[00:02] cya [10:56] woops, broke switching back [12:44] fyi http://bugs.debian.org/693158 may be somewhat relevant to ubuntu as well [12:44] Debian bug 693158 in xserver-xorg-input-penmount "Xorg: symbol lookup error: ...penmount_drv.so: undefined symbol: xf86XInputSetScreen" [Grave,Open] [12:48] proof that nobody uses it? :-) [12:48] bah [12:49] hm maybe we use a newer version [12:50] oh we don't [13:01] tjaalton: hey can we drop penmount from universe? :> [13:03] sounds like a plan [13:11] https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-input-penmount/+bug/1082015 weee [13:11] Launchpad bug 1082015 in xserver-xorg-input-penmount (Ubuntu Raring) "please remove xserver-xorg-input-penmount from universe" [Undecided,New] [14:51] Sarvatt and RAOF: I spotted a code merge for piglit that enables EGL, opengl es1 and opengl es2 tests. === Amaranthus is now known as Amaranth [15:52] waffle one? [16:01] mlankhorst: yes, waffle is required for opengl es 1/2 tests ( and to use egl w/ opengl on piglit ) [16:02] and there's a missing environmental variable... PIGLIT_SOURCE_DIR needs to be set ( most likely to /usr/share/piglit/ ) [17:07] tjaalton: hm can I add recommends xserver-xorg-{video,input}-all, libgl1-mesa-glx to xserver-xorg ? [17:08] should make stack switching slightly more robust [17:10] I fear that may end up installing xserver-xorg-(input,video)-all again on machines that previously removed them, though.. [17:11] yeah i think it'd have some side-effects.. [17:13] I'll try without.. [17:27] hey, how do I view and change my font path? I want programs to search ~/.fonts first and /usr/share/fonts later [17:27] is that up to X? [17:41] guessing fontconfig cares [17:43] tjaalton: can we recommend xserver-xorg-input-all at least? [17:56] mlankhorst: i'm not sure, and not able to check now what issues it might have [17:56] ok :/ [17:56] well in the worst case switching back will only leave you with evdev then.. [18:02] do you have a dl for the stack? [18:03] it's in my experimental ppa [18:05] but still testing a little [18:08] maybe things clear up if I re-add the conflicts/replaces to itself back in *-all [18:09] so no immediate deadline then. i should be able to test it tomorrow [18:09] the up/downgrade [18:10] it's still broken on -dev packages though [18:21] phoronix headline: AMD's New Catalyst Linux Driver Isn't Too Good [18:27] I don't think amd has to worry about catalyst drivers being too good [18:33] mlankhorst: Stupid fglrx makes it impossible to operate 3 monitors, while open-source gallium has no problem with it... [18:33] grmpf [18:38] tjaalton: forward switching should be ok though, still seem to hit some bug with multiarch [18:38] apt-get install xserver-xorg-lts-quantal libgl1-mesa-glx-lts-quantal:i386 [18:38] works around it [19:00] MCR1: that's not a issue. [19:00] Dandel: ? [19:00] MCR1: just add "Virtual 4096 4096" to the display subsection [19:00] where the other lines are "viewport 0 0" and "depth 24" [19:01] MCR1: use xrandr [19:01] I need more than 4096x4096 [19:01] it's so xrandr plays nicely with hotplugging [19:01] try higher, but my experience was not good [19:01] I need 1920+1280+1920 in x axis [19:02] and in that case fglrx fails with not enough vram error or something like that [19:02] which is ofc not true [19:02] care to give me the system information? [19:02] I know how to get the bug report to the right people [19:03] ATI HD 5750, Quantal, latest AMD fglrx beta [19:03] build id? [19:03] one moment [19:04] 9.01-121018a-148746E-ATI [19:04] Driver Packaging Version ^^ [19:05] k... care to get the output from? /usr/share/fglrx/atigetsysteminfo.sh [19:05] that will need a pastebin tho [19:05] it'll output to a text file ~/atisysteminfo-report.txt [19:07] grmpf, it is too long for pastebin.com [19:07] compress it and file a launchpad report with the file? [19:09] Dandel: Got it now (pastebin.ubuntu.com): http://pastebin.ubuntu.com/1377782/ [19:09] thanks [19:10] thanks to you - would be great to have 3 monitors running with fglrx also... [19:13] MCR1: i think i see the issue. [19:13] you need 3 seperate screens? or just one large one? [19:16] MCR1: that configuration is mangled. [19:16] 3 separate ones... combined to one big desktop, not mirrored [19:19] do you mind fixing it using xrandr? [19:21] MCR1, you have to admit that 3 monitors is really, really, really difficult for them [19:25] http://pastebin.ubuntu.com/1377815/ - try th is for a second. [19:25] don't expect immediate fix, because you will need to open up xrandr and move the screen around and disable mirror [19:26] for now, set the 1920x1200 at point (0,0), the 1280x1024 at point (0,1200), and the 1920x1080 at the point (1920,1080). [19:28] or if that's hard... two commands: "xrandr --output DFP2 --right-of DFP4" and "xrandr --output CRT2 --below DFP4" [19:28] if that gives you big desktop, but setup a little odd, it's ok... the next test is to change the virtual size once more. [19:29] change "Virtual 4096 4096" to "Virtual 6144 6144" then just move the crt and dfp to the positions you desire. [20:01] Dandel: Thanks a lot for your suggestions and help - I'll try to do that :) Great ! [20:02] Do you think it would be possible to make a 3-screen horizontal config also ? [20:02] yes. [20:03] just apply the change i mentioned about a half hour ago. [20:03] and then use xrandr to move displays around [20:05] it's mainly the use of the command: xrandr --output "Display to be effected" (--left-of, --right-of, --above, --bellow) "display to anchor by" [20:08] although, i also know that the gnome monitor preferences ( or mate monitor preferences ) should do a great job also. [20:09] since what this did is fix the built in desktop manager screen configuration utility on the system... what happens is the login screen will most likely get mirrored, but once you login, the setting will reapply. [20:15] ricotz: would it be possible to update libwaffle from 1.1.0 to 1.2.0? [20:23] Dandel: I am sorry for my slow responses, but I was attending a meeting on #ubuntu-desktop - Once again, thanks a lot 4 your help - It is highly appreciated. I will try and report success or failure to you in the next days... [20:27] bjsnider: Fact is that AMD/ATI advertise their cards for 3-monitor usage and it works with gallium... But maybe I get it to work with fglrx also, thanks to the help of Dandel... I hope so at least as I wanna watch movies on my TV ;) [20:33] bjsnider and MCR1, amd actually advertises support for up to 6 monitors on a single card, but many cards can only can handle up to 3 [20:33] You don' t get what 'up to' means then [20:43] Dandel, i guess i can do [20:43] 1.2.1 that is? [20:44] 1.2.1 works... it's mainly to get the latest fixes ( including the opengl es 3 support ) [20:45] \o/ OpenGL ES 3 [20:46] Dandel, pushed it to the ppa, but precise will probably fail [20:47] it should work just fine due to the fact that the last piglit worked fine. [20:47] (due missing cmake multiarch support in precise debhelper) [20:48] cmake in precise needs a blanket upgrade anyways [20:48] Dandel, is there any care for the oneiric/lucid builds in there? [20:48] the debian build using cmake is broken [20:48] it generates packages that can't be installed [20:48] i see [20:49] * ricotz meants to delete the oneiric/lucid packages there [20:50] oneiric and lucid, i don't care about... i have one machine using precise, the other using ocelot ( or w/e 12.10 is called ) [20:50] quantal ;) [20:50] actually, i just had to reinstall on the other machine because grub upgrade balked ><; [20:51] 10.04.1 ( lts) to 12.04.1 (lts) broke grub and i had to rework the install from scratch. [20:51] urgh [20:51] nothing like a grub 2 install without any modules :/ [20:52] MCR1: if fglrx crashes with the 8192x8192, it's because of driver bug ( it fails on the hd6520G i have ) [20:53] that's part of why i suggested 4096x4096 initially since that did not crash the test machine i use. [20:53] Dandel: I do not have the energy today anymore to try and fix if it fails, so I'll try tomorrow... but IIRC it crashed last time I tried 8192x8192 with some not enough videomem error [20:54] MCR1: your config was weird [20:55] yeah, something went wrong during last update I guess, but I did not look at it in detail before, because my 2-screen config was working perfectly - so I did not even know that it was bortched [20:57] but I could not change config by disabling one monitor and enabling the TV instead via AMDCCC anymore, I guess the weird config was causing this [20:57] so I still see a chance in your suggestions and will try to apply them ;) [20:58] Does Virtual always need the same x and y to work ? [20:58] Or is something like "8192 2048" also valid ? [20:58] should work [20:59] but it'd limit the screen orientation changes [20:59] I do not need to change orientation never, so I would not care [21:00] then it should be fine. [21:00] thx again - I'll promise to report the outcome ;) [21:01] afk [21:03] ricotz, pkg-config tool not found [21:03] on that waffle build [21:04] MCR1, why not use radeon instead of fglrx? [21:04] bjsnider, i know [21:12] bjsnider: Used gallium, was very satisfied and it was very stable, had no problems with 3 monitor-config, had a much smoother 2d-scrolling in chromium, no gfx-error with docky, but 3d performance was not comparable and unfortunately weak... I have a quite fast card with lots of fast DDR-5 VRAM and I expect it to max out Compiz framerate on 3 screens, which it wasn't capable of... [21:13] it's an exapnsion card? [21:13] bjsnider: ? [21:13] is the gpu built into the board or is it plugged into the mainboard [21:13] ATI HD 5750 PCI-E [21:13] plugged-in [21:14] i don't understand why you bought an amd card in the first place [21:14] 3 screens ? [21:14] why not nvidia? [21:14] not possible with nvidia 1 card [21:14] sure it isn't [21:16] and at the time this card offered a good value for the price and was 40 nm GPU tech [21:16] and DDR5, while Nvidia was still using DDR3 [21:17] other computer uses GTX-570-448 cores... [21:19] ricotz: there's a pull request for piglit packing that has had 1 approval, and just needs some review... mind looking at it? [21:31] Dandel, seems fine [21:32] the only other thing i can see is the missing opencl libs being an issue the stock repo for precise. ( i know there's a backport request for it