[01:55] Would you believe it? Destroying buffers before you've finished rendering from them makes for some crazy rendering. [01:55] * RAOF feels rather stupid. [01:59] Yup. There goes the flickering! [03:30] okay, so I got the magic trackpad... and it doesn't scroll or anything. [03:32] Is something I should go to #ubuntu for? [03:38] It should do 2, 3 & 4 finger gestures under Unity, IIRC. [03:40] ah, what I'd want is the same as regular Synaptics, at least. [03:40] You *should* be able to get synaptics to claim it by judicious editing of config files; you'll lose gesture support there, though. [03:41] Scrolling (2-finger and/or edge, though Gnome devs think nobody could possibly ever want both), and middle and right button are most important. [03:41] I _think_ gesture support should handle 2 finger scroll, though. Again, only under Unity [03:41] Stupid apple... 2-finger right-click, but where the heck's my MIDDLE button? [03:41] I use middle way more than I use right, when browsing. [03:42] Oh, and when I hold the lower edge, all movement stops. [03:43] Correction: as long as two fingers are on the board, literally nothing happens when I move the fingers. [04:32] DanaG, what device is this? [04:34] Magic Trackpad. [04:40] I won't go to +1 until I can get... ATI driver... to work with new X server. [04:50] fglrx? We haven't got a new X server in Natty yet. [04:53] hmm, is there anything else significantly new in natty, anyway? [04:53] Well, new kernel. [04:53] (Slightly) newer mesa. [04:53] New -intel,... [04:53] Radeon uses too much power... runs warm and noisy. [04:54] Random aside: why is it even called a "trackpad", anyway? You touch it, not track it! [04:54] I'm tracking my mouse.... oops, it got away! [04:54] Even if you set dynpm? [04:54] Yeah, last time I checked, dynpm didn't use lowest-power modes, and even with fixed low, it still uses like 8 watts more than fglrx. [04:55] Hey, if a trackpad is called such because it "tracks" you... then we should call webcams WATCHcams, because they WATCH you. =þ [04:56] ah, I'm switching the "magic" thing to Synaptics. [04:58] Regression with the change from hal to xorg.conf.d: changing .conf file doesn't take effect merely by reattaching device! [05:00] Well, either that, or the 60-magictrackpad.conf isn't what defines the driver. [05:00] I changed it to say synaptics, and it's still using udev. [05:01] evdev, you mean. [05:01] er, yeah. [05:02] Say, does using /etc/X11/xorg.conf.d still break everything by making it ignore /usr/share/X11/xorg.conf.d/ ? [05:03] I don't think so. [05:03] I created a new file (61) with identifier "Magic Trackpad, damnit"... and it doesn't use the file. [05:04] I used to be able to do this for synaptics: change the fdi file, then unload and reload psmouse. [05:05] Well, time to ctrl-alt-backspace, then. grr. [05:09] Okay, restart did it... though now i'm screwed by the "can't have both" limitation. [05:10] Built-in touchpad won't do multi-finger... so I get edge scrolling with magic. [05:10] Well, it WILL do multifinger with exactly one Windows driver that seems to patch the firmware.... and some people refuse to believe that that could possibly be true. [05:10] Magic should do two finger scroll with Synaptics I believe. [05:11] https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-input-synaptics/+bug/546697 [05:11] Although cnd is the guy with the hardware, who wrote the driver, and who's doing the multitouch implementation :) [05:11] Launchpad bug 546697 in xserver-xorg-input-synaptics (Ubuntu) "enable multitouch support on older touchpads, as supported by driver v15.0.9.0 (dup-of: 554980)" [Medium,Confirmed] [05:11] Launchpad bug 554980 in xserver-xorg-input-synaptics (Ubuntu) "Two finger scroll not working on all old touchPads (emulation approach). (affects: 30) (dups: 3) (heat: 141)" [Medium,Confirmed] [05:12] RAOF: the problem is that Gnome won't let me enable BOTH types of scrolling at once. [05:12] And those damn developers: the special windows driver IS NOT FREAKING EMULATION! [05:12] I'd guess you could probably twiddle the properties yourself with xinput [05:13] RAOF: tried that... gnome later resets it. [05:13] If I have gpointing-device-settings installed, I get enable...disable...enable...disable...enable every second or so. [05:13] Sweet! [05:16] I hate those devs that REFUSE to accept that a driver could ever possibly patch the firmware! [05:17] wait, in that thread, the person is not a dev. [05:18] But in the other discussion... the one where the person sniffed the PS/2 traffic of the new touchpads, all I wanted was them to try that sniffing with the special driver. [05:19] Really hackish thing I'm considering trying: sniffing PS/2 traffic with my Xilinx board, and dumping it over serial. [05:20] Sorry about the rant. [05:21] Anyway: https://bugs.launchpad.net/ubuntu/+source/gnome-settings-daemon/+bug/467258 [05:21] Launchpad bug 467258 in gnome-settings-daemon (Ubuntu) (and 2 other projects) "Touchpad: simultaneous two-finger and edge scrolling (affects: 3) (heat: 17)" [Wishlist,Triaged] [05:26] I must say, wouldn't Synaptics as default have been a better default for the Magic? [05:27] Anyway, I'm off to go try the thing with Windows. [05:27] Got it for work, mainly... and they use Windows. [07:00] I hope we can at least get drivers to the level of NVIDIA, or preferably, fglrx, in terms of packaging. [07:00] Some day, I mean. [07:01] In terms of packaging? [07:02] Yeah. Can't beat: ./ati-driver-installer-whatever.run --buildpkg Ubuntu/maverick [07:02] Ah. [07:02] Well, I think you can; the edgers PPA is pretty kickarse. [07:03] I mean for arm, though. [07:03] Does edgers do that? [07:03] When the ARM ppa buildds are up, yes. [07:06] Spiffy. So no more asking TI for permission? [07:07] For me, 99% of my 3D stuff in Linux is compiz. It just doesn't feel as nice without it. [07:07] Oh, permission to redistribute their drivers? Ah. Different question. [07:07] They *do* have a PPA for it already, I think. [07:07] At least for the OMAP boards. [07:08] Last time I used a beagle, you had to download a huge package from TI, and then use their failure of a build script. [07:09] "make" did "make clean" on a hardcoded target dir that didn't exist. [07:18] For me, 90% of the 3D stuff I do in Linux is Compiz, so GLX is mandatory. [07:19] Oh, and oddly enough, last time I checked power usage on battery with ATI binary, it was lower WITH Compiz and Aero than without! [07:22] Not totally odd; you get less redraw with a compositing manager. [07:23] I'll have to try Radeon on alpha... If power usage is tolerable, I may switch. [07:23] Actually drawing stuff and then blitting a texture costs more than blitting a texture :) [07:23] The difference between fglrx and Radeon was like 10 watts. [07:24] All those silly people who recommend disabling it... [07:24] * RAOF wonders if he's mentioned that fglrx has more lines of power management code than the radeon kernel module, mesa driver, and DDX in total? === yofel_ is now known as yofel [14:04] Sarvatt: it seems compiz is hanging at another X call now. [14:04] http://paste.debian.net/101860/ [14:33] hyperair: so what happens, it just hangs? [14:43] Sarvatt: yep, compiz hangs, X hangs. [14:43] no wait [14:43] compiz hangs [14:43] X doesn't hang [14:43] killing compiz brings X back [14:43] it might be useful to attach to the X process and get a bt there [14:43] ah [16:02] Sarvatt: does xorg-edgers have pixman dbgsyms? [16:02] ah it does [16:03] -dbg yeah [16:03] Sarvatt: it seems that Xorg is stuck in libpixman [16:03] and compiz is stuck waiting on XGetWindowProperty again [16:03] what libx11 again? [16:06] Sarvatt: http://paste.debian.net/101876/ [16:06] Sarvatt: the one from xorg-edgers maverick? [16:07] Installed: 2:1.3.4+git20100720.554da76e-0ubuntu0sarvatt4 [16:15] hyperair: you use a dock? [16:15] Sarvatt: yes i do. docky. [16:15] on the left side of the screen like unity? [16:16] yep [16:16] wait, how did you know? [16:16] O_o [16:16] yah looks like its just drawing that, nothing odd there [16:16] you could tell that from the backtrace? [16:16] yep [16:16] hmm [16:16] but it hung there? [16:17] the last thing it drew I'm guessing [16:18] li see. [16:19] how strange. [16:19] i'm not seeing any obvious locking problems at least :( [16:23] =\ [16:23] but XGetWindowProperty shouldn't block [16:23] meaning it's not compiz's fault, or is it? =\ [16:23] hmm [16:24] do you have a backtrace of the hang with libxcb1-dbg installed by any chance? [16:27] nope [16:28] oh yeah, it particularly likes happening when i'm busy killing indicator-applet [17:10] bryceh: been trying for a while to figure out how to get ubuntu-bug to work, tried asking in ubuntu-uk and they told me to use apport-collect which just says "no additional information collected" so I'm kinda stuck with providing the ubuntu-bug info you wanted :( [17:10] been trying to do it since yesterday [17:10] I did file the bug though, https://bugs.launchpad.net/xorg-server/+bug/687412 [17:10] Launchpad bug 687412 in ubuntu (and 1 other project) "X crashes after a while on 9500GT in arecord (affects: 1) (heat: 6)" [Undecided,New] [17:36] bryceh, Sarvatt, where are we with nouveau and 3d ? [17:37] its not going to be there any time soon i assume? [19:16] apw, it depends... we've been discussing it but it's be easier to decide, if upstream was supporting it [19:16] bryceh, does that mean it is even an option? [19:25] bryceh: https://bugs.launchpad.net/ubuntu/+bug/687412 I filed that bug, someone marked it as invalid for some reason [19:25] Launchpad bug 687412 in ubuntu (and 1 other project) "X crashes after a while on 9500GT in arecord (affects: 1) (heat: 6)" [Undecided,Invalid] [19:26] I couldn't get ubuntu-bug to work, spent a day trying and asked in ubuntu-uk, they told me to use apport-collect that didn't work either, not my day :( [19:29] you didn't specify the package [19:29] ah [19:29] don't remember it asking me to specify a package when filing [19:30] I did set it to also effects X.Org though [19:30] it's not a package, but a project [19:30] oh [19:31] tjaalton: how do I set a package? [19:37] choose the ubuntu project and enter the name? [19:38] tjaalton: I clicked also affects project typed ubuntu in the box and it says there is no project named Ubuntu [19:38] I tried using the search, it says there are too many projects with the name Ubuntu [19:38] nbo [19:38] no [19:38] it lists "ubuntu" there because you didn't say which package had the bug [19:38] forget the projects [19:39] :S [19:39] click ubuntu, enter the name of the package [19:39] ah [19:40] dunno what to think of the backtrace [19:40] it's looping somehoe [19:40] and the package name is xserver-xorg, should I change it's status too? [19:40] w [19:40] tjaalton: bryceh said it could be caused by arecord triggering an event which it hooks creating a loop [19:40] iirc [19:41] so the server didn't crash, but it's flooded? [19:41] yes it crashes, I drop back to gdm [19:42] is it immediate? [19:42] tjaalton: nope, it takes a while [19:43] so the real bug is what triggers the loop [19:43] the server then runs out of resources or so. that could be an nvidia bug [19:44] but the backtrace is pretty much useless [19:44] we never saw anything in the logs that it was actually a crash though the other day, its still possible that it could just be some oddball mythtv session stuff dropping it back to login for some reason [19:44] Sarvatt: surely if gdb picks up on the crash it's a crash and not a proper exit [19:45] also this issue doesn't happen with the old graphics card I had in there (6600GT) [19:45] the same driver? [19:46] I think the 6600 uses an older nvidia driver, so before I upgraded I used jockey to deactivate the driver, installed the 9500, then used jockey again [19:46] X sends lots of signals gdb stops on.. [19:48] I could try swapping the 9500 back out with the 6600 [19:48] to see if it still happens [19:49] you said on the bug that you don't know what triggers it? [19:49] that's right, it happens at random [19:49] just like the best bugs should [19:49] haha, I know it's a terrible description [19:50] but it happens when the keyboard is on the other side of the room with nobody nearby [19:50] there's no pattern to it I can see as of yet, I will keep trying to figure it out though [19:50] as I say I plan on stopping x11vnc which should stop arecord from being enabled so I can see if that stops it from happening, and switching back to the 6600 to see if that stops it [19:50] and I want to start logging the times of the crashes [19:51] apw, yes, it's an option [19:55] tjaalton: should I change the status from invalid to new again when specifying a package? [19:58] Azelphur, I already did [19:58] bryceh: ty :D [19:58] it probably got set as invalid because it didn't have a detailed enough description [19:59] yea, I admit the description was crap but I don't have a lot to go on [19:59] getting the files you requested now :) [20:01] we get so many bugs, that developers tend to focus on bugs that have good descriptions and complete log information, and it's fairly plain how the bug comes about [20:07] too bad it's not reliably happening 1 minute after login like it is for these people http://ubuntuforums.org/showthread.php?t=1612704 [20:07] they have nothing in their logs either [20:09] Azelphur: try passing -noxrecord to x11vnc? [20:09] no clue how its starting in mythbuntu [20:09] yea was just about to say I should try that :) [20:10] probably in init.d or something, I'll find it :p [20:11] grep -R x11vnc /etc/ [20:12] it's one attachment per comment right? [20:13] Sarvatt: it's not in there, easy enough for me to find I know the maintainer of mythbuntu quite well, I can prod him :) [20:13] apport-collect -p xorg 687412 ? [20:15] Sarvatt: haha, I tried to use that in the first place but didn't know that I needed to specify a package, that explains why I never got it to work. Running it now :) [20:15] yeah it's not documented very well :( [20:16] or at all [20:16] is it OK to keep getting "Program received signal SIGPIPE, Broken pipe." http://paste.ubuntu.com/541127/ ? it seems to be occurring when opening/closing animations with compiz.. [20:16] seems i am noticing this because i'm debugging X, but during normal usage it would just be like a fraction of a sec stuck and then starting of the animation [20:16] vish: handle SIGPIPE nostop [20:16] Sarvatt: that gives him all the files he asked for, right? [20:17] Azelphur: yea [20:17] cool [20:17] Sarvatt: ah , will all that arg , but is that normal? or should it not be doing that? [20:17] vish: yep its normal [20:17] cool.. [20:17] next on my list is to try with -noxrecord then to see if that stops it from happening. [20:18] vish: you want to nostop on SIGUSR1 and SIGUSR2 too or it'll stop on input events and stuff [20:19] Sarvatt: ok, sure, thanks.. this SIGPIPE seems more common for me. [21:00] Azelphur: https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/525066 [21:00] Launchpad bug 525066 in xorg-server (Ubuntu) "x11vnc able to segfault xorg (affects: 2) (heat: 23)" [High,Triaged] [21:21] Sarvatt: looks like the same crash but mine happens at random while the machine is logged in, nothing to do with the login screen [21:21] I can try to reproduce that bug though :) [21:22] they're just saying to boot it up to the login screen not that the login screen has any significance to it from how I read it [21:22] so they are saying X crashes as soon as the open a VNC session? [21:32] ola [21:32] ingles ou portugues? [21:32] Hi [21:32] English our Portuguese? [21:33] because the intel video card, hangs with compiz? [21:33] 82945G/GZ [22:13] apw: Nouveau 3D status is roughly what it was for Maverick - installing libgl1-mesa-dri-experimental will get you 3D, it'll probably work, but kindly don't bug anyone if it doesn't. [22:14] apw: One difference is that we've enabled the 3D driver for the old cards (TNT2 → Geforce3) in the most recent mesa upload. [22:15] RAOF, thanks [22:16] We should also look at jocky-ifying libgl1-mesa-dri-experimental, so it's presented to users as an alternative to the binary blob. [22:21] RAOF, thats an interesting idea [22:22] Should be relatively simple too, and allow users to try Unity on the livecd. [23:10] Azelphur: I think I might have a patch for you to try [23:16] Azelphur: uploading it to https://launchpad.net/~sarvatt/+archive/green, will take an hour or so to show up [23:17] http://patchwork.freedesktop.org/patch/2166/ [23:23] Azelphur: heading out for the night, let me know how it goes if you get a chance to test it [23:25] because the intel video card, hangs with compiz? 82945G/GZ [23:26] Patricia: That card should work; could you provide more details, like: What Ubuntu release are you using? How often does compiz hang? What, exactly, happens when compiz hangs? [23:27] Ubuntu 9.10 [23:27] our [23:27] Ubuntu 10.04 [23:27] our [23:27] Ubuntu 10.10 [23:27] latch completely [23:27] whatever version ubuntu 9.10 after hangs [23:28] to return the computer, just restarting the reset button [23:29] the period of catch is about 5 seconds after activating compiz [23:29] everyone who has this video card does not work, compiz [23:30] :/ [23:42] I have 2 laptops with (a variation of) that videocard which seem to work okay [23:43] it's probably the most-used GPU ever... [23:43] Unfortunately all they are activating compiz, this fighting, I collect data from other users of that board in #ubuntu-br [23:43] and does not work [23:43] Patricia: is this with the default compiz settings? [23:44] ¬¬ [23:44] or not the time to fiddle around [23:45] you activate the effect and already hangs [23:45] yes default settings