RAOF | cnd: | 00:00 |
---|---|---|
apw | bryceh, its constantly on my mind, and i am hoping to play with it one of these fine days ... its a pig to fix | 00:01 |
RAOF | cnd: http://paste.ubuntu.com/570833 | 00:01 |
bryceh | apw, ok thanks. | 00:02 |
=== wgrant_ is now known as wgrant | ||
cnd | RAOF, the first set of warnings is from the upstream cherry pick to fix the xi 1.x keyboard events | 00:02 |
cnd | so not from me :) | 00:02 |
cnd | but the second one does look like it's from me | 00:02 |
cnd | I'll take a look | 00:03 |
RAOF | :) | 00:03 |
cnd | hmmm, I don't get that warning when I build here | 00:04 |
cnd | RAOF, what's at line 1696 in your source tree | 00:04 |
cnd | I just have a declaration for the client variable | 00:04 |
cnd | RAOF, I can say that it can't be used uninitialized | 00:05 |
cnd | but maybe gcc didn't pick up on it | 00:06 |
cnd | the patch needs to be cleaned up more, but I didn't have time to make it sparkle and shine :) | 00:06 |
RAOF | cnd: 1696 is just TouchClientPtr client; here. I'm not sure what gcc thinks it's complaining about! | 00:07 |
RAOF | It would be rather more instructive if it pointed out *where* it thought it was used uninitialised :) | 00:08 |
cnd | RAOF, that's what it usually does I think | 00:08 |
cnd | so I think your gcc is confused :) | 00:08 |
cnd | which version of gcc do you have? | 00:08 |
cnd | I have 4:4.5.1-1ubuntu3 | 00:09 |
cnd | should be up to date | 00:09 |
cnd | unless a new one was released just today | 00:09 |
RAOF | 4.5.2\3umusnu2 | 00:18 |
RAOF | cnd: Ahem. 4.5.2-3ubuntu2 | 00:18 |
cnd | RAOF, where'd you get that version? | 00:32 |
cnd | I did an apt-get update | 00:32 |
cnd | and it's not listed | 00:32 |
RAOF | It's the version in my chroot. Let me check. | 00:32 |
RAOF | Oh! It's gcc-4.5 | 00:33 |
cnd | ahh, I've got 4.5.2-ubuntu2, and 4.5.2-ubuntu3 is a candidate | 00:34 |
RAOF | Ah. And you've also been running ‘apt-cache policy gcc’, rather than “gcc --version”, I'd wager ;) | 00:34 |
cnd | yeah | 00:35 |
cnd | RAOF, hope things are going well? | 02:18 |
RAOF | Yup. | 02:18 |
RAOF | Testing is going well. | 02:19 |
cnd | awesome | 02:19 |
RAOF | Synaptics seems to still work, evdev still works, the server doesn't crash… :) | 02:19 |
cnd | that's good :) | 02:20 |
RAOF | Everything rebuilds (apparently) correctly… | 02:20 |
RAOF | Shouldn't be long before I'm trolling for sponsors. | 02:21 |
cnd | RAOF, you're not a core dev? | 02:22 |
RAOF | No. | 02:22 |
cnd | I would assume you'll be shortly enough, right? | 02:22 |
RAOF | Yeah. | 02:22 |
* RAOF schedules tomorrow afternoon for finishing the bits of the process needed. | 02:23 | |
cnd | RAOF, I was playing some community built multitouch games :) | 04:15 |
cnd | and I found a crasher | 04:15 |
cnd | I think they use sdl, and sdl is actively grabbing the pointer | 04:15 |
RAOF | Hey, cool! They exist? | 04:15 |
cnd | and that's caused a bug | 04:15 |
cnd | yeah :) | 04:15 |
cnd | like 7 of them | 04:15 |
cnd | https://launchpad.net/~oxullo/+archive/libavg?field.series_filter=natty | 04:15 |
cnd | so I've got a patch | 04:15 |
RAOF | Ok. | 04:16 |
cnd | if you're too far down testing, you don't need to take it right now | 04:16 |
cnd | what would you like to do | 04:16 |
RAOF | How safe is the patch? | 04:16 |
cnd | it only affects touch event handling | 04:16 |
cnd | so it has no effect on non-multitouch hardware | 04:16 |
RAOF | I don't mind picking up a patch to fix a crasher bug :) | 04:16 |
cnd | and it only affects the code path where you are beginning a touch while there's an active pointer grab | 04:17 |
cnd | ok | 04:17 |
cnd | let me fix it up and push it | 04:17 |
RAOF | I'll check git is up to date. | 04:17 |
RAOF | I presume it's xserver? | 04:17 |
cnd | yeah | 04:18 |
RAOF | xserver's now up to date in git. | 04:19 |
cnd | k | 04:19 |
cnd | oh the crazy things people do with X device grabs... | 04:19 |
RAOF | And I'm going to go and grab some stuff from the shops while the world does its final test rebuild. | 04:20 |
cnd | RAOF, tis pushed | 04:22 |
cnd | thanks! | 04:22 |
cnd | and now I'm off to bed | 04:42 |
RAOF | Sleep well! | 04:43 |
RAOF | Tell me again: why am I wasting builder cycles rebuilding xserver-xorg-video-nv? :) | 05:13 |
RAOF | Anyone up for some X sponsoring? | 09:50 |
lilstevie | ? | 10:06 |
tjaalton | RAOF: yep | 10:08 |
tjaalton | downloading.. | 10:29 |
tjaalton | RAOF: uploading | 11:26 |
RAOF | tjaalton: Rockit rockin. | 11:26 |
RAOF | Thanks muchly. | 11:27 |
tjaalton | ...aand done | 11:27 |
RAOF | Feel free to add to https://wiki.ubuntu.com/ChrisHalseRogers/CoreDevApplication ; I'll start finishing that process soon :) | 11:28 |
tjaalton | hehe :) | 11:28 |
tjaalton | wonder if there's a way to batch-debsign a bunch without having to enter the gpg passphrase a hundred times | 11:36 |
seb128 | tjaalton, use a gpg agent? | 11:40 |
tjaalton | seb128: ok, didn't know of such a beast.. | 11:41 |
seb128 | tjaalton, seahorse-plugins does it | 11:42 |
seb128 | or you can use gnupg-agent | 11:42 |
tjaalton | seb128: yeah, one more thing to fix on my setup, thanks.. dunno why this is the first time I actually thought about it :) | 11:47 |
tjaalton | tormod: hey, got my t23 reinstalled with natty. do you have problems with getty not working right, ie. the vc:s don't work? | 13:30 |
tormod | tjaalton, no but haven't tried natty on it for some weeks | 13:32 |
tormod | tjaalton, btw https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-savage/+bug/723035/comments/2 | 13:32 |
ubot4 | Launchpad bug 723035 in xserver-xorg-video-savage (Ubuntu) "please merge xserver-xorg-video-savage 1:2.3.2-3 from Debian unstable main (affects: 1) (heat: 8)" [Undecided,New] | 13:32 |
tormod | tjaalton, btw2 for your signing agent, see also http://bazaar.launchpad.net/~xorg-edgers/xorg-server/xorg-pkg-tools/view/head:/README.ppa-update | 13:33 |
tjaalton | tormod: thanks, I'll upload that | 13:34 |
tormod | tjaalton, I have verified that it builds fine on natty | 13:36 |
tormod | (after installing the new xserver-xorg-dev) | 13:37 |
tormod | (which it build-depends on :) ) | 13:37 |
tormod | tjaalton, you're blaming getty? just flashing cursor on the consoles? | 13:38 |
tjaalton | tormod: yep | 13:38 |
tjaalton | restartin them doesn't do anything | 13:38 |
tjaalton | thought if it's a general issue with ums | 13:39 |
tormod | tjaalton, historically these have been upstart issues, upstart related I mean, there are no bugs in upstart :) | 13:41 |
tjaalton | yeah it's faultless | 13:41 |
tjaalton | getty's are running though | 13:41 |
tormod | and the corresponding /dev/tty? exist? | 13:42 |
tjaalton | yep | 13:43 |
tormod | can you try something like: echo yooo | sudo tee /dev/tty3 | 13:44 |
tjaalton | nothing shows up there | 13:45 |
tjaalton | uhm | 13:45 |
tormod | and boot with "text" to see if X is possibly at fault | 13:45 |
tjaalton | you mean if I get something back? yes | 13:45 |
tjaalton | right | 13:46 |
tormod | back? | 13:46 |
tjaalton | the terminal where I ran that | 13:47 |
tjaalton | displayed what was echo'ed | 13:47 |
tormod | tee always gives back - your best friend | 13:47 |
tjaalton | ah right.. | 13:47 |
tormod | next I would try reading from /dev/vcs3, too see if the kernel thinks something is there | 13:52 |
tjaalton | removing 'quiet splash' did nothing | 13:53 |
tjaalton | all the vt's have the same text printed by grub | 13:54 |
tjaalton | so maybe it's grub's fault | 13:54 |
tormod | but you see the kernel/etc messages while booting? | 13:54 |
tjaalton | no | 13:54 |
tormod | sometimes, after an X crash, my savage card is messed up and the consoles do not refresh correctly, until power cycle | 13:55 |
tormod | well they refresh only when switching consoles | 13:56 |
tormod | you could try booting a cd -> no grub | 13:57 |
tjaalton | ha, it's the gfxpayload-stuff | 13:57 |
tjaalton | removing that line fixed the console | 13:58 |
tormod | was there an incorrect gfxpayload line?= | 13:59 |
tjaalton | it doesn't seem to work right with ums | 14:00 |
tjaalton | hmm probably was incorrect | 14:02 |
tjaalton | because it had the variable, not a value in it ("set gfxpayload=$linux_gfx_mode") | 14:03 |
tjaalton | I'll try with =text | 14:04 |
tormod | google gave me this http://forums.debian.net/viewtopic.php?f=5&t=41881 same story with wrong variable names | 14:04 |
tormod | well if you don't see boot messages, booting with "text" will leave you with black screens | 14:05 |
tormod | I guess you would have had the same loss of console text with KMS *until* KMS kicks in and resets the modes | 14:06 |
tjaalton | perhaps | 14:07 |
tjaalton | with "text" i do indeed get the vt's | 14:07 |
tjaalton | er, vc's | 14:07 |
tormod | yes but now without that gfxpayload? or still with? | 14:08 |
tjaalton | with | 14:08 |
tjaalton | swapped the variable with 'test' | 14:08 |
tjaalton | text | 14:08 |
tormod | oh I see, I was thinking about the boot option "text" (for gdm) | 14:09 |
tjaalton | that didn't seem to work at all | 14:09 |
tjaalton | oh well | 14:09 |
tjaalton | actually yeah it did | 14:09 |
tjaalton | just that the console didn't :) | 14:09 |
tormod | this was a clean install? so grub conf is broken? | 14:10 |
tjaalton | yes | 14:10 |
cnd | bryceh, I think there's a problem with the input modules RAOF uploaded | 15:46 |
cnd | I think they got built against the previous xserver-xorg-dev | 15:46 |
cnd | which means they popped out with a dependency on xserver-input-abi-12.1 instead of xserver-input-abi-12 | 15:46 |
cnd | tjaalton, ^^ (in case bryceh isn't up yet) | 15:48 |
tjaalton | gah | 15:53 |
tjaalton | so a rebuild? | 15:53 |
tjaalton | +needed | 15:53 |
cnd | I think so | 15:54 |
cnd | xorg-server is built now for all archs | 15:54 |
cnd | so we should be good to go now | 15:54 |
tormod | tjaalton, can you do the mga sync also? LP 722877 | 15:58 |
ubot4 | Launchpad bug 722877 in xserver-xorg-video-mga (Ubuntu) "please sync xserver-xorg-video-mga 1:1.4.13.dfsg-3 from Debian unstable (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/722877 | 15:58 |
tjaalton | tormod: yes, though I'll sort the rebuilds first, to see what's needed | 15:59 |
tjaalton | cnd: give an example of a driver that built against the wrong version? | 16:01 |
cnd | I'm providing info from pmcgowan | 16:02 |
cnd | he saw bad deps on evdev | 16:02 |
tjaalton | at least joystick is good | 16:02 |
cnd | I don't have an i386 machine, and I haven't manually downloaded the i386 debs to check them | 16:03 |
cnd | and amd64 isn't built yet | 16:03 |
tjaalton | https://launchpad.net/ubuntu/natty/i386/xserver-xorg-input-evdev/1:2.6.0-1ubuntu7 | 16:03 |
tjaalton | looks normal | 16:03 |
tjaalton | maybe he looked at the old package? | 16:03 |
cnd | tjaalton, ahh, didn't know you could see deps there | 16:04 |
cnd | it does look right | 16:04 |
cnd | maybe he does have some stale state | 16:04 |
tjaalton | *phew* | 16:04 |
tjaalton | so no need for rebuilds.. | 16:04 |
cnd | he was still getting ubuntu6 evdev | 16:07 |
cnd | so I think his mirror was just slow to sync | 16:07 |
tjaalton | yep | 16:07 |
cnd | he was piecemeal giving me info cause he was on another computer and couldn't copy/paste | 16:08 |
cnd | and I just didn't check that... | 16:08 |
tjaalton | great, my dns is broken | 16:21 |
tjaalton | can't dput anything | 16:21 |
tjaalton | had to copy elsewhere | 16:25 |
tjaalton | tormod: you didn't build savage with -sa | 16:25 |
apw | is the threat to uninstall ubuntu-desktop a temporary thing? | 16:29 |
tjaalton | so it got rejected because of missing tarball | 16:29 |
tjaalton | apw: due to? | 16:29 |
tormod | tjaalton, possibly, but do you not need to debuild it again yourself? | 16:29 |
tjaalton | tormod: no, I'll just debsign it | 16:30 |
tormod | aha | 16:30 |
apw | tjaalton, actually how do i tell? | 16:30 |
tjaalton | apw: aptitude should show | 16:30 |
tormod | tjaalton, should I fix it? | 16:30 |
tjaalton | tormod: well I could rebuild it, either is fine | 16:31 |
apw | tormod, ahh xserver-xorg-vide-geode | 16:31 |
apw | which is depending on abi-0.9 | 16:31 |
tjaalton | yeah it's missing from the rebuilds | 16:32 |
tormod | tjaalton, I fixed it | 16:32 |
apw | tjaalton, so i assume its being fixed ... and i just wait a bit | 16:33 |
tjaalton | apw: yep | 16:33 |
tjaalton | tormod: thanks | 16:33 |
Sarvatt | mga sync? does it have the xinerama fix in it or are we regressing https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-mga/+bug/292214 ? | 16:47 |
ubot4 | Launchpad bug 292214 in xserver-xorg-video-mga (Ubuntu Natty) (and 2 other projects) "Xinerama broken since intrepid on MGA (affects: 14) (dups: 1) (heat: 110)" [High,Fix released] | 16:47 |
* Sarvatt looks | 16:47 | |
jcristau | it's there | 16:49 |
Sarvatt | phew, its all good | 16:49 |
jcristau | xserver-xorg-video-mga (1:1.4.11.dfsg-4+squeeze1) unstable; urgency=low | 16:49 |
bryceh | morning | 16:55 |
ricotz | bryceh, hi | 18:35 |
bryceh | ricotz, hey | 18:39 |
ricotz | bryceh, i had a look at wayland | 18:39 |
ricotz | bryceh, i think your package will get rejected | 18:39 |
ricotz | lintian complains about a lot fsf copyrights | 18:40 |
ricotz | i packaged the latest git http://people.ubuntu.com/~ricotz/wayland/ | 18:41 |
ricotz | it is better using debsrc3 and newer compat level | 18:41 |
cnd | bryceh, (or anyone else familiar with mesa), can you check out this bug: https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/259219 | 18:44 |
ubot4 | Launchpad bug 259219 in mesa (Ubuntu) "Broken TLS support in libGL.so (affects: 1) (heat: 8)" [Undecided,Expired] | 18:44 |
cnd | one of the multitouch games developers is hitting this | 18:45 |
cnd | we're wanting to push them to universe, but this is causing issues | 18:45 |
bryceh | ricotz, thanks but aiui that won't work with the natty mesa. Did you test it? | 18:46 |
ricotz | bryceh, i didnt test it, but it compiles | 18:46 |
ricotz | bryceh, this is only the library, the demos are in a new branch wayland-demos which would be a new source package | 18:47 |
bryceh | ricotz, yeah it'll probably crash when you run it | 18:47 |
ricotz | but as i said there are a lot copyright complains | 18:48 |
bryceh | ricotz, yeah probably would be too late to get wayland-demos into natty anyway, given that FF is tomorrow | 18:48 |
bryceh | regarding the gpl issue, I already uploaded a fixed version yesterday | 18:48 |
ricotz | ok, then i might be an issue with the newer source | 18:49 |
bryceh | well, the 'fix' was just to mention gpl in debian/copyright | 18:49 |
bryceh | I don't think at this point we care whether it's mit or gpl really | 18:50 |
ricotz | ok | 18:50 |
ricotz | bryceh, did you take a peek at the debian.tar.gz? | 18:50 |
bryceh | cnd, ironically wayland needs --enable-glx-tls (at least, the snapshot version I'm trying to get into natty needs it) | 18:50 |
bryceh | ricotz, no | 18:51 |
jcristau | whether debsrc3.0 is better is.. arguable. | 18:51 |
ricotz | i think it is easier to maintain | 18:51 |
cnd | bryceh, hrm... so does wayland need to be fixed? | 18:52 |
bryceh | cnd, I don't know if we needed it for anything else besides wayland, will have to check with raof | 18:53 |
cnd | bryceh, it's not a show stopper, we can hack around it | 18:53 |
Sarvatt | very very very arguable :) | 18:53 |
bryceh | cnd, rather than reopen such an ancient bug ;-), could you file a new bug and explain what the current breakage is in more detail? | 18:53 |
cnd | install wrapper scripts to export LD_PRELOAD before calling the real binary | 18:54 |
jcristau | can they talk to mesa upstream? | 18:54 |
cnd | bryceh, tbh, I'm just following what others were saying | 18:54 |
Sarvatt | goes well if you like the packaging workflow libxcb and cairo use I guess | 18:54 |
cnd | but I'll ask the people who know more to do that | 18:54 |
bryceh | ricotz, thanks, yeah that looks better than my cargo culted rules file ;-) | 18:56 |
bryceh | ricotz, btw I've contacted kibi about wayland packaging, and am planning on coordinating with him to get the packaging consistent with how debian wants it | 18:57 |
ricotz | bryceh, yeah ;), but the copyright needs an rewrite to fit the new specs | 18:58 |
ricotz | bryceh, it should be formatted this way http://paste.debian.net/108504/ | 18:59 |
jcristau | err. no. | 19:00 |
ricotz | jcristau, no? | 19:01 |
jcristau | it can if you're that bored. it doesn't need to. | 19:01 |
jcristau | debian/copyright is free-form | 19:01 |
ricotz | oh, i thought it should follow these dep5 quideline | 19:02 |
apw | tjaalton, no sign of the archive becoming cnsistant ... does a build need a rescore? | 20:33 |
tjaalton | apw: at least geode is ready | 20:34 |
tjaalton | apw: so is it some other driver now? | 20:35 |
apw | tjaalton, oh you are off the hook it seems, now language-selector is to blame | 20:41 |
tjaalton | apw: whee :) | 20:42 |
apw | stupid how apt-get doesn't tell you why | 20:42 |
=== Amaranth_ is now known as Amaranth | ||
cnd | apw, if you update to the latest x bits, you'll now have multitouch through X from your dell mini trackpad :) | 21:46 |
cnd | though you can't easily test it with anything yet | 21:46 |
jcristau | xinput! :) | 21:47 |
apw | cnd, if the archive is every consistant again i would update ... though my mini is a little sick right now | 21:47 |
cnd | so the trackpad from hell has more use now | 21:47 |
cnd | jcristau, xinput hasn't had touch events added :( | 21:47 |
cnd | apw, sorry to hear that, hope she gets well soon | 21:47 |
cnd | RAOF, I hit this bug: https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/259219 | 21:47 |
ubot4 | Launchpad bug 259219 in mesa (Ubuntu) "Broken TLS support in libGL.so (affects: 1) (heat: 8)" [Undecided,Confirmed] | 21:47 |
cnd | I'm going to duplicate a clean bug for it | 21:48 |
cnd | but I started digging into it | 21:48 |
cnd | I'm thinking we should be going with solution #2 as stated in the bug description | 21:48 |
cnd | if you want background reading: https://docs.google.com/viewer?url=http://people.redhat.com/drepper/tls.pdf | 21:48 |
cnd | I'm still trying to digest it myself :) | 21:48 |
cnd | btw, thanks for the uploads! everything seems to have gone smoothly | 21:49 |
jcristau | cnd: yeah reading drepper's tls paper :) | 21:55 |
cnd | jcristau, if I read it right, it seems like mesa has it completely wrong :) | 21:58 |
cnd | it's selecting the initial-exec model, which seems to only be valid if you statically link against mesa | 21:58 |
Sarvatt | I like option 3) ignore libavg :D | 21:58 |
jcristau | cnd: so your app dlopens libGL? | 21:59 |
jcristau | (or dlopens something linked against libGL) | 21:59 |
cnd | jcristau, libavg does it appears | 21:59 |
jcristau | ok | 21:59 |
cnd | I don't know if there's a difference between implicit dynamic loading or dlopen for the tls model | 22:00 |
jcristau | bah and docs/dispatch.html doesn't explain why it's using that tls_model | 22:03 |
cnd | yeah | 22:05 |
cnd | I'm actually about to troll through git history | 22:05 |
cnd | to see if there are any clues | 22:05 |
jcristau | it does seem wrong.. | 22:05 |
cnd | I don't think people usually use tls models | 22:05 |
cnd | so I think this is really an effort to be as fast as possible | 22:05 |
cnd | and it's just stepping too far | 22:05 |
RAOF | I'm trying to remember whether I've seen that discussed upstream. I've got a vague recollection of it. | 22:07 |
cnd | this was the bug that started it all: https://bugs.freedesktop.org/show_bug.cgi?id=1822 | 22:10 |
cnd | and this is the commit that resulted: https://bugs.freedesktop.org/show_bug.cgi?id=1822 | 22:10 |
cnd | oops | 22:10 |
ubot4 | Freedesktop bug 1822 in libGL "libGL (and DRI drivers) should support TLS" [Enhancement,Resolved: fixed] | 22:10 |
cnd | 25fe93f0a11e6f4c8d470441ff91b9cddf7b3023 | 22:15 |
* jcristau just got there | 22:17 | |
jcristau | and reading the bug it references.. | 22:19 |
=== merriam is now known as 5EXAB8WYE | ||
jcristau | cnd: so the initial-exec stuff was already in http://marc.info/?l=dri-devel&m=105187260618425&w=2 | 22:29 |
jcristau | which seems to be where this stuff came from | 22:29 |
jcristau | silly netsplits.. | 22:33 |
=== mdeslaur_ is now known as mdeslaur | ||
=== RAOF_ is now known as RAOF | ||
bryce_ | [ubuntu/natty] wayland 0.1~git20101129.ac93a3d3-0ubuntu1 (Accepted) | 23:42 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!