Sarvattvish, that was fast - http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/commit/?id=77b13a02c70842a58e0590d0243f0ae016c5a64001:38
Sarvattlet me know if its working for you01:38
tjaaltonhmm, do we even have an fglrx that works in lucid?08:29
tjaaltonkinda funny to ask people test it..08:29
vishSarvatt: sure, it seems to be working fine.. [although i didnt know what triggers it to happen] I'll watch the gdm log for a week for any activity09:09
vishalso , shouldnt the errors go to Xorg.0.log ? i think that would need to be corrected too?09:10
* Ng spies an ubuntu job advert for an X integration engineer11:39
Sarvattyeah i was thinking the thing tjaalton :D14:06
Sarvattlibdrm_radeon and libdrm_nouveau are missing from ia32-libs14:46
Sarvattbryceh: http://paste.ubuntu.com/367635/ :)15:10
Sarvatt(with 100_rethrow_signals.patch)15:10
Sarvattuncommented line 302, stops segfaulting when things FatalError  http://git.debian.org/?p=pkg-xorg/xserver/xorg-server.git;a=blob;f=debian/patches/100_rethrow_signals.patch;h=97545e5e5e75151537e40ccd611ae3c5067d0797;hb=refs/heads/ubuntu15:14
tjaaltonSarvatt: yeah, maybe I should merge mesa/libdrm so -ati can be synced. then we'd have preliminary r800 support..15:56
=== yofel_ is now known as yofel
hyperairhmm ubuntu-x-swat regression testing eh16:22
hyperairbut nvidia geforce 2 and above?16:22
hyperairare we testing nouveau?16:22
hyperairi mean nvidia hasn't exactly been very nice with their -96 drivers.16:23
jcristaucatching fglrx regressions early.  is that a joke?16:23
hyperairthey claim they maintain it, but it's still a pile of crap that makes my screen quiver like jelly.16:23
bjsnidernvidia lacks the resources to properly maintain the old drivers16:31
Sarvatttjaalton: you dont even need to update libdrm or mesa, the r800 support is 2D non KMS only. i doubt r800 dri will make it until mesa 7.9 even since 7.8 is going to branch in a week or two to get ready for release17:05
Sarvatti was just guessing they have a fglrx that works with xserver 1.7 for people to beta test, don't know why they said geforce 2 or higher also since there is no nvidia proprietary driver that works for geforce 2's on xserver 1.7 as far as I know17:08
Ngask ara :)17:08
hyperairtoo late.17:09
hyperair8 seconds too late ;-)17:09
hyperairi can't wait for nouveau to spawn proper 3D support for nv1X17:09
hyperairthen i won't have to put up with a jelly-like screen17:10
Sarvatti dont think it ever will?17:10
Sarvattthey were doing mesa classic for that17:10
hyperairit said it was in progress17:10
hyperairi mean..17:11
hyperairin the feature matrix17:11
Sarvattah I mean I dont think it'll be gallium based like nv20+ ones are, they're doing a classic mesa driver for those old ones as far as I know17:15
jcristauSarvatt: yep, looks like they're moving that way17:16
hyperairi see.17:17
hyperairbut why separate them?17:17
hyperairare the nv10+ ones so old that gallium just won't work or something?17:17
* hyperair sighs =\17:18
Sarvattits a shame they sold nv1x cards for so darn long, people were buying up those geforce mx's for years and years after they were junk :D17:19
hyperairheh lol17:20
hyperairi think mine was bought while it was still one of the better cards around17:20
jcristaui don't even know what family mine is.  10de:018a.17:21
jcristaubut then it's old enough that it may not see something newer than lenny, so.17:21
indusis this the ubuntu X team?17:28
tjaaltonSarvatt: oh right, -ati only depends libdrm 2.4.17, not -117:39
indusi read on phoronix about the fglrx driver and i would like to test it17:44
indusi have a radeon 485017:45
superm1i dont think that article is accurate17:47
superm18.660 is the same driver that was in karmic17:47
indusi read the mailing list 17:47
indusfrom ara17:47
superm1did ara ever say there was a new driver though?17:48
industhis one https://lists.ubuntu.com/archives/ubuntu-qa/2010-February/000775.html17:48
indushmm i guess no17:48
araindus, can you send me an email, instead, please17:49
indusara, yes sorry i wil;17:49
araindus, thanks17:49
indusbut just want to make sure if this is for lucid17:49
indusok nvm ill send an email17:50
araindus, yes ,it is for lucid17:50
indusok mail coming up17:50
indusara, sent17:54
araindus, thanks, I will get back to you later this week17:54
indusok no problem17:55
indusbye then17:55
industhank you17:55
indusone question about the open source driver, does it support the 4000 serries ATI17:56
indushmm didnt think it did, i mean straight from the lucid repos or with ppa17:58
jcristauno ppa.18:00
indusok cool i test lucid today then18:00
indusbye and thanks18:00
tjaaltonara: so there's a new fglrx for the testers to try?18:27
aratjaalton, not sure, but the point is that they don't get properly tested, anyway18:28
tjaaltonara: the current one doesn't work with the xserver in lucid18:30
tseliotara, tjaalton: there's no new fglrx to test yet but we'll have something in time for the release. I don't think I'm allowed to say more about it18:56
tjaaltontseliot: too late for testing though19:00
tseliottjaalton: I can't tell you when but yes, there might be little time to test it19:01
tjaaltonsince I doubt they'll be able to fix it in time when issues pop up19:01
tseliottjaalton: well, they have their own schedule..19:07
tseliotbut of course I agree with you on this19:07
tjaaltonslackers :)19:07
bryce_tjaalton, yeah the purpose of the testing is not to do a one-off test of the latest, but rather to do regular weekly testing, so if some change elsewhere in the system breaks the proprietary drivers, we'll have a chance to catch it19:29
bryce_tjaalton, this is to help us avoid the situation we hit last release where we didn't really notice -nvidia had been badly broken by the upstart changes until the week after release19:29
bryce_so it's good they'll start testing before we get new -nvidia/-fglrx, so they can get a good baseline to compare against19:30
bryce_ara, I noticed that this morning 5 people joined the Ubuntu-X team :-)19:31
arabryce_, I will pass you later on the list of volunteers... it's crazy... I have been all morning processing emails19:31
bryce_ara, good work :-)19:33
arabryce_, hopefully some of them will stick around :D19:33
tjaaltonbryce_: ok19:33
bryce_tjaalton, btw for the shlibs stuff, what would be a good example package for me to look at?  I was modeling it after libxrandr, is there a better package to look at?19:34
Duke`grumbl... google street map is really a killer feature19:35
tjaaltonbryce_: libdrm19:35
Duke`"killer", that's the word19:35
tjaaltonbryce_: but maybe it's overkill if libeagle is going away anyway19:36
superm1tseliot, have you started any of the work to port fglrx packaging over to the new way of doing things?  I made some changes on phorogit, so just wanted to make sure you are set up and good to go with operating on there so we dont step on each other's toes if one another does a few more pieces19:52
tseliotsuperm1: not yet. I'm in Portland at a sprint and I should be able to do it soon20:06
superm1tseliot, okay well just make sure to check phorogit in case i get any of that in order before you do20:07
superm1i switched the package name over to fglrx already20:07
tseliotsuperm1: also, I will make additional changes to both jockey and nvidia which will apply to fglrx too20:07
tseliotsuperm1: sure20:07
tjaaltonwhy does jockey offer fglrx?21:02
tjaaltoneven if it doesn't work21:02
tjaaltonthere are several bugs against xorg-server because of that21:02
tjaaltonlike bug 50886021:03
ubottuBug 508860 on http://launchpad.net/bugs/508860 is private21:03
tjaaltonbug 50886021:03
ubottuBug 508860 on http://launchpad.net/bugs/508860 is private21:03
tjaaltonlike hell it is21:03
bjsnidertjaalton, getting a bit upset with fglrx?21:07
bjsniderit's the best option, after all of the others21:08
tjaaltonbjsnider: eh? I just don't like seeing dozens of crashers against xorg-server21:10
jcristaupretending fglrx doesn't exist works fine for me so far :)21:10
superm1tjaalton, maybe it's a good idea to improve the xorg-server apport hook to go and change the package to fglrx-installer immediately if it sees *fglrx* installed on the system21:11
superm1and likewise for nvidia21:11
superm1that'll at least keep the bug reports clean21:11
tjaaltonsuperm1: true21:12
tjaaltonsuperm1: though now you can have them both installed at the same time, so why bother? :)21:14
superm1haha. well then at that point - go and query if fglrx is modprobe'd or nvidia and adjust package accordingly21:15
tjaaltonyeah that's a better metric21:16
tjaaltonsame goes for the virtualbox crashers21:26
tjaaltonthere are maybe hundreds of dupes21:26
tjaaltonmost moved to vbox of course21:26
superm1tjaalton, okay well just piggyback that on https://bugs.edge.launchpad.net/ubuntu/+source/xorg-server/+bug/51626421:28
ubottuUbuntu bug 516264 in xorg-server "Xorg-server apport hook should detect fglrx/nvidia" [Undecided,New]21:28
superm1might as well solve all those in one swoop if possible21:28
tjaaltonsuperm1: done21:39
bjsniderdoes vbox take down the kernel that much?21:41
superm1regardless of how often it's taking down xorg-server, if it's providing external pieces in, it shouldn't be lodging it's bugs in xorg-server's set of bugs21:44
RAOFbjsnider: You were after 3D support in nouveau for your ancient nvidia card?  The DDX already registers a different glx provider on those old cards, and it looks like a sorta-functional nv0x - nv2x driver is on its way to be merged back into mesa.21:51
bjsniderRAOF, i wasn't21:52
RAOFAh, sorry.  'twas hyperair.21:52
hyperairRAOF: =O21:53
hyperairthat's great news!21:53
hyperairwhat's DDX?21:53
RAOFThe X component of the driver stack.21:54
RAOFCurrent status of that 3D is apparently, and I quote: “However the killer feature is 'it actually draws stuff'”21:54
hyperairwell isn't that nice.21:55
RAOF(And some simple games may even run at a decent FPS)21:55
hyperairoho, really?21:55
hyperairit won't be outclassing my i965 anytime soon eh =p21:56
bjsniderit wouldn't be very fast even if the code was a perfect as if god has written it, i imagine21:56
hyperairwell if compiz works i'm happy.21:56
hyperairor would that be expecting too much? =p21:57
bjsnidercompiz, that thing that's being replaced by gnome-shell?21:57
hyperairgnome-shell's not replacing compiz anytime soon on my macine(s)21:58
RAOFGiven that the much, much better developed nv4x & nv5x gallium drivers *mostly* run compiz, I think you may be expecting a liiiitle too much.21:58
hyperairRAOF: heh figures. i'll just scrap GNOME and jump to LXDE or something on that machine.21:58
hyperairbjsnider: i don't suppose nouveau would run the shell either, would it?21:59
bjsniderit's opengl either way21:59
bjsniderit's  not xrender21:59
hyperaircompiz is opengl too, isn't it?22:00
bjsnideryeah, that's what i mean22:00
hyperairso both won't work eh22:00
hyperairwell, at least compiz++ brings support for crippled drivers22:00
hyperairgnome-shell doesn't, and probably will never22:01
hyperairnot to mention that it eats eats more RAM than firefox does >_>22:01
bjsniderhow do you know?22:01
hyperairbjsnider: iirc there was an argument on some mailing list or other.22:02
hyperairsomething about the shell not getting a backward compat mode, because of the way it's designed, blah blah blah drivers should have been fixed by then, blah blah something about we shouldn't hold ourselves back because of a few crap drivers22:03
bjsnideri don't think anything's written in stone22:03
hyperairbut the most fun part is how they keep convincing themselves that compiz is a pile of bloated crap that brings nothing but eyecandy and no usability features when gnome-shell takes up 5x the amount of memory compiz does.22:04
bjsnidercompiz is bloated crap22:04
bjsniderthere's no question about it22:05
hyperairwell then, so is GNOME.22:05
hyperairand the shell is 5x more bloated than compiz, so where does that put it?22:05
hyperairbloated, incomplete, unusable crap?22:05
bjsniderit's certainly incomplete22:05
hyperairright, and while it's this incomplete, it's already eating up 5x the amount of memory compiz does.22:06
bjsnideri'm using it right now, so it can't be considered unusable22:06
hyperairwhen it's complete, how much will it take?22:06
hyperairright, so it's "usable", but nowhere near the usability compiz gives me.22:06
jcristauhyperair: when it's complete you'll have 1TB memory.  and a flying car.22:06
hyperairjust for the record, i like having my screen edge triggers at the bottom of my screen.22:06
hyperairjcristau: that sounds about right, yeah.22:06
bjsniderusability, like, animated transparent gears, snow, and fire on the desktop?22:07
hyperairjcristau: no actually, i don't think it'll ever be completed. halfway along the line, they're going to decide that it's become very unmanageable, akin to gnome-panel, and scrap it for a... GNOME SHELL2!22:07
hyperairbjsnider: there. that's the problem with you gnome-shell guys.22:07
hyperairbjsnider: when i say usability, you pick the most useless plugins out of compiz and slam the entire compiz for it22:07
hyperairbjsnider: open your eyes damnit.22:08
bjsniderso you admit that there is feature bloat in compiz22:08
hyperairbjsnider: i disabled those plugins.22:08
hyperairbjsnider: but that's all they are. "plugins"22:08
hyperairif i toss them out the window, it means that they aren't running, they don't eat up space on my hard disk, and all in all, there is no bloat.22:08
jcristaua few compiz plugins should be builtin, the rest should be thrown away.  and then you get something halfway sane.22:09
hyperairi like my scale plugin.22:09
hyperairit beats anything gnome-shell can offer me at the moment.22:09
bjsnideri can't make out the useful ones from the bloat because they're all piled in there together22:09
hyperairbjsnider: that's why i told you to open your eyes.22:10
hyperairthe more you look at those crap demo videos uploaded, the more you think "compiz is all bloat"22:10
hyperairbecause those demo videos attempt to showcase all these useless effects while not actually showcasing any of the really useful ones22:11
bjsniderit's not all bloat, only 95% of it is bloat22:11
hyperairbjsnider: better than gnome-shell, which can't even match up to 5% of compiz's functionality.22:11
hyperairassuming your figures are correct22:11
bjsnidersure it does22:12
hyperairoh right, if you want to count in memory usage, i suppose gnome-shell achieves 500%22:12
hyperairbecause some random joker decided "let's use javascript!"22:13
hyperairand everyone was drugged so heavily that they all agreed >_>22:13
bjsniderthere's a gnome-shell irc channel where you can voice your complaints directly to the developers...22:14
hyperairi nearly got kicked for "trolling"22:14
bjsnideri don't know *why* they would think that22:15
RAOFFWIW, I quite like many of the ideas embodied in gnome-shell.22:15
hyperairit's utterly ridiculous. they'll bar out your choices, in favour of theirs, and then when you don't agree with it, they blame you for trolling and ignore you.22:15
RAOFI don't think that's an accurate description; certainly not of the mailing list.22:15
hyperairRAOF: that was on IRC.22:15
CobaltEvery new idea is not necessarily a good idea. Gnome-shell stinks a bit.22:16
bjsnidernew features have been added that you can check out in rico's ppa22:16
RAOFUbuntu-X: All gnome-shell, all the time. :/22:16
bjsnideri think that's the one22:16
bjsniderwell, they kicked us out of the gnome-shell channel because they thought our hectoring and badgering was trolling, but we really have good ideas22:18
hyperairdon't get me wrong, i love GNOME and it pains me to have to move to anywhere else, but i want my panel, and i want my compiz. otherwise you let me have back my bottom screen-edge triggers, my infinitely customizable keybindings, and something akin to scale + scale addons + scale text filtering22:18
Cobalthyperair: Problem is, they'll dumb it down enough to the point of not giving you a choice of having it back.22:18
hyperairCobalt: yeah, exactly.22:19
CobaltAnd bill it to the price of progress. And brilliant new ideas. Of how to interact with your data. Never mind that over the years you already streamlined the way you do things, and are quite happy for things to go on the same way.22:20
bryce_ tjaalton, superm1, I added a thingee to apport to detect if the user is running vbox and if so tell them the issue is unreportable22:20
bryce_tjaalton, superm1, that's been in there for some months, so if we're *still* seeing vbox bugs filed against xorg-server we should see how those are coming in22:20
bryce_I agree we should exclude them22:20
RAOFbryce_: Did you get in contact with nvidia re: nouveau ctxprogs?  Have you heard back?22:31
bryce_RAOF, not yet, I did just touch base with a couple of the kernel engineers22:31
RAOFDid they have anything interesting to say?22:33
RAOFI'm currently preparing something approximating a debian/copyright for a nouveau-firmware package to add to the testing packages in xorg-edgers/nouveau.  Should I stop?22:37
bryce_RAOF, the kernel team is looking at how it should be packaged, probably in some separate package.  andy will get back to us on it22:37
bryce_RAOF, continue on with it22:38
bryce_we're not taking it as a blocker at this time22:38
RAOFOk.  Once I've got that bit, I was thinking of writing a call-for-testing to ubuntu-x & ubuntu-devel; those bits in xorg-edgers/nouveau should (a) not break anything on non-nvidia hardware, and (b) work.22:39
bryce_that sounds great22:40
bryce_RAOF also andy said he can update the kernel22:40
bryce_we also talked about supportability post-release22:40
bryce_he said that the kernel team has an exception from having to do srus for linux-backport-modules22:41
bryce_so if we take the approach of having nouveau in l-b-m then it gives the flexibility of putting out new git snap shots of the driver post-release if it looks sane enough to do so22:41
RAOFThat would be good.22:42
bryce_I also asked if we can also include the nvidia binary package from l-b-m on the livecd, and he said that should be doable; we'd need to go through the foundations team for that22:43
RAOFSo we wouldn't have to try to backport commits.  Great.  I notice that there are a bunch of nv5x fixes pending on the nouveau mailing list already :)22:43
RAOFWould that mean that l-b-m-nouveau would be pulled in by the kernel metapackage?22:43
RAOFOk.  I'm going to stop stressing about debian/copyright and just put in all the information I think is pertinent.  Someone else is going to need to go over it anyway!22:47
bryce_sounds good22:47
bryce_thanks for raising it as an issue, I've added to my todo list to follow up on it later on and make sure everything's square22:48
bryce_RAOF, mirco's borrowing the desktop today but I think tomorrow I'll get nouveau going on it again and re-test stuff.  I'd really like to get compiz running successfully, and sort out why it's not vt-switching22:52
RAOFbryce_: You'll need newer dri2 protocol headers if you want to build mesa from git (and it's really the only way to fly ;)).  If it's a GeForce8 or newer you'll also need the nouveau-firmware package before any acceleration works.22:54
bryce_ok thanks, yeah it's a G86, so perhaps that's why compiz failed22:56
RAOFYes.  It might also be why VT switching is failing, but I'd hope not.22:56
bryce_tseliot, <brian> bryce_: so my the usb id for my tablet isn't in 69-xserver-xorg-input-wacom.rules fwiw23:13
tseliotbdmurray, bryce: does it work correctly if you add your id in the udev rule?23:15
bdmurraytseliot: I haven't tried yet, add it and then what?23:16
superm1bryce_, not sure if the vbox stuff is still showing up.  just know the fglrx/nvidia is23:17
bryce_superm1, yeah I didn't fuss with that in the apport rule23:18
tselioton next boot (as you can't unplug it) the device should be detected and should show up in "xinput list"23:18
tseliotand work23:18
bryce_however I do have a script to process the xorg queue and look for nvidia/fglrx bugs and move them to fglrx-installer or nvidia-{mumble}23:18
jcristauyou can fake unplug it23:19
bryce_superm1, I think we could probably improve the apport script itself to file the bug directly23:19
jcristauudevadm trigger --action={add,remove} or something23:19
jcristauor stuff an event in sysfs23:19
superm1bryce_, so can't you just change the Package in the crash report in those scenarios?23:19
bdmurraytseliot: so what would I use for SYMLINK in the udev rule then?23:20
bryce_superm1, yep that's exactly what I meant23:21
superm1cool, glad we're on the same page :)23:22
tseliotbdmurray: I guess "input/tablet-$MODEL_NAME" should be fine23:22
tseliotwhatever your model iss23:23
tseliotor you can simply call it "input/tablet-brian-test"23:23
bdmurraytseliot: no change afaict23:45
tseliotbdmurray: if that didn't do it, you might want to ask the kernel team about it23:52

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!