[00:24] <RAOF> cnd: It probably needs a ping; I'm not aware of anything that needs doing on my end.
[00:25] <cnd> yeah, looks like it
[00:26] <cnd> I just pinged kees in case he could look at it
[00:27] <cnd> he's the only one that I know is in a reasonable timezone right now
[00:30] <cnd> RAOF, approved :)
[00:31] <RAOF> Excellent.  Now, we wait for the freeze to be lifted, then upload.
[00:31] <RAOF> And by “upload”, I mean “upload an Xserver with a with build-depends on libgtest”
[00:31] <RAOF> -xorg
[00:31] <cnd> yeah
[08:07] <eruditehermit> is gesture support still broken in unity?
[08:07] <eruditehermit> tseliot, hey
[08:08] <tjaalton> wouldn't know about gestures since I've no hw for it
[08:09] <tseliot> hi eruditehermit
[08:09] <eruditehermit> tseliot, I was told you were the person to talk with about fglrx and hybrid graphics. Can you help me?
[08:10] <tseliot> eruditehermit: what's the problem?
[08:10] <eruditehermit> tseliot, first of all, is hybrid graphics supported with the packages in precise? the new fglrx 12.3 release?
[08:11] <tseliot> eruditehermit: not officially
[08:11] <eruditehermit> I have always had more problems with the ubuntu packages than the ati packages from their site
[08:11] <tseliot> eruditehermit: I maintain both and they share the same source
[08:13] <eruditehermit> in order to make it work, I have to manually mess around  with update-alternatives --config x86_64-linux-gnu_gl_conf
[08:13] <eruditehermit> tseliot, so if I try to make it work with the ubuntu packages, I should have the same success as with the official packages?
[08:15] <tseliot> eruditehermit: yes, I think so. It all depends on your laptop/BIOS
[08:15] <eruditehermit> ok
[08:15] <eruditehermit> are you around for a bit today?
[08:15] <eruditehermit> can I restart and find out what was going wrong?
[08:16] <eruditehermit> ideally I want to be able to use amdcccle to switch GPUs
[08:20] <eruditehermit> tseliot, how does the xorg.conf work?
[08:20] <eruditehermit> does the fglrx require it?
[08:20] <tseliot> eruditehermit: it's all handled automatically
[08:21] <eruditehermit> so if I install the package
[08:21] <eruditehermit> I don't need to create it?
[08:21] <eruditehermit> currently it doesn't exist
[08:22] <tjaalton> the server uses fglrx if it's installed
[08:23] <eruditehermit> ok
[08:23] <eruditehermit> well brb
[08:28] <eruditehermit> tseliot, http://paste.ubuntu.com/903536/
[08:28] <eruditehermit> I installed fglrx, and now I have no 3D capabilities
[08:30] <tjaalton> intel wins :)
[08:31] <eruditehermit> so my card requires both intel and fglrx to work with fglrx
[08:31] <eruditehermit> since it is muxless
[08:31] <eruditehermit> I've had it working in hte past sporadically and fglrx loads intel
[08:31] <tjaalton> no, you can disable the intel part
[08:32] <tjaalton> i think
[08:32] <eruditehermit> nah
[08:32] <eruditehermit> it requires it
[08:32] <tjaalton> then you can't use fglrx
[08:32] <eruditehermit> hrm
[08:33] <eruditehermit> I can
[08:33] <eruditehermit> its worked in the past
[08:33] <tjaalton> i'll just shut up, haven't touched intel/ati hybrid myself..
[08:33] <eruditehermit> its tricky
[08:33] <eruditehermit> not sure exactly what is going on ever
[08:33] <eruditehermit> lol
[08:34] <eruditehermit> brb
[08:44] <tseliot> eruditehermit: unless you can set your AMD card as primary in the BIOS, I'm afraid you can't use hybrid graphics
[08:44] <eruditehermit> ok
[08:44] <eruditehermit> got it working
[08:44] <eruditehermit> but I have to manually attack it
[08:44] <eruditehermit> to get fglrx to play nice with my GPUs I have to
[08:45] <eruditehermit> manually set x86_64-linux-gnu_gl_conf
[08:45] <eruditehermit> and I have to have an /etc/X11/xorg.conf from fglrx
[08:46] <eruditehermit> tseliot, is there a way to make it just work?
[08:48] <tseliot> eruditehermit: are you sure it works?
[08:48] <eruditehermit> tseliot, yep, fgl_glxears works
[08:49] <eruditehermit> tseliot, also glxinfo says OpenGL vendor string: Advanced Micro Devices, Inc.
[08:50] <tseliot> eruditehermit: and did you change anything in the BIOS?
[08:50] <eruditehermit> nope
[08:50] <eruditehermit> there are no GPU settings in the bios
[08:50] <tseliot> I'm surprised it loaded the driver
[08:51] <eruditehermit> my discrete GPU requires the intel GPU to be on
[08:51] <eruditehermit> it uses the intel GPU for a mux
[08:51] <tseliot> eruditehermit: I have a laptop with the same problem, I guess. If I can get it to work, as you say, I can probably make it just work ;)
[08:52] <eruditehermit> ok
[08:52] <eruditehermit> want to test it now?
[08:52] <eruditehermit> I have a Sony VAIO SA 290x
[08:53] <eruditehermit> 2011
[08:53] <eruditehermit> hmm
[08:53] <eruditehermit> let me try with the iibgl alternative at auto
[08:53] <eruditehermit> maybe it just needs an xorg.conf
[08:53] <eruditehermit> brb
[08:54] <mlankhorst> morning
[09:02] <eruditehermit> tseliot, any luck?
[09:02] <tseliot> eruditehermit: unfortunately I can't test it right now (I have some other work to do first)
[09:03] <eruditehermit> tseliot, when I try to switch to intel GPU only, my window managers won't start for some reason
[09:03] <eruditehermit> I have to switch to a VT
[09:04] <eruditehermit> type DISPLAY=:0 gnome-shell --replace
[09:04] <eruditehermit> and then it loads gnome-shell or unity or whatever
[09:15] <eruditehermit> is there a place where unity or gnome-shell report why they didn't start properly?
[10:08] <eruditehermit> tseliot, seems to be that if fglrx is installed and I set it to use the intel GPU. It initially uses the sw rasterizer. Unity and gnome shell get confused and crash. However if I start a gnome-shell from my VT later, it correctly sets the rasterizer to be the intel dri driver and it works
[10:11] <tseliot> I don't think that's how it's supposed to work
[10:15] <eruditehermit> for some reason it is failing to load my window managers the first time around
[10:15] <eruditehermit> if I load fluxbox, it shows it is using the software tungstem llvm driver
[10:16] <eruditehermit> however, if I load gnome-shell
[10:16] <eruditehermit> it fails to load the wm
[10:16] <eruditehermit> but then I can load it from my VT
[10:20] <eruditehermit> hrm perhaps its my kernel?
[10:20] <eruditehermit> brb
[10:33] <dholbach> hiya
[10:33] <dholbach> hiya
[10:33] <dholbach> it seems my girlfriend has an interesting problem with keyboard layouts
[10:33] <dholbach> and I'm wondering if you've seen anything like this before
[10:34] <dholbach> 1
[10:36] <dholbach> http://paste.ubuntu.com/903691/
[10:36] <dholbach> and the active layout is actually German
[10:36] <dholbach> where else could I check if something is broken or set in a wrong way?
[10:36] <dholbach> pitti sent me here :)
[10:48] <dholbach> "setxkbmap no" seems to make it work, but after restarting the session it's gone again
[11:35] <dholbach> ok, for now we resorted to filing this bug to keep track of the issue: https://bugs.launchpad.net/ubuntu/+source/gnome-settings-daemon/+bug/967034?
[14:07] <Darxus> tseliot: Nvidia posted a solution to their (cairo-gl triggered) memory usage problem, you were assigned the previous related bug, so thought you might be interested.  Gah, copying and pasting on my girlfriend's ChromeBook is miserable.
[14:08] <tseliot> Darxus: sorry, where?
[14:10] <Darxus> tseliot: Bug 967149.
[14:11] <tseliot> Darxus: very interesting, I'll have a look at it, thanks
[14:11] <Darxus> Awesome, thanks.
[14:11] <tseliot> RAOF, Sarvatt ^
[14:12] <Darxus> There is a followup post from nvidia in that thread also.
[14:12] <tseliot> good
[17:30] <mlankhorst> hey :)
[17:35] <bryceh> heya mlankhorst
[17:39] <mlankhorst> hello
[17:47] <cnd> bryceh, I want to use arsenal for creating utouch bug reports
[17:47] <cnd> I see you recently added a search for subscribers of bugs for distro tasks
[17:47] <cnd> I want the same, but for all tasks
[17:48] <cnd> I'm new to arsenal, so any pointers you can give are appreciated :)
[17:48] <bryceh> hi cnd
[17:49] <bryceh> cnd, actually that was someone else's patch which I merged
[17:49] <cnd> oh, ok
[17:50] <cnd> ahh, I see it's chris arges
[17:50] <cnd> I know him well :)
[17:50] <bryceh> cnd, you can leave out distribution and instead specify the project(s) you want to pull from
[17:50] <cnd> bryceh, does arsenal support bug searching and reporting for non-distro tasks?
[17:51] <bryceh> yeah see the hwe reports for example
[17:51]  * bryceh -> breakfast
[17:51] <cnd> bryceh, do you know if the timeouts we see when using launchpad.net will also occur using arsenal?
[18:04] <bryceh> cnd, unfortunately launchpad is flaky like that even at the api layer.
[18:04] <bryceh> cnd, you might not get those time outs, but there are various other issues that can crop up
[18:04] <cnd> ok
[18:23] <cnd> bryceh, I'm having a hard time understanding the lplib api
[18:23] <bryceh> cnd, join the crowd :-)
[18:23] <cnd> I know there are http methods documented here: https://launchpad.net/+apidoc/1.0.html
[18:24] <cnd> but I don't know how to map an http method to an arsenal or launchpadlib method
[18:24] <cnd> for example, I want to get all the tasks by doing  searchTasks on a person
[18:25] <cnd> how do I follow the toolkits and libraries to figure out what python code to call?
[18:25] <bryceh> cnd, ok here is a simple place to start
[18:25] <bryceh> apt-get source python-launchpadlib-toolkit
[18:25] <bryceh> cd into the scripts directory
[18:25] <bryceh> look at ls-assigned-bugs
[18:26] <bryceh> that's a very trivial implementation of one single LP call, but shows how to use searchTasks
[18:26] <bryceh> ls-series in the same dir is a slightly more sophisticated example, look at that one next
[18:27] <cnd> bryceh, what's the upstream project for that package?
[18:27] <bryceh> cnd, arsenal
[18:27] <cnd> hmm.. why do I not see scripts/ls-assigned-bugs in lp:arsenal?
[18:28] <bryceh> cnd, http://bazaar.launchpad.net/~arsenal-devel/arsenal/python-launchpadlib-toolkit/files/head:/scripts/
[18:29] <cnd> oh, I'm in the wrong scripts directory
[18:29] <bryceh> yes, it needs to be moved out into it's own project rather than be a branch of arsenal
[18:29] <bryceh> some day...
[18:29] <cnd> oh, it's a *branch* of arsenal...
[18:32] <cnd> hmm... I want to call searchTasks without specifying the project or distro...
[18:33] <bryceh> I think you have to do one or the other
[18:33] <bryceh> however, you could retrieve a list of all projects and distros and then iterate over those, calling searchTasks on each and building a collection
[18:34] <bryceh> might take a while to run...  ;-)
[18:35] <cnd> the http api seems to show that it's possible
[18:35] <cnd> I need to figure out how to code it though
[18:41] <bryceh> cnd, ls-team-subscribed-packages.py shows how to get a person (aka team) object.  You can then call searchTasks() on it
[18:41] <cnd> I'll try it
[18:55] <cnd> bryceh, well, I think I'm trying to search for all bugs subscribed by utouch-bugs :)
[18:55] <cnd> it's taking a while and just sitting there...
[18:56] <cnd> I assume if I did it wrong I would get an immediate error
[18:56] <cnd> hmm, 503: Service Unavailable
[18:57] <cnd> timeout error
[18:57] <cnd> so I guess this doesn't let us get around the timeout
[18:59] <bryceh> cnd, can't say I didn't warn you upfront.  ;-)
[18:59] <cnd> yeah
[18:59] <cnd> it was worth a shot
[19:00] <bryceh> cnd, I do wish they had a different timeout setting for command line apps than for the web
[19:00] <bryceh> cnd, often with those time outs the problem is that you're retrieving too much data in one go.
[19:00] <bryceh> cnd, there's a trick I know of to batch it up... dunno if it'd make any difference though.  hang on
[19:01] <cnd> in this case, they just have a backend sql query that takes too long
[19:01] <bryceh> s/batch it up/parcel it out/
[19:01] <bryceh> ah, yeah then in that case you're just screwed
[19:03] <bryceh> cnd, what I was thinking of is in the Collection section on https://help.launchpad.net/API/launchpadlib
[19:04] <bryceh> cnd, you can retrieve just a fixed number of items, so if the problem was just too much data, that might help in dividing it into smaller chunks that would not trip the time out
[19:04] <cnd> ok
[19:04] <bryceh> however I don't think that'd do the trick here.  might be worth trying, dunno
[19:05] <cnd> I think we'll stick to all bugs in specific projects and ubuntu packages, and add in all tasks of ubuntu packages that are directly subscribed by the utouch-bugs team
[19:05] <cnd> both of those (usually) work without timing out
[19:05] <bryceh> cnd, *nod* makes sense
[19:45] <bdmurray> does ppa-purge with xorg-edgers and precise?  I discovered I had that ppa enabled and don't want it anymore and ppa-purge created a list of 550 things for me to review
[19:47] <bryceh> bdmurray, I've not heard anything about it not working.  but 550 things sounds a bit excessive
[19:49] <bdmurray> is there some other way to get back to precise besides ppa-purge?
[19:55] <Sarvatt> bdmurray: ppa-purge doesn't work with multiarch if you're on amd64
[19:55] <Sarvatt> basically you want to apt-get install package/precise for every package you have installed from edgers, dpkg -l then grep sarvatt or ricotz or edgers to find them all
[19:55] <Sarvatt> http://paste.ubuntu.com/904473/ should be most of it
[19:56] <Sarvatt> (after disabling the ppa and doing an apt-get update)
[19:57] <bdmurray> Sarvatt: ah, yep I'm on amd64
[20:00] <Sarvatt> removing wine1.4 first helps
[20:43] <FernandoMiguel> boas