[04:51] bryceh, why does nvidia-current keep getting pulled on my intel box. it keeps ****ing up my libgl. [04:54] i believe nvidia-current is always installed by default. it is only activated if done manually through jockey [04:57] not true [04:57] I don't have it [04:58] maybe he has a metapackage or something [04:58] nvidia-common? [04:59] that's installed on every system, but it doesn't pull -current [05:00] LLStarks: purge it, it should say what depends on it [05:00] restricted modules? [05:00] apt-cache rdepends nvidia-current shows the deps [05:00] i do, but somehow it ends up back on my system once in a while. [05:00] one of those is the culprit [05:00] i don't know how to keep my libgl alternates from changing [05:01] well trace the deps from aptitude [05:01] well, how i do use the proper libgl without uninstalling nvidia-current? [05:01] uncheck it from jockey [05:03] i don't see why nvidia-current couldn't be installed on any system, as well as fglrx, as long as it isn't activated and isn't in the xorg.conf file, if there is one [05:04] but i thought that was supposed to be the policy in lucid [05:05] yes, it _can_ be installed and not used, but they are not installed by default [05:05] it's getting activated somehow and i forgot the update-alternatives command for libgl. [05:06] so then jockey prompts the user to not only activate it but to actually install it too [05:06] i don't use jockey driver manager [05:06] bjsnider: at least the first time, yes [05:08] is nouveau installed by default? [05:08] yes [05:08] oh my goodness that's funny [05:08] why? [05:08] nvidia is the preferred driver [05:08] by whom? [05:09] typical nvidia users [05:09] could be [05:09] ubuntu has never installed it (or fglrx) by default [05:09] and doing so wuld be a pr loss among the devs [05:10] would [05:10] installed but _not_ activated [05:10] would be a pr loss? [05:10] yes [05:10] but it would be easier on users [05:11] why? they still need to enable it [05:11] besides there is no space on the cd's [05:11] true [05:11] there would be on the dvd [05:11] and the network install [05:12] I don't think it's ever going to be installed by default [05:12] this boils down to a FOSS issue correct? [05:13] it's already made as easy as possible, no need to go further [05:13] lucid+1 will have nouveau 3D-support [05:14] so even less of a reason to push for the blob [05:33] speaking of nouveau, starting a second session fails when checking the drm, and then the first session breaks as well [05:34] it's like the vt's were on top of the X session, since the mouse cursor works and changes as if there were windows under it [15:57] Is there a reason xorg depends on x11-apps instead of simply recommending or suggesting it? [15:58] hysterical raisins i'd think [15:58] I doubt people really use xlogo, xbiff, xedit, xeyes, xgc & co any more [15:59] otoh what harm are they doing? [16:00] jcristau: They take up space on the LiveCD that doesn't need to be taken up :) [16:01] If someone has been long around to use xbiff or xman or xclock, they've been long around enough to apt-get install x11-apps [16:01] oh. i guess there are people who care about that.. === yofel_ is now known as yofel [21:29] @saravatt / bryceh / jcristau : Hi [21:30] I was told you could help regarding some problems Cairo-Dock has with some graphic cards [21:33] s/saravatt/Sarvatt [21:33] I'm sure: They watch TV (USA vs. Canada) ^^ [21:34] (but I don't say the current result :P ) [21:35] fabounet: It's probably a good idea to just state the problems cairo-dock has with some graphics cards; there are others who could help, too. [21:35] yep, I'm looking forward to solve this, if possible [21:36] actually it seems some drivers still have an incomplete support of OpenGL 2.0 [21:37] old ATI, but also some Intel cards [21:37] the problem seems to come from the lack of pbuffers support [21:37] and? [21:37] afaik pbuffers are deprecated [21:38] they are part of opengl though [21:38] and supported on most cards [21:38] not on the free drivers iirc [21:38] sometimes indirect rendering gives good results [21:39] do you mean pbuffers are banned from Intel drivers now ? [21:40] probably not. [21:41] but i still don't know what your problem is [21:41] the situation varies depending the card [21:41] for instance on nettbooks (GMA450), using indirect rendering makes everything works, but on not GMA800 [21:41] what's gma800? [21:42] maybe it was an other number, don't remember exactly. [21:42] some more evoluted model [21:42] had this bug recently [21:43] anyway all Intel cards don't behave the same [21:43] Isn't it pretty much always the client's responsibility to check for needed GL extensions & fallback if necessary? Can't you detect the presence of the necessary GL support & react appropriately? [21:43] I was expecting they would eventually have the same level of support for OpenGL2.0 [21:44] i915 can't go above 1.4 or 1.5 according to wikipedia [21:44] well I do detect the presence of pbuffers. [21:44] bt soetimes indirect rendering is needed to make them work [21:44] and sometimes it's the opposite (Gallium ATi R3xx drivers) [21:45] i965 can do 2.0 or 2.1 [21:45] seems there is no definite method to be sure [21:45] I do recall compiz wanting a GLX call to determine what GL extensions a particular context supports, so it could discriminate between extensions supported with a direct context & those only supported with an indirect context. [21:46] that would be helpful [21:47] currently I check that GLX is >= 1.3 [21:47] if not, then it's unclear [21:47] anyway from what i understood (i know nothing of GL) the recommended method is to use fbos instead of pbuffers. [21:47] do you remember the way to be sure whether indirect rendering would be needed ? [21:48] That's my (somewhat second hand) understanding, too. [21:48] fbo is younger than pbuffer, is it more supported though ? [21:49] #dri-devel might be a better place for this, it has people who know gl and write those drivers.. [21:50] ok, well I'll go and ask them then. [21:50] thanks for your advices ! [22:03] seems like i didn't misremember too much :) [22:04] yep, seems I'll have to rewrite some part with fbo ^^ [22:43] What's meant to be shipping drm.h at this point? I can't build nouveau because I can't find drm.h :) [22:46] linux-libc-dev [22:49] which interestingly also carries driver specific headers. [22:51] Ah. Why is linux-libc-dev not providing drm.h for me? [22:52] if you're using edgers its libdrm-dev and you have to reinstall it every time theres a linux-libc-dev update [22:53] hope jcristau's libdrm header patch goes upstream :) [22:53] Hm. Maybe that did'nt get replaced when I un-edgered. [22:53] yeah i have a note on the page saying you need to reinstall linux-libc-dev when you switch away from edgers [23:01] hmm looks like xserver-xorg-video-intel 2.10.901 requires >= xserver 1.7 now? http://launchpadlibrarian.net/39848281/buildlog_ubuntu-karmic-i386.xserver-xorg-video-intel_2:2.10.901%2Bgit20100226.a0ee9c3d-0ubuntu0sarvatt~karmic_FAILEDTOBUILD.txt.gz