Sarvattnvidia 260.19.x seems to have killed most gt310 and gt330 users, and the CustomEDID option some people with optimus are using is broken to boot :(03:22
RAOFsarvatt: That's re: the current #ubuntu-desktop discussion?03:23
Sarvattoh? didn't see, I just see tons of bugs and posts on nvnews about it03:23
Sarvatteveryone seems to be on one specific model sony vaio actually03:24
Sarvattah the vpc f11 bugs are all using CustomEDID which is hanging03:25
RAOFNot bug #65544603:25
ubot4Launchpad bug 655446 in xorg (Ubuntu) "Xorg won't start (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/65544603:25
Sarvattlooks like its all vaio gt3xx's, vaio z which needed a huge amount of hackiness to even use 256.xx is broken too03:45
bjsniderthere are also a lot of +1 channel discussions from blob users who have a black screen on boot03:46
Sarvattthese crappy vaio's are popular machines it looks like :(03:47
bjsnideri wonder if nvidia is going to release another blob in the very near future03:48
Sarvattvaio gt2xx's had the same problem dating back a year ago too and it was only fixed in 256.44 - http://www.nvnews.net/vbulletin/showthread.php?t=14048203:50
bjsniderperhaps the 256.53 would be the lesser of two weevils03:50
Sarvatti wonder if the fix didn't make it to 260.19.x because they developed them in parallel and 256.44 was newer or something03:50
RAOFOr we could just blacklist it, and give them vesa!  That'll be popular :)03:51
bjsnideri don't know how you go back to the 256 now that you've already got the 26003:51
RAOFThe 260 also (apparently) fixes a pretty major performance regression in text rendering, so it's not exactly a win-win.03:52
RAOFIt'd technically be very easy to go back to 256.  :)03:52
bjsniderthat would be ugly03:53
Sarvatt256 would mean backporting 260 with the fix wouldn't be possible and we'd be stuck with the slow 256's03:54
bjsniderbut then a whole bunch of people wouldn't be able to use it at all03:54
bjsniderthat's one way to get people on the nouveau driver03:55
SarvattRAOF: if nouveau works I think I'd lean towards keeping 260 and updating later for sure because the impact is limited to vaio's with quad core cpu's, dual core optimus ones didn't work out of the box as it was anyway04:01
Sarvattfor gt3xx generation gt2xx generation did and it is a regression there though04:02
Sarvattugh :(04:02
bjsnidergt2xx can't be all bad. i'm ok on it04:03
RAOFBut do you have a VPCF11 sony viao laptop?04:03
Sarvattjust vaio04:03
bjsnideryou mean all the people who have come into the +1 channel today bellyaching about this are using vaios?04:03
SarvattVPCCW is the big gt2xx vaio fixed in 256.44 and broken again in 26004:03
Sarvatti *only* see reports from vaio users04:04
RAOFsarvatt: I don't think there's any way we'd drop back to 256 on the basis of a specific single make of laptops having a problem.04:04
bjsniderwell, i'll ask the ones tomorrow about it04:04
Sarvatthere's a Sager Np5125 but thats an optimus laptop and is broken because the CustomEDID option is broken 04:08
hallynis https://help.ubuntu.com/community/BinaryDriverHowto/Nvidia/Nouveau still the right way in maverick to try out nouveau if i'm currently using nvidia drivers?13:46
hallyn(only have one laptop handy so prefer to keep experimentation to a minimum right now, but OTOH i'm getting these random 0-10 second delays in screen updates that are just not conducive to getting work done)13:47
Sarvatthallyn: to use nouveau you just disable the binary drivers13:48
Sarvattif you want 3D support with nouveau you install libgl1-mesa-dri-experimental13:48
hallyn'disable' the binary drivers by removing the nvidia-glx package?13:50
hallyn(sorry, the whole thing about nto having an xorg.conf just has me all confused these days :)13:50
Sarvatthallyn: in system - administration - additional drivers13:57
Sarvattthe disable button next to nvidia-current there13:57
hallyndo you know the name of the program that brings up?13:58
Sarvattyou can just purge nvidia-current and remove your /etc/X11/xorg.conf by hand if thats easier14:01
hallynthanks - i'm running jockey-gtk right now.  didn't see a disable tickie-mark, so i hit 'remove'14:01
hallynmaybe not what i wanted :)14:01
hallynwas sort of hoping for a symlink i could redirect14:01
Sarvattyou can switch with update-alternatives/ldconfig but its a pain in the rear14:03
hallynSarvatt: cool, seems to be done, i'll re-log-in in a bit to test - thanks!14:03
Sarvatthallyn: thanks, let me know if it works out because I'm curious14:03
bjsniderSarvatt, i've got a guinea pig here who's got a blob issue but not on a vaio14:17
Sarvattbjsnider: 256.xx was fine for them and 260.xx doesn't work?14:18
bjsnideri'm trying to get that info14:18
Sarvattbecause i'm not doubting there's blob issues14:18
hallynSarvatt: worked out great, actually.  In fact, now I get the high-res console to see the rest of bootup messages before x actually starts.  (with glx it just goes blank).  and (crosses fingers) so far none of the random hangs.  still have my 1600x900 resolution, not using any 3d right now so not sure if that's working14:55
Sarvatthallyn: need to install libgl1-mesa-dri-experimental to get accelerated 3D in case you want that :)14:55
Sarvattit may have problems and isn't supported by upstream yet so its not installed by default but it works great here14:56
hallynSarvatt: yup, i installed that.  just don't have any 3d apps ahandy :)14:57
hallynoh, look at that, got a hang14:57
Sarvatthah! :)14:57
* hallyn starts to wonder if xterm is the problem14:57
Sarvattif you install it compiz starts by default and that may be a problem14:58
hallynno, i'm running wmii...14:58
hallynat first is uspected compiz (i was using that with glx very pbriefly)14:58
hallynthis is crazy.  switching to gnome-terminal...14:59
Sarvattare there any errors in dmesg?15:00
Sarvattworst case scenario, you may need to boot with nouveau.noaccel=1 added to the kernel command line15:00
hallynSarvatt: but it also happened with nvidia... still worth trying noaccell in that case?15:01
Sarvattoh? odd..15:01
hallynnothing in dmesg :(15:01
* Sarvatt isn't familiar with wmii15:03
Sarvattwhat hangs, just individual terminals or the whole screen? how do you recover?15:03
hallynit just recovers after awhile.  often jsut the terminal hangs, i *think* sometimes the whole screen does, but i could be wrong about that15:03
hallynSarvatt: well, it's pretty random, so it not having happened doesn't mean much, but it really may be xterm hanging15:20
Sarvatthallyn: a wmii problem would be my first guess, i see a few bugs on their tracker about some apps hanging with a blank window that might be related15:21
hallynSarvatt: no, it happens with gnome+metacity as well15:23
hallynheck, could be the keyboard driver for all i know15:23
hallynneed to write a little app that read kbdinput and prints out timestamps next to each char it reads15:24
hallynthen have my kids sit there and type 2 chars per second for a few minutes :)15:24
hallyna lego robot, maybe15:24
CobaltHey guys. I just got a Apple Magic Trackpad. I was wondering how I could force it to use the Synaptics driver in Xorg.16:14
bjsniderSarvatt, the guy i was mentioning earlier has an 8400 gs and it won't work with the 260 blob16:28
SarvattCobalt: sudo mv /usr/share/X11/xorg.conf.d/60-magictrackpad.conf{,.bak}16:30
jcristauor just override it in xorg.conf16:30
CobaltSarvatt: Sorry, on Karmic.16:30
Cobaltjcristau: That's what I wasn't too sure how to do. The Ubuntu Wiki had a blurb on adding a section in Xorg. They edited out of the page yesterday and I can't find it anymore.16:31
jcristaui don't remember what was in karmic.16:31
CobaltI don't want anything fancy, I was just wondering if I could get single-click to work.16:31
Sarvattnot sure about karmic, i'm surprised it works at all there16:32
CobaltIt points, but no click actions except with the nubs. And only button1 too.16:32
CobaltHence my wish for exploring a bit further. Oh I will upgrade, but since everything else works fine, I'm not in too much of a hurry.16:32
SarvattCobalt: it might just work™ if you try a maverick kernel but that might cause other problems16:34
CobaltI tried a Lucid kernel on this once. It messed up the wireless pretty bad. Eee PC, you know. I have reservations about a bolt-on Maverick kernel.16:35
Sarvattyeah rt2600 is no fun16:36
CobaltWhat's the command to dump a sample Xorg config file to standard output?16:36
CobaltWell, RT nothing is fun, actually. I've had only grief with their products.16:37
Sarvattsudo X :1 -configure ?16:39
CobaltAh yeah, just Googled that actually. Seemed a bit too simple. Gah sometimes, something obvious is just too obvious.16:39
Sarvattshould be one at ~/xorg.conf.new after that16:39
raymondjtoth2any one know how to install Mesa 3D GL driver for intel grahic16:47
raymondjtoth2i never did icpiling like to get all new stuff i just install Mesa 3D GL driver16:47
raymondjtoth2any one here16:48
raymondjtoth2any one know how to install Mesa 3D GL driver for intel graphic card16:52
raymondjtoth2i did install Mesa 3D GL driver16:53
raymondjtoth2i mean16:53
raymondjtoth2i need help install mesa 3d gl driver i did install libgl1-mesa-dri-experimental16:53
raymondjtoth2any one16:54
raymondjtoth2 any one know how to install Mesa 3D GL driver for intel grahic16:56
raymondjtoth2<raymondjtoth2> i never did icpiling like to get all new stuff i just install Mesa 3D GL driver16:56
raymondjtoth2<raymondjtoth2> any one here16:56
raymondjtoth2<raymondjtoth2> ?16:56
raymondjtoth2 i did install libgl1-mesa-dri-experimental16:56
raymondjtoth2<raymondjtoth2> any one16:56
raymondjtoth2any one17:02
raymondjtoth2 any one know how to install Mesa 3D GL driver for intel grahic17:03
raymondjtoth2<raymondjtoth2> <raymondjtoth2> i never did icpiling like to get all new stuff i just install Mesa 3D GL driver17:03
raymondjtoth2<raymondjtoth2> <raymondjtoth2> any one here17:03
raymondjtoth2<raymondjtoth2> <raymondjtoth2> ?\17:03
CobaltYou need to be patient, I think.17:06
jcristauit's installed by default, you don't have to do anything.17:07
tseliotmaybe he meant gallium?17:07
raymondjtoth2how i upgade to new build of it17:09
raymondjtoth2i ike to have all new stuff in it17:09
raymondjtoth2i need 3d for 945GM Graphics how i get the driver for that one17:09
raymondjtoth2flash video seam slow 17:09
raymondjtoth2on it17:09
raymondjtoth2and laging17:09
jcristautseliot: there is no intel gallium driver17:09
raymondjtoth2what i do17:09
raymondjtoth2jc i got a 945GM Graphics17:10
raymondjtoth2intel 945GM Graphics17:10
jcristauyou said that already.17:10
raymondjtoth2o ok 17:10
raymondjtoth2jc any thing i can do17:10
raymondjtoth2never did conpiling befor17:11
raymondjtoth2or hand install driver17:11
raymondjtoth2jcristau any idea17:11
raymondjtoth2any i dea17:14
CobaltSee, with this Bluetooth pairing thing, is it guaranteed every time I reboot my Magic Trackpad will be on /dev/input/event13? I have a funny feeling the answer might be 'NO'.17:14
CobaltAlso, I have a trackpad on my laptop which does not come up with X -configure.17:15
CobaltI wonder if its configs lie elsewhere.17:15
raymondjtoth2any one eles see my q17:15
Cobaltraymondjtoth2: Not me.17:16
penguin42Cobalt: I bet it'll always be /dev/input/by-id/something17:17
Cobaltpenguin42: Yeah, but it would be a bummer to have to manually change xorg.conf each time manually to factor that in.17:17
penguin42Cobalt: No, I mean the bit after by-id/ will always be the same17:18
CobaltStrange, it mentions something from 'Broadcom'. Which is not what it is. :S17:19
CobaltI think it's grabbing the BT adapter.17:20
CobaltStill, something to try next time.17:20
CobaltFirst I need to restart X to see if any of that actually works.17:20
penguin42Cobalt: Yeh so there is by-id and there is a by-path, but the path might change depending where you plug it in17:20
CobaltIt's an internal Bluetooth adapter. Some devices don't show up. But event13 which is the Apple Trackpad shows up as Broadcom something.17:21
CobaltAnd Broadcom is the make of the BT adapter.17:21
CobaltI've got it already.17:24
raymondjtoth2is jc here'17:28
Sarvattman, i want to see the bug count graph after today's over because i'm going crazy on the intel apport bugs17:50
Sarvattgot a pretty good routine going on how to spot dupes fast. start by picking a generation, run intel_error_decode on i915_error_state.txt, and you can categorize them by EIR PGTBL_ER and IPEHR matching to spot the dupes17:53
Sarvattliterally hundreds of these go through though and intel_error_decode in maverick's intel-gpu-tools isn't good enough to grab the info for all generations17:54
Sarvattwhen chipset generation EIR and PGTBL_ER match the triggers described in the bug are almost always matching up17:56
brycehSarvatt, why don't you add that logic to the apport hook so it dupes things up properly?17:57
Sarvattjust matching EIR isn't enough, on 965 class hardware i'm seeing 3 different PGTBL_ER's with EIR: 0x00000010 so far17:58
bryceh(when you're done with the awesome cleanup)17:58
Sarvattbryceh: because I'm just figuring this out now, will do :)17:58
Sarvattquite a few are just the same people submitting lots of the same thing18:00
brycehaside from being dupes, are you finding the collected crash reports are actionable or upstreamable?18:02
Sarvattnot at all18:03
Sarvattpeople just hit submit and dont describe whats happening 99% of the time18:04
brycehthey probably think the magic of apport collection gathers 100% of the needed info18:05
brycehin which case ironically the apport script might be doing more harm than good18:05
Sarvattthe info is useful when i find a real bug report and have the big database of potential dupes to get more info from though  :)18:05
brycehwish we had a crash database separate from the bug tracker18:05
Sarvattnow *that* would be nice18:06
SarvattBlackZ: what's up?18:06
brycehit might be possible to have the apport script prompt the user with a dialog to provide more info or something18:07
BlackZSarvatt: oh, if you want, we can talk in #ubuntu-motu or here, both channels are OK for me18:07
BlackZSarvatt: but I think the question I was about to do you is a bit OT for this channel :)18:08
Sarvattconsidering us X guys made ppa-purge for xorg-edgers I dunno about that :)18:09
CobaltSarvatt: It autodetects the Magic Trackpad then automatically loads evdev. I can't seem to disable evdev and make it use Synaptics instead, to find if there's some better usability there. :S18:12
Sarvattwhy does this have to be so complicated? need to create a team so other people can commit to ppa-purge since its just using xorg-edgers now18:19
brycehSarvatt, didn't it get added to the repo that contains add-apt-repository?18:22
Sarvattnot that i'm aware of?18:23
brycehhmm there is https://bugs.edge.launchpad.net/ubuntu/+bug/44621618:26
ubot4Launchpad bug 446216 in software-properties (Ubuntu) (and 1 other project) "add-apt-repository should have an option to remove ppa from sources.list (affects: 11) (heat: 48)" [Wishlist,Fix released]18:27
brycehand https://code.edge.launchpad.net/~bryceharrington/software-properties/rm-apt-repository/+merge/2598818:27
brycehanyhoo, time for some breakfast18:29
Sarvattit's hard working up the will to go back to these other 40+ bug tabs I had open now that I got distracted :)18:33
raymondjtoth2any one see my q18:34
cndjcristau, hi there18:44
cndas I'm trying to do some development on X, I'm wishing the debian packaging worked more like ubuntu's kernel packaging18:44
cndwe keep everything in git, and the debian packaging and our own patches we apply on top of the upstream kernel are maintained as separate commits to the tree18:45
cndwhen a new upstream kernel is released, the whole tree is rebased18:45
cndit makes patch work much simpler, and allows us to easily track upstream development18:45
Sarvattbooo, hiss!! :)18:45
cndwhat would be your thoughts on doing something similar for debians X packaging?18:45
cndSarvatt, are you not in favor of the ubuntu kernel approach, or are you just trying to incite a riot :)18:46
Sarvattvery much in favor of using the upstream tarballs and patching on top of that, I love the way it's set up now personally :)18:47
raymondjtoth2how i install 915resolution18:47
raymondjtoth2i dont see it18:47
jcristauyou don't18:47
raymondjtoth2why that18:47
raymondjtoth2tell me i need it if 8 or 9 card18:47
cndSarvatt, it'18:47
cndit's still patching on top of upstream18:48
cndjust in git form instead of tarball + quilt patches18:48
cndthe orig tarball can still exist as is, the debian diff would look odd though18:49
cndmy gut instinct says who cares what the debian diff looks like if we can do our development much faster/better18:50
jcristauwhy would the debian diff look odd?18:50
jcristauwouldn't it just be the same as it is today, with the patches applied inline instead of in debian/patches?18:51
cndjcristau, yes, that's what I'm thinking18:51
cndit would just look odd if you expect it to have stand alone patches18:51
jcristau(today it's a mix)18:51
cndahh, I didn't know that18:51
cndI'm actually working on a tree right now18:51
cndsort of a rough draft of this work flow18:52
cndlet me push it somewhere18:52
jcristauwhen a patch is already upstream we normally just cherry-pick it to the debian branch18:52
cndahh, that makes sense18:54
cndwhich is sort of half of the work flow I'm envisioning18:54
Sarvattxorg-edgers will be... fun.. if that change happens, only as easy as it is now because ubuntu changes are self contained patches on top of debian. guess i'd have to add some logic to maintain lists of commits to revert instead of patches to disable like it works now18:54
cndSarvatt, it should be easier actually18:55
cnduse git revert18:56
cndwhich creates a new commit that undoes the debian patches we don't want18:56
Sarvattand hope the commit didn't touch the changelog so it actually reverts?18:56
cndwhen you rebase onto the latest X, it will come along too18:56
cndthe ubuntu kernel commits don't actually change the changelog18:56
cndthe changelog is generated at build time18:56
cndwell, actually it's generated at upload time18:57
cndwe'd need to replicate what the ubuntu kernel does there18:57
cndok, I'm remembering things now, the ubuntu kernel workflow adds a new git commit for each upload to the archive18:57
cndand the contents of the commit is just the changelog entry18:58
cndso all the patch commits leave the debian changelog alone18:58
cndso you can easily revert things18:58
jcristauwould you mind taking that to email e.g. on debian-x?19:04
cndjcristau, sure19:04
cndI'll make a tree and send an email about it19:04
Sarvattbarely put a dent in them - http://www2.bryceharrington.org:8080/X/Reports/ubuntu-x-swat/totals-maverick-workqueue.svg19:14
SarvattRAOF: asking about the vaio 260.19.xx problem in #nvidia to see if its a known problem, bug reports are still flooding in and they are all vaio gt2xx or gt3xx systems so thats holding up19:16
tjaaltonSarvatt: that's.. a depressing graph :)19:39
bjsniderit still looks to me like there's more broken hardware than justt he vaio19:42
penguin42Sarvatt: I think the good news is the free Radeon driver is actually getting useable and I suspect more people are using it so finding more bugs19:53
Sarvattthe radeon problems are almost all in the kernel, intel is the one thats making it nuts19:54
cndSarvatt, jcristau: how are ubuntu changelog entries handled when debian moves from one release to another, but there were ubuntu entries in between19:57
cndit seems like we interleave them19:57
Sarvattyeah we merge them back with merge-changelog19:58
cndSarvatt, ok, I'll take a look at it19:58
cndSarvatt, based on the merge-changelog code (I just glanced at it, man page was no help), it looks like it merges the changelogs and orders entries based on the package version?20:05
cnddoes that sound right?20:06
cndSarvatt, so is this the process when you make a new ubuntu release:20:07
cnd1. make changes20:07
cnd2. run merge-changelog20:07
Sarvattour goal is to not have a diff over debian as much as possible in the first place20:07
cnd3. dch -i20:07
cndof course :)20:07
cndactually, I'm guessing 2 and 3 are swapped20:08
Sarvattwhich package are you using as an example? it varies a lot, some are perpetually ubuntu versioned like the server20:09
cndI'm interested in xorg-server right now20:09
cndwhat do you mean by perpetually ubuntu versioned?20:09
Sarvattnever need to run merge-changelog, i fix up the changelog diff with git mergetool when merging the debian branch into ubuntu20:10
Sarvattcnd: ./auto-xorg-git -t + -H hooks -d origin/ubuntu -u git://yourgitrepo.git -b yourbranch -g -p xorg-server -a 0ubuntu0cnd would work for the server if you have master and your changes in that branch, just need to add the patches to drop in hooks/xorg-server.prepatch, if you're just testing locally its easier to not bump the abi's in hooks/xorg-server.prebuild and just rebuild what you care about against it20:40
Sarvatti just run it once and let it fail then fakeroot debian/rules clean and retry after adding the patch to drop in the hook to work out what patches in ubuntu need to be fixed or dropped20:41
Sarvattthere's only 3 or 4 ubuntu specific patches that are really needed anyway out of that huge mess in the server20:43
cndSarvatt, thanks for the pointers20:43
cndI'll give it a go20:43
cndjcristau, Sarvatt, to follow up, I think Sarvatt's scripts fulfill my needs, and I'm not sure the gitorized packaging branches would have worked out the way I intended in the end21:17
Sarvattwoohoo 3 of the arrandale/clarkdale bug reporters with (IPEHR: 0x01820000) are saying its fixed, need to ask for retesting on this mass of other bugs with the same error after duping them21:29
Sarvattthat error state seems common to lid closing or dpms events21:33
Sarvattmaybe i shouldn't dupe them and just ask for testing individually instead, all it takes is one person to say no to keep a bug with 50 dupes open :)21:35
jcristaujust close it when one of them say it's fixed :)21:36
Sarvattwell at least I learned to say hibernate in 4 languages looking at PGTBL_ER: 0x00000003 bugs :)21:50
Sarvattthat one looks to be common across gen3 and gen4 but i'm not duping cross chipset just in case, got separate master bugs21:55
Sarvattlol I was amazed I found an actual bug that could go upstream, then I see it's in karmic and I'm sure its fixed https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/51073822:00
ubot4Launchpad bug 510738 in xserver-xorg-video-intel (Ubuntu) "[g45] Stormbaan Coureur crashes: 'render error detected' (EIR: 0x00000010 PGTBL_ER: 0x00800000 IPEHR: 0x780a0101) (affects: 2) (heat: 10)" [Undecided,Confirmed]22:00
Sarvattgot the apport bugs condensed to 60 so far and I'm stopping for the day, there's still about 30 left to run intel_error_decode on not counting the ones where the people didn't use the apport generated title - https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bugs?field.searchtext=PGTBL_ER22:14
Sarvattbryceh: weren't you looking for examples of when fdo bug statuses dont show up right in launchpad?22:30
Sarvatthttps://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/634683 https://bugs.freedesktop.org/show_bug.cgi?id=3009722:31
ubot4Launchpad bug 634683 in xserver-xorg-video-intel (Ubuntu) (and 1 other project) "[gm45] Running piglit/occlusion-query-discard locks GPU (affects: 4) (heat: 163)" [High,Triaged]22:31
Sarvattverified fixed critical shows up as confirmed critical22:31
Sarvattguess the fdo bug should be resolved instead of verified or does that look right?22:32
RAOFHigh, Triaged?  That got closed in the 20100924 mesa snapshot upload.22:33
brycehhang on, gotta reboot22:34
SarvattRAOF: oh I didn't even look at the launchpad side of the bug, I swear you closed it in the mesa changelog!22:34
RAOFIt was probably during that time where LP wasn't closing bugs for us.22:35
brycehSarvatt, it's possible (likely) that the watch just didn't get triggered to update22:38
brycehI'll reset it22:38
brycehSarvatt, ok, check it again in about an hour and ping me if it's not set correctly then22:39
Sarvattwell I guess it could be worse - http://www2.bryceharrington.org:8080/X/Reports/ubuntu-x-swat/totals-lucid-workqueue.svg22:50
Sarvattoh thats only a little over a month, whoops22:53
penguin42Sarvatt: What was the release date back then?22:54
penguin42so you get a whole raft of fglrx-installer ones in the last few weeks until it just works by release?22:55
Sarvattlooks like someone cleaned those up right before october 1st for maverick22:56
Sarvattmid september rather22:57
penguin42Sarvatt: What do you reckon for something I've suggested as bug 63641822:57
ubot4Launchpad bug 636418 in jockey (Ubuntu) "update should clean up/warn about jockey (affects: 1) (heat: 176)" [Undecided,New] https://launchpad.net/bugs/63641822:57
Sarvattthe warning would be nice, but as of the latest packages it already cleans up the installed ones right so that wont happen since tseliot added abi checks22:59
penguin42ah ok23:00
penguin42we just get so many people arriving on +1 suddenly finding they can't get graphics in an alpha/beta/rc23:01
Sarvattfglrx->natty xserver 1.10 will remove fglrx on upgrade, the checks weren't there this cycle though and it was a mess23:01
Sarvattwait, I may be wrong23:01
dandelI'd say contact amd first, and check to see when/if xserver 1.10 will be supported before just blindly removing fglrx.23:02
Sarvattnot sure if there are blacklists for specific drivers or if it removes everything wth the old abi23:02
Sarvattthey'll gladly not support it for years if no one ships it23:03
dandelxserver 1.10 is slatted for 11.x right?23:04
Sarvattthat was just an example, hasn't been decided or anything23:04
Sarvattif 1.9.x rocks as much as 1.7.x did maybe not23:05
Sarvatthaven't seen a 1.10 release schedule yet23:06
dandelThe only real concern i have is the fglrx kernel bugs.23:06
Sarvattoh wait, its 6 months, pretty much for sure it'll be 1.9.x23:07
Sarvatthad a brain fart there, intel guy is the release manager so was thinking 3 months :)23:07
dandelI just hope that libva will get fixed correctly.23:07
Sarvatti doubt it, libva is a real mess23:07
dandelmaverik is missing packages.23:07
dandelit'd be good if libva could get ramped semi-frequently.23:08
RAOFWhat are we missing, libva-wise?23:08
Sarvatt965 "support"23:08
RAOFSarvatt: You mean the 965 support that results in an assert as soon as something tries to use it? :){23:08
dandelfor one... libva-glx or libva-glx1 23:08
Sarvatti'm sure the debian maintainer could use the help if you want to help out with it23:08
dandeli just flat installed the libva 1.0.4 packages from the experimental debian repo.23:09
Sarvattyeah it's not even released in debian23:09
Sarvattjust been brewing in git for a long time23:09
dandellibva is a pain, but is important and should of been properly put into the 10.10 release (but packages are missing)23:10
Sarvattso why not help out with it?23:11
RAOFIt'd be more important if anything used it.23:11
RAOFAnd a reasonable selection of hardware supported it :)23:11
Sarvatti want to stab my eyes out every time i look at that code23:13
RAOFIt'd be nice if there was a (free) gstreamer libva element.  That'd make me care about libva a lot more.23:14
RAOFFluendo's gstreamer elements are kick-arse.23:15
dandelRAOF, it's in use by newer releases of vlc, ffmpeg, xbmc, mplayer, gnash and is picking up support increasingly on the api's. 23:15
jcristauif it doesn't support any hw that's kind of moot23:17
dandelIt's supports nvidia video cards via vdpau.23:17
RAOFAnd fglrx via the (non-free?) xvba thingy.23:18
dandelan amd dev said that xvba got too much news coverage early in the dev cycle and it has no set ready date.23:19
Sarvattjcristau: poulsbo and fglrx isn't good enough for you? :)23:19
jcristauSarvatt: surprisingly, no23:19
RAOFI mean, I'm in no way against having it in the archive, and will give it some support, but it's hard find the motivation to invest time in it when it'll only run on top of the existing APIs of blobs.23:20
RAOFWhich means that it'll become more interesting as it actually starts to support Intel hardware.23:21
dandelanyways, xvba doesn't work for me (yet), but that's somewhat anticipated... i have a evergreen based video card.23:23
RAOFDoesn't fglrx support evergreen properly yet?23:24
dandelit supports evergreen properly... it's just i don't have video accelerated decoding yet.23:25
Sarvatthttp://wiki.debian.org/DebianMultimedia has some places to go to see what the status is for libva23:25
RAOFSo, not full support for evergreen then.23:26
dandelhowever, i do get to run unigine heaven with tessellation enabled.23:26
dandeland the rendering for that program is correct... however, gluxmark2 (last i checked) had some defects in the 10.8 and 10.9 driver.23:27
dandelit's just that ati has no proven bug report methods that get quick results and testing... The location i use to report bugs has long delays. (a month just for confirmation of the bug as reproduce-able)23:28
dandelRAOF, do you want to actually see what i get when i try to run a vaapi based decode?23:30
RAOFNot particularly, no :)23:32
dandeli get a moziac of the output image.23:32
dandelthe output is decoded with the gpu, but it's just that the output images are incorrect.23:33

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