[08:39] wgrant: ok, I'll do that [09:34] * Ng boggles, this morning my scrollwheel emulation came back when i re-ran the xinput stuff after suspend [09:34] I tested multiple suspend cycles and it was definitely not doing that! [10:12] hey there [10:12] seb128: my mouse wheel doesn't work anymore... is it a known thing? I cannot find anything relevant on launchpad [10:12] is there some known issue on the topic since the xinput changes? [10:13] seb128: shouldn't be [10:13] hum ok [10:15] xinput shouldn't be doing much mouse wheel stuff... [10:18] tjaalton: How would you feel about applying a syndaemon patch of mine to make it work without SHMConfig? I've had positive reports from others, and it will still work with SHMConfig if properties aren't available or -S is given. [10:20] wgrant: sure [10:20] * wgrant locates a diff. [10:20] Is out -synaptics in git somewhere? [10:20] s/out/our/ [10:21] yes, git.debian.org [10:21] I also think we need to revert that two-finger-scrolling disentanglement patch, as it certainly doesn't click as easily as it used to. [10:21] ubuntu-branch [10:21] Hmm, couldn't see it there. [10:21] hmm, I'll check [10:21] The only branch I could see was debian-unstable... or is it in another location? [10:21] it is there for sure [10:22] Not branch in the normal git sense? [10:22] you cloned it? [10:22] origin/ubuntu [10:22] wgrant: 'git branch -r' lists remote branches [10:22] * tjaalton is hung over.. [10:22] Haha. [10:23] not the best day to work [10:23] :P [10:23] Probably not, no. [10:24] jcristau: On which URL should I be trying that? [10:24] * wgrant has very limited git skills. [10:24] in your local repo [10:25] Oh. [10:25] I see. [10:25] So clone only grabbed the default branch? [10:25] Right, there we go. [10:26] clone grabbed all branches, but only created a local branch for the default [10:27] 'git checkout -b ubuntu origin/ubuntu' to create a local ubuntu branch [10:27] That worked. [10:27] Thanks. [10:27] I often wonder if some features of git were designed to be completely counter-intuitive. [10:54] Am I meant to use git-buildpackage to build it, or do you X people use some other setup? [10:55] dpkg-buildpackage works [10:55] Won't that leave a .git? [10:56] not with -i [10:56] True. [10:56] is it normal that the nvidia driver from jockey works. and the driver from nvidia.com has huge issues? [10:57] the kwin effects are all totally broken:( [10:58] I guess there are some special patches ubuntu is using? [11:22] elmargol: define what you mean by "works" [11:22] and specify the version of the driver you're using, please [11:24] tseliot: 177.80 from nvidia.com does not work. [11:24] 177.80 from ubuntu does work [11:24] If I enable Desktop effects. The window manager does not work anymore. (I can not switch properly between windows) [11:25] Show desktop etc. does not work too [11:25] elmargol: please don't use the driver from nvidia.com or you might break some symlinks [11:25] Some problem using 177:76 [11:25] tseliot: well If I use the ubuntu drivers I get a hard freeze [11:26] thats the bug I'm complaining for weeks now [11:26] elmargol: it's the same driver [11:27] tjaalton: Any changes that you want for -synaptics? [11:27] elmargol: therefore I think that you're not getting the hard freeze with the one from nvidia.com simply because that driver is not loaded [11:27] what else to do.. noone from nvidia/ubuntu cares. experimenting is the only thing left to do i think [11:28] elmargol: what's the bug report again? [11:28] bug #270617 [11:28] Launchpad bug 270617 in nvidia-graphics-drivers-177 "Display corruption after short period" [Undecided,Confirmed] https://launchpad.net/bugs/270617 [11:30] next think I try is reverting to a old kernel. if this does not help :( i have to downgrade to hardy [11:32] elmargol: the log doesn't show the problem. Maybe it's a bug in the nvidia driver that affects kernel 2.6.27 and it's triggered by some module which is specific to your hardware [11:33] nothing that we can fix [11:33] did you try updating the BIOS? [11:33] This is a dell inspiron 9400 there have to be 1000 users out there [11:34] tseliot: it worked just fine using hardy and vista [11:35] elmargol: yes, you mentioned that [11:36] i check if there is a newver bios [11:43] elmargol: if that doesn't work you can try to underclock your card: http://www.phoronix.com/scan.php?page=article&item=197&num=1 [11:44] wgrant: nope [11:45] tjaalton: OK, see http://people.ubuntuwire.com/~fujitsu/git/xserver-xorg-input-synaptics [11:48] Hmm, there's a bug reference I should have in there. [11:48] wgrant: what should I be looking at?-) [11:48] tjaalton: The ubuntu branch there. It has some changes. [11:57] wgrant: ok, pulled and merged [11:58] tjaalton: Thanks. [12:05] eww. somebody made shmconfig default to true? [12:05] jcristau: yeah, for some model [12:05] jcristau: Yes... [12:05] In the fdi file. [12:06] I'm considering reverting all of the changes there except the edges and maxtapmove, but I presume there's a reason for them. [12:06] hm, should I be able to build the source package even though I'm not the one who touched the changelog? [12:06] tjaalton: Yes, with -kyourkeyid [12:06] wgrant: cool [12:07] that was.. obvious [12:07] wgrant: yeah, i don't think the fdi should set any options, really. if the defaults are wrong, that should be fixed in the driver itself imo [12:07] but, shmconfig on by default is.. weird. [12:07] I think those were dropped upstream [12:07] jcristau: They are largely now, but this particular model had strange dimensions. The driver autodetects it now. [12:08] They were. [12:08] ok. cool. [12:09] tjaalton: Did you mean to leave the xinput, -synaptics and -evdev tasks of the property API update bug open? [12:11] tjaalton: Thanks. [12:13] jcristau: Did you end up looking into that GetDeviceProperty issue any further? [12:13] wgrant: no [12:15] OK. [12:20] wgrant: nope, will close now [12:23] tjaalton: Great. [13:04] ok have a new bios version now [13:04] lets see if this helps [13:12] nothing about this in the bios changelog... [13:29] elmargol: of course you won't find anything relevant in the changelog [13:31] 30 min uptime now [13:31] while true; do kate&kate&kate&kate&kate&kate&kate&kate&sleep 5;killall kate; done [13:31] :D [13:33] stresstest [13:36] elmargol: if the problem is solved, please add the solution to the bug report [13:37] tseliot: i crashed after 2 hours once :/ [13:37] I have to try this 1-2 days to be sure [13:37] elmargol: yes, of course [13:40] tjaalton: http://picasaweb.google.co.uk/cmsjtenshu/Misc#5256616902573945746 - ever seen something like that on boot? [13:40] Ng: I believe so, yes [13:41] yes thats wenn your nvidia card is broken [13:41] at least my was gone after this :( [13:41] the X300 has intel gfx [13:41] no idea then [13:41] I don't even think it's X that's doing it (I just asked tjaalton because his laptop has similar guts to mine) [13:44] heh, just got that screen [13:44] it seems fairly early in the boot to me [13:44] I wonder if it's the kernel crashing [13:45] I've had it once or twice when booting without usplash and I didn't see an oops on the console :/ [14:00] tjaalton: did you apply the patches which are mentioned here? https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/146859 [14:00] Launchpad bug 146859 in mesa "No dri for virtual screen greater than 2048x2048" [High,In progress] [14:03] nope [14:05] tjaalton: too bad. Those patches should get around the current limitations of the intel driver [14:05] for some models [14:11] ok crashed again [14:27] elmargol: if it has to do with overheating then you might try what I suggested ^^ [14:40] tseliot, I can play an hour quake4 on this machine (vista) and have no issues [14:41] elmargol: ok but you're using a different kernel and a different driver [14:54] while I'm throwing problems around all over the place, Compose doesn't seem to be working since I upgraded to Intrepid [14:54] I don't use it very often, so I can't say exactly when it broke [14:55] (I have the gnome keyboard capplet set to make Alt-GR a compose key) [14:57] what keysym does xev report when you press that key? [14:58] state 0x0, keycode 108 (keysym 0xff20, Multi_key), same_screen YES, [14:59] so that part works [14:59] Ng, i've been seeing that a lot [14:59] Ng, on a variety of Dell boxes. it only happens on intrepid usplash [15:00] superm1: I get it on a seemingly random basis with or without usplash [15:00] Ng, have you caught a kernel trace then without the usplash? [15:00] superm1: no, there doesn't seem to be one :/ [15:01] Ng, well when it freezes up like that though, you would think you would have one available :( [15:01] superm1: unless it hangs so deep inside the kernel that it can't trigger the bits that write an oops [15:01] also even if it did, the display corruption would make it impossible to read [15:01] Ng, does your thinkpad have a serial port then perhaps? [15:02] all of my dell boxes exhibiting it are serial free [15:02] superm1: nope :/ [15:02] I have vga and ethernet and nothing else ;) [15:02] Ng, have you brought this to the attention of the kernel team yet? [15:02] yeah [15:02] what'd they say? [15:02] that they'd like me to try and get more information [15:02] I'm going to disable usplash tonight and do a few boot loops to trigger it and hopefully be able to tell from the timing, where it happens [15:03] I filed bug 282700 but I suspect it may end up being a duplicate of bug 263782 [15:03] Launchpad bug 282700 in linux "[Intrepid] occasional hangs with screen corruption while booting" [Undecided,New] https://launchpad.net/bugs/282700 [15:03] Launchpad bug 263782 in linux "intrepid hang with screen corruption during boot with 2.6.27-2-generic on x61" [High,Triaged] https://launchpad.net/bugs/263782 [15:03] Ng, I didn't realize it happened without usplash ever, so I'll be glad to set up a reboot loop today too then on at least one of the boxes that exhibited the behavior [15:04] that would be handy :) [15:05] I've a similar crash on boot often too [15:06] seb128: what kinda hardware? [15:06] mine is all-intel [15:06] dell D630 [15:06] which is basically all intel too [15:09] hum come to think of it, the other platforms i've seen this, I think they were all intel, but i'd have to double check one of them [15:10] Ng, http://picasaweb.google.co.uk/cmsjtenshu/Misc#5256006922907903138 mmm btw :) [15:11] I'm on a weird baking trip at the moment ;D [15:29] superm1: I'm dealing with the bug about nvidia-xconfig and nvidia-glx-173. Do you think it would be ok if I added python-xkit to "Recommends" of nvidia-glx-173? [15:30] tseliot, I would think it should be depends [15:30] and do I need a FFE for this? [15:30] tseliot, since you can't use nvidia-xconfig without it [15:30] ok [15:31] FFE or not? [15:31] it wouldn't hurt just to be safe. [15:31] it is indeed a new feature [15:31] i'd attach the changes to one of the existing bugs talking about the issues [15:32] and subscribe ubuntu-release to that [15:33] ok, thanks [16:17] Ng, ah I have reproduced it on a d630 as well now without usplash turned on. [16:17] superm1: and you get a garbled mess with the characters flickering? [16:18] Ng, yeah [16:18] superm1: out of interest, did the machine spontaneously reboot/powercycle after a few seconds? [16:18] Ng, hard drive light flickering when it happens, nothing useful in syslog post reboot. [16:19] Ng, no, I did that myself after a few seconds though [16:19] I was trying to take a picture and mine power cycled itself [16:19] I don't know a picture would have been that helpful, couldn't make much sense of everything on the screen [16:21] yeah :/ [16:21] wish I had a video camera, but I think I'm going to have to try and guess what it is from the timings [16:24] I would have thought it was S90console-scren.kbd.sh, but that's a little too late in the boot process. this happens far earlier [16:24] yeah [16:24] oh that's still in rcS.d [16:25] superm1: hmm, so one clue would be if it's before or after the font changes [16:25] perhaps settings the console font itself then. you tried removing console-setup from your boot and see if it goes away? [16:28] not yet, but I suspect that is a prime candidate [18:11] superm1: could you maybe confirm https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/282700 and add your hardware details? [18:11] Launchpad bug 282700 in linux "[Intrepid] occasional hangs with screen corruption while booting" [Undecided,New] [18:12] I think it's pretty clear we're talking about the same thing and that others are seeing it too [18:14] hi, I am trying to build intel in a PPA, but it looks for /usr/lib/libdrm_intel.so - what gives? [18:14] tormod: you have libdrm from git? [18:15] jcristau: not really, it's from git20080926 [18:15] libdrm_intel was probably introduced later than that [18:16] is there a way to easily disable it? I need to check an upstream 2D fix. [18:16] no [18:16] hmm the change was made on 20080910 [18:17] I see in the libdrm build log it was built... [18:18] maybe it was built and then not added to the package :) [18:23] maybe the file is there after all, the build fails with: [18:23] dpkg-shlibdeps: failure: no dependency information found for /usr/lib/libdrm_intel.so.1 [18:24] missing shlibs file [18:25] what is an shlibs file? what would the name be? [18:26] also i think libdrm_intel should be in its own package, so i (or chris) will update drm-snapshot packaging soonish [18:29] should the shlibs file be in the libdrm-dev package? [18:29] no [18:29] it's part of the metadata of the package containing the lib [18:30] metadata like in /var/lib/dpkg stuff? [18:30] yes [18:31] see dh_makeshlibs [18:33] I have dh_makeshlibs in debian/rules though [18:34] but libdrm2.shlibs doesn't say anything about libdrm_intel.so.1 [18:37] right. but that file is autogenerated if I understand man dh_makeshlibs [18:38] no. maybe it should be, but right now it isn't [18:40] doesn't "dh_makeshlibs" create the shlibs file? or does it just create contents and you have to copy it in? [18:44] it creates debian/$package/DEBIAN/shlibs [18:45] but then that's overwritten by dh_installdeb copying debian/$package.shlibs afaict :) [18:46] I think I understand. So I have to edit debian/$package.shlibs. what's the syntax for adding drm_intel? [18:47] 'libdrm_intel 1 libdrm-intel1' and then add the package in debian/control, fix debian/*.install, ... [18:47] thanks [18:47] or 'libdrm_intel 1 libdrm2 (>= foo)' [18:48] add which package? [18:49] libdrm-intel1 [18:51] aha the clean way of making a new package I suppose. but if I just want to hack the old packages, that would be your second option 'libdrm_intel 1 libdrm2 (>= foo)' ? [18:51] yup [18:51] replace foo with the actual version number, obviously :) [18:52] yeah I got that, but I hope the rest was rather verbatim :) [18:53] yeah. the format is '$libname $somajor $dependency' [18:54] in this case the soname is libdrm_intel.so.1 -> shlibs contains 'libdrm_intel 1 $whatever' [18:55] $whatever should be >= my libdrm version? [18:57] $whatever is what goes in the Depends field of anything linked against libdrm_intel (through ${shlibs:Depends}) [18:57] so in this case 'libdrm2 (>= yourversion)' [20:02] jcristau: thanks a lot, it built successfully. now to see if it runs... [21:01] Hi, which package should I file a bug about my capslock light on my laptop not turning on when enabled by an external keyboard [21:02] it does turn on when X is disabled [21:02] and it worked in previous versions [22:39] superm1: so, I had an idea of what this crash could be.... intel_agp [22:40] I just reproduced it twice and looking at the text I can just about see in the console, it gets a little way past the point where that loads [22:40] and I guess it's an obvious choice for something which would immediately mess up the display [22:48] * wgrant takes votes on who uses multi-finger tapping and finds it somewhat broken in Intrepid.