[00:02] Sarvatt: Yup; that got changed recently. People now shouldn't be able to switch from wont fix if they can't switch to won't fix. [00:03] sweet, just wish you could keep things marked fix released at the same status because there are *tons* of people reopening and reclosing the gem memory leak bug [00:04] not saying anything, just changing the status to fix released - fix commited and fix commited - fix released in one go, probably spamming for karma :) [00:08] Sarvatt, yeah I filed a usability bug against launchpad about that one [00:13] good god, I've never seen a bug with so many dupes - #507062 [00:14] sooo [00:14] linux-headers-generic | linux-headers [00:14] nvidia-current requires that [00:14] I *think* there may possibly be a fix for that… why haven't I pointed that bug at my aubergine ppa? [00:15] if someone has -generic headers installed but is using the -generic-pae kernel without -generic-pae headers installed they are screwed :) [00:15] https://bugs.edge.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/569237 [00:15] Launchpad bug 569237 in nvidia-graphics-drivers "package nvidia-current 195.36.15-0ubuntu2 failed to build on pae kernel: Could not determine kernel version" [Undecided,New] [00:15] Sarvatt: Right. I'm pretty sure there is no way of ensuring that the user has the appropriate headers installed through dpkg. [00:16] RAOF, you mean #507062 ? [00:16] indeed, another one for the prop drivers notes [00:17] bryceh: Yeah. #507062. I'm not sure if it's a fix, but the upstream “make the bits of Xlib that are meant to be threadsafe *actually* threadsafe” branch fixes a bunch of similar errors. [00:17] ah [00:17] yeah this sounds familiar, we've had threading issues in xcb before [00:18] So, not at this point a fix that's remotely SRUable :) [00:18] RAOF, shall I assign this one to you? [00:18] bryceh: Yeah, go for it. [00:18] we had one just like this in hardy, I'm having flashbacks [00:19] there were workarounds that could be applied to foss apps, but we never did get the proprietary apps sorted [00:19] and that bug stuck around like a thorn forever. maybe it's still open [00:19] Possibly as a dupe of this one? :) [00:19] * bryceh goes ahuntin [00:21] 419501 maybe? [00:21] “Selected text item” + “Find launchpad bug by nubmer” in Do is Love. [00:21] bug 507062 [00:21] bug #419501 [00:21] Launchpad bug 507062 in libx11 "synaptic assert failure: synaptic: ../../src/xcb_io.c:385: _XAllocID: Assertion `ret != inval_id' failed." [High,Incomplete] https://launchpad.net/bugs/507062 [00:21] Launchpad bug 419501 in gasp-core "apport-kde assert failure: python: ../../src/xcb_io.c:242: process_responses: Assertion `(((long) (dpy->last_request_read) - (long) (dpy->request)) <= 0)' failed." [Critical,Fix released] https://launchpad.net/bugs/419501 [00:21] * Sarvatt just wanted a link [00:22] ah yeah bryceh thats because the error is basically, X crashed underneath me and I couldn't connect [00:22] mm [00:22] bryceh: Yeah. I've posted the PPA on that, and it doesn't seem to totally fix everything :). It *does* fix the “ico -threads 10” testcase. [00:23] ah yes I see [00:24] actually 419501 isn't the one I was thinking of [00:24] however I see shuttleworth asked me a question there that I didn't answer. whoops. [00:24] whoops! [00:25] :) [00:25] xorg-edgers also has all of the libs and xcb updates that were added to possibly fix it, that was the main reason i rebuilt everything in there [00:26] Sarvatt: You built from the branch, or has that been merged now? [00:29] ah no i didnt build the multithreaded xcb stuff, just the mass of libs that got updated to only call XAllocID with the display lock held [00:30] there were 8 or 9 libs that were potentially where the problem was coming from [00:32] heeeeere we go. bug #232364 [00:32] Launchpad bug 232364 in xfce4-utils "dbus-launch hangs at session start waiting on socket output in libxcb" [Critical,Fix released] https://launchpad.net/bugs/232364 [00:32] libXi libXext libXtrap libXrender libXfixes libXdamage libXcomposite -- guess I overcounted [00:34] we figured out that with open source stuff that the order that you start things up could mask the bug (comment #48) so we did that where we could (i.e. no solution for proprietary apps) and left it at that [00:34] they all had non-threadsafe functions though [00:38] http://lists.freedesktop.org/archives/xorg-commit/2010-April/date.html -- search the page for libXcomposite: Changes to 'master' Jamey Sharp [00:38] thats where the updates were [00:39] sheesh, friend can't boot the livecd on 3 of his machines with older nvidia cards, gogo nouveau! [00:42] 7900GS 7800GT 7800GTX all dont work :( [00:45] Sarvatt: That makes me a sad panda. [00:47] Is modesetting so hard that it needs to break across cards in the same generation?! [00:52] those cards aren't particularly old [00:53] i wonder what ben skeggs would say about that [00:54] use fedora, probably [00:54] I wonder if you can boot fedora on a MacBook Pro 5,x :) [00:55] RAOF, only if you contribute to upstream, you insensitive clod!! [01:04] I asked him to boot fedora 13 to see if it worked, I'm going to go out on a limb and say it will since we have a crapload more nouveau backports than they do :) [01:04] err we have less I meant [01:05] * Sarvatt is eyeing the phantom tv out problems on 7xxx series that they are working around.. [01:07] people can criticize ubuntu for that and also poulsbo [01:08] Sarvatt: They don't so much have nouveau backports as they've pulled in nouveau/linux-2.6 :) [02:01] That's a valid point though: we currently don't distinguish between [02:01] "connected but with broken EDID" and "disconnected". I think it sounds [02:01] reasonable to treat the former as connector_status_unknown instead. [02:01] That way, if nothing else is connected, X will still light you up at the [02:01] fallback size. OOOH YES PLEASE. [02:04] *That's* something to track and ensure it gets into Maverick :) [05:01] bjsnider: at least thats a valid criticism, poulsbo isn't exactly something ubuntu specific.. [05:58] hmm, wonder if i should munge the drm symbols and module name for psb-kernel-source so it doesn't interfere [06:08] noone here using psb eh..? [06:08] * Sarvatt wants to make sure the kernel module is right before doing all of the rest of the packages [06:10] https://edge.launchpad.net/~sarvatt/+archive/psb [06:36] oh hmm packaging needs a lot more work, no point shipping psb-modules and building them [16:09] Hmm is anybody running any gallium based DRI drivers on Ubuntu 9.10. [16:09] ? [17:03] there's probably people on phoronix doing it, xorg-edgers doesn't have it at least [17:04] Hmm ok [17:04] I'm just trying to diagnose a very weird bug I'm seeing. [17:04] Thanks anyway [17:06] sorry, I dont have any systems on karmic to test :( === BUGabundo is now known as BUGabundo_dinner [21:08] woo 6 day queue for PPA builds === BUGabundo_dinner is now known as BUGabundo [21:09] Sarvatt: by then people will have started flooding the archive with m stuff, so ppa will fall even more behind? :) [21:10] I know I will be adding to that too :) [21:11] already switched to maverick here [21:12] Sarvatt: maverick! [21:13] Sarvatt: say, how did you get around your initrd+make-kpkg issues? [21:13] * hyperair is compiling a new kernel right now and is interested to know [21:14] sudo update-initramfs -c pointing at the installed kernel and update-grub [21:14] ALREADY? [21:14] think i have to add the initrd line to grub manually too, can't remember since i haven't done it in a few months [21:14] there's no tool chain yet [21:14] is there?!? [21:15] its all copied over from lucid right now [21:15] ahhh [21:15] * BUGabundo seds source.list [21:15] here I go [21:15] Sarvatt: lolwut [21:15] BUGabundo: er there's no maverick in the repository.... [21:15] Sarvatt: oy april fools was last month. [21:15] http://archive.ubuntu.com/ubuntu/dists/maverick/ [21:15] whut [21:16] eheeh [21:16] yes there is [21:16] you have to be kidding m [21:16] e [21:16] I changed to lucid on day two [21:16] bah! [21:16] 36h after karmic was released [21:16] so I'm on time to move to maverick [21:16] pfft [21:16] my mirror hasn't updated. [21:16] very few will [21:17] only archive.ubuntu.com yeah [21:17] I must use main until alpha1 [21:17] heh? [21:17] why wouldn't thye? [21:17] hyperair: when you do it that way you have to manually remove the /boot/initrd-whatever for that kernel when you remove the package too [21:18] don't they all rsync from archive.ubuntu.com? [21:18] nope [21:18] Sarvatt: oh hell. that sounds really bothersome >_> [21:18] very few sync devel archive [21:18] probably needs manual intervention to add the new release to each mirror, no idea though [21:18] Sarvatt: should I keep a lucid repo for security updates [21:18] it's only been up a few hours [21:18] or will everything be pushed to maverick repo ? [21:19] i just changed the main ones to maverick and left all the -updates -security and stuff on lucid, wont be anything for a few days probably on maverick :) [21:19] okay [21:19] its what I usually do [21:19] I run both repos for several weeks [21:20] then I must take them out [21:20] on 7.x something I had clash of packages deps cause of that [21:20] but I've learned to deal with most of them, by now [21:20] looking forward to gcc 4.5 though with -march=atom support for kernel building [21:20] heh [21:20] my mirror has it http://mirrors.fe.up.pt/pub/ubuntu/dists/maverick/ [21:21] hyperair: honestly it's probably easier to just learn the ubuntu kernel build system for those mainline kernels, thats what i plan on doing next time i need a kernel built. theres a kernel-team git repo with the scripts to build them and doesn't look that complicated [21:22] Sarvatt: er actually what i'm worried about is disk space. [21:22] Sarvatt: the mainline kernels compile with debug symbols, then strip them [21:22] that alone can blow my /home to hell [21:22] you compile kernels on your encrypted /home?? [21:23] that sounds painful! [21:23] yeah thats true though, i wasn't thinking about that and it'll be a problem here too [21:23] 0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded. [21:23] only have 10GB on my VPS where I do it and make-kpkg already uses 4.5 or so [21:23] boooo [21:24] Sarvatt: er everything's encrypted and stuffed in lvm [21:24] Sarvatt: you don't mean to say you compile everything on tmpfs [21:24] probably adding 3x to the compilation length there :D [21:24] meh [21:25] i've got bfs [21:25] so kernel compilations do nothing but heat up the comp [21:25] compiz, X, and everything else runs as smoothly as ever \o/ [21:30] BUGabundo: oh yeah the maverick kernels have been coming out for awhile now too but just in a PPA - https://edge.launchpad.net/~leannogasawara/+archive/ppa/+packages [21:30] fine by me :) [21:30] nvidia aint going to work for awhile though no doubt [21:31] let me know when you guys have nvidia blobs or nouveau to test [21:31] hyperair: http://kernel.ubuntu.com/git?p=ubuntu/kteam-tools.git;a=summary theres those scripts that make the mainline daily kernels [21:39] I really hope we get better IO with the new kernel [21:40] the changes done in the last part of the cycle to lucid [21:40] killed all my IO boost [21:40] copying stuff is very slow [21:40] I dropped from over 55MBs ovver e-sata [21:40] with now impact to my system [21:41] to 26-30MB/s with my disk IO rendering the system useless === BUGabundo is now known as BUGabundo_bones === BUGabundo_bones is now known as BUGabundo_Bones [21:50] BUGabundo_Bones: use ionice [21:51] BUGabundo_Bones, since when does esata work? [21:51] BUGabundo_Bones: and btrfs, or something that doesn't suck as hard with fsync. [21:51] hyperair: I already do [21:51] =\ [21:51] bjsnider: for ever? never had *any* prob with it [21:52] other then being asked GKSUDO for it [21:52] huh? gksudo? [21:52] yeah to mount it [21:53] er don't you always need root perms to mount anything that's not an external device? [21:53] just stick it in fstab [21:54] no [21:54] usb sticks automount [21:54] no prob there [21:54] my e-sata disk, all stay umounted [21:54] and clicking on the icon, to mount, asks password [21:54] so stick it in fstab [21:54] why should a regular user need root permission to mount a filesystem? [21:54] but why would I need it for some, and not for others [21:54] no idea [21:55] I filed it a while ago [21:55] no one ever touched it [21:55] I stop caring [22:00] well, at least it's not as bad as fedora, where you have to type in a password every time you move the mouse [22:01] i just chmod/chgrp the mount point to my account [22:03] ahahhahaahahaha [22:05] giving the regular user ownership over /media would solve some of the mount permission annoyance i guess [22:06] I can try [22:06] * BUGabundo_Bones goes hacking FS [22:07] $ sudo chown bugabundo.bugabundo /media -R [22:07] chown username:groupname /media [22:08] oops i didn't mean chmod there [22:11] vish: looks like your problem isnt the same as any of these bugs [22:12] would changing /media's ownership persist after a reboot? [22:13] vish: 2.6.33.3 didn't fix anyone else on any of the bugs and no one else in them is having the problem on lvds it seems as far as I can tell (they 5 bugs I'm looking at are all external monitor related) [22:14] bjsnider: ill know tomorrow [22:37] Sarvatt: yay i got it working, with update-initramfs and all. just ln -s everything from /usr/share/kernel-package/etc/kernel/*.d [22:37] * hyperair is off to reboot into new kenrel [22:37] hyperair: woohoo! [22:38] it's funny i've been using kernel-package 12.x for over a year and still haven't set it up right :) stopped caring about building my own around the same time [22:39] if i used my turion machine more i'd be doing it because PHC rocks [22:39] undervolting atom as low as it'll go makes almost no difference [22:41] i'll be glad pcie aspm will finally be enabled in MM though, that and force quirking my ata controller into SATA mode were the main reasons I built my own kernels but I figured out how to just edit the efi variables in the bios to change it to SATA mode directly [22:45] psb is close - http://pastebin.com/yjDNXS0u [22:47] Sarvatt: I die a little bit inside every time somebody says psb. === BUGabundo_Bones is now known as BUGabundo_Naruto === BUGabundo_Naruto is now known as BUGabundo