=== wgrant_ is now known as wgrant [02:57] woo baby, this amd z430 proprietary driver makes psb look like a dream - insmod: page allocation failure. order:12, mode:0xd1 [02:58] What's the order of an allocation? I'm vaguely aware that it's something to do with its size? [02:58] people are going to love it when arm netbooks get popular since poulsbo is a nice taste of what everyone using arm has to put up with :) [03:01] Aren't we just going to assume that something's put a framebuffer somewhere we can scribble to and just use the super-fast CPU to do all the rendering? :) [03:13] RAOF: tell that to the people complaining about poulsbo :) [03:14] omapfb looks really nice though, even has KMS now [03:16] RAOF: when are you arriving in brussels? [03:17] bought a second 10 hour battery for the trip :) then I found out lucid is using ~2 watts more than it was a few months ago [03:17] Sunday, early evening. [03:18] brightness on minimum and everything disabled and i'm using around 8.8 watts idle, ugh [03:19] was down to 6.7 at one point [03:19] Yeah, I think it did pick up a bit. [03:23] shuttleworth says he's going to have the menubar integrated into the panel like on osx. i didn't know that was in the works. [03:24] in UNR anyway [03:24] that GMA600 is odd, looks like its based on the same powervr 535 as poulsbo but its right on the cpu instead of external now, hope it doesn't make its way into netbooks :) [03:25] osx doesn't have the menu bar in the panel [03:25] does it? [03:25] it does [03:26] i thought he was saying it was going to go a step further and put the menu bar in the panel [03:26] is that a snow leopard thing or something? i'm stuck on 10.5.x because my only mac is a powerpc [03:26] when you o pen a new app, the menu bar does not appear in the top panel? [03:28] nope [03:28] i don't know how to explain that, since it does in every mac i've seen [03:29] maybe you've got it all hacked up or something? [03:30] http://www.bleedyellow.com/blogs/Simon/resource/NotesDesignerinParallels5.jpg [03:31] you can see in that shot lotus notes has its menu bar int he panel, not in the windows [03:31] the active windows does anyway [03:31] Man, windows 7 chrome looks terribly out of place on OS X. :) [03:32] sheesh, shows how much I use osx, I really dont remember that [03:33] i hope the whole title bar goes up there, close/maximize/minimize and all [03:33] i guess it's a gui design philosophy where you have only one location for the menubar no matter what window you're looking at [03:33] i love maximus and window picker applet on my netbook [03:34] just wish i could ditch the title bars on most apps, thats the main reason i use chromium [03:35] next pc is going to be a 11.6" 1366x768 for sure, finally starting to feel cramped on this 8.9" 1024x600 after 2 years :D [03:36] * RAOF holds off on a “why can't they provide me with a decent screen resolution” rant. [03:37] 11.6" is probably a good size, though. [03:38] i'm addicted to the portability, can fit this thing in a cargo pocket and i spend most of my time online out of the house :) [03:40] if they turn the resolution way up on a physically small screen, things like window titlebars and images and so forth would end up being too small to be practical [03:43] yeah 8.9" 1024x600 looks amazing compared to how crappy 15.4" 1280x800 on all my other laptops are, looks sharper than my 24" 1920x1200 screens too [03:46] RAOF: thanks for signing me up for some of the sprints, i'm still trying to work out what to do there [03:46] bjsnider: Which is why 200DPI mobile phone screens suck? :) [03:48] GNOME is already reasonably good at resolution independence, which means that, by and large, the higher the resolution the easier it is to read stuff. [03:54] ok the power problems are because of ondemand now, I'm convinced :) [03:54] switched to conservative and my idle power usage dropped from 9.1 to 7.6 watts [03:59] RAOF, i wuldn't know, because i don't have a mobile phone or any other handheld device [03:59] i seriously doubt they're 200dpi though [03:59] I think the Nexus One is ~250DPI? [04:00] They're like 860x480 in 2" [04:00] They're high DPI displays. [04:01] And the sooner we get 300DPI screens on my desktop, the sooner we can stop futzing around with crazy antialiasing schemes. [04:01] wouldn't that take an enormous amount of power to run? [04:02] the sooner we get duterium-powered fusion reactors inside computers, the sooner we won't have to worry about power anymore [04:02] I don't think the pixel-switching hardware takes anything but a fraction of the power of the backlight on LCD displays; that's one reason why LED backlights are such a win. [05:40] good morning === Bernardo|away is now known as Bernardo === Bernardo is now known as Bernardo|away [08:56] apw: Yo! [08:56] RAOF, hi ya [08:57] You rang, m'lud? [08:57] RAOF, must have been sometime ago :) i think i sent you all an email in the end [08:58] i was thinking though that we need to interlock between graphics and kernel a bit more regularly while we clean up KMS [08:58] Yeah. [08:59] We'll be able to cover some of that ground in the X/kernel “video bugs in the KMS world” session I've signed you for :) [09:00] Gah. Signed you up for! [09:02] The video drivers do try to work against arbitrary kernel versions (with the exception of nouveau), but that doesn't necessarily work :) [09:03] heh indeed :) they have made the world a complex place for us [09:22] Ah! *There's* your mail; hiding in all the bugmail evolution is struggling to filter! [09:31] is failsafe X stuff still supposed to work? Booting a lucid machine without a monitor seems to lead to X not starting [09:43] In an ideal world that *should* work, but I think it doesn't (may be driver dependent, though); X bails if there's nothing to display on. There's a patch on the xorg ML to make it work better. [16:51] ricotz: huh, why are the opengles specs even building without --enable-gles1 and --enable-gles2, they're supposed to default to no [16:51] Sarvatt: kick krh in #dri-devel [16:52] probably already been talked about, gotta read the scrollback :) just saw python-libxml2 was required to build mesa now [16:53] not that I know about. [16:57] Sarvatt, hi, yeah i added this dep [17:02] ricotz: did you try explicitly disabling gles1 and gles2 instead by any chance? [17:03] Sarvatt, no [17:23] darn, still cant copy everything to maverick [17:25] got a llvmpipe enabled mesa here waiting to build, want to be sure its right before pushing the changes to git - https://launchpad.net/~sarvatt/+archive/xorg-testing/+build/1719309 [17:28] hi [18:11] Sarvatt, when is the updated nvidia blob going to be in lucid? [18:13] no idea, hopefully not never :) [18:16] anyone know where psb progress is? [18:18] check the toilet [18:18] herp derp [18:25] bryceh: lol [18:26] heya tseliot [18:27] ;) [18:27] hi [18:27] hi [18:27] bjsnider: I think it would be something for the x-updates ppa, unless the new nvidia driver fixes some bugs which are worth an SRU [18:28] should'nt xf86Resources.h be profided by xserver-xorg-dev [18:28] it adds support for a new line of chips that are being released this month [18:29] tseliot: yeah 195.36.16 doesn't support the whole new generation of GPU's just released [18:30] outlaw45: not anymore [18:30] RAC has been removed [18:31] @jcristau: pretty much read the same in a bug report [18:31] any workaround? [18:31] Sarvatt: I think it's common practice not to do SRUs to update nvidia drivers unless they fix bugs. I don't think the lack of support for new models is a bug. It would be worth asking the desktop team about this though [18:33] @outlaw45 how can i help? any news? [18:33] outlaw45: don't include it. [18:33] outlaw45: and don't call the functions that used to be declared there [18:33] @lucazade: can you build xserver-xorg-video-psb? [18:33] going to try it.. [18:34] Bernardo said on the forum he was able to build them, but I get errors.. [18:34] @jcristau: I'm not a C programmer so not including and solving bugs is pretty hard for me [18:35] then learn C or get someone else to fix it.. [18:35] am a Java programmer and C and especially X drivers are a completely different story.. [18:35] outlaw45: you aren't applying xorg-x11-drv-psb-0.31.0-xserver17.patch then [18:36] Sarvatt, stick -nvidia updates in x-updates [18:36] bryceh: I have been! :D [18:37] @Sarvatt: you're probably right, I'll clean up.. ;) [18:37] Sarvatt, perfect [19:00] outlaw45: still have 'libdrm not found' building xserver-xorg-video-psb [19:03] you have to symlink /lib/pkgconfig/libdrm.pc to /lib/pkgconfig/libdrm.pc [19:03] sorry last path is /usr/lib/pkgconfig [19:04] ln -s /lib/pkgconfig/libdrm.pc /usr/lib/pkgconfig [19:04] ? [19:05] yep [19:09] boo - /usr/bin/ld: cannot find -lffi [19:09] xorg-server-psb succesfully built (also dbg package) [19:09] i'll try xpsb-glx [19:11] i guess llvm-dev isn't enough for llvmpipe, was going by what nano src/gallium/drivers/llvmpipe/README said [19:11] err minus the nano :D [19:14] @lucazade: installed everything and the result's the same [19:15] X always crash? [19:15] looked a minute at Xorg.log but nothing caught my eye, [19:15] no, am able to get into gdm [19:15] and that's stable, but it crashes if you login [19:16] ok [19:16] but have to hook my keyboard to the tablet [19:16] back in a min.. [19:16] dum de dum [19:17] does chase douglas hang around here? [19:17] he has a pretty interesting patch in bgo#610482 but I have some questions [19:23] federico1, try #ubuntu-kernel maybe [19:23] bryceh: cool, thanks [19:56] outlaw45: Xorg.0.log looks like normal [19:56] is there any debug log? [19:56] I have a backtrace in Xorg.0.log.old [19:57] but can't find anything else [19:57] ok seen [19:58] hi === Bernardo|away is now known as Bernardo [19:58] hi [19:58] hi [19:59] I see that you guys haven't made much progress also [19:59] no... [20:00] as anyone tried to remove the libGL redirections? [20:00] has [20:00] no [20:00] the dpkg --divert you posted on the forum? [20:01] I'll try if it makes any difference [20:01] trying [20:01] yes, even though I posted it for karmic [20:01] we are building libGL in the lucid package [20:02] I tried it after building lbdrm package [20:02] but I didn't have any .xlibmesa files [20:03] the redirection is done by xpsb-glx, so until you install it you won't have any .xlibmesa files [20:04] I'm trying to change xpsb-glx now to generate debug packages also [20:04] I rebuild and installed libdrm, xpsb and psb driver [20:04] so we can get anything valid out of the backtrace [20:04] but /usr/lib/psb/ is empty [20:05] here too [20:05] Backtrace: [20:05] 0: /usr/bin/X (xorg_backtrace+0x3b) [0x80e937b] [20:05] 1: /usr/bin/X (0x8048000+0x61c7d) [0x80a9c7d] [20:05] 2: (vdso) (__kernel_rt_sigreturn+0x0) [0xdcc410] [20:05] 3: /usr/lib/libpixman-1.so.0 (0x6c8000+0x4cf10) [0x714f10] [20:05] 4: /usr/lib/libpixman-1.so.0 (0x6c8000+0x147e3) [0x6dc7e3] [20:05] 5: /usr/lib/libpixman-1.so.0 (pixman_blt+0x78) [0x7021e8] [20:05] 6: /usr/lib/xorg/modules/libfb.so (fbCopyNtoN+0x24d) [0xba905d] [20:05] 7: /usr/bin/X (miCopyRegion+0x19e) [0x819ca5e] [20:05] 8: /usr/bin/X (miDoCopy+0x44d) [0x819cffd] [20:05] 9: /usr/lib/xorg/modules/libfb.so (fbCopyArea+0x78) [0xba85c8] [20:05] 10: /usr/lib/xorg/modules/libexa.so (0xc8d000+0xed90) [0xc9bd90] [20:05] 11: /usr/lib/xorg/modules/libexa.so (0xc8d000+0x95c6) [0xc965c6] [20:05] 12: /usr/bin/X (miCopyRegion+0x19e) [0x819ca5e] [20:05] 13: /usr/bin/X (miDoCopy+0x44d) [0x819cffd] [20:05] 14: /usr/lib/xorg/modules/libexa.so (0xc8d000+0x7a4a) [0xc94a4a] [20:05] 15: /usr/bin/X (0x8048000+0xda9c3) [0x81229c3] [20:05] 16: /usr/bin/X (0x8048000+0x28df5) [0x8070df5] [20:05] 17: /usr/bin/X (0x8048000+0x2a477) [0x8072477] [20:05] 18: /usr/bin/X (0x8048000+0x1ed7a) [0x8066d7a] [20:05] 19: /lib/tls/i686/cmov/libc.so.6 (__libc_start_main+0xe6) [0x74ebd6] [20:05] 20: /usr/bin/X (0x8048000+0x1e961) [0x8066961] [20:05] Segmentation fault at address 0x9fc [20:06] Caught signal 11 (Segmentation fault). Server aborting [20:06] got that backtrace in Xorg.0.log.olg [20:06] Xorg.0.log.old [20:06] mine backtrace is similar [20:06] my [20:08] applied divert.. same crashes [20:09] mmm [20:09] but /usr/lib/psb/ is empty and xpsb-glx is installed [20:10] strange, I have those filled, and the redirection is there [20:10] is your a clean install of lucid or upgrade? [20:11] I'm trying now to get it to build debug packages [20:11] clean [20:11] mm strange [20:11] i could try installing your xpsb-glx file? [20:11] yep [20:11] one sec [20:11] ok [20:12] * Bernardo has to find how to send files in kvirc... [20:13] (EE) PSB(0): First SDVO output reported failure to sync or input is not trainded!!!(EE) PSB(0): First SDVO output reported failure to sync or input is not trainded!!! [20:13] oops.. (EE) PSB(0): First SDVO output reported failure to sync or input is not trainded!!! [20:13] outlaw45: I get that sometimes even in karmic [20:13] reboot and try again [20:13] me too [20:14] X starts even with that error [20:15] Bernardo file transfer died.. don't know y [20:18] I'll start again now, might be my firewall [20:18] i see file request.. but transfer doesn't start (maybe because im on pidgin irc) [20:18] ah [20:18] :) [20:20] could you mail it to me? [20:20] yep, I was just asking your email address in private [20:28] did you get it? [20:28] yes yes [20:29] have to go.. I'll check the forum for your progress.. [20:29] cya [20:29] ok see u [20:30] nice posts on phoronix today libv [20:31] /usr/lib/psb still empty [20:43] lucazade: that is strange [20:43] yes [20:43] that is the exact package I installed here [20:43] :) [20:43] and it did the redirection [20:44] i've no idea why is different here [20:46] Bernardo: reisntalling psb-kernel-source i get 2 named psb.ko in /lib/modules/2.6.31 [20:46] and also 2 drm-psb.ko [20:49] where? [20:50] in /lib/modules/2.6.31 during dkms building [20:50] got a kernel panic! [20:50] ugh [20:52] ouch [20:53] I know we need to clean the build process as it is trying to install in /lib/modules//extra [20:54] but never had a kernel panic in dkms [20:55] i was playing modprobing psb and drm_psb by hand and startx caused a kernel panic [20:59] I've finished rebuilding xpsb-glx [20:59] with debug symbols [21:00] hope thius will help [21:00] installed, rebooting the netbook [21:10] in kde4, waiting for it to crash [21:14] and it won't... :) [21:14] ehhhhh [21:14] yep [21:14] now I want it to crash, and it won't [21:15] hehe better [21:16] jcristau: dri2.proto isn't getting installed with xtrace, is it just a matter of adding dri2.proto to dist_pkgdata_DATA in Makefile.am? [21:16] really, it won't crash. What is wrong with the damn drivers? [21:16] lucazade: I'm sending you now the latest attempt [21:17] only xpsb? could you send me all the packages updated? [21:19] sure, I'll up them on svn again [21:19] ok [21:20] Sarvatt: ah probably [21:20] ok, finally crashed [21:20] err Makefile.in sorry [21:20] with a ls in a console [21:20] gh [21:22] Sarvatt: fixed [21:23] ok, ls crashes it regularly [21:23] a simple ls... [21:23] \o/ thanks again for doing that [21:28] boo, even more now :D robert@asuka{~}:xtrace -D :5 > dri2.xtrace [21:28] Unable to find 'genericevents.proto' in search path! [21:28] Bernardo trying debug debs [21:30] genericevents.proto is the last one, works adding that one too [21:31] lucazade: I'm also installing debug for libpixman-1-0 [21:32] jcristau: looks to be working good here, you rock :) http://sarvatt.com/downloads/glxgears.xtrace.txt [21:32] the backtrace is becomming more detailed [21:32] Sarvatt: fixed again.. [21:33] looks like there is some issue with icons, in gdm and in desktop (before it crash) [21:34] not able to load icons [21:34] I'm also installing xserver-xorg-core-dbg [21:35] in kde it loads most of them [21:35] but they are svg [21:35] yes [21:35] png are broken [21:35] notify osd uses svg and it works [21:37] libpixman is about icons? [21:38] Description: pixel-manipulation library for X and cairo [21:38] so not directly, but maybe related [21:39] and now it is refusing to crash again... Adding debug seems to delay the crash [21:39] here (on gnome crash the same as before) [21:43] ok, seems to be in miDoCopy in X, which then calls libexa [21:45] unfortunately, even installing xserver-xorg-core-dbg doesn't add detail to that so [21:46] arg [21:46] was trying to disable aiglx and composite [21:48] Bernardo what else can we do? [21:48] (crying) [21:50] gdb, add debug to psb, lets go hunting needles in a haystack [21:51] at least we now it is inside fbCopyArea [21:52] so we check routines in the psb driver that might be related to copying bitmaps to/from the framebuffer [21:52] right... i bet on pitmaps [21:53] Bernardo: i'm going out to make my brain get fresh air.. i'll back in a couple of hour.. i'll look here or on the forum for any news [21:53] ok [21:54] I'll also leave soon, maybe someone will have an idea [21:54] ok see you soon [21:55] 'night [21:55] good night === Bernardo is now known as Bernardo|away [22:07] well thats something I didn't expect, quadrapassel actually works with i915 gallium and not mesa classic :D hasn't worked since upgrading xorg-edgers to xserver 1.8 [22:10] not sure what to do with the libgl1-mesa-dri-gallium stuff, aiglx expects things to be in /usr/lib/dri and doesn't respect --with-dri-searchpath paths [22:13] need to figure out what else llvmpipe needs too [22:13] g++ -L/usr/lib/llvm/lib -lpthread -lffi -ldl -lm lp_test_format.o lp_test_main.o -o lp_test_format -Wl,--start-group -lX11 -lXext -lXxf86vm -lXdamage -lXfixes -ldrm -lm -lpthread -ldl -L../../auxiliary/ -lgallium libllvmpipe.a -lLLVMBitWriter -lLLVMX86CodeGen -lLLVMX86Info -lLLVMSelectionDAG -lLLVMAsmPrinter -lLLVMInterpreter -lLLVMJIT -lLLVMExecutionEngine -lLLVMCodeGen -lLLVMScalarOpts -lLLVMInstCombine -lLLVMTransformUtils -lLLVMipa [22:13] -lLLVMAnalysis -lLLVMTarget -lLLVMMC -lLLVMCore -lLLVMSupport -lLLVMSystem -L/usr/lib/llvm/lib -lstdc++ -Wl,--end-group [22:13] /usr/bin/ld: cannot find -lffi [22:20] duh, libffi-dev :D [23:13] During Lucid's development my wacom device lost some evdev properties like axis inversion and axes swap. [23:14] Is there a particular package I should try reverting or a patch that might help? [23:19] if it's using wacom it won't have evdev properties [23:20] jcristau: it used to during Lucid's development that was a mistake then? [23:21] bdmurray: you want to do it through xsetwacom [23:21] something like - http://www.uluga.ubuntuforums.org/showthread.php?t=1447925 [23:22] if was using evdev instead of wacom for a short period when there was no xf86-input-wacom that worked with xserver 1.7 yeah [23:23] Sarvatt: sweet, that looks like what I was doing [23:24] yeah I remembered seeing your blog post and wondering what comic reader you were using :) [23:24] I was using qcomicbook but I'm trying out comix now [23:25] * Sarvatt is a diehard comix fan [23:25] I ran into a bug with it recently though [23:25] not rendering bmp's? [23:25] no, page ordering for multi-page scans [23:26] ah yeah thats pretty much fixed in svn too [23:26] i noticed that with some naming schemes people used too but haven't had a problem with it since i built it last in october [23:27] and that's a later version than is packaged I take it? [23:27] yep but it's python, dead simple to install. i think i might have packaged it actually, lets see [23:28] right, I forgot it was python [23:31] comix is by far the most used program here, I'm a manga addict :) looking at the revision log it might not be fixed though and I just remembered I've had a few volumes i've had to manually extract first to view them right recently [23:31] here is the upstream issue I found [23:31] http://sourceforge.net/tracker/?func=detail&aid=2456807&group_id=146377&atid=764985 [23:32] Error: Sourceforge bug 2456807 not found [23:37] uploaded 4.0.5 here but it'll take awhile to build - https://edge.launchpad.net/~sarvatt/+archive/ppa [23:37] oh yuck that stinks [23:38] people really name things that way? [23:38] so it seems [23:38] they must all use cdisplay on windows where it actually works with that naming scheme :) [23:39] I thought maybe one could make a list of page numbers and if the differnce between items was greater than 2... [23:41] did you see the comments? [23:42] yes, the last one? [23:43] yeah, that would be really annoying if you have a lot of archives like that :( [23:43] using the editing dialog would be spoilerful too [23:44] so bug 570523 should be invalid then? [23:44] Launchpad bug 570523 in xinput "several evdev properties went missing" [Undecided,Incomplete] https://launchpad.net/bugs/570523 [23:48] yep, closed it with an explanation, the evdev properties were only there until mid feb. or so until xf86-input-wacom started working for your tablet [23:48] you could use them now if you wanted to ditch wacom for evdev again [23:54] bdmurray: have you seen https://edge.launchpad.net/~thjaeger/+archive/tabletpc/+sourcepub/1076480/+listing-archive-extra ? [23:54] nope [23:55] automatically rotates the tablet when the display is rotated [23:56] I'll have to play with that then [23:56] Do you know the right way to turn on "Synaptics Palm Detection" regularly? [23:57] Option "PalmDetect" "1" in xorg.conf [23:58] or synclient PalmDetect=1? [23:58] don't know what you mean by regularly [23:58] for every boot