[00:15] chrisccoulson: check both the vidmode lib and server side code [00:16] chrisccoulson: really easy to debug this when you have the code and gdb ready [00:18] ah, i'd say randr and driver issue. [00:36] chrisccoulson: any way to force gnome-screensaver to activate from the command line? (-a doesn't work) [00:39] CAD not good enough? [00:39] :p [01:24] does xgamma have the same result btw? [01:51] chrisccoulson: hmm yeah its using FADE_TYPE_GAMMA_NUMBER since it says it doesn't support ramps according to that xtrace, and it goes down to test_number: because it only has a gamma parameter not ramps, res = XF86VidModeGetGamma (GDK_DISPLAY (), screen, &info [screen].vmg); should be failing in his case and making him not use fade at all but its not [01:53] maybe just add a check if kvm is in use to gs-fade.c and use FADE_TYPE_NONE for it :D [01:55] try using vesa yourself like the cirrus adapter in kvm is using due to the xserver patch we have and see if it does the same thing maybe? [01:56] imagine there would be alot more bug reports similar if it happened to everyone using vesa and its something kvm specific [01:59] oh i was looking at old source, this changed things alot http://git.gnome.org/browse/gnome-screensaver/commit/?id=f423a56d536b4ce19383ed74c1550ebd12f88541 [02:13] tjaalton, yeah tried doing a hardy->lucid upgrade on my webserver but it crashed and burned so had to do a reinstall [02:14] tjaalton, however most of the X content gets generated on another machine and pushed to my webserver, so all that is good to go right now [02:15] some of the older scripts like the package version report aren't running yet [02:23] chrisccoulson: check out hw/xfree86/dixmods/extmod/xf86vmode.c to see why it would return BadValue btw [02:34] bryceh, when you started that upgrade, what did you expect? i mean would success have been a reasonable expectation? [02:35] a lot of lts users are going to be faced with this decision soon [02:36] bjsnider, what are you getting at? [02:36] i just wondred if you thought it was going to work, and if so how much confidence you had in it [02:36] it would have been nice to see it upgrade cleanly, but I didn't have expectations that it would [02:37] i see [02:37] it's barely after alpha-3 [02:37] ok, wrong question [02:37] do you think it will be possible to upgrade from hardy to lucid once the final release happens? [02:38] yes [02:38] with a high amount of confidence that it will work? [02:39] well you're asking entirely the wrong person there ;-) [02:39] are efforts underway to test that scenario? [02:39] i mean other than the one you just went through [02:41] ah, here's the question: is that upgrade officially supported by canonical? [02:41] in fact there's people at canonical for whom testing lts->lts upgrades is the primary part of their job [02:41] yes it is [02:41] wow, that's a realy tough upgrade though [02:42] that's a lot of deprecated packages, new packages, changes in the xorg/mesa system etc. [02:42] new grub [02:42] * bryceh nods [02:43] surprisingly I didn't have too many problems with the X packages, although it did get stuck on the -r128 transition package [02:43] system was radeon and came up with kms ok, purple screen and everything [02:44] something was bugged in the kernel though [02:45] bryceh: purple is the new brown [02:45] it even replaced my gnome-terminal black with _this_ fushia [02:45] ew [02:45] terminals gotta be white with black text. anything else is abomination [02:46] black with green text is ok if you're old school [02:46] WHAT? [02:46] black with white text [02:46] *shudder* [02:46] next you'll be saying terminals should be transparent [02:46] ehe [02:46] err [02:46] mine is :D [02:47] 70% opace [02:47] you know my kind, it seems [02:48] ;-) [02:48] oh that's better [02:48] back to 70% [02:49] this new look theme changed that too [02:49] was on 90% [02:49] bjsnider, I just wish we had a good tool for doing a smart backup of a hardy system [02:49] $ dd ? [02:49] bjsnider, like for instance instead of just copying the mysql db files, to do sql dumps of them so they can be loaded in the new version [02:49] etckeeper? [02:49] how do you mean smart? only changed files are backed up and so forth? [02:49] no, like backup in a form that can be restored onto lucid [02:50] i solve that problem by mounting my home directory as a separate partition [02:50] like for instance my drupal stuff for my blog I backed up but I can't really just copy it back into place, I kind of have to recover it chunk by chunk [02:50] i don't do upgrades, only clean installs [02:50] ehe [02:50] yeah me too. I was half hoping the upgrade would do all the magic of converting all the files over [02:51] until my disk died early this year [02:51] but forgot to copy some of the drupal themes and modules and stuff [02:51] I had 3 cycles of upgrades here [02:51] point your webserver stuff at your home directory so it will stick around [02:51] well I have /var/www on a raid partition and that's all there fine [02:51] it's the stuff that had been under /usr I didn't think to back up [02:51] bryceh: that's why I started doing clonezilla of my discs prior to format [02:51] I *always* forget something [02:52] BUGabundo, vloopback.c:676: error: ‘TASK_INTERRUPTIBLE’ undeclared [02:52] package missing? [02:52] find where that function is declared and install that package and you'll be fine [02:52] ahh [02:52] task_interruptible [02:52] and all of the others [02:52] forum says to use another package [02:52] that's why make is failing [02:52] cause the one in the tar is too old for this kernel [02:53] yeah, if those functions were removed in the newer kernel, that would cause it too [02:53] !info vloopback-source [02:53] vloopback-source (source: vloopback): vloopback modules for Linux. In component universe, is optional. Version 1.1-1 (karmic), package size 15 kB, installed size 60 kB [02:53] that's what moprons like me have to worry about when making deb packages [02:53] let me try to get that source package 1st [02:54] BUGabundo, check the vloopback.c file. at the beginning there will be a bunch of #includes statements. make sure it has those files [02:54] The following NEW packages will be installed: [02:54] cvs{a} debhelper{a} gettext{a} html2text{a} intltool-debian{a} libmail-sendmail-perl{a} libsys-hostname-long-perl{a} module-assistant{a} [02:54] po-debconf{a} vloopback-source [02:56] my idea didn't help [02:56] trying yours [02:56] try adding the ppa, then doing an apt-get build-dep flashcam [02:57] ok [02:59] !find sched.h lucid [02:59] 0 upgraded, 0 newly installed, 0 to remove and 1 not upgraded. [02:59] didn't take me no where [02:59] File sched.h found in asterisk-dev, dietlibc-dev, fftw-docs, libace-dev, libc6-dev (and 31 others) [02:59] did I tell you the ppa only has 32bits paclages? [03:00] BUGabundo, try installing libc6-dev [03:00] then do make again [03:01] already instaled [03:01] no go [03:02] which files are in the include statements? [03:02] why did he only build i386 packages? [03:03] http://paste.ubuntu.com/390042/ [03:03] cause that's all he needed for him? [03:05] what a numbskull [03:06] you have the linux-headers package for your kernel installed? [03:06] because almost all of those headers are from the kernel [03:06] checking [03:07] is this popovetsky the developer of the software? [03:08] $ dpkg -l | grep linux-headers | pastebinit [03:08] http://paste.ubuntu.com/390046/ [03:09] what kernel are you currently running? [03:09] no, this is https://sourceforge.net/users/swift-tools/ [03:10] Linux BluBUG 2.6.32-15-generic #22-Ubuntu SMP Tue Mar 2 02:23:29 UTC 2010 x86_64 GNU/Linux [03:11] Copyright © 2008-2010 Olivier "Swift" Debon [03:11] Kernel compiled with CONFIG_VIDEO_V4L1_COMPAT [03:11] is that the case with the ubuntu kernel? [03:11] yes, I expected as much [03:11] since Nol did an heck of a job getting my cam driver upstreamed [03:12] and since karmic it works OTB [03:12] let me take the current release and throw it into pbuilder to see if it builds [03:14] okay [03:14] I'll be asleep before it finishs [03:14] probably not [03:14] I can already barelly keep my left eye open [03:15] brb, getting some cookies [03:23] back [03:25] first pbuilder run failed due to missing build-depends [03:25] doing second run [03:27] ehe [03:27] which ones? [03:28] build system missing /lib/modules/2.6.31-20-generic/build directory [03:38] hmm, we have the cirrus patch to make it work with KVM/qemu, why are we defaulting it to vesa? [03:42] 100_fedora_libpciaccess.patch in x-x-v-cirrus makes it work for KVM use, 184_virtual_devices_autodetect.patch in xorg-server is defaulting it to vesa [03:46] BUGabundo, just tried running make here and got an error related to that option i mentioned above that the kernel has to have [03:46] i am asking in the kernel channel if it has that option enabled [03:47] thanks [03:48] i'll bet the 32-bit flash already has full v4l2 support and if you were using the i386 version this wouldn't be an issue [03:48] :p [03:48] maybe! [03:50] if it's a usb webcam you could probably try it in a vm [03:51] its a mounted cam, but uses internal usb [03:53] so you might be able to pass it off to the vm [03:53] I guess [03:53] I just don't use windows [03:53] and use it with the native 32-bit flash plugin [03:53] don't even have a licence around for it [03:54] what does windows have to do with it? [03:54] the 32-bit linux plugin [03:54] it's very different from the 64-bit version [03:57] ahhh [03:57] I tried BOTH [03:58] 64 bits from adobe labs [03:58] flash player from adobe [03:58] 32bit [03:58] and archive 32bits too [03:58] none seems to work with my webcam [03:58] how do you test that? [04:00] close all browser, install lib (either in .mozilla/plugins or system), open browser, make sure its in about:plugin, open a site that uses cam [04:00] what site? [04:04] bambuser.com , ustream,com, chatroulette.com/ [04:13] BUGabundo, not going to do anymore tonight. [04:13] np [04:13] thanks for the time [06:45] bryceh: yeah I noticed the stats came online later [09:02] bryceh, just a quick question. I'm running Lucid + xorg-edgers and I get this with glxgears: "IRQs don't seem to be working correctly" apparently this is fixed in 2.6.33 (according to this chat log: http://bit.ly/dpP7mH) is that something that will be backported? [09:09] already is [09:09] just that the image is not built yet aiui [09:09] ok cool :) [10:04] hi guys. my x crashes after wakeup from sleep [10:24] what driver === hypera1r is now known as hyperair [17:52] anyone able to make this guy a kernel with the patch i attached included? https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/507196 I tried using the ubuntu debian/ build system and failed a few times, dont think he'd want one of my kernel-package ones without an initrd [17:52] Launchpad bug 507196 in linux "[i945gme] Ubuntu falls back on vesa driver (no screens)" [High,Confirmed] [17:55] not sure how to properly bump the version in the ubuntu build system, making it 2.6.32-15.22+bug507196 or 2.6.32-15.23~bug507196 in the changelog didn't build the package properly [17:57] i asked upstream if they would consider adding the ignorelid module option to ignore the acpi lid status because i've made 6 bad lid status quirks in the past week for people and it would make it possible to have a usable system until the quirks went upstream (and they could verify they needed the quirk with a module parameter instead of rebuilding a kernel) [17:58] so many chinese knockoff netbooks with screwed up acpi popping up [17:59] Sarvatt: you should try poking people in #ubuntu-kernel [17:59] i'd build one, but mine's a kernel-package one as well [18:00] do ya know how to add the update-initramfs hooks back to the new kernel-package package? [18:00] the documentation is really confusing about it [18:02] tempted to make a new package thats 11.017 kernel-package with the updated stuff to work for .33+ kernels since that looks self contained [18:03] Sarvatt: --initrd [18:03] $ alias buildkernel [18:03] alias buildkernel='AUTOBUILD=1 CONCURRENCY_LEVEL=${CONCURRENCY_LEVEL:-2} ionice -c 3 schedtool -D -e make-kpkg --initrd --rootcmd=fakeroot --append-to-version=-hyper${KERNELREV:-1} kernel_{headers,image}' [18:03] doesn't work, they ripped out the initrd hook scripts [18:03] i think there's something in /etc/kernel-pkg.conf [18:03] er maybe not [18:04] ah /etc/kernel-img.conf [18:04] you need to copy the hook script from /usr/share/kernel-package/SOMETHING to somewhere in /etc/kernel/ but i couldnt work out where [18:05] $ tree /etc/kernel [18:05] /etc/kernel [18:05] |-- postinst.d [18:05] | `-- dkms [18:05] `-- prerm.d [18:05] `-- dkms [18:05] nothing. [18:05] for the --initrd option to work [18:05] it's just /etc/kernel-img.conf [18:05] just change the line [18:05] yeah theres nothing there now, you have to manually link it according to the man [18:05] i think you need some CONFIG_INITRD or so as well [18:06] hey i'm compiling my own kernels, and i've been using initrd [18:06] and i've got nothing in that directory! [18:06] i'm telling you, it's not necessary [18:06] ahhhhh here we go /usr/share/kernel-package/examples/etc [18:07] you're on karmic hyperair [18:07] oh. [18:07] seriously? [18:07] 12.x kernel-package is completely different now [18:07] * hyperair facepalms [18:07] you cant compile .33 kernels with the 11.015 you have in karmic either [18:07] oh yes i acn. [18:07] $ uname -r [18:07] 2.6.33-hyper1 [18:08] hmm what version is kernel-package? [18:08] 11.015 [18:08] i think i know what you're talking bout [18:08] i changed some .mk scripts [18:08] the header location changed alot in .33 [18:08] ah yeah [18:09] with 12.x you dont need to make-kpkg clean if compiling fails anymore though [18:09] /usr/share/kernel-package/ruleset/misc/version_vars.mk FAILED [18:09] eh that's cool [18:09] you can keep debian/ around, saves alot of time [18:09] yes it does [18:09] i've been doing rm -rf conf.vars debian [18:09] to keep the compiled objects in place [18:10] that saves time [18:11] Please note that unless there are hook scripts in /etc/kernel or added into the hook script parameter of /etc/kernel-img.conf. no initrd will be created (the bundled in example scripts are just examples -- user action is required before anything happens). [18:11] i wish http://git.debian.org/?s= worked [18:12] its either wait 10 minutes hoping that'll work or wait 10 minutes for the whole index of git repos to load with the non-caching gitweb debian is using.. [18:18] the repo index is supposed to be cached [18:22] ahhah /usr/share/doc/kernel-package/README.gz has good info on how to add initrd and update-grub hooks to /etc/kernel/ [18:24] maybe its just that gitweb is slow as heck to show that many repos on my machine [18:44] hmm https://bugs.edge.launchpad.net/ubuntu/+source/xorg/+bug/533418 [18:44] Launchpad bug 533418 in xorg "[xorg-edgers] ...dri/r600_dri.so: undefined symbol: _glapi_tls_Context" [Undecided,New] [18:44] anyone else using xorg-edgers on karmic with an r600+ by any chance? [18:45] i dont know why its using this -- file=/usr/X11R6/lib/modules/dri/r600_dri.so [0]; needed by /usr/lib/libGL.so.1 === radoe_ is now known as radoe [18:57] yeah it looks like he has fglrx crap still hanging around, fglrx installs stuff to /usr/X11R6/lib/modules/dri/ [19:09] is fglrx getting dropped or something? [19:09] isn't it the only decent 3D driver around for ATI? [19:12] fglrx drops itself [19:12] takes them a year to update to trivial api changes [19:32] its at the point where I wouldnt even use fglrx at the moment unless I had an R800 anyway even if they didnt take >10 months to adjust to a new xserver video abi, but you really need newer than distro packages to keep up with all the progress [19:32] should take bets on how long it'll take to support xserver 1.8 :) [19:35] r600+ really needs a bleeding edge kernel for power management on laptops [19:37] thats one of the other sucky things about nouveau at the moment too, geforce 8xxx+ is just as reliant on power management support to run at lower clocks and voltages and there isnt any yet [19:39] people with passively cooled desktop 8xxx+'s are going to be in a sucky situation with lucid's nouveau too :D [19:39] my 8600gt seems fine so far [19:40] my passively cooled radeon hd2400 pro triggered gpu resets from overheating all the time in windows when I forced it to max clock speed [19:41] so pull more upstream crack for .1 :) [19:41] +1 is going to have 2.6.35 or .36 anyway, will be alot better by then i'm sure :D [19:42] no I meant 10.04.1 [19:42] which is probably late june [19:44] BUGabundo, can the android play all of the files here: http://www.linnrecords.com/linn-downloads-testfiles.aspx [19:44] ill have to test [19:44] why do you ask? [19:44] flac support depends on what rom he's using [19:44] looking to buy a new super phone? [19:44] i dont think the stock ones under 2.0 support flac [19:44] Sarvatt: using CM [19:44] AFAIK it does have extra codecs support [19:45] let me ask [19:45] gave my htc dream to the wife 6 months ago so i havent kept up with android stuff :( [19:45] i'm interested in the 24 bit and studio master 5.1 files [19:46] on a phone? [19:46] (2010-03-07 19:45:50) sunjon: BUGabundo: flac support is in afaik, haven't ever used it myself [19:46] Sarvatt: time to update to the last stable cyanogenmod [19:47] Sarvatt, just want to see what its limitations are [19:47] one sec i'll point you to the dsp source so ya can see [19:47] andoid.git.k.o [19:47] what device are you looking at bjsnider? [19:48] and http://github.com/cyanogen [19:48] http://android.git.kernel.org/?p=platform/external/opencore.git;a=summary [19:50] xine and mplayer have trouble with the 5.1 flac, but gstreamer and vlc can play it [19:50] what type of container do you have the 5.1 flac in? [19:51] flac [19:51] i put mine in mkv and mplayer doesnt have problems with it [19:55] i just reencode all my 5.1 flac into ac3 to be passed straight to my receiver over digitial spdif though, dont have one of these fancy hdmi receivers so its reencoding anyway somewhere [19:56] and ac3 can have tags and stuff in a mka container [19:57] bjsnider: FYI player plays does samples really fine [20:04] Sarvatt, what version of mplayer are you using? [20:07] i play that stuff on my htpc which runs windows, was an mplayer i built myself about 1.5 years ago when I last cared enough to mess with anything windows related besides updating my htpc front end :) [20:10] BUGabundo, here's the problem with the flashcam compile: The version of vloopback included with Flashcam is too old for your kernel. [20:10] I know [20:11] I told you that yesterday [20:11] lots of places it could go wrong though, you could have it resampling to a format your speakers cant handle or the audio backend its using is screwing it up or any number of things [20:11] also the developer said that he is updating vloopback soon, or whatever [20:12] Sarvatt, why do you need a windows htpc? [20:14] because linux video acceleration was years behind windows and i dont care enough to switch it over since it took years to get everything working just the way i wanted it :) [20:14] and last i looked my tv tuner doesn't even work under linux [20:15] well, those are very good arguments [20:16] this 2.6.32-16 kernel is using ~2 watts more than -15 :( [20:17] looks like its all chromium's fault though [20:18] sheesh, wish i had a touchpad hooked up internally over USB instead of going through the EC, I hate how using the touchpad shoots power usage and wakeups/second through the roof [20:20] power usage almost doubles and i get an extra 600 wakeups from idle per second if i use the touchpad.. lowering the psmouse module rate option to 30 from 100 makes it a little better [20:22] closed so many bugs against -synaptics about that since its a hardware limitation, at least we dont get huge cpu usage spikes while using the touchpad like windows :) [20:23] BUGabundo, alright, go it [20:23] BUGabundo, still have your flascham directory? [20:24] guess so [20:24] head in there [20:24] there is a driectory called vloopback. see it? [20:24] yes [20:24] blow it away [20:25] done [20:25] download this tarball to the flashcam directory: http://www.lavrsen.dk/foswiki/pub/Motion/VideoFourLinuxLoopbackDevice/vloopback-1.3.tar.gz [20:25] unpack it. rename the resulting directory to vloopback [20:25] then run make [20:26] done [20:26] make on the root dir, right? [20:26] yes [20:26] its going [20:26] erros [20:26] which kernel is this? [20:26] http://paste.ubuntu.com/390582/ [20:26] .31 or .32? [20:27] 2.6.32-15-generic [20:27] this is tested up to .31, but do make clean [20:27] done [20:28] only run make ont he root dir. do not go into the vlloopback directory and run make [20:28] ok [20:28] make again ? [20:28] yes run make [20:29] http://paste.ubuntu.com/390583/ [20:31] well, it looks like vloopback is not compatible with the .32 kernel at this point [20:31] hmm I think I found this guys r600 dri problem, am I crazy or does it look like xorg-driver-fglrx not remove /etc/X11/Xsession.d/10fglrx after you remove it? [20:31] it did work here with the karmic kernel [20:31] thanks bjsnider [20:32] that's the latest version of vloopback but not the latest development code [20:32] oh wait. he says svn trunk will work with .32 [20:33] cool [20:33] what's the subversion command that only downloads the latest code and not the entire history of it? [20:33] superm1: sorry to bug you, but do you have any idea about that? the 10fglrx Xsession script is exporting the fglrx specific LIBGL_DRIVERS_PATH and I dont see anywhere where it is being removed when the package is removed [20:33] * BUGabundo pulls up git [20:33] svn ? hummm rev ? [20:33] or head [20:33] nooo that's git [20:34] "rev last" I think [20:34] checkout gets you the whole history [20:34] google [20:35] BOO [20:36] google fails me [20:36] export [20:36] alright, blow away the vloopback directory again [20:36] ahh [20:36] done [20:36] then run this: svn export http://www.lavrsen.dk/svn/vloopback/trunk/ vloopback [20:36] then run make [20:37] Exported revision 22. [20:37] right [20:37] make clean 1st [20:37] making now [20:37] DONE [20:37] it worked? [20:37] it created the executable? [20:37] http://paste.ubuntu.com/390589/ [20:38] it did [20:38] Found V4L2 Capture device: /dev/video0. gspca_gl860/USB 2.0 PC Camera. Current Size: 1600x1200 Format: GBRG [20:38] it has the wrong prefix, but it worked [20:39] ? [20:39] the make run is set to install the files to /usr/local [20:39] ok, back to MAN [20:39] and get this working [20:39] ahh [20:40] thanks, once again [20:40] no prob [20:41] you can run make install at this point [20:41] FATAL: Module vloopback not found. [20:41] PATH? [20:42] make ; su - ; make install [20:42] /sbin/modprobe videodev [20:42] /sbin/modprobe vloopback [20:45] oi [20:46] heyo, thanks because its a distro specific issue most likely and not an upstream radeon one most likely :) [20:46] does /etc/X11/Xsession.d/10fglrx exist? [20:46] I'm glad to have found to correct place :) [20:47] No, neither does this env. [20:47] darn [20:47] I do have lots of fglrx stuff floating around, I'm cleaning it up. [20:48] i'm almost positive thats the issue [20:48] /usr/share/app-install/desktop/fglrx-driver.desktop /usr/share/apport/package-hooks/source_fglrx-installer.py [20:48] Well, if env dosen't show it then it's not the variable... I can try and clear it. [20:48] its looking for r600_dri.so at the location fglrx installs its s [20:48] err [20:49] wrong window when i was finishing that sentence :) [20:49] can you pastebin your /var/log/Xorg.0.log? [20:50] maybe its loading libdri.so (or libglx.so) from fglrx or something [20:53] https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/533418 <-- xorg.log [20:53] Launchpad bug 533418 in xorg "[xorg-edgers] ...dri/r600_dri.so: undefined symbol: _glapi_tls_Context" [Undecided,New] [20:53] sweet. [20:54] Sarvatt: yeah that sounds like a bogus loader [20:54] $ objdump -T /usr/lib/xorg/modules/extensions/libglx.so |grep glapi_tls [20:54] 00000000 g D .tdata 00000004 Base _glapi_tls_Dispatch [20:54] 00000004 g D .tbss 00000004 Base _glapi_tls_Context [20:54]