/srv/irclogs.ubuntu.com/2008/05/28/#ubuntu-x.txt

tjaaltonbryce: I've filed a bug to remove fonttosfnt from the archive, bug 235411. you could confirm that if you agree ;)08:04
ubottuLaunchpad bug 235411 in fonttosfnt "Please remove from the archive" [Undecided,New] https://launchpad.net/bugs/23541108:04
brycetjaalton: okies08:06
tjaaltonhmm, a couple of pending sync requests.. guess they'll be processed soon08:08
tjaaltonso, there'll be libdrm 2.3.1 without any memory-manager goodness08:10
tseliottjaalton: I'm uploading the source code of my packages to my website. NOTE: I haven't touched the control files (as regards dependencies) and haven't carefully checked the diversions, however I would like you to review what I have done so far.08:29
tjaaltontseliot: ok, cool08:32
tseliottjaalton: I haven't removed the dependency on linux-restricted-modules-common yet since we are not sure about its removal and I had to add a "-modaliases" package so that we don't screw up jockey08:34
tjaaltonwow, Wolfram answered to bug 19716309:21
ubottuLaunchpad bug 197163 in xorg "Mathematica renders fonts incorrectly in Hardy" [Undecided,Won't fix] https://launchpad.net/bugs/19716309:21
tjaaltonso we are buggy at least in some respect, because there are no usable Courier and Times fonts (msttcorefonts not helping here)09:22
tjaaltonhm, though it should help09:29
tjaaltonfedora seems to have urw-fonts which is a free bundle of tt-fonts09:33
brycetjaalton: true09:36
tjaaltonI'll see what it takes to package that09:37
tseliottjaalton: here is a txt file which contains all the links to the source code: http://www.albertomilone.com/ubuntu/newlrm/lrm-links.txt09:45
tjaaltonapparently gsfonts should have the fonts..09:45
tjaaltontseliot: thanks09:45
tjaaltonI'll just grab the directory09:46
tseliotin the meantime I'll work on the xorg parser for bryce 09:46
tjaaltonuh, forbidden09:47
tseliotyes, I know09:48
tseliotyou should wget the links in that txt file09:49
tjaaltonworks09:49
tjaaltonok, the version number is wrong :) -1 is reserved for debian, this should be -0ubuntu109:50
tseliotyes, I (temporarily) used exactly the same version as the debian package09:51
tjaaltonok, that just makes no sense :)09:52
unggnuhi all09:53
tjaaltonhi unggnu 09:53
unggnuAccording to mail [ubuntu-x] Xorg Triaging Projects it is not possible for non developers to set Wont' fix09:53
unggnuhi tjaalton09:53
unggnuOnly solution is to set to invalid I guess or get the permission to use Wontfix.09:54
tjaaltonor join ubuntu-bugcontrol09:55
unggnuDo you have a link?09:56
tjaaltonhttps://edge.launchpad.net/~ubuntu-bugcontrol09:56
unggnuthx09:57
unggnuBtw. isn't it possible with a script to mark all displayconfig-gtk bug reports which have only one package association with Wontfix and add the description?09:59
tjaaltonshould be, but I'm not familiar with the mail-interface10:01
unggnuOk10:02
unggnuIs something planned for people with hardware which send wrong data or something like that? Some can't even fixed with a quirk afaik. It is possible to change xorg.conf manually but it isn't so easy I guess for most people.10:06
unggnue.g. https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/19100010:06
ubottuLaunchpad bug 191000 in xserver-xorg-video-intel "[Hardy] X Server fails to start [845g] (dup-of: 194760)" [Unknown,Fix released] 10:06
ubottuLaunchpad bug 194760 in xorg-server "EDID fail" [High,Triaged] 10:06
unggnuI guess a tool with a "problem monitor" db would be great or the ability to import windows monitor drivers and change the xorg.conf settings with this information. Afaik displayconfig-gtk had the second ability. Maybe only the monitor part can be extracted.10:17
unggnuAnother problem are CRT displays afaik. It is possible to change the user resolution with the screen resolution tool but not the login one and afaik X uses the highes possible one which doesn't look so well on CRTs.10:55
unggnuAre there any plans for this problems?10:57
unggnubbl11:05
tjaaltonCRT's are dying anyway ;)11:06
jcristauunggnu: X uses the mode that your monitor reports as preferred11:07
jcristauunggnu: if that doesn't work then it can be quirked11:07
tjaaltonyeah the first problem seemed like it should be able to be quirked, since it tries to use a wrong mode11:08
tjaaltonmaybe the inted engineer meant that it's not possible to quirk the driver11:21
tjaalton*intel11:21
jcristauthe initial mode selection code has changed a lot recently though11:25
jcristauso maybe the server from git does a better job11:25
jcristauin any case this looks like a bug in the server, not xf86-video-intel?11:25
tjaaltonyeah11:27
unggnure11:59
unggnujcristau: so a CRT sends the data that he supports resolutions up to 1600x1200 but prefers 1280x1024?12:00
unggnusorry, it supports :)12:00
jcristauyes12:00
unggnuBut why are so many people with CRT complaining about it?12:01
unggnuAnyway, if xserver uses this preferred mode it is Ok I guess. I can't test it since I haven't any CRT anymore. Yes, they are dying :)12:04
unggnuI guess in future the biggest part of the driver are quirks ;)12:07
jcristauno12:07
unggnubbl13:11
brycetjaalton: do you know of any reasons we should put a newer fglrx into 8.04.1 (deadline's coming up)16:32
tjaaltontjaalton: I'll read the relnotes to see if there's anything16:41
tjaaltonI didn't realize there was a deadline :/16:42
bryceyeah sunday is the cutoff for getting things into 8.04.116:43
tseliottjaalton: https://a248.e.akamai.net/f/674/9206/0/www2.ati.com/drivers/linux/catalyst_85_linux.html16:43
tseliotBTW they broke the API16:43
brycetseliot: erf16:46
tjaaltontseliot: meaning?16:48
tjaaltoneh, the list of known issues keeps getting longer..16:50
tseliottjaalton: it means that you will have a lot of fun packaging it16:50
tseliotfurthermore the ati installer doesn't include Mario's latest scripts16:51
tseliottherefore I suggest you to get such scripts from git and see what's changed16:51
tseliotI'm planning to make this driver available through EnvyNG too (sooner or later)16:52
tseliotbryce: have a look at my email when you have the time16:52
brycetseliot: ok just replied16:53
tseliotbryce: thanks16:53
bryceheya unggnu17:25
unggnuhi bryce17:25
brycetseliot, tjaalton: so sounds like the new fglrx is too sketchy to think about including for 8.04.1?17:25
unggnuCan't someone ask an Launchpad expert or maintainer if this is possible to close all displayconfig-gtk bugs since they don't seem to need user interaction.17:26
unggnuI thought updating fglrx or nvidia isn't possible or at least feasible? At least I have read it somewhere.17:26
tseliotbryce: yes, I guess so17:26
bryceunggnu: normally yes, but I got a dispensation for it for 8.04.1 -- but it makes sense to do it only if we know it will fix issues and we're confident it won't bring regressions17:27
unggnuok17:28
bryceunggnu: I was going to just close the dcg bugs manually, so I could check that they're actually dcg bugs, and give additional answers where appropriate17:32
unggnuOk, that's a point. I could be minimize through closing simply the displayconfig-gtk only bugs but the risk is still there.17:34
unggnuSo kde-guidance should be closed too?17:36
bryceI believe so, but check with scottk17:36
tjaaltonbryce: yep17:39
johanbrbryce: About the Radeon lockup bug (bug #234811), "xrandr -s 1440x900" works (in contrast to doing it with the GUI, which causes a hard lockup).17:44
ubottuLaunchpad bug 234811 in xserver-xorg-video-ati "[Hardy] system freezes after setting resolution to 1440 x 900 with Radeon 7500 graphics card, when using DVI output, Acer AL1916W monitor" [Undecided,Incomplete] https://launchpad.net/bugs/23481117:44
brycejohanbr: *nod* that's too bad, since if it could be reproduced with xrandr we could forward the report upstream17:45
brycejohanbr: as it is, it may be some issue with gnome-display-properties' randrwrap code17:45
unggnuWhat happens with Bullet-Proof-X in Intrepid?17:45
bryceunggnu: timo and I discussed it and are thinking we'll drop it17:46
johanbrEven so, there's an X server or driver bug in there somewhere, right? I'd think the GUI shouldn't be able to bring everything down.17:46
bryceright17:46
bryceit's definitely a legit X bug, just not clear yet how to troubleshoot it17:47
unggnuI liked the idea very much but was never lucky with displayconfig-gtk.17:47
johanbrbryce: I haven't yet tried asking him to rebuild the Gutsy version of the driver. I'll do that.17:47
unggnuTheoretically it isn't needed if the Vesa fallback of Xorg works fine.17:48
brycejohanbr: basically what needs to happen is to dig through the code and identify which Xrandr protocol calls are generating the problem, make a test case that reproduces the issue, and forward that upstream17:48
brycejohanbr: but I think requesting a user code up a test case in C for us is probably asking much ;-)17:48
brycepossibly there is a way we could gather more detailed logs or something from g-d-p, but I don't recall offhand17:49
johanbrI guess running g-d-p under gdb and putting in some carefully chosen breakpoints would do the trick.17:50
brycejohanbr: possibly17:55
brycejohanbr: although I suspect it probably just happens during the protocol call17:56
brycemight be able to dig out some state info and such, although it would be difficult to talk a user through all that in gdb17:56
brycebut perhaps breaking on that call and getting a full backtrace would be doable17:57
bryceit would be good to have a documented procedure on this, as there's been a few other such reports and I daresay we may be getting more in the future17:57
johanbrHe's new to linux, but interested in helping to resolve the bug and not afraid of reading man pages and such. I'll see what I can do.18:00
unggnuI guess it is Ok to drop bulletproof. X still is very picky about things in xorg.conf but it could be easily repaired through Recovery Mode. Maybe it would be great to start X without an xorg.conf instead of Bulletproof if X couldn't start.18:04
unggnuMaybe with a question after login for the admin users if the xorg.conf should be recreated like xfix does.18:06
unggnuOk, could be a loop but a simple check could prevent it. I don't know. People like their GUIs :)18:08
bdmurraytjaalton: is bug 113679 specific to nvidia hardware?19:13
ubottuLaunchpad bug 113679 in xorg-server "xorg freezes when running openoffice" [High,Fix committed] https://launchpad.net/bugs/11367919:13
bryceunggnu: yeah... ideally I'd like to replace displayconfig-gtk with something, but there's not really anything viable at the moment19:19
unggnubryce: Xfix would do the job I guess if there is not a general Xorg/vesa problem.19:21
bryceunggnu: yeah that's what I'm thinking too19:22
bryceanother option would be to strip down displayconfig to remove the multi-screen stuff and other bits that could cause confusion19:23
brycewe really don't have a good solution for the corner case of edid failures or where edid is not being provided by the monitor19:23
brycei.e., http://bugs.launchpad.net/ubuntu/+bug/edidfail19:24
unggnuWould be great too for wrong edid data but afaik the conventions have changed. Modelines are ignored at lest for the intel driver. Mainly it would be a rewrite I guess.19:24
* bryce nods19:24
brycewow I didn't know they're ignoring modelines now19:25
unggnuAfaik of course :)19:25
unggnusomething with PreferredMode19:27
unggnuhttp://www.intellinuxgraphics.org/dualhead.html19:27
unggnualso described here in the latest comment https://bugs.freedesktop.org/show_bug.cgi?id=1472619:29
ubottuFreedesktop bug 14726 in Driver/intel "[845G] X Server fails to start with "intel" driver, works with "i810"." [Major,Resolved: fixed] 19:29
unggnubbl19:30
tjaaltonbdmurray: shouldn't be AFAIK19:33
bdmurraytjaalton: Okay, I saw a couple of references to nvidia in it.  Is there a test case at all?19:34
tjaaltonbdmurray: I don't think there is an easy way to reproduce it.. we have a couple of hundred desktops (all nvidia) and don't remember anyone reporting issues with OO.o :)19:37
tjaaltoneven with dapper19:37
bdmurrayhrm, that'll be interesting then19:37
tjaaltonapparently the fix works, since the google guys haven't seen the problem since upgrading to the version on my ppa19:40
bryceunggnu: hmm I read that as meaning that the order of the resolutions is ignored, but the resolutions themselves are accepted, and I read that as modelines are still accepted19:51
bryceunggnu: the issue seemed to be that the user added a mode, but couldn't get it to be taken as the default resolution in the new x server by simply listing it first in the list - he had to add it as the preferred mode19:52
bryce(I could be wrong though, but that was my interpretation)19:52
brycegood gem vs. ttm article:  http://lwn.net/Articles/283793/20:04
brycecopy of it here:  http://pastebin.com/m1a80b77b20:05
tjaaltonwow, the topic suggested by Mirv :)20:06
unggnubryce: I never got Modelines working with the -intel driver but I am not sure.20:22
unggnuMaybe it only accepts the modeline if preferredmode is used.20:24
unggnuIt seems so http://lists.freedesktop.org/archives/xorg/2008-February/032791.html but I don't understand why they use a new option for an old feature?20:25
tjaaltonit's randr-1.2, same thing with radeon AFAIK20:27
tseliotunggnu: something like this would work:20:29
tseliotModeline "1280x1024_75.00"  138.54  1280 1368 1504 1728  1024 1025 1028 1069  -HSync +Vsync20:29
tseliotOption "PreferredMode" "1280x1024_75.00" 20:29
unggnuI think so. So the tool have to add a Modeline and PreferredMode.20:29
bryceunggnu: probably the xrandr changes altered how things worked internally20:32
unggnuYeah, Xorg devs are very fast forward :)20:32
tseliotbryce: thinking of the special case which you suggested in the email, we can test such case with my program, which I will keep separate from the parser, and decide which approach we should adopt.20:33
bryceI think there didn't used to be the idea of a PreferredMode so adding that feature broke old behavior20:33
brycetseliot: good to hear.  it could help prevent xorg.conf's from getting chugged up with extraneous cruft (which tends to exacerbate troubleshooting issues)20:34
tseliotbryce: I would like to make both parts maintainable and easy to use so that at least the parser can be shared between different applications. We'll see how it goes...20:35
tseliottjaalton: quoting from Phoronix "NVIDIA this morning has released the 173.14.05 driver,  which marks the return to their old naming convention"20:36
brycetseliot: heh20:37
tseliottjaalton: this release supports kernel 2.6.25 therefore I will have to update the driver and drop the patch for 2.6.25 for driver 169.1220:37
tjaaltontseliot: cool, it only took them a couple of months..20:42
unggnuBtw. I have asked this some time ago. Is there any doc how to compile svn Intel driver for normal users without messing up their system?20:49
brycethere's a few such documents but from my experience users just whine when asked to build from git20:50
bryce(not that I blame them, I probably would too)20:50
unggnuBut upstream often asks for it.20:50
bryceright20:51
unggnuThis time the debian sid driver works but it is not often the case afaik.20:51
unggnue.g. https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/20788120:52
ubottuLaunchpad bug 207881 in xserver-xorg-video-intel "[Gutsy, Hardy] Black screen with mouse pointer on i830, intel driver" [Unknown,Confirmed] 20:52
unggnuthis one wanted a patched driver20:52
unggnuI have tried it but there is one hunk.20:53
unggnu+failed20:57
bryceoften with patches from upstream I have to recode them a little to make them patch cleanly against the driver we carry20:58
bryceunggnu: tell you what, I'll take care of 20788121:00
unggnuthx :)21:01
bryceunggnu: btw, fyi we have the 8.04.1 cut off date coming up sunday, so what I'm working on now is looking for patches that are complete, verified, and ready for backporting to hardy21:02
bryceunggnu: so if you could search around for good looking patches that look like they could be put to hardy, please assign them to me21:02
unggnuOk, what is the best way? Searching for bug reports with an upstream report and fix released?21:04
unggnuOr just patches which are already in Intrepid?21:04
brycesure both21:05
unggnuok21:05
bryceactually there are probably going to be exceptionally few in the latter category21:05
bryceso focus on the first21:05
tjaaltonbryce: what was the firefox addon that allows you to open files in firefox instead of it proposing to save it first?21:08
brycemy notes say "Open In Browser"21:12
brycehowdy tormod21:15
tormodheya bryce21:16
brycewow, don't think I've seen this many people on #ubuntu-x.  :-)21:16
tjaaltonbryce: hm, either I'm blind or can't find it21:17
brycemaybe this?  http://www.spasche.net/mozilla/21:18
brycelooks like it's only for ff221:18
bryce(I finally gave up on ff3 yesterday and reverted back to 2)21:18
tjaaltondamn :)21:19
tormodbryce, due to the disk activity?21:19
brycetormod: yeah the periodic freeze-ups that are reportedly due to sql-lite size limits21:19
brycetormod: plus a scattering of crash issues (less common now than they used to be though)21:20
tormodthey need a full RDBS in there I guess21:20
unggnubryce: What should I do if there is already an assignee?21:20
brycetjaalton: hrm, didn't install onto ff2 for me21:21
bryceunggnu: if it's already assigned to me, then I've already looked at it and it may already be in my queue21:21
unggnubryce: no, not you :)21:21
bryceif it's assigned to the ubuntu-x swat team, you can consider those unassigned21:21
unggnubryce: bryce: this one looks good https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/21175921:21
ubottuLaunchpad bug 211759 in xserver-xorg-video-ati "Live-CD (7.10, 8.04) fails to start X server (ATI Xpress 1250)" [High,In progress] 21:21
bryceassigned to anyone else, just mention the bug id to me here and I'll review.21:22
unggnucannonical-xorg21:22
tormodwhen is launchpad getting per-package bug filing instructions, like attach your Xorg.log...21:22
bryceok, treat canonical-xorg to be unassigned as well; that's just a generic team assignment21:22
unggnuok21:22
bryceI'm the only person in canonical-xorg anyway, so all those can just be re-assigned to me.  I don't know who added that team or why, but I don't use it21:23
brycetormod: in fact I mentioned exactly that to kiko last week at UDS during one of the "improving launchpad to improve bug reporting" sessions21:23
brycetormod: I didn't get a solid commitment to implement such a thing, but favorable noises were made, so we'll see21:24
tormodbryce: excellent. but it's not like a new idea... it has been requested before I think. hope it helps this time.21:24
tormodhas it been decided that intrepid gets xserver 1.5 and mesa 7.1?21:25
brycetormod: this was in particular related with how to improve being able to get bugs upstreamed, which kiko is very passionate about.  I said that if launchpad could be set to require attaching Xorg.0.log to every package managed by the xorg team, it'd probably double our triaging effectiveness21:25
brycetormod: yes21:25
tormodyay for that 21:26
tormodcan someone look at my workaround patch for bug 87661, before I ask upstream?21:27
ubottuLaunchpad bug 87661 in mesa "[apport] polytopes crashed with SIGILL in _mesa_x86_64_transform_points4_perspective()" [Medium,Confirmed] https://launchpad.net/bugs/8766121:27
brycetormod: looks roughly sane to me (although I didn't read the whole bug report)21:29
tormodbryce: I don't know so well how 64-bit processors work with different code and kernels etc21:30
unggnubryce: I am not sure with this one. It only seems to be a dependency issue so easy fixable but I don't know ... https://bugs.edge.launchpad.net/ubuntu/+source/mesa/+bug/4411521:51
ubottuLaunchpad bug 44115 in mesa "xserver-xorg-video-tdfx must depend of libglide3 to get 3d aceleration in 3dfx cards" [Wishlist,Confirmed] 21:51
tjaaltonwould mean installing libglide3 on every machine21:53
tjaaltonnot something to be done on a stable release21:53
tjaaltonor ever21:53
brycetjaalton: can it be closed as wontfix then21:53
bryce?21:53
tjaaltonimho yes21:54
unggnuOr put it to suggest like in Debian21:54
tjaaltonit is?21:54
tjaaltonyes21:55
jibelhi21:56
tjaaltonso it's the same in ubuntu too -> closing the bug21:56
unggnuhi jibel21:56
jibelbryce: bug 229043 is fixed in v2.3 of the intel driver. Do you known if it will be fixed in 8.04.1 or earlier ?21:57
ubottuLaunchpad bug 229043 in xserver-xorg-video-intel "System freezes when login out" [Undecided,Fix committed] https://launchpad.net/bugs/22904321:57
brycejibel: I'll take a look22:19
jibelbryce: Thanks. I'll keep informed the reporter.22:27
unggnubryce: I guess huge patches should be integrated?22:55
unggnu*shouldn't22:55
bryceright22:56
bryce(generally)22:56
unggnuso no need for assigning?22:56
bryceunggnu: right22:59
unggnuhm, haven't found so much23:08
bryceok good.  I did a patrol for quirks and found only one23:08
unggnufour and one seems to be dirty but was confirmed several times https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-input-synaptics/+bug/13306023:08
ubottuLaunchpad bug 133060 in xserver-xorg-input-synaptics "Synaptic touchpad fails to register some of the taps for tap-to-click" [Undecided,Confirmed] 23:08
tormodactually update-manager uses more memory than firefox, lol23:21
tormodcan I trust System Monitor - firefox only 43MB?23:22
tormodI wonder if I add up those numbers and compare with "free"23:22
unggnubryce: another quirk possibility https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/201032 23:36
ubottuLaunchpad bug 201032 in xorg-server "[hardy][regression]xorg gets wrong screen dpi" [Unknown,Confirmed] 23:36
unggnuand another one https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/20080523:39
ubottuLaunchpad bug 200805 in xorg-server "xorg uses laptop video BIOS 'Panel Size' for external monitors, so apps end up confined to small space in top left of screen" [Undecided,Incomplete] 23:39
tormodoops that was wrong channel - time for bed QED23:40
unggnu:)23:42
unggnuI am not sure with them so I haven't assigned you.23:43
unggnuI am off too, ciao23:44
bryceexcellent23:55

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