james_whey, someone's requested a sync of input-mutouch, to pull in jcristau's Y axis fix, are you ok for this to be synced?00:05
james_wbug 275650#00:05
ubottuLaunchpad bug 275650 in xserver-xorg-input-mutouch "[UVFe] Please sync xserver-xorg-input-mutouch (1.2.0-2) from Debian unstable (main) - fixes: mutouch driver in hardy is Y axis Inverted" [Medium,Confirmed] https://launchpad.net/bugs/27565000:05
brycejames_w: got a debdiff?00:07
james_wbryce: it's a sync request00:09
brycejames_w: right - the diff between what we currently have and what's upstream00:11
james_wah, give me one moment please00:12
brycein any case, I'd not be predisposed to syncing it.  at this late stage in the release cycle we need to take care to do code reviews on changes coming in, for sanity check since there's limited time left for extensive testing00:13
brycealso, understanding what changed can help if people report seeing regressions00:13
brycejames_w: thanks.  Yes that looks fine.  +1 from me to sync00:25
james_wbryce: thanks, I'm just doing a test build/install00:25
wgrantShould bcm5974 tocuhpads really have SHMConfig on by default?02:34
wgranttjaalton: Apart from evdev in your PPA having broken deps (probably due to the xserver-xorg-core shlibs), it all looks good to me. I've got g-c-c and g-s-d working with it all.07:27
wgrantAnd it no longer hangs.07:27
wgrantWhich is excellent.07:28
tjaaltonwgrant: goody, so it's upload time07:34
wgrantwgrant: Just watch out that your new shlibs are correct.07:35
wgrantThey look like they should be good for upload, but I'm not entirely sure.07:35
tjaaltonwgrant: what's wrong and where?07:36
wgrantI am obviously doing multiple things.07:36
wgranttjaalton: evdev wants xserver-xorg-core 0ubuntu3.07:36
wgrantOnly 0ubuntu2.2 is in your PPA.07:36
wgrantBut 0ubuntu3 for be correct for upload to the primary archive.07:36
tjaaltonyeah, right07:37
tjaaltonit's fixed here07:37
* wgrant loves being able to test changes to input drivers without restarting the X session.07:38
tjaaltonain't that cool07:38
tjaaltonbryce: around? we've now finished what needs to be done for bug 27472807:40
ubottuLaunchpad bug 274728 in xorg-server "Update the input properties API to current version" [Undecided,In progress] https://launchpad.net/bugs/27472807:40
tjaaltonwgrant: have you contacted seb128?07:41
tjaaltondoesn't seem to be around yet07:41
wgranttjaalton: I haven't - I wasn't sure if I would have time to finish everything before the weekend until this morning.07:42
brycetjaalton: yep I'm here07:42
brycetjaalton: btw I sorted out that brightness key issue mdz had07:43
bryceturned out to be a gnome-power-manager bug.  I did a patch and posted it to the bug.  Probably needs to be more cleverly thought out, but that's for the GNOME guys to nut out I guess.07:43
tjaaltonbryce: whee07:44
bryceoh, and the thinkpad-specific aspect (numlock not being able to be turned off) seems to be a kernel issue07:45
wgrantI need to get my g-p-m bug sorted out, too.07:45
tjaaltonbryce: so, do you ack the uploads? they update the API to the current/"finished" one. If we don't do this, we'd be incompatible with upstream, and backporting properties related stuff from upstream would be more work07:46
tjaaltonbryce: I'd also add some patches to the xserver from upstream 1.5-branch07:46
wgrantI hope they fix that BadDevice issue that many are having.07:47
brycetjaalton: hm, I'd like to see what the diffs are07:47
tjaaltonbryce: http://cgit.freedesktop.org/xorg/xserver/?h=server-1.5-branch07:47
bryceI don't have anything against it in principle, although it sounds like a significant change so is scary from that perspective07:47
tjaaltonnow that the devPrivates-commit was reverted, it should be safe to just sync to head of that branch07:48
tjaaltonthere aren't that many commits07:48
tjaalton11 after 1.5.107:48
tjaaltonwgrant: that's properties behaving badly, and hopefully my backport fixes that07:49
wgranttjaalton: That's what I was referring to.07:49
tjaaltonwgrant: ah, ok07:49
wgrantMaybe I can convince RAOF to check it out now...07:49
tjaaltonwgrant: please do, would make me even more comfortable with this :)07:49
brycehmm, still looks like a lot to digest07:50
tjaaltonwhat does? the xkb fixes?07:51
brycethe 11 changes in the branch... give me a few minutes, still reading07:52
tjaaltonlook at the lines changed :)07:52
wgrantOnly the XKB ones are non-trivial.07:52
tjaaltonand those are long-standing bugs07:53
wgrantOw. The synaptics driver's logic hurts.07:54
bryceyeah looks good07:56
tjaaltonbryce: great, thanks07:56
bryceif you can associate at least some of those changes with bug id's in some bug tracker, that would be much preferred07:57
tjaaltonin fear of the wrath of slangasek ;)07:57
bryceyeah, the b595b65e change took time to get through but I see it's just a revert of an addition that was made in this same set of patches07:58
tjaaltonyes, ajax said it'd be an ABI change07:58
brycelet me look at the input properties change07:58
tjaaltonit's on my local tree, let me push to git.d.o first07:59
bryceer, wait, where's the diff for that07:59
bryceah ok07:59
brycecan you pastebin a diff?07:59
brycealso, what needs the updated input property stuff exactly?08:01
wgrantg-c-c, g-s-d and xinput are the only clients at the moment.08:01
bryceI'm guessing we're not going to be doing a lot of sruing of stuff like after hardy08:01
tjaaltonbryce: http://users.tkk.fi/~tjaalton/dpkg/xserver-prop.diff08:02
wgrantI presume not.08:02
tjaaltonbryce: the diff looks messy, since the patches were merged to keep my sanity08:03
tjaaltonbryce: but in a nutshell; a couple of functions were deleted08:03
wgrantAnd some args changed.08:03
wgrantBut those are trivial.08:04
tjaaltonhmm, the patch header is not right anymore08:04
tjaaltonmight just as well delete it08:04
* wgrant prepares g-s-d and g-c-c debdiffs.08:04
bryce"stuff->data"?  someone had a bout of uncreativity there ;-)08:04
wgrantHmm. No seb128 to ping.08:04
tjaaltonheh, that's old though08:05
brycewhat functionality in g-c-c/g-s-d/xinput does it enable?  08:09
wgrantxinput - properties.08:09
wgrantg-c-c, g-s-d: Synaptics configuration.08:09
brycewell, I see it adds an api for deleting properties, and tweaks how change properties works, but I'm not understanding how this would be required for enabling synaptics configuration or properties in general08:19
wgrantseb128: Hi. We're currently discussing fixing bug #274728, which will involve a g-s-d and g-c-c upload.08:20
ubottuLaunchpad bug 274728 in gnome-settings-daemon "Update the input properties API to current version" [Low,In progress] https://launchpad.net/bugs/27472808:20
wgrantbryce: It isn't.08:20
wgrantbryce: It is required to avoid using an API that never existed in a release.08:20
seb128wgrant: don't forget to commit to bzr if you upload g-c-c otherwise no objection08:20
wgrantIf we release Intrepid like this, and somebody tries to build something that uses properties (not unlikely, given their newness), they will go WTF when they see that we use an API that was never in a release.08:20
wgrantseb128: Thanks.08:21
wgrantAnd there's something broken with the current Intrepid backport.08:22
tjaaltonbryce: http://lists.freedesktop.org/archives/xorg/2008-September/038758.html08:22
tjaaltonthe following discussion lead to some minor changes08:22
brycetjaalton: alright, I've commented my concerns/review on the bug, and given a +1 on uploading it08:40
tjaaltonbryce: great, thanks08:41
wgrantJust found something strange with the Synaptics driver.08:42
wgrantSetting properties works.08:42
wgrantBut the change isn't reflected in the properties themselves.08:42
wgrantSo I can disable something using Synaptics Off, and it will disable.08:42
wgrantBut it will show as enabled.08:43
wgrantPlain evdev is fine.08:43
wgrantI guess -synaptics doesn't know that it needs to say it's OK now.08:43
* wgrant will look into that after eating.08:44
tjaaltonwgrant: I wonder if this would help.. http://lists.freedesktop.org/archives/xorg/2008-October/039251.html08:46
wgranttjaalton: Hmmm, I thought that would more cause it.08:46
wgrantBut I guess we don't have it yet, so it can't be to blame...08:46
wgrantI guess that would fix the incompleteness, but there is still some buggy driver.08:47
tjaaltonno we don't08:47
tjaaltonanyway, it should be a bug in the current version as well?08:48
wgrantIt returns TRUE (1), when it's meant to return Success (0).08:49
tjaaltonthe driver?08:49
* wgrant checks git.08:49
tjaaltondidn't see anything related there08:51
wgrantAll FALSEs were changed to BadMatches.08:52
wgrantBut the TRUE wasn't changed to Success.08:52
wgrantI'm testbuilding now.08:52
wgrantRight, that works.08:57
wgrantI think I might review the rest of those two diffs, just in case...09:01
wgranttjaalton: src/properties.c:507, s/TRUE/Success/09:03
wgrantI'll poke through the evdev and synaptics patches just to check there's nothing else, and submit to xorg@l.fd.o09:04
tjaaltonwgrant: great, thanks09:04
wgranttjaalton: Shouldn't 355e845 be in debian/patches somewhere?09:09
tjaaltonwgrant: no, I pulled it09:09
wgrantAh, right.09:09
tjaaltoncherry-picked to be precise09:09
wgrantI really should learn how to use git properly at some point.09:10
tjaaltonbasically you only need a handful of commands09:11
wgrantI know how to use it locally.09:11
tjaaltonoh, "properly" :)09:11
tjaaltonok, I'll upload inputproto now09:12
wgrantTo primary?09:12
wgrantSounds good.09:12
tjaaltonand libxi, since it b-deps on that so won't build before inputproto09:12
wgrantThen xi?09:12
wgrantThen xinput, g-c-c, g-s-d and server in no particular order, then drivers after server.09:13
tjaaltonwell, those could be uploaded all together, because we have build-deps09:14
bryceI've filled in the -fglrx section on https://wiki.ubuntu.com/X/Drivers09:20
bryce(trying to make the best of the sad situation)09:20
wgrantI added more versioning information to the mess that is X/Config earlier.09:21
bryceneed to get the rest of that page updated09:21
brycewgrant: which portions do you think are a mess?09:22
wgrantbryce: It's a big mess of versions. Some stuff only applies to Intrepid, some only to Hardy and before, and they're almost completely different.09:23
wgrantSome of the info was also wrong.09:23
tjaaltonpeple seem to be confused and believe that no input driver can be configured with xorg.conf anymore, when it's only the ones that hal knows about09:24
wgrantWhich doesn't it know about?09:25
brycemaybe we should make it Intrepid-only09:25
wgrantbryce: An Intrepid-only version would probably be good.09:25
wgrantBut replacing it would be bad.09:25
wgrantOr maybe shifting the pre-Intrepid stuff to another page.09:26
wgrantPeople still ues Hardy.09:26
wgrantAnd will for years.09:26
bryceI think it should also be broken into separate pages for video configuration vs. input configuration09:26
wgrantIt's fairly big and confusing as is, and there's little reason to squish them together.09:27
tjaaltonwgrant: serial devices for instance09:28
tjaaltonI'm wondering if the current wacom hal fdi is useful at all09:28
wgranttjaalton: That's what I thought. Hmm.09:28
wgrantWouldn't some wacom devices be USB?09:28
tjaaltonsince to be able to configure wacom fully, you need to add another fdi file to _not_ load the driver, and use the old xorg.conf you had for ages09:28
wgrantSo hal-detectable?09:29
tjaaltonmost attachable are09:29
wgrantWhy not use an fdi file with the appropriate config info?09:29
tjaaltonbut tablets tend to be serial onew09:29
tjaaltonwgrant: not possible09:29
tjaaltonsince you can only load the driver once, when the old method does that three times09:30
tjaaltonfor the different "devices", stylus, pen etc09:30
wgrantThat's special.09:30
wgrantReally special.09:30
tjaaltonthe driver is just lacking here09:30
tjaaltonbut who knows how long it'll take for them to fix that09:30
tjaaltonbryce: you have an opinion about this?09:31
wgranthttp://www.qeuni.net/f/1/2008/g-s-d_2.24.0-0ubuntu2.diff is the g-s-d debdiff (a segfault fix and some deconfused warnings are also there, which I guess aren't strictly necessary)09:32
brycetjaalton: I haven't looked into wacom for a few weeks but I wasn't able to get it configured using i-h; only by disabling it and doing it the old fashioned way with xorg.conf09:32
brycetjaalton: so yeah I don't know that the current wacom fdi is of use to anyone09:32
tjaaltonbryce: yeah, maybe I'll drop it then09:33
wgranthttps://code.edge.launchpad.net/~wgrant/gnome-control-center/bug-274728 is the g-c-c change.09:33
brycewgrant: anyway if you feel like breaking out / rewriting the intrepid input config stuff, knock yourself out.  I did the current stuff back before we really knew how it would work; perhaps it would have been better to not write at all.  ;-)09:34
wgrantbryce: I was planning on at least refactoring it when I had time.09:34
brycewgrant: great thanks09:35
tseliottjaalton: how can you use an fdi file to prevent a driver from loading?09:36
tjaaltontseliot: use the same logic and leave x11_driver empty09:36
tseliottjaalton: but you can still set properties, can't you?09:38
tjaaltontseliot: if the driver is set, yes09:39
tseliotoh but this would involve adding a section for each device in xorg.conf09:39
tseliotwith the driver09:39
tseliotgood night bryce09:40
wgrantNight bryce.09:41
tseliottjaalton: if application has to configure a wacom tablet it should 1) add 3 InputDevice sections to xorg.conf and set the driver 2) customise the fdi file. Is this correct?09:42
* tseliot will buy a wacom tablet for testing sooner or later09:42
tjaaltontseliot: I'll drop the fdi file from the driver, since it leaves the device partly unconfigured09:43
tjaaltonso you only need to do 1)09:43
brycetseliot: msg me your postal address and I'll send you one09:44
tseliotbryce: really? Thanks a lot :-)09:44
tjaaltontseliot: is this for intrepid or jaunty?09:47
tseliottjaalton: jaunty09:48
tjaaltontseliot: would be nice if it was backportable to intrepid09:52
tjaaltonto shut up all the wacom-nuts :)09:52
tseliottjaalton: we would have to backport the input module of python-xit too09:53
tseliotbut it would be trivial to do so09:53
tjaaltonwgrant: you need a sponsor?10:25
elmargolstill no solution for my nvidia problem. Is there a way to contact a nvidia developer?10:26
tjaaltonelmargol: nvidia-bug-report.sh10:26
elmargoltjaalton: i have to use the binary drivers for this?10:27
tseliotelmargol: try here: http://www.nvnews.net/vbulletin/forumdisplay.php?f=1410:27
tjaaltonelmargol: you need to have it installed, yes10:28
wgranttjaalton: I do - I'm only a MOTU.10:32
tjaaltonwgrant: I can upload g-s-d, but don't know what needs to be done for g-c-c10:33
wgranttjaalton: You need to merge my branch into ~ubuntu-core-dev/gnome-control-center/ubuntu10:33
wgrantAnd then upload as normal.10:33
wgrantseb128: Are there instructions for that around?10:34
wgrantAll packages seem to do it differently, and g-c-c is particularly odd.10:34
=== crevette_ is now known as crevette
seb128mvo: that's for you ;-)10:34
tjaaltonI reviewed the merge10:34
tseliotmaybe you can ask mvo to merge from your branch10:34
seb128wgrant: my bzr foo is limited, mvo made me use it ;-)10:34
* wgrant knows how to use it in general, but this is a very strange way to do things that I've not seen before.10:35
seb128I do how to get, pull, push basically10:35
seb128mvo: do you think you could do that upload? and maybe sneak the patch on bug #159996 too which is waiting for sponsoring? ;-)10:35
ubottuLaunchpad bug 159996 in gnome-control-center "Appearance Preferences offer Visual Effects without installed Compiz" [Low,In progress] https://launchpad.net/bugs/15999610:35
tjaaltong-s-d uploaded10:41
tjaaltonalong with xinput and the drivers, so only g-c-c missing now :)10:41
wgrantExcellent. Thanks tjaalton.10:43
elmargol2.6.27 was released 11 hours ago. This means we get a new kernel now?10:47
elmargolI think we are using a RC at the moment?10:47
tjaaltonelmargol: uploaded already10:51
tjaaltonrebased to 2.6.2710:51
elmargolah cool maybe this fixes my problem10:52
tjaaltonwgrant: hmm, you didn't bump the build-deps on g-s-d / g-c-c?10:52
tjaaltong-s-d failed to build on lpia10:53
tjaaltonbut we can retry it when the new versions have landed10:53
tjaaltonfails on other archs too, of course10:53
tjaaltonwgrant: seems that g-s-d does not directly build-depend on libxi-dev, so there was nothing to bump :)10:55
wgranttjaalton: Sorry, forgot to tell you that.11:01
tjaaltonwgrant: no problem, I added it and uploaded a new version11:02
tjaaltong-c-c needs the same?11:02
wgranttjaalton: It will probably be uploaded sufficiently late.11:05
mvowgrant: sure, I can merge that branch11:10
wgrantlibxi-dev is published on most archs, so it should be fine without the build-dep bump in a few minutes.11:12
mvowgrant: where is your branch?11:13
wgrantmvo: It should be visible on your branch page as proposed for merging... but it's lp:~wgrant/gnome-control-center/bug-27472811:14
mvowgrant: thanks, merging now11:15
wgrantmvo: Thanks.11:17
wgrantmvo: Did you end up uploading g-c-c?12:53
wgrantIt's uninstallable now, and people have already started noticing.12:56
seb128users are annoying, they are using an unstable version they should wait when something is transitioning and not start making noise13:02
mvowgrant: I have not uploaded it yet, I was collecting more changes13:02
mvobut if it is uninstallable (why?) I can do the upload now13:02
tjaaltonlibxi Breaks the current version13:03
wgrantseb128: True.13:03
mvoseb128: I disagree with the patch for the appearance applet13:03
mvoseb128: I think it should set the option to insensitive but not hide them13:03
wgrantHiding them sounds strange...13:03
mvoI will put that into the report13:04
seb128mvo: that's fair enough, I'm not such in such cases, when it's unsensitive users wonder why and how to enable it13:04
seb128it's done this way for menus to not change the layout dynamically which would be confusing13:05
seb128mvo: that might be worth pinging mpt just for having his opinion13:05
mvoseb128: agreed, I will do that13:05
mvoideally we would have a "click here to install" box13:05
seb128mvo: right13:07
mvoseb128: the gnome-panel upload is now ready too, do you want to review the debdiff or can I just go ahead?13:09
seb128mvo: you know what you are doing just upload ;-)13:10
mvo(most of the time :P)13:10
mvoI just need to add this Vcs-Bzr header, then its ready13:11
* seb128 hugs mvo13:27
wgrantmvo: Thanks.13:29
tjaaltonjcristau: turns out it's not that simple to add fallback-drivers to non-autoconfig setup, since if there's a conffile present, autoConfigDevice returns only the one matched driver15:28
tjaaltonat least it should be possible to hack listPossibleVideoDrivers not to run videoPtrToDriverList if matchDriverFromFiles succeeds15:30
tjaaltonand if mDFF fails, use vesa15:30
tjaaltonum, if they both fail15:30
tjaaltonbah, I give up for the day15:39
mvotseliot: your script is still not merged, sorry. I had to do a "emergency" update-maanger upload because there were a lot of people with hangs in u-m17:04
tseliotmvo: no problem17:05
tseliotit's not urgent17:05
crevettedoes the xorg-server update include the performance fix ?17:30
crevette(the underline question, should I restart my session :))17:31
tjaaltoncrevette: no, it was dropped upstream17:36
crevettetjaalton: you mean it was commited or not approved18:10
tjaaltoncrevette: committed and reverted18:15
tjaaltonABI change18:15
crevetteah 18:15
tjaaltonwell, sort of18:16
crevetteso it is not for 1.5 18:16
crevettetoo bad :/18:16
tjaaltonbryce: btw, the serial-mouse bug.. I think it belongs to casper18:26
brycetjaalton: oh?  why's that?18:28
tjaaltonsince the livecd is pretty hard to use without a mouse18:29
tjaaltonalthough there would still be the problem of getting the same information to the installed system18:29
bryceiirc there was a question on using the approach mandriva or fedora uses for this situation (wiggling mouse or whatever); I take it that's something that could be done entirely within casper?18:31
superm1is serial mouse really something that is supported on machines in this day and age though?18:34
pwnguintabletPC is a serial wacom18:35
tjaaltonsuperm1: apparently yes, in less-developed countries..18:35
pwnguindoes that count?18:35
superm1tjaalton, ah i see18:35
tjaaltonsuperm1: I asked the very same question, and got a response :)18:35
tjaaltonbryce: they have a different way to install the system...18:36
tjaaltonfor us the primary way to install is via a live-session18:36
brycetjaalton: ah too bad18:36
brycewell at least that answers that18:36
tjaaltonevand/cjwatson know better how to tackle the issue..18:38
johanbrI did some further experimenting with the projector I was using (-ati driver). This time it refused to detect *anything*. If I switched to a console, the projector would sync and I could see the console text. When I switched back to X, the projector lost sync.18:38
brycehmm, mandriva chose to ship with xorg 7.3 - http://wiki.mandriva.com/en/2009.0_Notes22:09
tseliotbryce: I don't know how accurate this page is but it says that Mandriva's xorg-server is still 1.4.2: http://distrowatch.com/table.php?distribution=mandriva23:12
bryceprobably about right23:13
brycetseliot: I'd been looking to see how they handled the -fglrx / xorg 1.5 issue, and see they've just punted on it23:13
tseliotbryce: yes, I was curious about their solution too23:15
tseliotbut they couldn't revert the ABI changes (the screenprivateindex thing)23:16
bryceit sort of makes me wonder what life would be like if we'd also stuck with 1.4.x23:16
tseliothmm... I wouldn't know23:17
tseliothopefully AMD and NVIDIA will release a "fix" soon23:18
wgrantWe can't let proprietary drivers delay us too much.23:26

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