[03:27] bryceh, have you and RAOF looked into gtest? [03:27] I'm looking at your unit tests for xrandr-utils [03:27] Which one is gtest? [03:27] RAOF, google-test [03:27] I'm in love with it [03:27] Ah, ok. That one. [03:27] just thought I'd mention it [03:27] I'm not really familiar with it. What does it do to make you fall in love with it? [03:28] it's based on c++, which means test environments, fixtures, and such can all be managed in a hierarchical way [03:28] and you don't have to write a main() or register any test cases [03:28] * RAOF sees that it's written in C++, which may result in madness. [03:29] RAOF, you can check out xorg-gtest, which we posted to xorg-devel today [03:29] Hm. How does automatic test-case detection work in a language with no introspection? [03:29] we're using it in utouch-frame [03:29] RAOF, magic? [03:29] Aieee. [03:29] RAOF, check out http://people.freedesktop.org/~cndougla/xorg-gtest/ [03:30] look at the example [03:30] that's literally all you need to code if you want to have something that builds and tests [03:30] there's no other code that has to be written [03:31] That looks like something that could indeed be useful. [03:31] We should really get rid of the "need to be root" bit, there. [03:32] RAOF, I don't think it's possible if you want to fire up an Xorg server [03:32] It should be possible with the dummy drivers, but it isn't at the moment. [03:32] because any config file you specify must be in /etc/X11 or /usr/share/X11/xorg.conf.d, *unless* you are root [03:32] yeah, that's true [03:33] it still doesn't help us for the touch team because we need to use uinput to replay device events [03:33] which requires root to interact with the kernel [03:33] but I'm not really suggesting that xorg-gtest would be worth your while as-is [03:34] because the dummy video driver probably won't get you very far with xrandr [03:34] but gtest in general is going to be much better than rolling your own [03:34] I was skeptical at how much better it could make my life, but I really love it after having developed tests and frameworks with it [03:35] Yeah, I'm a big fan of not rolling my own. [03:36] Actually, aren't there patches to enable xrandr on the dummy driver? [03:36] I wouldn't know [03:36] if there were, then xorg-gtest would be perfect :) [03:36] The documentation there also suggests that it can hook you up with an existing X11 server, which would be the next best thing. [03:36] yeah [03:37] by default when you run tests with --no-dummy-server it will use $DISPLAY [03:38] and I suppose --no-dummy-server can be used when not root [03:38] anyways, time to eat dinner and stop working! [03:38] back tomorrow! [03:41] :) [05:04] cnd, interesting. I'm trying to keep xrandr-utils close to X standard practices; what I've got in there parallels the xserver test suite [05:04] I like C++ myself but know keithp is not a fan. [05:06] heya RAOF, btw I tossed the xrandr-utils patches up on xorg-devel@ a couple hours ago; thanks for your review earlier, I took out that perl test. [05:07] bryceh: Sweet; I'll throw some reviewed-by tags your way. [05:09] RAOF, thanks! [05:10] I had another chance today to learn just how much git rebase kicks ass [05:11] Sometimes your own :) [05:23] RAOF: cool, I'll give it a go today :) [05:24] tjaalton: It seems to do weird things to wine; I'll be interested to see if you also notice that. [05:24] after going through the xmas shopping mayhem.. [05:24] ah, does wine work on amd64 currently? [05:27] You can't install ia32-libs-multiarch, but equivs will let you install ia32-libs which will let you install wine. Then it's just a matter of hunting down libfoo:i386 :) [05:35] ugh, ok [06:11] RAOF, I stuck the ppa on a radeon box. Boots to the ubuntu logo, shows a mouse cursor, but no greeter [06:12] Huh. Odd. [06:12] Anything interesting in the logs? [06:12] bryce@dorset:~$ apt-cache policy lightdm-gtk-greeter [06:12] lightdm-gtk-greeter: [06:12] Installed: (none) [06:12] Candidate: 1.0.6-0ubuntu4 [06:12] Heh. [06:12] Oh, no. You're after unity-greeter rather than lightdm-gtk-greeter. [06:13] ah right $ apt-cache policy unity-greeter [06:13] unity-greeter: [06:13] Installed: 0.1.1-0ubuntu1 [06:13] Candidate: 0.1.1-0ubuntu1 [06:14] skimming through logs... nothing interesting so far [06:14] bryceh, are you becoming a git convert? [06:14] root@dorset:/var/log/lightdm# more x-0-greeter.log [06:14] (unity-greeter:1864): Gtk-WARNING **: cannot open display: :0 [06:15] ah, potentially I had fglrx installed [06:15] Ah. I haven't uploaded an fglrx to the ppa. [06:15] hmm, no evidence so far [06:16] There's a new -ati in there (I think); I may have bollocksed that up. [06:20] xserver-xorg-video-ati: [06:20] Installed: 1:6.14.99~git20111219.aacbd629-0ubuntu1 [06:20] Candidate: 1:6.14.99~git20111219.aacbd629-0ubuntu1 [06:22] Yeah, that's a newer snapshot than in the main archive. [06:23] Which should be fine. :) [06:23] yeah it feels like X is basically ok, it's just the greeter is refusing to do its thing [06:24] actually when I restart lightdm I see a brief glimpse of my desktop [06:25] then it comes back to a black screen with movable mouse cursor. [06:26] switching to vt1 and back to vt7 gives me the console text with mouse cursor. maybe something's not painting [06:26] [ 914.068497] lightdm[915]: segfault at aaaaaaae ip 08051ee8 sp bff44560 error 4 in lightdm[8048000+28000] [06:26] [ 914.070871] init: lightdm main process (915) killed by SEGV signal [06:26] ooh [06:27] hmm that might be irrelevant though [06:27] root@dorset:/var/log# ps aux | grep lightdm [06:27] root 2684 0.0 0.1 40844 3540 ? Ssl 22:24 0:00 lightdm [06:27] root 2688 0.6 1.8 63240 38320 tty7 Ss+ 22:24 0:01 /usr/bin/X :0 -auth /var/run/lightdm/root/:0 -nolisten tcp vt7 -novtswitch [06:30] ah the lightdm segfault occurs when restarting the service [06:31] Dec 21 22:30:11 dorset kernel: [ 1274.512212] type=1400 audit(1324535411.385:27): apparmor="DENIED" operation="open" parent=1 profile="/usr/lib/telepathy/mission-control-5" name="/home/bryce/.cache/dconf/user" pid=3445 comm="mission-control" requested_mask="rwc" denied_mask="rwc" fsuid=1000 ouid=1000 [06:31] wonder if that's relevant [06:35] That's empathy being denied access to dconf; I should report that bug. [06:36] It does, however, indicate that you appear to have made it all the way through to your session. [06:36] yeah [06:36] this is what is printed to syslog when I do a sudo service lightdm restart: [06:36] http://pastebin.ubuntu.com/778322/ [06:38] intriguing... ppa-purge of the ppa results in working X session. [06:38] let's try again [06:41] hmm, apport caught something... crash in "fts.py" [06:41] part of zeitgeist [06:41] guess coincidental [06:44] lp #907661 [06:44] Launchpad bug 907661 in zeitgeist (Ubuntu) "fts.py crashed with DatabaseLockError in __init__(): Unable to get write lock on /home/bryce/.local/share/zeitgeist/bb.fts.index: already locked (affects: 1) (heat: 8)" [Undecided,New] https://launchpad.net/bugs/907661 [06:46] RAOF, bryceh, fyi, I plan on pushing mt versions of xorg-server and evdev to the x-staging ppa tomorrow [06:47] cnd: Awesomesauce. [06:47] I was hoping keithp would pull Peter's mt branch today, but I can cherry-pick it as-is if he doesn't get to it [06:48] maybe I'll look into forward porting the qt multitouch patch after that :) [06:48] anyways, bed time for me [06:48] cnd, sounds good [06:49] I'm running the ppa on a devoted test box, I can update and test whatever whenever [06:50] hmm this is weird, after ppa-purging and re-adding, it doesn't want to install the new X bits [06:50] Oh! Are you trying to run Unity? It still won't start without mt/gestures. [06:52] RAOF, heh ok that'd do it [06:55] RAOF, that hasn't been fixed yet? [06:55] RAOF, ok, I assumed it just required the new input stack. I take it cnd's upload tomorrow is required then? [06:55] grrr... [06:56] bryceh, it's a unity bug [06:56] cnd, mm [06:56] there was a "fix" for it that was obviously wrong on further inspection [06:56] a week ago I assigned it back to the original fixer, hoping they would resolve it [06:56] ok, and I have auto-login enabled so it just logs me into unity automatically; presumably if I switch that off I could go into unity2d or classic [06:56] maybe they've been on vacation? [06:58] at least I've learned ppa-purge isn't reversible with apt-add-repository (as it probably should be) [06:59] bryceh, https://bugs.launchpad.net/ubuntu/+source/unity/+bug/860707 [06:59] Launchpad bug 860707 in unity (Ubuntu) (and 1 other project) "Unity crashes when started in an environment without utouch support (affects: 58) (dups: 2) (heat: 270)" [High,Triaged] [06:59] I might have to fix that myself [06:59] * cnd tries to stay away from unity development [07:00] they've probably all been repurposed into cloud developers or something [07:01] ok yep switching to gnome classic (no effects) boots in fine [07:30] cnd: btw, debian wants multitouch soon, so there's no need for our own upstream branches, could've used -experimental there [07:30] but it's no biggie [07:37] oh unity still busted? maybe i won't test the ppa afterall :/ [08:29] tjaalton, definitely busted [08:29] tjaalton, classic seems to work [08:32] bryceh: yeah, i'll hold off then.. though it's just my laptop to break, hmm [08:48] *nod* === dzan|away is now known as dzan [14:55] damn.... my touchpad is a MESS [15:25] tjaalton, were people getting upset about me pushing ubuntu specific packages? [15:25] KiBi seems a bit miffed that I would do that [15:50] cnd: yeah we discussed it on debian-x the other day, and told that we didn't know what their plans were and forgot to ask. no harm done :) [15:51] I just chatted with him on #debian-x, I think he's ok once I mentioned this is a one-time, time-sensitive deal [15:51] I'm trying to get things into the ppa before I go on vacation [15:51] yeah he's cool [15:51] Sarvatt told him that we (you) were on a deadline [16:21] I just pushed a fix for the utouch-unity crash: https://code.launchpad.net/~chasedouglas/unity/utouch-crash/+merge/86727 [16:21] hopefully it is reviewed and uploaded soon [16:22] but I don't know if it will be before the holiday break === yofel_ is now known as yofel [17:10] rebooting seems to have fixed my touchpad probs.... for now [17:30] I decided to create a unity package with a fix for the utouch issue [17:30] after I see that it builds, I'll push it to the staging ppa [18:11] cnd: oh great, thanks [18:22] it's now uploaded [18:22] I haven't had a chance to test it yet [18:22] I'm just now updating behemoth to precise [18:22] but it's a simple fix, so it should work [19:43] cnd: hmm don't see unity on the ppa yet? [19:43] tjaalton, it probably hasn't built yet [19:44] cnd: ah, right.. [19:45] start in 23min [19:45] i can wait :) [20:02] cnd, hi, you probably want to put fglrx in there too [20:03] ricotz, RAOF (and maybe others) are handling anything non-input related [20:03] I don't really know about anything outside of input :) [20:03] alright ;) [20:08] ricotz: the version in precise should already handle the new abi [20:09] unless it needs a rebuild to fix the deps [20:10] tjaalton, yes, just a rebuild is needed [20:10] but an update would be nice too since there is a newer upstream [20:11] same goes for nvidia [20:12] tseliot will handle updates [20:12] in january :) [20:12] ie. before those are pushed to precise [20:13] alright === dzan is now known as dzan|away [21:05] tjaalton, unity is built [21:06] have you had a chance to test? [21:07] cnd: upgrading.. [21:07] tjaalton, there *may* be an issue with evdev in the ppa right now [21:08] but I think it should work if you aren't using a multitouch device [21:08] it needs to be rebuilt after xorg-server is built in the ppa [21:08] seems to work fine otherwise, but compiz takes all the cpu [21:10] I get: [21:10] Xorg: symbol lookup error: /usr/lib/xorg/modules/input/evdev_drv.so: undefined symbol: InitTouchClassDeviceStruct [21:10] which is what I expect [21:11] huh, works here [21:13] unity-2d loads, apart from the dock [21:14] ok [21:18] ah, unity-2d-launcher taking 100% cpu :) [21:18] tjaalton: Yeah, unity-2d-launcher *also* hates missing multitouch/gestures it seems :) [21:19] at least the upgrade itself was rather uneventful [21:20] ricotz: There's already a build on nvidia-current in there; there will be fglrx & the rest of the drivers in there before they land in precise. [21:23] RAOF, hi, yeah i figured that, just didnt know when it will be moved [21:26] Not before Christmas :) [21:34] yeah no more breaks for this year! :) [22:32] once evdev rebuilds in the ppa it should be all ready for testing [22:32] and it should have multitouch for touchscreens [22:32] \o/ [22:33] cnd, sweet, ping us once it's ready [22:53] tjaalton, unity-2d probably crashes because qt's multitouch code is now incompatible with the upstream XI implementation [22:53] I'll probably upload a new qt with the patch commented out [22:53] i'll just wait for evdev though :) [22:53] to build [22:59] though maybe I'll just check it in the morning [23:05] oh, building.. [23:27] cnd: ok so the same still happens with these, will check in the morning if there's a new qt available :)