I was suspicious and wondering why Xwayland was half the size of Xmir. From what I read recently though that might be because Xwayland is missing features
[03:18] <duflu> Not sure
[03:19] <duflu> robert_ancell: I might have an extra crucial fix for Xmir today. In an hour or two
[03:19] <duflu> Maybe not
[03:19] <duflu> Or 30 min
[03:22] <duflu> robert_ancell: BTW I think you can rebase Xmir in future without losing the commit history
[03:22] <duflu> That would be nice
[03:22] <robert_ancell> duflu, how can you do that?
[03:22] <robert_ancell> have a separate branch?
[03:22] <duflu> robert_ancell: I forget but it's a git feature
[03:24] <duflu> robert_ancell: Oh I just did it by accident. Repushed and everyone's commits are there in chronological order now
[03:25] <robert_ancell> duflu, you git push -f?
[03:25] <duflu> robert_ancell: Nope
[03:26] <duflu> robert_ancell: It was after your merge and includes you merge so I hope everything's there
[03:31] <robert_ancell> duflu, oh, it just has an empty merge commit
[03:31] <duflu> robert_ancell: Yeah the beauty of not having revision numbers is that you can merge things and each commit retains the same identifier it always had
[03:33] <duflu> robert_ancell: YES. Fixed the missing cursors. Let me get that fix in
[03:33] <duflu> It's a Mir bug, but I can work around it in Xmir
[03:36] <robert_ancell> I have to head out for 20-30 mins, bbl
[04:12] <Hock> https://bugs.launchpad.net/ubuntu/+source/ubuntu-ui-toolkit/+bug/1493185
[04:12] <Hock> any chance for this bug to be look at?
[04:16] <RAOF> Looks like someone *is* looking at it? It's assigned to someone.
[04:16] <RAOF> (It might be nice to capture the shaders that your GL is choking on. I wonder if apitrace works on the phone...)
[04:17] <Hock> how do I capture the shaders?
[04:23] <RAOF> You *might* be able to install apitrace on the phone, then modify the unity8 upstart job to run ‘egltrace unity8’ rather than just ‘unity8’.
[04:23] <RAOF> That'd get you a trace of all the EGL + GLES calls unity8 makes, including the sources of the failing shaders.
[05:23] <Hock> no wifi at the moment. :(
[05:23] <Hock> guess have to wait for update to the lp bug
[05:26] <RAOF> You could adb push the relevant .debs, I think.
[05:38] <Hock> sure. given that it might not help as much. it might make more sense to wait for Loïc to find the time to delve into it.
[05:38] <Hock> if he ever get to it as there seem to be quite a number of bugs going on in the toolkit
[05:39] <RAOF> Yeah, always more bugs than people ;)
[05:45] <Hock> sad hard truth
[05:57] <duflu> Sometimes more projects than people :)
[05:59] <Hock> the downside to opensource
[06:00] <Hock> similar projects in some cases fortunately and yet unfortunately
[06:16] <anpok_> apitraces does work yes..
[06:16] <anpok_> -s
[09:26] <duflu> alan_g: Could you do me a favour and audit the SessionMediator? It's starting to look like it forgets to release some resources at least on unclean disconnects. And such bloat would not show up as a leak either, because destructors would clean it up...
[09:26] <duflu> Some of those resources may be fds related to Android graphics
[09:27] <alan_g> I'll add it to the list.
[09:27] <duflu> As usual, not one bug, but multiple bugs in multiple projects... maybe
[12:31] <guest123124> hi all
[12:32] <guest123124> i'm on wily, if i run mir_demo_server in a tty it locks my pc
[12:44] <alan_g|lunch> guest123124: sorry, that's a poor failure mode (there should be better cleanup and a better error message). Mir needs root access in that configuration: just add "sudo".
[12:46] <guest123124> alan_g|lunch, thanks :D shoud i open a bug? maybe other people will try mir and they'll end up with a locked pc and maybe get angry or stuff
[12:48] <guest123124> alan_g|lunch, also, do you know if i "mir_demo_server --vt 1" from terminal would work? should it run mir_demo_server in tty1 and i can then ctrl alt f1?
[12:48] <alan_g|lunch> guest123124: check if there is a bug filed already (it's quite likely)
[12:49] <guest123124> will do
[12:49] <alan_g|lunch> You'll need something like --platform-graphics-lib /usr/lib/x86_64-linux-gnu/mir/server-platform/graphics-mesa-kms.so.4 to avoid it defaulting to x11
[12:53] <guest123124> alan_g|lunch, like this mir_demo_server --platform-input-lib server-mesa-x11.so.4 --vt 1?
[12:53] <guest123124> alan_g|lunch, like this mir_demo_server --platform-input-lib server-mesa-x11.so.4 - works.. it opens a window (mir on x)
[12:54] <guest123124> i still get Unknown command line options: --vt 1
[12:55] <alan_g|lunch> guest123124: --vt only makes sense when using the kms graphics platform.
[12:57] <guest123124> alan_g|lunch, like this?  mir_demo_server --platform-graphics-lib /usr/lib/x86_64-linux-gnu/mir/server-platform/graphics-mesa-kms.so.4 --vt 1
[12:58] <guest123124> alan_g|lunch, thanks that worked :P
[12:59] <alan_g> guest123124: yw