[00:05] <racarr> thomi: Sorry we don't assure quality so you'll have to talk to the QA guy...
[00:06] <racarr> :p
[00:06] <racarr> I don't have any ideas though
[00:06] <racarr> there is a mismatch between your libdri and mesa and kernel!
[00:06] <racarr> that was my only idea and I think
[00:06] <racarr> it's pretty unlikely
[00:07] <RAOF> thomi: Oh, yeah.
[00:07] <RAOF> thomi: You've got an XMir-enabled video driver (most likely xserver-xorg-video-intel) but not an XMir-enabled server.
[00:08] <thomi> RAOF: that would be -nvidia in mcase
[00:08] <thomi> RAOF: so... I can't run unity3d with the /staging PPA added, and I can't run mir without it. How do you guys do mir development? run KDE or something?
[00:09] <RAOF> My unity3d actually works, but with software rendering.
[00:09] <RAOF> But I'll upload a new xserver to the staging PPA and fix the cause.
[00:10] <thomi> RAOF: thanks, can youplease ping me when the dput is done?
[00:11] <RAOF> Sure.
[00:17] <kdub> RAOF, could it be because of http://bazaar.launchpad.net/~mir-team/mir/trunk/view/head:/include/client/mir_toolkit/mir_client_library.h#L194
[00:17] <kode54> is the correct procedure for testing mir as a compositor for X to add the staging PPA and install mir_demos?
[00:17] <kode54> also dist upgrade
[00:18] <RAOF> kode54: Add the staging PPA, install unity-system-compositor, set “type=mir” in /etc/lightdm/lightdm.conf.
[00:18] <RAOF> But do this after I've unbroken X :)
[00:18] <kode54> ah
[00:18] <kode54> yeah, I did that last night, and X was apparently broken
[00:18] <RAOF> kdub: Well, I'll need to update for that too, clearly :)
[00:18] <kode54> and the docs said type=unity
[00:19] <kode54> my current installation with bzr678, X starts, but only without a type= argument
[00:19] <kode54> I didn't know about the type=mir
[00:19] <RAOF> The docs are more likely to be right than me.
[00:20] <RAOF> :)
[00:20] <kode54> ah
[00:20] <kode54> docs said type=unity, and that fails to start and attempts low graphics mode
[00:20] <kdub> RAOF, i got it to compile: http://pastebin.ubuntu.com/5662955/
[00:20] <kode54> leaving a permanent unity demo server red cursor on screen
[00:20] <kdub> haven't gotten xmir to run though
[00:21] <RAOF> kode54: Oooh, that probably shouldn't happen :)
[00:21] <kode54> er, mir demo server
[00:21] <RAOF> kode54: Just to check - you *have* installed unity-system-compositor, yes?
[00:21] <kode54> yes, but I had to mess with the package
[00:22] <RAOF> thomi: Oh, just to check - you're seeing this problem on Saucy, right?
[00:22] <kode54> unity-system-compositor : Depends: libmirserver0 (= 0.0.2bzr676raring0) but 0.0.2bzr678raring0 is to be installed
[00:23] <kode54> that was the state it was at yesterday
[00:24] <kode54> and there's also a "mir" package that Depends: libmirserver0 (= 0.0.2bzr643raring0)
[00:24] <kode54> but that didn't get installed, I guess it's a meta package?
[00:25] <RAOF> I thought we'd dropped the “mir” package?
[00:25] <kode54> oops
[00:33] <kode54> huh
[00:34] <kode54> the unity-system-compositor in the repository still says it depends on that version of libmirserver0
[00:37] <kode54> and then libmirserver-dev depends on libmirserver0 = 0.0.2bzr681raring0
[00:37] <kode54> only straggler appears to be unity-system-compositor
[00:43] <RAOF> Yeah. unity-system-compositor needs the exact version of libmirserver0 it built against, so it needs a rebuild for every Mir change.
[01:06] <racarr> Off for now! Will be back this evening to write some emails
[01:07] <kode54> oh I see
[01:07] <kode54> the "mir" package is necessary after all
[01:07] <kode54> it's the source package that is targeted for any source fetches for the subpackages it contains
[01:30] <kode54> tried latest packages, including building my own unity-system-compositor against the current version of the dev packages
[01:30] <kode54> the only error I get is
[01:30] <kode54> init: lightdm main process (16930) terminated with status 1
[01:31] <RAOF> kode54:  /var/log/lightdm/*.log should have more details.
[01:31] <RAOF> But, as I said, you should wait for me to unbreak X. Because it's not going to work until I do :)
[01:31]  * RAOF is test-building packages now.
[01:35] <thomi> RAOF: no, raring
[01:36] <RAOF> thomi: Hm. Could you pastebin your /var/log/Xorg.0.log too?
[01:37] <RAOF> thomi: XMir won't work under raring (care of mir API change), but I wasn't aware raw X was broken.
[01:42] <kode54> right
[01:42] <kode54> it's X alright
[02:00] <thomi> RAOF: http://paste.ubuntu.com/5663150/
[02:01] <thomi> RAOF: hmm, Im not sure that log file is the right one any more - - how often is it rotated?
[02:01] <thomi> I've purged the /staging PPA and rebooted since, so I once again have a working desktop
[02:05] <RAOF> thomi: It's deleted each time you start X :)
[02:05] <thomi> huh, well darn
[02:05] <thomi> RAOF: no log file for you then :(
[02:06] <thomi> I've already wasted most of a day trying to get a desktop  - I'll upgrade the second laptop to saucy and use that for mir development
[02:06] <thomi> xmir should work on saucy?
[02:10] <RAOF> Once this damn build builds, yes.
[04:12] <kode54> so, how's that build going?
[04:17] <RAOF> Uploading now :)
[04:19] <RAOF> thomi: Enjoy a freshly dputted build.
[04:20] <kode54> done now?
[04:31] <RAOF> kode54: In the PPA. Won't have built yet.
[04:32] <kode54> ah
[04:32] <RAOF> Gah! WTF, PPA builder? This built fine in my schroot!
[04:33] <RAOF> thomi: Are you aware that the saucy Mir builds are failing?
[04:34] <thomi> RAOF: no, I wasn't - in launchpad, or....?
[04:34] <RAOF> In Launchpad. Because of !g++-4.7
[04:34] <thomi> we don't support 4.8?
[04:34] <RAOF> Allow me to fix that.
[04:34] <thomi> RAOF: please do
[04:34] <thomi> :)
[04:35] <RAOF> We explicitly override the default g++ to 4.7, which isn't installed in the saucy chroots.
[04:35] <RAOF> Also we fail to build with g++ 4.8
[04:36] <RAOF> You know what? I'm going to fix that, and then remove our inane GCC version lock.
[04:36] <thomi> RAOF: yes, please do - packaging hacks are .... bad
[04:38] <RAOF> make universe-explode-in
[04:38] <RAOF> -delight
[04:39] <kode54> so I suppose I shouldn't jump on this ship just yet
[04:39] <RAOF> XMir in raring should work soon :)
[04:45] <thomi> wooo!
[05:24] <racarr> around for a while if I can be useful :)
[06:02] <RAOF> racarr: Isn't it like 100 o'clock there?
[06:09] <tvoss> RAOF, aka crazy'o'clock
[06:15] <racarr> RAOF: 11:15!
[06:15] <RAOF> racarr: You could rubber-stamp https://code.launchpad.net/~raof/mir/fix-g++-4.8-build/+merge/163635 and https://code.launchpad.net/~raof/mir/remove-gcc-version-forcing/+merge/163641 if you really wanted :)
[06:17] <racarr> ahahaha
[06:17] <racarr> CMP0015
[06:17] <racarr> of course...
[06:17] <racarr> what
[06:18] <racarr> hmm...
[06:20] <RAOF> thomi: WTF? https://jenkins.qa.ubuntu.com/job/mir-raring-amd64-ci/62/console
[06:26] <kode54> looks like an error with the coverage report generation
[06:27] <kode54> maybe gcov output syntax changed between 4.7 and 4.8?
[06:29] <kode54> https://bugs.launchpad.net/gcovr/+bug/1086695
[06:32] <kode54> or maybe that's not the problem
[06:32] <kode54> meh
[06:38] <kode54> yay, now I have working xmir
[06:39] <kode54> but the red cursor won't go away
[06:41] <tvoss> kode54, is it an "ugly" arrow cursor?
[06:41] <kode54> yes
[06:41] <tvoss> kode54, that's the hw-cursor and the policy to hide it is set to "always show" :)
[06:42] <kode54> is there any way to hide that?
[06:43] <RAOF> unity-system-compositor should be doing that.
[06:43] <RAOF> Although it might not be :)
[06:45] <kode54> wtf, now plymouthd is eating 100% CPU
[06:47] <RAOF> Ah, we're not shutting that down correctly then :)
[06:49] <kode54> also, alt-f1 through alt-f5 toggles mouse input on and off (and also hides the hardware cursor when it's disabled)
[06:50] <RAOF> That sounds like VT switching?
[06:50] <kode54> should be, but no switching seems to occur
[06:51] <kode54> oh weird
[06:52] <kode54> when I alt-f4, it doesn't just stop watching mouse input, it also stops updating the display altogether
[06:57] <RAOF> That would be you quitting unity-system-compositor, I think :)
[06:58] <kode54> it stops updating display if I hit alt-f1 through alt-f5, and resumes if I hit any of those again
[07:03] <RAOF> That's odd :)
[07:04] <kode54> where is the latest source for unity-system-compositor?
[07:04] <kode54> just the source package in the ppa?
[07:07] <RAOF> bzr branch lp:unity-system-compositor, I believe
[07:07] <kode54> ah, found it
[07:07] <kode54> yeah
[07:22] <RAOF> kdub: Is there a particular reason why your cross-compile solution doesn't build a minimal armhf chroot using debootstrap? That way you'd get all the funky dependency resolution for free.
[10:05] <alan_g> RAOF: you about?
[10:12] <alf__> mmrazik|afk: Hi! I noticed we are still setting -DMIR_DISABLE_INPUT=OFF in the jenkins hook scripts (H15enable_testing), but this has no effect any more. Feel free to remove it.
[10:33] <RAOF> alan_g: Briefly. What's up?
[10:33] <alan_g> Just wondered about proposing a fix to remove-gcc-version-forcing or rolling it into a fixed MP
[10:34] <alan_g> RAOF: ^
[10:34] <RAOF> Yeah; I got sidetracked into getting the android build to work on my saucy machine, and then EOD hit.
[10:35] <RAOF> Feel free to roll it into a fixed MP.
[10:35] <alan_g> ack
[10:36] <RAOF> And now, sleep.
[12:09] <alan_g> mmrazik|afk: @https://jenkins.qa.ubuntu.com/job/mir-android-raring-i386-build/686/console - can we please reconfigure Jenkins to grab g++-arm-linux-gnueabihf/raring instead of g++-4.7-arm-linux-gnueabihf/raring?
[12:50] <mmrazik> alan_g|lunch: done and retriggered the ci job on that branch (seems to be working; ~53% compile ATM)
[13:01] <kode54> I saw nothing in the unity-system-compositor about the hardware cursor
[13:01] <kode54> so I'm not sure where that's supposed to be disabled
[13:24] <kgunn> alf__: curious, i've been trying for a bit now to use prebuilt mir bins
[13:24] <kgunn> problem is always the same
[13:25] <kgunn> when i make set lightdm.conf type=unity
[13:25] <kgunn> and i restart lightdm
[13:26] <kgunn> it just gives a blank screen with cursor
[13:26] <kgunn> (some thought it was a mismatch for old mirserver)
[13:26] <kgunn> but its still happening
[13:26] <kgunn> after a reboot i can always get back to a prompt
[13:27] <kgunn> to remove the type=unity
[13:27] <kgunn> and then get the ui back ok
[13:27] <kgunn> alf__: ^
[13:27] <kgunn> any ideas?
[13:27] <kgunn> i'm on a macbook/intel integ gfx
[13:28] <kgunn> side note, when i try to launch unity from command line it whines about
[13:28] <kgunn> not being able to load the gl plugin in compiz
[13:33] <alan_g> kgunn: I don't know if it is related, but I had problems with one of my boxes over the weekend - the compiz config had some conflicts that prevented unity loading.
[13:33] <alf__> kgunn: (note: I haven't tried running unity/xmir for some time now.) Is the cursor the X cursor? Is mir definitely running?
[13:34] <kgunn> alf__: ignorace is bliss...what is the "X cursor"?
[13:34] <kgunn> there's no prompt, i can type characters...but they do nothing (seemingly)
[13:35] <alf__> kgunn: is the mouse cursor you normally get under X? I am just trying to figure it if X starts at all, i.e. is this an xmir issue or a unity issue
[13:36] <kgunn> alf__: lemme give it a shot & be more aware....i'll grab the lightdm log as well
[13:37] <kgunn> alan_g: thanks...i blanked all my ccsm settings(like no more wob windows :)
[13:38] <alf__> kgunn: if you can ssh to your laptop you could try launching an xterm/xcalc to see if X(Mir) is running properly
[13:38] <kgunn> alan_g: now that you mention it...just did apt-get dist-upgrade....got a bunch of x/compiz updates...
[13:39] <alf__> kgunn: could it be that you overwrote some XMir packages from the PPA?
[13:40] <kgunn> alf__: nope...just doing that update now
[13:40] <alf__> kgunn: ok
[13:40] <kgunn> alf__: maybe it'll cure it :)
[13:42] <kgunn> brb
[13:51] <alan_g> mmrazik: thanks
[14:03] <thomi> RAOF: ugh - coverage breaks with gcc 4.8
[14:10] <kgunn> alf__: sneaking time from UDS :)
[14:10] <kgunn> but i just retested
[14:10] <kgunn> same result
[14:10] <kgunn> & i know i wasn't clear
[14:10] <kgunn> order of events are
[14:10] <kgunn> change type=unity, restart lightdm
[14:10] <kgunn> get screen with seemingly useless cursor
[14:11] <kgunn> reboot machine
[14:11] <alf__> cursor as in mouse cursor/pointer ?
[14:11] <kgunn> get a useful prompt but still no ui
[14:11] <kgunn> on useless cursor screen no mouse
[14:11] <kgunn> no mouse on useful cursor screen either
[14:13] <kgunn> alf__: oh, and when i try to launch unity manually after reboot with type=unity in lightdm
[14:13] <kgunn> it says compiz(core) fatal: couldn't open display:0
[14:14] <alf__> kgunn: is an X process running ?
[14:15] <kgunn> alf__: how would i go about verifying that?
[14:15] <kgunn> alf__: note, the lightdm log says WARNING: Failed to create default seat
[14:18] <alf__> kgunn: ps -Af | grep X
[14:19] <kode54> bwahahaha
[14:19] <kode54> "clang: error: unknown warning option '-Werror=unused-local-typedefs'"
[14:19] <kgunn> alf__: thanks! will steal more uds time later today to try :)
[14:21] <alf__> kgunn: probably robert-ancell would be of more assistance here, since he knows lightdm and its interactions with mir and X better
[14:21] <kgunn> alf__: cool
[14:22] <kgunn> alf__: i had pestered a bit last week....just know he's on baby watch this week
[14:24] <kode54> I thought the issue with xmir breaking was due to something wrong with the Xorg package in the mir staging ppa
[14:25] <alf__> kgunn: kode54: could be, so I guess RAOF is the one to pester :)
[14:25] <kode54> I thought he fixed it
[14:25] <kode54> it was working here
[14:26] <kode54> but the unity-system-compositor package still hasn't been rebuilt, and the current version doesn't hide the hardware cursor
[15:02] <alf__> status: investigating/fixing bugs, experimenting with lttng and tools
[15:09] <kdub> status, mostly starting performance work with android devices
[15:10] <kdub> greyback, lp:mir should have the code needed for the shell on phone
[15:11] <greyback> kdub: it does. I'm having a problem with the qpa plugin that I need to understand/ask racarr about
[15:12] <alan_g> status: trying to understand the current design well enough to add composition bypass cleanly
[15:37] <racarr> Morning
[15:39] <kdub> hello racarr
[15:40] <greyback> racarr: hey
[15:40] <racarr> kdub: greyback: Did you guys see my email?
[15:40] <racarr>  / did it make sense
[15:40] <greyback> racarr: yep, have been working off it today
[15:40] <racarr> I wrote it at like midnight right before going to sleep aha
[15:41] <greyback> :)
[15:51] <racarr> kdub: So what do you think about
[15:52] <racarr> letting the shell get at the EGLNativeDisplay without a surface yet (i.e. like in my email)
[15:52] <racarr> Qt wants to choose configurations and initialize EGL etc when QScreen is created
[15:52] <racarr> so ubuntu_application_ui_get_native_display
[15:52] <racarr> can't take any arguments
[15:53] <kdub> racarr, makes sense to me
[15:54] <racarr> kdub: p.s. do your eglconfig for qtubuntu fixes on phone exist somewhere
[15:54] <racarr> greyback is trying to get mirclient on phone going
[15:54] <kdub> i'll have to find them
[15:54] <racarr> Ok. Thanks :)
[15:54] <greyback> kdub: thanks!
[15:54] <racarr> I...can't figure out how to expose this native display...
[15:54] <racarr> doe it jut go back to the graphics platform?
[15:59] <kdub> racarr, we could have an InternalClient that takes the surface in the creation of the eglNativeWindowType
[15:59] <kdub> as a parameter to the function, as opposed to in the constructor
[16:00] <racarr> kdub: Oh. That's a good idea
[16:03] <kdub> greyback, this is how i got the qpa plugin to work over the sprint... http://pastebin.ubuntu.com/5664882/
[16:03] <kdub> just to give some ideas :) probably not the universal solution
[16:03] <greyback> kdub: thanks, will give it a go
[16:05] <racarr> I think this qeglconvenience
[16:05] <racarr> is copied from a qt built with EGL (which ours isn't or wasn't or something?)
[16:05] <racarr> a qt5 beta from a long time ago
[16:05] <racarr> or something
[16:05] <racarr> I think Qt can do this for us now but will have to figure out how
[16:12] <racarr> The coffee shop on my new corner has insanely caffeinated coffee oO
[16:18] <alan_g> that's what has you writing at midnight?
[16:19] <greyback> racarr: kdub: excellent, we have stuff on screen at last
[16:19] <kdub> greyback, great :)
[16:19] <racarr> Yay!
[16:19] <kdub> input working too?
[16:20] <racarr> alan_g: No. What has me writing at midnight is that I've actually been sleeping lately so im not tired as soon as it's 6pm haha
[16:20] <racarr> which sounds kind of counterintuitive but writing emails at midnight is much less of a problem than being kept up by housemates at 2 am :p
[16:20] <greyback> kdub: I'm getting the events printed in my terminal, so appears to
[16:21] <greyback> kdub: bringing up something more intelligent now
[16:21] <kdub> greyback, when i got it working at the sprint, i had to use qml touch events, mouse events weren't going through for some reason
[16:21] <kgunn> greyback: kdub racarr aweseome!!
[16:22] <greyback> kdub: interesting, will look into that
[16:26] <greyback> kdub: input working
[16:26] <kdub> cool
[16:26] <kdub> it should be multi-touch too
[16:26] <greyback> kdub: checking...
[16:28] <mlankhorst> ]
[16:28] <greyback> kdub: yep
[16:28] <racarr> whee
[16:28] <racarr> kdub: Good news, it wasn't a shared hallucination!
[16:32] <kdub> great :)
[16:33] <tvoss_> greyback, ping
[16:33] <greyback> tvoss_: pong
[16:34] <racarr> ok almost done with this native display stuff....
[16:35] <racarr> we have about a half dozen duplicated instances of StubPlatform : mg::Platform ;)
[16:37] <kdub> racarr, i noticed that too
[16:40] <alan_g> racarr: is there something wrong with that?
[16:45] <alan_g> Different tests may need different behaviour from stubs of the same interface.
[16:45] <kdub> alan_g, yeah, the disadvantage is the typing :)
[16:47] <racarr> alan_g: In this case they are all the same really ;)
[16:47] <racarr> just null tubs
[16:47] <racarr> stubs*
[16:47] <racarr> I think
[16:47] <racarr> I dunno
[16:47] <racarr> it's cumbersome to have to update a half dozen files
[16:48] <racarr> including unexpected ones
[16:48] <racarr> so it becomes
[16:48] <racarr> compile
[16:48] <racarr> wait for error
[16:48] <racarr> etc
[16:48] <alan_g> racarr: kdub: if they are necessarily (as opposed to coincidentally) duplicate there are grounds to refactor
[16:49] <racarr> I made a note of it in my: "Things to do when its not obvious what else to do"
[16:49] <racarr> list ;)
[16:50] <alan_g> But surely the motivation for a change to the interface comes from some new tests - maybe a different interfaces is a better solution?
[16:50] <racarr> maybe in this case...I'm not sure
[16:50] <racarr> kdub: https://code.launchpad.net/~robertcarr/mir/emancipate-egl-native-display-from-surface/+merge/163775
[16:50] <racarr> When you have a chance.
[16:51] <alan_g> racarr: maybe...maybe not - but I'm sure that the wrong choice has been made sometimes.
[16:53] <racarr> Perhaps create_internal_client belongs on the Display
[16:53] <racarr> greyback: Want to test to see if it works now while I try and get unity/phablet building?
[16:54] <racarr> I think if you have: https://code.launchpad.net/~robertcarr/mir/emancipate-egl-native-display-from-surface/+merge/163775
[16:54] <racarr> and revision 62 to ~robertcarr/platform-api/mir (just pushed)
[16:54] <racarr> qtubuntu/mirserver should build fine now
[16:55] <greyback> racarr: sure, once I do a few hacks to get shell working without the mir application module
[16:55] <alan_g> Goodnight all!
[16:55] <alf__> alan_g: goodnight!
[16:57] <racarr> Great!
[16:57] <racarr> night alan :)
[16:57] <racarr> who...already left.
[16:58] <racarr> im purging my qt ppas atm
[16:58] <racarr> *taps foot* aha
[17:05] <racarr> ugh now it finds the cmake files from qt but then they claim that nothing exists...
[17:05] <racarr> why did I have to add /usr/lib/x86_linux_gnu/cmake to CMAKE_PREFIX_PATH
[17:05] <racarr> ?
[17:06] <tvoss_> kgunn, ping
[17:12] <greyback> racarr: should/can the mir server run as phablet user, or must be root?
[17:14] <racarr> Can be run as phablet user if you have input permissions
[17:14] <racarr> probably ;)
[17:14] <racarr> and should be
[17:14] <racarr> then a system mir runs as root I guess
[17:15] <kdub> racarr, greyback, i'm unsure about graphics though... i'd suggest running as root if it doesn't work :)
[17:16] <greyback> racarr: kdub: it is working as root, was just curious
[17:17] <racarr> ok got unity/phablet building. going to go for a walk around the block while it runs, back in a bit
[17:23] <philipballew> Just saw a demo of Mir without X. Looking nice!
[17:24] <racarr> :D
[17:25]  * philipballew smiles back at racarr 
[17:25] <racarr> hmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm now ./build -s tells me I may have held broken packages :(
[17:25] <racarr> probably...
[17:27] <racarr> perhaps ubuntu-unity/experimental-certified conflicting with unity-next-desktop
[17:27] <racarr> ppa-purge is a real godsend
[17:28] <kgunn> tvoss|dinner_lun: late pong...sorry, had to grab some grub
[17:29] <olli> kgunn, tvoss|dinner_lun just told me that he might be late for the OSK session due to dinner
[17:29] <kgunn> olli: ack, thanks
[17:31] <racarr> That is in 30 minutes yes?
[17:33] <olli> racarr, yep
[17:33] <racarr> Great
[17:35] <racarr>  indicator-appmenu-tools : Depends: indicator-appmenu (= 13.01.0daily13.03.28-0ubuntu1) but 13.01.0daily13.04.03-0ubuntu1 is to be installed
[17:35] <racarr> greyback: Latest build -s conundrum, any ideas?
[17:36] <greyback> racarr: hmmm
[17:37] <racarr> trying downgrading the package
[17:37] <racarr> ok its working now
[17:38] <racarr> not sure if I will get the qt5 errors later aha..realized the error involved build -s failing early and me going in to build and running cmake myself ;)
[17:38] <racarr> which I just thought was how build -s worked
[17:38] <racarr> building lots of branches and stuff
[17:39] <greyback> racarr: no idea where the duplicate packages are coming from. apt-cache showpkg helps?
[17:39] <racarr> greyback: Mm. Im just going to use the downgrade for now
[17:40] <racarr> need to upgrade to saucy soon anyway
[17:40] <racarr> and do some housecleaning
[17:40] <greyback> racarr: ok. let me know if you still have problems
[17:43] <racarr> thanks :)
[18:01] <racarr>  how do I get lower third set up
[18:01] <racarr> I cant find out how to install it and it is not showing up in featured apps
[18:02] <racarr> Ok here we go
[18:17] <tskorte> Hello everyone! I've followed the "Installing pre-built packages on a PC" on unity.ubuntu.com/mir and the tried to run it natively. Problem is that I get this "failed to create GBM buffer". Has anyone encountered this or have I missed something?
[18:20] <kdub> having one of those days where i was like 'this looks like a simple cleanup...' and then it cascaded into many clean ups :)
[18:21] <kdub> tskorte, you installed the special mesa package?
[18:23] <tskorte> kdub: Well, I installed all the pre-built packages.  That's not enough? :)
[18:24] <kdub> tskorte, it /should/ be, however, i'm not up on what desktop gpus work with it... RAOF will know when he comes back
[18:25] <tskorte> Great
[18:26] <tskorte> In the mean time I'll try to install it from source
[18:26] <kdub> it works with my intal gma-x3100
[18:27] <tskorte> Should've mentioned that I have  a amd radeonhd 5750
[18:43] <racarr> yay unity finished building...test time
[18:43] <racarr> p.s. can someone look at https://code.launchpad.net/~robertcarr/mir/emancipate-egl-native-display-from-surface/+merge/163775
[18:46] <racarr> oh just the dependency build finished :(
[18:53] <hikiko_> tskorte, i had the same problem you need the special mesa pkg from the mir ppa
[18:54] <racarr> greyback: Should I expect that run -m should work if qtubuntu mirserver is working?
[18:54] <racarr> getting some stuff about a binding loop and hud complaints, mir starts but then a few second laters it all quits/dies
[18:54] <racarr> still figuring out
[18:55] <tskorte> hikiko_: Nice! What are the package names?
[18:55] <greyback> racarr: you might also need the "-f" flag to work around the lack of the application API
[18:56] <hikiko_> hmm :) I don't remember one second to check tskorte :)
[18:57] <hikiko_> tskorte, did you try apt-get update apt-get build-dep mir after adding the mir ppa to your apt/sources.list? it should install the correct libraries I think
[18:58] <tskorte> hikiko_: No, I ran update && dist-upgrade. Will try apt-get build-dep now
[18:58] <greyback> racarr: I've to go afk for about 90 mins, good luck!
[18:59] <racarr> greyback: Ok see you! got a stack trace so hopefully can fix it
[18:59] <racarr> looks like my fault ;)
[19:00] <greyback> oO
[19:08] <racarr> eglplatform.h strikes again
[19:09] <racarr> yay it works
[19:09] <tvoss> racarr, success? :)
[19:10] <kdub> racarr, when you say eglplatform.h strikes again
[19:10] <kdub> you mean that we don't have a #mir_platform ifdef that defines our types?
[19:10] <kdub> for the mesa platform?
[19:12] <racarr> tvoss: Yep phone shell running again on g bm and 95% sure it should run on android ;)
[19:12] <racarr> kdub: Yep.
[19:12] <racarr> I will remember...to talk to chris about that today
[19:12] <tvoss> racarr, make it running :)
[19:12] <hikiko_> hey why nobody is in the hangout?
[19:12] <hikiko_> :s
[19:12] <kdub> i've noticed that too... on mesa, our ipc package has morphed into the native buffer type
[19:13] <racarr> hmm?
[19:13] <racarr> tvoss: it will be today :) greyback has a set up all ready though that just needs to plug in this fixed platform-api branch so
[19:13] <tvoss> racarr, go ahead
[19:13] <racarr> im going to wait for him to test it and work on other things :)
[19:13] <tvoss> racarr, :)
[19:14] <racarr> the video lens looks awesome on 27 inch screen haha
[19:14] <racarr> excluding the big ugly cursor
[19:16] <tvoss> racarr, ;)
[19:17] <kdub> racarr, once RAOF logs on, lets talk eglplatform types for mesa :)
[19:18] <racarr> sounds good!
[19:53] <racarr> Lunch!
[19:54] <tvoss> racarr|lunch, enjoy
[19:56] <racarr|lunch> literally next door to a curry place now...hopefully this doesn't get out of control ;)
[20:41] <racarr> back!
[20:46] <FunnyLookinHat> Saw the demo w/ Unity 8 running on Mir - looks good guys!
[20:48] <olli> so who of you guys did the U8 / Mir on ARM demo
[20:48] <olli> that was a nice surprise
[20:48] <kdub> olli, probably greyback, haven't seen it yet
[20:49] <olli> https://plus.google.com/u/0/116997345010659023379/posts
[20:49] <olli> can't tell from the fingertips
[20:49]  * olli needs to get to know the team better
[20:50] <kdub> yay, performance looks ok in the video too :)
[20:51] <racarr> how
[20:51] <racarr> is that working XD
[20:51] <racarr> maybe out of process stuff or did someone just make it?
[20:52] <kdub> racarr, i'd guess out-of-process...
[20:56] <olli> racarr, how often do you and greyback sync
[20:59] <kdub> greyback, your video's famous :)
[21:01] <greyback> kdub: lol my finger will be recognized in public now
[21:01] <kdub> yep :)
[21:04] <racarr> olli: About 4 times a day ;)
[21:04] <greyback> racarr: any luck with in-process mir+unity?
[21:05] <racarr> greyback: Hehe was just going to say yes, its working
[21:05] <greyback> racarr: nice!
[21:05] <racarr> greyback: If you build ~robertcarr/platform-api/mir with https://code.launchpad.net/~robertcarr/mir/emancipate-egl-native-display-from-surface/+merge/163775
[21:05] <racarr> it should work, I havent' tested on android yet
[21:05] <racarr> but it's the "same code path" now so
[21:05] <greyback> racarr: ok, I'll give it a go
[21:05] <racarr> Awesome
[21:05] <racarr> it looks great on 27 inch screen :D
[21:05] <racarr> the video lens...*wide eyes*
[21:06] <greyback> xD
[21:06] <greyback> I hope you get a nice big value for GRID_UNIT_PX, so things scale nicely
[21:07] <racarr> greyback: It all look pretty appropriate.
[21:07] <racarr> Maybe I'm thinking too far ahead, but I think there will be some mir work (mostly exposing new interfaces, etc) required as well as shell work
[21:07] <racarr> so I was wondering during lunch...
[21:08] <greyback> write them down, do share!
[21:08] <racarr> how many pieces are missing from getting the launcher to some point like "displays a list of desktop files, can launch applications, can switch between them"
[21:09] <racarr> I'm just thinking about little mini milestones to try and get some basic surface management in place
[21:10] <greyback> sure. We need a concrete list, prioritized, and then decide where each bit belongs
[21:10] <greyback> and who does it
[21:12] <racarr> What do you mean?
[21:14] <greyback> racarr: just like you said, to identify the missing pieces and make work items out of them
[21:14] <racarr> Haha ok, sorry I wasnt sure if you were referring
[21:14] <racarr> to the work items
[21:14] <racarr> or the desktop files
[21:14] <racarr> and I was like...well...I dont think it really matters which desktop files...
[21:14] <racarr> nvm :p
[21:17] <greyback> ah, understood
[21:17] <racarr> hmm.
[21:18] <racarr> ill think about this more tomorrow actually :p because there is not really much useful you can do
[21:18] <racarr> without the window stacking/sorting interface in mir
[21:18] <racarr> So that's a missing piece that's a work item
[21:18] <greyback> yep
[21:18] <racarr> I guess the first missing piece ont he shell end, is mir has a SessionListener you can implement
[21:19] <racarr> that is told session_appeared/vanished with the application id (desktop id I guess right now)
[21:19] <racarr> so the session needs to be assosciated with the launcher icon
[21:19] <racarr> so it can be switched to later
[21:20] <greyback> well each session is related to a desktop file, no?
[21:20] <racarr> yes (well, if qtubuntu passed this through)
[21:20] <greyback> I think it did
[21:20] <racarr> I just mean there needs to be an implementation of this interface in unity
[21:20] <racarr> that exposes it out to qml
[21:21] <racarr> the matching should be trivial
[21:21] <greyback> For the launcher, I think all Mir needs to tell us is what apps are starting/open, how many windows they have (desktop-only) and things like, it is looking for attention, etc
[21:22] <racarr> ok
[21:22] <racarr> ill fix the desktop file pass through
[21:23] <greyback> for now I think that's all we really need
[21:25] <racarr> I'm really in to getting a sort of functional (if super bare bones) shell together so I can start dogfooding the input stack on myself
[21:25] <racarr> I imagine between /dev/input and every application there are probably a few quirks to iron out :p
[21:28] <greyback> yep, but you're right - we gotta start somewhere
[21:41] <racarr> I actually dont have a whole lot planned to do for the end of the week because my highest priority stuff is blocked on the platform API refactoring
[21:41] <racarr> so maybe I will try and do the start of the launcher integration in unity I think it will help me to start to get
[21:41] <racarr> a mental map of the unity 8 code
[21:45] <racarr> RAOF: Ping?
[21:49] <racarr> thomi: Ping?
[21:50] <racarr> Raof isn't around so I'm going to ask you about something different instead ;)
[22:48] <RAOF> racarr: Pong.
[22:56] <racarr> RAOF: Can we fix this eglplatform.h thing so I can stop telling people to edit eglplatform.h?
[22:59] <kdub> racarr, RAOF we're in a pretty good position for  you to define whatever you want for the mesa/mir native types
[22:59] <kdub> good position, from a mir-internals perspective
[22:59] <kdub> i was wondering... could we even use the defines in __GBM__ ?
[23:00] <racarr> EGLNativeDisplayType for mirclient mesa is not gbm_device * though
[23:00] <racarr> its MirEGLNativeDisplayType (i.e. void* i.e. MirMesaEGLNativeDisplay)
[23:00] <kdub> racarr, right, i don't really know the diff between the two
[23:00] <kdub> (namely, what's in gbm_device)
[23:01] <kdub> but it doesn't sound like they're close to each other
[23:02] <kdub> seems like MirBufferPackage has just become the de-facto buffer type for the platform
[23:02] <kdub> when, we could actively choose what we want in that struct now
[23:02] <racarr> thats not exposed out through the eglplatform.h
[23:02] <racarr> its just display, window, and pixmap
[23:03] <kdub> well, sure
[23:03] <kdub> but while we're defining our types for mesa, may as well take care of all the issues :)
[23:03] <racarr> mm
[23:08] <kdub> like, we have our EGLNativeDisplayType and EGLWindowType as just 'the most convenient internal types'
[23:08] <kdub> which, isn't bad, but we have the opportunity to actively define what we want out of the interface :)
[23:08] <racarr> I think we just want opaque pointers
[23:09] <kdub> that would be ok
[23:14] <kdub> racarr,  with emancipate-egl-.*, small needs fixing, otherwise looks good
[23:14] <racarr> I hope to one day name a class <Noun>Emancipator
[23:14] <racarr> im not sure what it does
[23:14] <racarr> escalates priveleges perhaps
[23:17] <racarr> kdub: Ok :) Just pushed a fix in r686
[23:17] <racarr> thanks.
[23:23] <racarr> RAOF: http://pastebin.ubuntu.com/5666057/ perhaps?
[23:23] <racarr> kdub: ^
[23:23] <kdub> racarr, i think it should be forward declared probably
[23:25] <RAOF> Looks plausible.
[23:36] <racarr> RAOF: Can we merge it?
[23:38]  * RAOF fires up the mergeotron.
[23:39]  * kdub is imagining one of those spinning carnival ride things
[23:40] <racarr> RAOF: Savor the moment, likely to be the first and last EGLNativeDisplayType we will define XD