[03:21] <ricmm> duflu: ping
[03:21] <duflu> ricmm: pong
[03:22] <ricmm> duflu: hey daniel, msh::Surface::resize() is meant to do something?
[03:22] <ricmm> or is it just stubbed
[03:22] <duflu> ricmm: It does work, providing...
[03:22] <duflu> (1) Desktop; you have the latest Mesa
[03:23] <duflu> (2) You don't expect events to be sent about it (they're in Mir 0.1.3)
[03:23] <duflu> Functionally it should work in Mir 0.1.2
[03:24] <duflu> ricmm: Also (3) In a nested server, I'm not sure. Nesting still has too many bugs to test it properly
[03:26] <ricmm> hmm
[03:26] <ricmm> tested on phone?
[03:27] <duflu> ricmm: Yes, for quite some time now
[03:27] <duflu> ricmm: You can pinch/zoom with demo_shell
[03:27] <ricmm> ah great
[03:28] <duflu> ricmm: I sent a "feature complete" email to Gerry. I'mm find it for you
[03:28] <duflu> I'll find it...
[03:29] <ricmm> oh nvm this surface was drawing weirdly
[03:30] <ricmm> led me to think it was in a different placement
[03:30] <ricmm> need to move it to 0,0 too
[03:30] <duflu> ricmm: Oh, yes, we don't have support for localized input coordinates yet
[03:31] <duflu> ricmm: 3 fingers to pinch/zoom in demo_shell. It's not a nice perfect 2-fingers because we're reserving 1-2 finger gestures for apps
[03:31] <duflu> Or Alt+middlebuttondrag on desktop
[03:34] <ricmm> ah perfect now :)
[03:39] <duflu> ricmm: Fun, right?
[03:40] <ricmm> well I wouldnt go that far
[03:40] <ricmm> but at least gallery seems to be correctly placed when it changes from maximized to fullscreen
[03:40] <ricmm> ;)
[03:41] <duflu> ricmm: That reminds me... we do have mir_surface_set_state(s, mir_surface_state_maximized), but it's not wired up to *do* anything with resizing. At last discussion the rest of mir-team thought it was appropriate to make that the shell's responsibility. I'm still not sure
[03:41] <ricmm> thats what im doing right now
[03:41] <ricmm> state_fullscreen triggers a set of actions
[03:41] <duflu> Good, I think
[03:42] <ricmm> that end up in a move/resize operation if policy allows
[03:42] <ricmm> well, that was the original intention of that message passing
[03:42] <ricmm> just that back then there was no resize support
[03:42] <duflu> ricmm: In that case beware of: https://bugs.launchpad.net/mir/+bug/1261647
[03:43] <ricmm> oh nice :)
[03:43] <ricmm> lol
[03:43] <duflu> ricmm: Don't shoot the messenger. I just found it this week
[03:44] <ricmm> thats probably very obvious in server_shell
[03:45] <ricmm> not as bad here as the original geometry only changes by about 60 pixels on the top
[03:45] <ricmm> which is usually useless
[03:45] <ricmm> but yea, big bug :)
[08:20] <anpok_> ok strange.. the mesa issues from yesterday are just gone
[08:23] <duflu> anpok_: Happy Friday then
[08:26] <anpok_> hi
[08:34] <alan_g> duflu: in the past we've merged development-branch directly to lp:mir - any reason for using lp:~vanvugt/mir/version-0.1.3 this time?
[08:34] <duflu> alan_g: We stopped doing that a few releases ago
[08:34] <duflu> alan_g: The head of dev is no longer the 0.1.3 release, that's why. See the tags :)
[08:35] <alan_g> I guess this is the first time I'm awake since then.
[08:35] <duflu> If the merge happens before the upstream releases, then certainly just merge from dev. But that's not how it's happened lately. And ideally not how it should
[09:47] <anpok_> hm is mir_image.h a nice drawing of the space station?
[09:48] <alan_g> anpok_: I had to be told what it is
[09:49]  * alan_g wonders why jenkins hasn't reviewed ~andreas-pokorny/mir/allow-transparent-server-buffers
[09:52] <RAOF> alan_g: Oh, you wanted me yesterday? I happen to be around now.
[09:52] <alan_g> RAOF: yeah, I wondered if you knew anything about alf_'s MP fiddling with linkage
[09:53] <alan_g> https://code.launchpad.net/~afrantzis/mir/mir-client-ensure-global-symbol-resolution/+merge/196110
[09:55] <duflu> alan_g: I believe Jenkins checks if the author is in some certain group (engineering teams or an employee) and only kicks in if so
[09:55] <duflu> anpok_: Yes it is a nice drawing (the artist is on mir-team I think) :)
[09:56] <alan_g> duflu: but it reviewed ~andreas-pokorny/mir/fix-missing-system-path
[09:56]  * duflu shrugs
[09:56] <anpok_> but only after somebody added jenkins to it
[09:56] <anpok_> i think kevin did
[09:56] <anpok_> before that it ignored the branch
[09:57] <duflu> The CI elves will know
[09:57] <anpok_> duflu: since I am all into transparency and glasnost. I thought about adding some non-255 alpha values to it
[09:58] <duflu> Indeed Glasnost is about transparency... (?)  http://en.wikipedia.org/wiki/Glasnost
[09:59]  * anpok_ nods 
[09:59] <anpok_> at least some weirdness of yesterday still exists
[10:00] <duflu> Sorry, have to run. Expecting company any second...
[10:00] <anpok_> see you .. next year?
[10:01] <duflu> anpok_: No, Monday. I'm around almost every day
[10:20] <anpok_> is the order of supported buffer formats passed to the client random?
[10:24] <RAOF> I don't think we have any semantics defined there.
[10:28] <RAOF> alan_g: Commented on that MP.
[10:29] <alan_g> RAOF: thanks for looking at that MP - I'm still unclear as to whether the right problem is being solved
[10:31] <RAOF> At a high level the right problem is being solved - libEGL.so.1 can't resolve Mir symbols when libmir* are loaded by something else with RTLD_LOCAL.
[10:32] <RAOF> I think the *most* correct solution would be to have something that's actually linked with libmirclient and libmirserver, but that's also the least trivial solution.
[10:37] <RAOF> alan_g: Actually, now that you mention it - why isn't EGL linked against libmirclient and libmirserver? Is it only because we suck at libmirserver ABI?
[10:37] <alan_g> RAOF: I don't know. I assume it is so that it runs on the mir-less stack
[10:38] <anpok_> because you dont have to update libEGL for each mirclient/server update?
[10:39] <alan_g> RAOF: I assume only a small part of the API is used? We could write a stable wrapper
[10:57]  * alan_g goes to talk to the CI elves
[11:04] <alan_g> anpok_: did you track down the problem with nesting render_surfaces on android?
[12:09] <anpok_> alan_g: not yet i will do that after lunch got.. trying to reproduce the weirdness i had this morning with mesa.. but it works now..
[12:10] <alan_g> ok. It occurred to me that I'd only used it on N4. N10 might be different. (Haven't tried yet)
[13:58] <alan_g> anpok_: Over on #ubuntu-ci-eng (where the CI elves hang out) fginther has promised to fix it so that Jenkins reviews your MPs. (He's a useful chap while our timezones meet.)
[14:52] <kgunn> alan_g: distracted helping out with some ui testing this morning...but later was going to do the promotion of dev branch....is everything ok?
[14:52] <kgunn> or any mp's that need to be waited on?
[14:53] <alan_g> kgunn: all good.
[15:25] <anpok_> that weirdness from this morning on mesa was that in some occasions the render_surfaces example was unable to find a matching egl config
[16:09] <anpok_> error was between chair and keyboard..
[16:49] <tvoss|lunch> anpok_, lol