[07:57] Anyway to test current work of "pixel perfect" scroll in Chromium? I'm on MacBook Air. Interresting work! :) [08:03] speakman, we do not have a ready to use ppa yet, but stay tuned :) === MacSlow is now known as MacSlow|lunch [12:05] tvoss: I'm staying very tuned (I just tuned you guys in) - just let me know if there's anything I can do to help or test! :) [12:06] speakman, awesome, thanks for your help :) I'll let you know once we have packages available [12:07] The upstream merge request is here: http://chromiumcodereview.appspot.com/9773024 [12:07] If you feel adventurous, you can try compiling it. [12:08] Thanks! Satoris; I'll give it a shot! [12:09] Just so you know: building Chromium is not a trivial task... [12:09] I figured :) [12:10] The patch works only on Precise. Other releases are too old. [12:11] k, so apt-get source chromium && patch ... && build? [12:11] No, you need Chromium trunk. [12:12] ok [12:12] Chromium wiki has build instructions which are quite good. [12:12] Yep, I though you meant the chromium version in Precise. I think I misinterpreted you. Sorry. :) [12:13] Also note that Chromium can not be built on 32-bit machines. It is too big. [12:13] The platform you build on needs to be 64 bit Precise and the source the newest head. [12:17] k, thanks! [12:27] gclient sync'ing now. Wonder if my 40GB SSD partition on MBA will be enough? [12:28] Probably yes. [12:30] ok, lets see. In this speed (~200kB/s) it will take a couple of weeks... === MacSlow|lunch is now known as MacSlow [12:32] btw, I'm happy this is happening on chromium and not firefox since the later has very slow rendering in Linux. Do you know why Ubuntu keep delivering with Firefox preinstalled and not Chromium? [13:23] btw - I don't have any particual gestures in Ubuntu Precise at all. Are there supposed to be any per default, or do I have to install any packages manually to have the guestures mentioned here https://wiki.ubuntu.com/Multitouch ? [13:24] Running 12.04 newly updated on MacBook Air 11" [13:24] a 2012 MBA that is [13:34] 12.04 has 3- and 4-touch gestures in Unity by default, have you tried those? [13:39] bregma: do you mean 3- and 4-finger guestures? I've tried those listed in the link, but none seem to work. [13:39] well, that doesn't sound right [13:40] I know Satoris has everything working on a Macbook, but evidently he's gone for the day [13:40] cnd might have some insight, but I'm not sure if he's around today [13:41] Ok, any idea which tool should be installed and running for gestures to work? I can always try to debug it myself if I could only get a hint where to start. [13:49] no tool, they should just work [13:49] ok, built-in into Unity? [13:50] yes [13:50] k, thanks [13:50] if you install the utouch-geis-tools package, you should be able to run the geisview tool to see if other gestures are recognized by the system [13:51] you need to give it the ID of a window you'll do gesture over, otehrwise it will assume the root window and will fail because Unity has already grabbed the gestures in the root window [13:54] Classes added: Flick, Drag, Pinch, Rotate, Tap, Touch [13:54] didn't complain about root window being busy, weird? [13:55] might be related: https://bugs.launchpad.net/ubuntu/+source/unity/+bug/889144 [13:55] Ubuntu bug 889144 in unity (Ubuntu) "Three-finger window drag no longer works" [Medium,Triaged] [13:56] speakman, so when you try a four-tap the dash doesn't show? [13:56] cnd: nope [13:56] speakman, and you're in unity 3d? [13:57] Are there any packages that might conflict? If I remember correctly, I think I tried to installed a few packages while trying to get the multitouch working. [13:57] cnd: unity 3d, yes [13:57] there shouldn't be any conflicts [13:57] ok [13:57] speakman, can you pastebin your /var/log/Xorg.0.log file? [13:58] More reports with the same issue:; https://bugs.launchpad.net/ubuntu/+source/unity/+bug/931171 [13:58] Ubuntu bug 931171 in unity (Ubuntu) "no multitouch gestures?" [Low,Confirmed] [13:58] cnd: sure, w8 [14:03] It's huge: http://paste.ubuntu.com/945725/ [14:03] cnd: ^ [14:04] ("RandR disabled" hm?) [14:04] ok, that looks fine [14:04] oops, five lines above: "RandR 1.2 enabled, ignore the following RandR disabled message." :) [14:04] cnd: look at the bottom - still looks fine? [14:05] oh right [14:05] Row 555 and below [14:05] looks like there have been problems :) [14:05] :D [14:06] I've not seen that bug before [14:06] it's saying that it reached touchscreen pointer emulation code, but your device isn't a touchscreen [14:06] so that's a bug [14:06] quick summary of my day: restoring my harddrive [14:07] Ok. According to the bug reports on Launchpad I'm not alone with gestures not working in Precise. Not sure they have the same bug though. [14:07] speakman, if you can reproduce it, install xserver-xorg-core-dbg and then log out and back in [14:07] tvoss: crash? [14:07] speakman, there are many reasons why gestures may not work, some of them expected :( [14:07] speakman, yeah ... forgot my power cable yesterday and wanted to get up too quickly [14:07] cnd: It's absolutely reproducable; I've never got gestuers working :) Installing... [14:08] tvoss: :( [14:09] Hm. Still downloading Chromium source. Wonder if it will restart from scratch if it gets interrupted? But I consider this a bigger issue so canceling the source clone. [14:15] I'm going to verify that the utouch-qml sru is correct, then help speakman with his troubles, then work on architecture documentation and a crazy idea for legacy trackpad gesture handling [14:15] dandrader, bregma: standups! [14:15] preparing an SRU for utouch-geis [14:16] and working on a new port of the geisv2 api in unity using the accept/reject stuff [14:16] cnd: is logging out enough or should I reboot? [14:16] speakman, logging out is enough [14:16] it restarts the X server [14:16] I'm working with dandrader on the SRU, playing with test cases, and looking at #987539 [14:17] bug 987539 [14:17] Launchpad bug 987539 in utouch-geis (Ubuntu Precise) "Grail backend does not populate device X and Y axis attributes" [Medium,Triaged] https://launchpad.net/bugs/987539 [14:18] and redoing my Debian package proposals, maybe I can get a sponsor at UDS [14:21] bregma, since I'm gone tomorrow, can you handle the weekly report? [14:22] cnd: I've restarted X, was tail -f'ing my Xorg.0.log and trying to perform some gestures, and then Xorg suddenly died and left me at login screen. [14:23] speakman, hrm... [14:23] that's not good [14:23] speakman, pastebin your /var/log/Xorg.0.log.old [14:23] Got some similiar errors in Xorg.0.log again, but it's the same backtrace as before -dbg package was installed. How do I make sure I'm running the dbg version? [14:23] speakman, btw, you can use pastebinit to quickly pastebin a file [14:24] speakman, the dbg package merely installs the symbols [14:24] what version of the xserver are you running? [14:24] what package version [14:24] Already using pasteinbit - don't think it would work else way :D [14:25] heh [14:25] cnd: xserver-xorg-core 2:1.11.4-0ubuntu10.1 [14:26] ok [14:26] there's a possibility it is a regression introduced in that package [14:26] Must leave office atm, back in a few minutes. I'll post pastebin urls before leaving [14:27] cnd: this is .old: http://paste.ubuntu.com/945751/ [14:27] speakman, are you on x86_64? [14:27] cnd, no problem on the report, especially if people copy me on their reports (hint hint everyone) [14:27] oh right, it's in the log [14:27] cnd: this is current: http://paste.ubuntu.com/945753/ [14:28] bregma, I'll forward them if not [14:28] I'm on x86_64 [14:28] I really have to leave, will be back in a ~15minutes or something. [14:29] k [14:29] I'll be analyzing your log to see what might be going wrong [14:52] speakman, for some reason your touches appear to be marked as pointer emulating touches [14:52] which should only occur for touchscreens [14:53] speakman, please run "xinput list bcm5974" and pastebin the output [14:56] hmm.. no, that wouldn't fully explain it [14:57] I wonder if you're hitting some sort of memory corruption bug [14:57] back... [15:00] here's xinput list in case of interrest: http://paste.ubuntu.com/945792/ [15:01] that looks good [15:01] speakman, when you try to perform a gesture, what applications are open? [15:01] what windows are you performing the gesture over? [15:08] i just got terminal window full screen right now for example [15:09] speakman, gnome-terminal? [15:09] yes [15:09] anything in particular you want me to try? [15:10] speakman, I might give you a package to install to give us more info [15:10] sure [15:10] will you be around for another 20 mins? [15:10] yep [15:11] ok === dandrader is now known as dandrader|lunch [15:19] speakman, please install this: http://people.canonical.com/~cndougla/utouch/xserver-xorg-core_1.11.4-0ubuntu10.1_amd64.deb [15:19] reproduce the issue [15:20] then pastebin your Xorg.0.log [15:23] on my way... [15:29] http://paste.ubuntu.com/945839/ [15:30] Can you see that Xorg.0.log is created by this new instance? Or will it always start over when Xorg is restarted? [15:33] it will be overwritten [15:34] speakman, but that particular paste is identical to your previous one [15:34] it may have switched DISPLAY numbers [15:35] you can do: ps aux | grep X [15:35] k [15:35] and look for the argument that is like ":" [15:35] that's the display number [15:35] then you want to paste the /var/log/Xorg..log [15:35] yep, daniel@mba:~$ env | grep DISPLAY [15:35] DISPLAY=:1.0 [15:36] yeah [15:36] so pastebin Xorg.1.log [15:36] http://paste.ubuntu.com/945846/ [15:36] should I try some gestures too? [15:37] yes [15:37] cause there's no errors in that log :) [15:44] I can't make any errors in the log no matter what [15:44] but gestures still doesn't work :) [15:45] hmmm [15:45] [ 11462.492] Build Date: 23 April 2012 10:28:29AM [15:45] didn't you just build a new xorg? [15:45] that might be a red herring [15:45] I had a build lying around [15:45] I added a patch [15:46] so it didn't rebuild the whole thing [15:46] speakman, I built with optimizations disabled [15:46] let me rebuild with optimizations [15:46] maybe that's affecting things [15:46] ok [15:47] otherwise, the change I made was merely to print out a stack trace when a bug occurred [15:48] it should have had identical functionality to the version in the archive [15:49] Ok, maybe I just can't trigger the bug anymore? [15:49] Maybe I havn't had the bug for a time, but kept pushing the wrong log to pastebin? [15:50] (still no errors...) [15:52] well, the log with the errors shows the -0ubuntu10.1 version [15:53] ok? [15:53] so it is likely a real bug [15:53] yes, Xorg.0.log is only ~1.5h old [15:53] whats your timezone btw? [15:54] PDT [15:54] speakman, have you dragged a window around using the titlebar? [15:54] I may have an idea [15:56] no, i havn't. Using three fingers or just single-click and hold while dragging? [15:57] hmm [15:57] speakman, well, download and install the xserver again [15:57] it's been rebuilt with optimizations [16:00] ok [16:03] speakman, I'll be back in about 20-30 mins [16:05] ok, I'm running the new xserver now and still can't trigger any bugs :( [16:29] speakman, you can try reinstalling the xserver from the archive to see if you can reproduce [16:30] we could also move on and try to figure out why gestures aren't working [16:30] for that, we need to start unity with a special debugging environment variable [16:31] in a terminal, do: [16:31] GRAIL_DEBUG=-1 unity 2>&1 | tee /tmp/unity.log [16:31] usually that will kill and restart unity [16:32] sometimes unity may die again [16:32] so you have to re-run the command [16:34] Now running with the grail debug stuff [16:36] What do you want me to do? :) [16:36] speakman, try performing gestures [16:36] like a four touch tap [16:36] you should see tons of output spewing down your terminal [16:37] lol, no nothing. :) [16:37] hmmm [16:37] that's odd [16:37] any errors in your X log? :) [16:37] except for "e"'s. For each four finger tap it will put an 'e' in the terminal window :) [16:38] hmm... that's very strange [16:38] Nope, silent... [16:38] speakman, is ginn running? [16:38] ps aux | grep ginn [16:38] yes [16:38] interesting [16:38] yes again :) [16:38] do you use ginn? [16:39] how do I know that? [16:39] you don't then :) [16:39] sudo apt-get remove ginn [16:39] then log out and back in [16:39] see if gestures work [16:39] on my way :) [16:40] Ok, now back in unity [16:40] (but not with the GRAIL_DEBUG set) [16:41] now things are happening [16:41] good :) [16:42] Now I can move a window using three fingers [16:42] but its VERY sensitive! [16:42] we need to stop punishing people who install ppa:utouch-team/daily with a ginn that autoruns at startup and may grab ahead of Unity :( [16:42] does not relate to the speed of the cursor! it's WAY faster [16:42] yeah, that's sort of a known issue [16:42] ok [16:43] I don't have a good answer for that, tbh [16:43] I've thought about it [16:43] the problem is that we are trying to emulate the speed of trackpad to cursor movements [16:43] but that's embedded deep in the X server [16:43] using algorithms that can be replicated outside the X server [16:44] our gesture code for moving windows is currently too simplistic [16:44] and will change speed depending on the resolution and size of your trackpad [16:44] speakman, I suggest filing a bug against unity and subscribing the utouch-bugs team [16:45] filing a bug for what? :) [16:46] for the window moving gesture being too fast [16:46] if it's an issue for you [16:46] I've kind of been hoping that while it's not good, it's not a huge deal for most people... [16:50] ah k, yes it's a big issue :) [16:51] So, is there a gesture to move back and forward in the browser? [16:54] speakman, no, not yet [16:58] speakman, so about your X issue, I still don't know what could be wrong [16:58] if you can reproduce it again, let me know [17:57] merge proposal emails seem really delayed today === dandrader|lunch is now known as dandrader [18:15] bregma, why is export-dynamic needed? [18:16] otherwise the mock X11 symbols are not used to satisfy the symbols in subsequently loaded libraries like libutouch-frame (and... bang) [18:16] why not? [18:17] isn't that how normal libraries work? [18:17] no [18:17] or are you saying that the utouch-frame libs are attempting to load the real x libs? [18:17] yes, that [18:17] ok [18:18] normally symbol resolution is a DAG [18:36] we seem to have broken UDD :( [18:51] bregma, hmm? [18:54] the automated UDD uploader has been balking on some utouch packages for a while now... it's not just our stuff, evidently the UDD uploader is pretty buggy [18:55] so, the code in lp for Ubuntu does not match the packages in Ubuntu [18:56] not _our_ problem [18:57] http://package-import.ubuntu.com/status/ [19:02] yeah [19:02] I've heard there are some issues there [19:04] we don't usually use UDD so we didn't notice [19:18] the desktop team has also given up on UDD [19:27] UDD is useful if you're packaging an external source but is more limited if you're packaging something with the upstream maintained in launchpad [19:30] .. holy crow, worms ate my day, I got nothing done I set out to do this morning [19:39] heh [19:39] I feel like that every day === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === therion__ is now known as therion_