=== ripps|sleep is now known as ripps [11:47] tseliot, ping [11:48] ara: pong [11:49] tseliot, hey!, when I try to activate compiz it gives me an error "desktop effects could not be activated", but I don't get any errors in Xorg.0.log or syslog [11:49] tseliot, any ideas where to look? [11:50] ara: what does "glxinfo | grep direct" say? [11:50] tseliot, Error: couldn't find RGB GLX visual or fbconfig [11:51] ara: I need the output of "update-alternatives display gl_conf" and your /var/log/Xorg.0.log [11:52] tseliot, alright [11:53] tseliot, http://paste.ubuntu.com/352857/ [11:53] tseliot, http://paste.ubuntu.com/352859/ [11:55] ara: did it work when you first tried my packages? [11:56] "it" being compiz [11:56] tseliot, no, I just tried it today [11:56] s/no/I hadn't tried before :) [11:58] ara: what do "lsmod" and "dmesg" say? [11:59] the log looks fine... [11:59] tseliot, http://paste.ubuntu.com/352861/ [12:00] tseliot, http://paste.ubuntu.com/352863/ [12:01] the right kernel module seems to be there too [12:03] ara: let's see your xorg.conf [12:04] tseliot, http://paste.ubuntu.com/352866/ [12:04] ara: and the output of this command: ldd /usr/bin/glxinfo [12:05] tseliot, http://paste.ubuntu.com/352867/ [12:12] ara: and the output of this command: cat /etc/ld.so.conf.d/GL.conf [12:12] if the file is empty then I know what to fix [12:14] also see if /usr/lib/nvidia-current/ld.so.conf exists [12:15] ara@sushirider:/var/log$ cat /usr/lib/nvidia-current/ld.so.conf [12:15] /usr/lib/nvidia-current [12:15] ara@sushirider:/var/log$ cat /etc/ld.so.conf.d/GL.conf [12:15] /usr/lib/nvidia-current [12:18] that's correct too [12:18] hmm... [12:19] ara: does nvidia-settings work? [12:19] or does it only show a limited number of options [12:19] ? [12:20] tseliot, it works [12:20] tseliot, but "Fail to query the GLX server vendor." [12:21] * bryyce waves [12:22] ara: ok, let's see if glx exists and what it points to (let's pretend that the X log lied) [12:22] bryyce: hey [12:22] hey bryyce [12:22] NFI why I'm awake at 4:30am [12:24] ara: which means that you should check "ls -l /usr/lib/nvidia-current/xorg/libglx.so", "ls -l /etc/alternatives/libglx.so" and "ls -l /etc/alternatives/libglx.so" [12:24] bryyce: :-) [12:26] what about just running glxinfo [12:27] tseliot, those 3 are there [12:27] (with 777 permissions) [12:27] tseliot, http://paste.ubuntu.com/352880/ [12:27] the 2nd and the 3rd are the same [12:28] I meant "ls -l /usr/lib/xorg/modules/extensions/libglx.so" [12:28] * tseliot blames it on the clipboard [12:31] * ara blames Ctrl+C/Ctrl+V [12:31] tseliot, same thing [12:31] ara@sushirider:/var/log$ ls -l /usr/lib/xorg/modules/extensions/libglx.so [12:31] lrwxrwxrwx 1 root root 27 2010-01-05 14:44 /usr/lib/xorg/modules/extensions/libglx.so -> /etc/alternatives/libglx.so [12:33] ara: ok, I must be something else then. I'll investigate the problem and let you know [12:33] thanks for reporting the problem [12:34] ara: what about the output of glxinfo? [12:34] * tseliot is having a hard time getting Lucid to work on his testing box [12:39] * tseliot -> lunch [12:39] you need more boxes ;-) [12:39] * bryyce -> bed [12:40] hmm, compiz-wrapper is no more, and compiz-core doesn't depend on mesa-utils [12:41] ara: install mesa-utils and try again [12:41] but it still uses glxinfo? i thought that would be gone and it'd do the GL calls itself? [12:41] maybe [12:41] compiz works here anyway :) [12:42] but glxinfo is in mesa-utils [12:42] or just running 'compiz --replace' should show the error [12:44] ara@sushirider:/var/log$ compiz --replace [12:44] compiz (core) - Fatal: Root visual is not a GL visual [12:44] compiz (core) - Error: Failed to manage screen: 0 [12:44] compiz (core) - Fatal: No manageable screens found on display :0.0 [12:44] Launching fallback window manager [12:44] glxinfo and xdpyinfo plz? [12:47] ara@sushirider:/var/log$ glxinfo [12:47] name of display: :0.0 [12:47] Error: couldn't find RGB GLX visual or fbconfig [12:47] http://paste.ubuntu.com/352892/ [12:49] weird. [12:51] ara: You have an nvidia card? [12:51] soren, yes [12:52] ara: And it actually works for you? I thought all nvidia users were screwed by bug #494627. [12:53] Launchpad bug 494627 in xserver-xorg-video-nv "nv driver crashing with segmentation fault in libpthread.so.0" [High,In progress] https://launchpad.net/bugs/494627 [12:53] soren, I am using tseliot's PPA, I am reporting back to him, to try to upload it to Lucid asap :) [12:54] Aha! [12:54] :) [12:55] Perhaps I should do the same. [12:56] Which PPA is this? [12:56] tseliot: You have a whole bunch :) [12:57] proprietary-video-improvements? [12:57] That one seems to have recent stuff in it. [12:59] that's the one [12:59] hrm. the backtrace in 440323 doesn't have libGL symbols… [13:00] also the ext_name is weird [13:00] #6 0x00007fa4bc76e7b2 in XextAddDisplay (extinfo=0x17153c0, dpy=0xf1f340, [13:00] ext_name=0x7fa4ae23a550 "ueryObjectuiv", hooks=0x7fa4ae35c2a0, [13:00] nevents=17, data=0x0) at ../../src/extutil.c:112 [13:00] dpyinfo = (XExtDisplayInfo *) 0xec0010 [13:00] "ueryObjectuiv" :) [13:03] and no qt symbols in 419501. bleh. [13:08] and 278261 doesn't look like a libxcb bug. [13:08] a new bug triager \o/ :) [13:09] i was looking whether your sync request had been processed, and there were 4 bugs in https://launchpad.net/ubuntu/+source/libxcb/+bugs so i figured i'd look :) [13:09] hehe [13:10] not going to look at the 292 xorg-server ones ;) [13:10] 334264 should probably be closed as fixed in 1.7.x though [13:10] it's rather depressing going through them [13:15] (only looked at the title for the Xdmx bug, didn't look at the details) [13:15] I'll ask the reporter to test on a lucid livecd [13:20] bah 435686 has no symbols for the server [13:20] weird though, caught a signal in EvdevOn [13:21] might be the bug fixed in evdev 2.2.6 [13:29] yep, asked to verify [14:02] bryyce: btw, what happens when the log time counter hits 99999.999999? :) [14:36] I'm trying Lucid on my laptop with nvidia 8600M GT. Isn't nouveau supposed to be automatically used? I can't even install it: [14:36] xserver-xorg-video-nouveau: Depends: xserver-xorg-core (>= 2:1.6.2) but it is not going to be installed [14:36] E: Broken packages [14:36] Any clues? [14:37] it hasn't been updated [14:37] and no, nouveau is not used by default [14:38] tjaalton: so the only solution for nvidia on Lucid for the moment is vesa, right? [14:38] beginning to doubt it will (for lucid= [14:38] today, yes [14:38] Thank you [14:39] tjaalton, the .33 kernel nouveau driver would not be backported to the lucid kernel? [14:39] bjsnider: nothing happened so far [14:39] visible anyway [14:39] and after linus worked so hard to bully red hat into getting it in there [14:40] bjsnider: so if I install .33 I'll get working nouveau drivers? (with all the instability, of course, I don't mind...) [14:40] alkisg_: no [14:40] the ddx will still fail to install [14:40] you'll still need the userspace part [14:41] May I ask a last question? I tried passing "xforcevesa" as a kernel parameter, and it only seemed to work some of the times, I didn't get when exactly (was it only on the live CD?) [14:41] Is there any way to specify that I want to use vesa without creating a xorg.conf? [14:42] rm xorg.conf [14:42] it should fall back to vesa on its onw [14:42] ^t [14:42] though it'll use nv and that'll crash [14:42] ah [14:42] then uninstall nv, or create xorg.conf [14:42] i thought mvo had fixed nv btw? [14:42] or maybe not, but at least the colors are wrong. that's what mvo's patch would fix [14:43] ack [14:43] it's not yet in the server [14:44] Yeah nv had some weird green/black colors [14:44] Not usable at all [14:44] that is a side-effect of my initial crash fix, it fixes the crash, but that is not helpful (because of the colors) [14:46] jcristau: libxcb finally synced (manually) [14:46] are you guys going for intel 2.10? [14:48] probably [14:49] we need to install some headers from libdrm again [14:49] given the amount of complains about the funny colors I guess I should upload a revert -nv and a patched xserver this evening - unless there is feedback on ML that says otherwise- what do you think? [14:49] erm, add to libdrm-dev, and Replace linux-libc-dev [14:50] mvo: I can add it to the server [14:50] and merge it too [14:50] with the new version in sid [14:50] sounds good [14:50] but first I'd like to know why I always get failsafe (which doesn't work with kms) [14:51] i haven't actually tested what i uploaded to sid, i'm running 1.7.99.3-ish atm :/ [14:52] hehe [14:52] I believe it works [14:52] will test too [14:52] i had been running 1.7.3.901 for a while before though [14:53] * mvo will upload -nv with the reverted patch in a bit [14:55] * sebner waves [14:55] sebner: what does "update-alternatives --display gl_conf" say? [14:56] tseliot: http://pastebin.com/m1ae40c9 [14:57] sebner: try "sudo sh /bin/nvidia-bug-report.sh" and show me the log that it generates [14:59] tseliot: http://pastebin.com/m7032eae5 [15:01] sebner: try with "sudo apt-get install --reinstall nvidia-current" (as it's pointing to the wrong glx file) [15:01] tseliot: then restart right? [15:02] sebner: no, maybe just show me the output of "ls -l /etc/alternatives/libglx.so" [15:02] and then restart [15:02] tseliot: lrwxrwxrwx 1 root root 38 2010-01-07 15:42 /etc/alternatives/libglx.so -> /usr/lib/nvidia-current/xorg/libglx.so [15:03] [ 0.123515] (II) Loading /usr/lib/xorg/modules/extensions/standard/libglx.so [15:03] sebner: ok, now restart and then do a "sudo sh /bin/nvidia-bug-report.sh" again and give me the log [15:03] aye [15:03] Sarvatt: yes, that is why I asked him to reinstall the package [15:03] you've got the same problem I do sebner, moving /usr/lib/xorg/modules/extensions/standard/libglx.so away should fix it [15:04] he's having the same problem sarvatt was [15:04] heh [15:04] oh, he just said that [15:04] i reinstalled it 4 times, no luck without manually moving the .so out of the way :( [15:05] wait, i thought aptitude fixed it? [15:05] your last message said aptitude fixed the problem [15:05] not here [15:05] Sarvatt: it means that X looks for libglx.so in subdirectories and ignores the link in /usr/lib/xorg/modules/extensions/ [15:05] yeah thats what I got from it too, and couldn't figure out why [15:06] jcristau: does this make sense to you ^^ ? [15:06] tseliot: http://pastebin.com/m46823125 [15:06] or, is it a feature? ;) [15:06] it prefers the actual file even though its in a subdirectory over the link in the main directory [15:06] * sebner uses apt-get :P [15:06] right [15:07] ok, let me fix that [15:07] tseliot: dunno. should all be in hw/xfree86/loader/ somewhere. [15:07] ok [15:07] Sarvatt, i thought it was just that the alternatives hadn't taken effect [15:08] reinstalling fixed the alternatives but that didnt fix the problem [15:08] had to move the .so out of the way [15:09] but the .so in the /standard/ path is a link [15:09] nope thats the actual one that used to be in /usr/lib/xorg/modules/extensions/ before [15:10] hi. resuming after suspend leads me to a gray screen (this started happening in lucid pre-alpha. it was fine before. i'm using nvidia). if i have Rhythmbox playing before suspending, when resuming it continues playing.. i can only go to a tty after doing a Alt+SysRq+R, but always with a gray screen.. [15:10] Sarvatt, are you sure about that? [15:10] yeah 100% [15:11] libdri.so and libglx.so were just moved there in the xserver packaging [15:11] work calls :( [15:13] if it helps, Xorg.0.log: http://pastebin.com/d2e13acca [15:13] and pm-suspend.log: http://pastebin.com/d2b126605 [15:14] afv: power management with blobs is like gambling [15:15] blobs? [15:15] nvidia [15:15] ah [15:15] binary blob [15:15] * tseliot nods [15:20] someone in another channel was complaining that his ati card under the radeon driver was constantly at full throttle while under fglrx it was dynamic, usually low power mode [15:21] radeon doesn't support that yet [15:23] tjaalton: it does but it's highly experimental [15:23] ;) [15:23] right [15:25] I'll experiment with this when I have to time to create a radeon package with dkms [15:29] hum... "bonobo-activation-server (afv-6617): could not associate with desktop session: Failed to connect to socket /tmp/dbus-zUBPlHkmdZ: Connection refused" [15:32] forget it [15:39] i find nothing special at syslog... http://pastebin.com/d216d2830 [15:41] what I tried to say is that there's probably nothing we can do about it [15:41] isn't there a way i can debug it better? [15:43] or.. try with another driver, for example? [15:44] currently there is no other driver that works even that much [15:44] hmm, ok [15:44] thanks then [15:44] brb [16:02] tseliot: something went really wrong: sebner@ubuntu:~$ glxgears [16:02] Error: couldn't get an RGB, Double-buffered visual [16:03] sebner: what does "sudo sh /bin/nvidia-bug-report.sh" say? [16:04] tseliot: don't always forget poor /usr/ :P http://pastebin.com/m28c9838a [16:05] sebner: ok, it looks like the right glx is loaded now but dri is not: (EE) Failed to load module "dri" (module does not exist, 0) [16:05] grrrrr [16:05] sebner: which means that I should fix that in nvidia [16:06] heh [16:06] as nvidia still points to the old path [16:06] I was right [16:06] nvidia rebuild :P [16:06] no more X recompilation should be required [16:07] heh [16:07] * sebner happy [16:33] tseliot: baaaaad news, http://pastebin.com/m466a5e34 [16:33] haha, now glx fails again [16:34] tseliot: maybe reinstall xserver to pick new one up? [16:34] sebner: it doesn't hurt to try [16:38] tseliot: I'm sorry, still b0rken [16:39] sebner: ls -l /etc/alternatives/libglx.so [16:39] tseliot: lrwxrwxrwx 1 root root 31 2010-01-07 17:28 /etc/alternatives/libglx.so -> /usr/lib/standard-x11/libglx.so [16:40] sebner: update-alternatives --display gl_conf [16:40] tseliot: http://pastebin.com/m2e715488 [16:41] * tseliot screwed up again [16:41] heh [16:41] I mean, let me check ;) [16:42] ok, I overdid [16:42] haha [16:43] switching between desktops with the same screen (with different brightness) is killing my eyes [16:44] heh [16:45] let me fix that [16:50] :) [16:56] sebner: ~proprietaryppa22 should do the right thing [16:56] tseliot: I'll try :) [16:56] thanks [16:59] back. purged nvidia and installed nouveau [16:59] suspend is working :p [17:01] yes, that usually works ;) [17:01] afv: are you using Lucid? [17:02] kees: how's the libaudit main promotion doing? :) [17:04] superm1: speaking of mirs, any progress on the one for libvdpau? [17:04] tjaalton: well... I'd like to get it derooted. [17:05] tjaalton: which requires a fair bit of work/testing. [17:05] tjaalton: at present, it's a postponed workitem for Lucid. what needs it? [17:05] kees: we've disabled selinux from xserver due to libaudit being in universe [17:07] kees: so, it's just a diff to debian [17:09] tjaalton: ok, I'll try to get it done, but time isn't looking great. [17:10] kees: ok [17:19] tseliot: here we go again: http://pastebin.com/m13d1ad5a [17:21] sebner: ls -l /etc/alternatives/libglx.so [17:22] tseliot: lrwxrwxrwx 1 root root 38 2010-01-07 18:12 /etc/alternatives/libglx.so -> /usr/lib/nvidia-current/xorg/libglx.so [17:22] sebner: it should work now [17:22] does it? [17:22] tseliot: nope [17:22] tseliot: Error: couldn't get an RGB, Double-buffered visual [17:30] ERROR: Error while querying attribute 'GLXServerVersion' on ubuntu:0.0 [17:30] (Unknown Error). [17:37] tseliot: that means? [17:38] sebner: I think it's still using the wrong libGL libraries [17:39] sebner: cat /etc/ld.so.conf.d/GL.conf [17:40] tseliot: /usr/lib/nvidia-current [17:40] which is correct [17:41] hmmm [17:41] weird [17:41] more weirdness [17:43] sebner: maybe do a "sudo ldconfig" [17:43] and reboot [17:43] * sebner tries [17:49] tseliot: still the same issue :( [17:54] sebner: maybe try to move /usr/lib/libGL.so.1.2 away, then ldconfig and reboot [17:54] this shouldn't happen though [17:54] tseliot: away means .backup is ok? [17:55] yes, of course [17:56] tseliot: seems something is b0rken. still not working [17:56] This was my 32nd reboot today btw ;) [17:57] .backup means it's still visible for ldconfig [17:57] oh, I misread your sentence then [17:57] I missed the dot [17:57] ohhh [17:57] jcristau: thanks for spotting that. My eyes are way too tired [17:57] now it's /usr/lib/foobar [17:57] ^^ [17:58] tseliot: not only yours ^^ [17:58] sebner: that should be ok [17:58] now sudo ldconfig and reboot [17:59] maybe ldconfig got lazy or maybe (more likely) I'm doing something wrong in the packaging [18:00] Number 33 brought me luck [18:00] *hehehe* [18:00] working now [18:04] sebner: ok, what if you put the library back and type ldconfig -n /usr/lib/nvidia-current [18:04] tseliot: I'll test later. Now I'm giving my harddrive a break ;) [18:05] sebner: ok, no hurry [18:35] wb tseliot [18:35] :) [18:36] tseliot: is there a difference between ldconfig and sudo ldconfig? No output at all with both [18:36] sebner: it should be done with sudo [18:36] k [18:36] did you try the command that I suggested before? [18:36] sudo ldconfig -n /usr/lib/nvidia-current [18:36] tseliot: Guess why I asked :P I haven't rebooted yet though [18:37] sebner: aah [18:37] ok [18:39] tseliot: nvm, rebooting now [18:40] ok [18:41] tseliot: working [18:41] sebner: ldd /usr/bin/glxinfo [18:42] tseliot: http://pastebin.com/m646d3d0d [18:43] at this point I think that if something else calls ldconfig it can break things unless I move those libraries to a separate directory [18:43] sebner: ok, it seems to be ok now [18:44] jcristau: opinions on this? [18:44] hmm? [18:44] tseliot: :D, great thanks! Though I would have ~10 reboots instead of ~40 if I had moved the file at the beginning ^^ [18:45] jcristau: I was thinking of moving libGL*, etc. to a different directory [18:45] as currently I have to do ldconfig -n /path/to/the/right/libraries [18:46] even though the path is set in /etc/ld.so.conf.d/GL.conf [18:46] as ldconfig seems to look in /usr/lib first [18:47] and then in whatever path we put in /etc/ld.so.conf.d/GL.conf [18:48] sebner: oh, well, you did it for a good cause ;) [18:50] tseliot: new harddrive then plz :P [18:50] heh [20:10] hrm, tseliot didn't add anything to the changelog about moving libglx/libdri === verbalshadow_ is now known as verbalshadow