[09:51] right, so g-s-d doesn't distinguish tablets with touch support from touchsreens when it's resetting the touch device to relative mode [11:22] oh gee, one guy wanted nvidia-173 so bad he downgraded X to 11.10 [11:22] and posted the instructions on the bug report [11:22] .. [11:50] so, this might be slightly OT here but: did anyone try Precise with an AMD APU? i'm considering in upgrading my workstation but i fear fglrx drivers and stuff [11:50] btw, hello ;) [14:35] bryceh: pushed g-s-d to my ppa, so once it's built please test :) [14:35] tested that my intuos5 is still in relative mode after update.. [14:35] dinner -> [19:47] tjaalton, still there? [20:02] seb128: on an off [20:03] *and [20:04] tjaalton, g-s-d, do you want your fix uploaded? [20:04] tjaalton, I'm about to do an upload [20:06] seb128: well, does it fix things? :) [20:06] tjaalton, dunno, I will upload http://git.gnome.org/browse/gnome-settings-daemon/commit/?id=d07807ca68e7d3f0b94fdce9bc2a88a71d4324ef [20:07] tjaalton, which is another touchpad devices issue [20:07] tjaalton, so I was just wondering if you want extra testing through the archive, we can always revert next week if it's not good [20:08] seb128: ah yeah in that case why not [20:08] tjaalton, ok, can you push it to the vcs? [20:08] tjaalton, I will add my patch on top [20:08] i did test that it didn't break touch capable tablet support [20:09] but the touchscreen part is still a mystery, but the logic should be right now [20:10] tjaalton, ok, let's get extra archive testing during the w.e [20:11] upstream probably would like to add a new class for those, so the condition would be more like (type != SCREEN) or such [20:11] well until they do your patch might be better than what we have ;-) [20:12] yeah, hope so :) [20:25] tjaalton, can you push your changes then? ;-) [20:32] seb128: to the bzr branch? [20:33] tjaalton, yes [20:33] tjaalton, lp:~ubuntu-desktop/gnome-settings-daemon/ubuntu [20:34] ok let me get off the phone irc then :) [20:41] hmm, can't login to my laptop [20:43] ahahhaha [20:48] last message on lightdm.log is "user foo authorized" [20:51] seb128: ^ ideas? I upgraded it this evening, noticed there was a new lightdm too [20:51] just in time for the long weekend :) [20:51] tjaalton, no, is there anything special on your pam or login? [20:52] tjaalton, is there any session process started? [20:53] auth.log looks ok [20:53] no user processes started [20:53] i mean, owned by the user [20:54] greeter.log says auth complete [20:54] tjaalton, I've tested your g-s-d [20:54] bryceh: works? [20:54] tjaalton, here's what I did [20:54] *bites fingernails* [20:54] I re-set the gsettings active property back(?) to true [20:54] guessing no since its a story :) [20:54] then I added your ppa, updated, upgraded [20:55] rebooted [20:55] Sarvatt: yeah :) [20:55] now, on the *login* screen, it does work great [20:55] but after logged in, same behavior [20:55] meh [20:55] but [20:55] what I'm wondering is that since I set it to true specifically, does that override your changes? [20:55] maybe I need to UNset it? [20:56] or should it matter? [20:56] bryceh: no it's just the setting for the plugin, use it or not [20:56] true is the default [20:56] gnome-settings-daemon: [20:56] Installed: 3.4.0-0ubuntu3.1 [20:56] so the plugin does it's own thing, and the relative/absolute for the touch functionality is not configurable by gsettings [20:56] *** 3.4.0-0ubuntu3.1 0 [20:56] 500 http://ppa.launchpad.net/tjaalton/ppa/ubuntu/ precise/main i386 Packages [20:56] yeah, so it's not complete, bah [20:57] either the logic is all wrong after me looking at it for too long :) or it can't be called from there. compiles fine though and doesn't break my setup [20:58] tjaalton, ok, so pre-release freeze is in 3 minutes, do you want to commit your stuff or should I upload my patch without it? [20:59] tjaalton, I can mess around with the patch locally if you'd like. get some more data on what's going on, etc... [20:59] seb128: withouth, bryce tested and it's flawed in some way [20:59] tjaalton, bryceh: we aware of http://git.gnome.org/browse/gnome-settings-daemon/commit/?id=d07807ca68e7d3f0b94fdce9bc2a88a71d4324ef as well [21:00] we->be [21:00] not sure if that could impact on what you are testing [21:00] that's the patch I want to update, but now I'm late, it's freeze time, I shouldn't have waited for tjaalton :p [21:01] what prerelease freeze? [21:01] seb128: sorry :) [21:01] tjaalton, the one starting at 21utc today [21:01] i.e now [21:02] which also suprised me, they were discussing it today on #ubuntu-release, from now one archive should be frozen again and uploads will be approved [21:02] i thought it was next week [21:02] skaet argued that previous cycle we didn't unfreeze after beta2 [21:02] ok then [21:02] release freeze is next week [21:02] bug fixes should still be ok meanwhile [21:03] they just need to be approved [21:03] right [21:03] not sure there will be lot of approvers during the long w.e though [21:03] guess I should get uploads rolling and not fight with silly sponsors who question my skills to package a source package for debian :) [21:03] and then merge those [21:04] ;-) [21:04] tjaalton, btw while you are there [21:04] tjaalton, https://bugs.launchpad.net/ubuntu/+source/gnome-settings-daemon/+bug/934445/comments/19 [21:04] Launchpad bug 934445 in gnome-settings-daemon (Ubuntu) "hits g_assert (device->priv->styli) when my Wacom Bamboo 2FG 4x5 is plugged in" [High,Triaged] [21:04] tjaalton, what do you think about that? [21:05] tjaalton, it's a workaround until g-s-d learns to not get unhappy on unknown devices but seems it could still be worth to ship with libwacom until then? [21:05] seb128: can't check the details, but there are some fixes in the driver git which might help there [21:08] meh, i'll switch to the desktop so I'll have a browser [21:17] seb128: checking the code [21:20] seb128: yeah, the device-id is the same, so adding the new stylus id should be ok [21:20] adding it to the same data file that is [21:21] tjaalton, can you do it in the next upload? [21:21] seb128: sure thing [21:21] tjaalton, thanks [21:24] hmm, I could upload it to debian now, and sync [21:27] bryceh: you said something about a disconnect between where you touch on your screen and the cursor location? I've noticed the same thing [21:27] I think it's the same with my intuos too [21:27] Sarvatt: so it would cover your n-trig woes as well [21:29] bdmurray, yep [21:29] bdmurray, doesn't seem to affect the stylus, just touch [21:29] bryceh: yeah same thing with the stylus [21:30] bdmurray, you see the same misbehavior with the stylus? odd, I don't [21:30] tjaalton, would it be worthwhile for me to tinker with the patch locally? [21:31] bryceh: I meant yeah the stylus works for me too [21:31] bdmurray, aha [21:31] bryceh: if you have the time then yes. you've seen what it tried to do, so it shouldn't be that hard to fix :) [21:32] alright, adding to the todo stack for today [21:32] tjaalton: WACOM_TYPE_TOUCH is only for actual touchpad capable wacoms [21:32] aka bamboo touch [21:32] bryceh: the gsd plugin has the logic to check if the device is a screen tablet, so set_absolute should be true for those [21:33] Sarvatt: well that has been proven wrong now? :) [21:33] ? [21:33] guess i should read scrollback, i've been looking at g-s-d [21:33] the gsd code touches every device with the touch property [21:33] yeah which serial wacoms dont have [21:33] in libwacom [21:34] unless i missed something [21:34] http://git.gnome.org/browse/gnome-settings-daemon/commit/plugins/wacom/gsd-wacom-manager.c?id=5dd89a267e7c539e681851017b18261dc2e092f0 [21:34] this is where it broke [21:34] not libwacom [21:34] tjaalton: oh they dont have Touch=false in the serial tablet definitions either [21:35] tjaalton: most everything else has Touch=false if its not usable as a touchpad, serial wacom definitions in libwacom dont [21:35] well I changed that bit to use TRUE, and it turned my intuos to absolute mode, so I figured the condition was wrong [21:36] ./isdv4-e3.tablet:# this is for the Wacom pen + touchscreen as found in the HP Touchsmart tm2 laptop. [21:36] bdmurray's laptop is covered by a different set of definitions than generic serial wacom tablets [21:36] let me finish this changelog entry first :) [21:36] so might explain why he sees different behavior [21:37] oh nevermind, i'm gonna stop talking without reading the scrollback fully [21:37] bdmurray: so do you have the touch set in relative ("touchpad") or absolute ("pen tablet") mode? [21:38] hell it'd be better to just revert that commit [21:38] Sarvatt: this is on an x201 tablet [21:38] Sarvatt: no, then all bamboos would be set in absolute mode.. [21:39] or would they [21:39] tjaalton: either way I set it the behavior is the same [21:39] i think there might be more laptops with touchscreens than bamboo tablets [21:39] if only I could login to my laptop I could test it :) [21:40] bdmurray: which is what? :) [21:40] and they can change it in the gui if they prefer it the other way :P [21:40] Sarvatt: alas, no they can't :) it's not exposed for the touch [21:40] only for the pad [21:42] because of http://git.gnome.org/browse/gnome-settings-daemon/commit/plugins/wacom/gsd-wacom-manager.c?id=389276ef566342c3c61547326c070a0d6914622a [21:45] bdmurray: I mean, is the touchscreen functionality in absolute mode on both the login screen _and_ in the session, or just the login screen [21:46] the latter is what everyone else is seeing [21:46] ok touch is a weird thing here, some people are referring to it for touchscreens where you want absolute, other people are referring to it for touchpads where you want relative.. you'd only want relative on tablets with a touchpad on them which is a smallllllllll minority of them [21:47] like the definitions for the touchsmart tm2 have it listed as a touch device [21:47] when its just the screen [21:47] well the only one complaining about it upstream is a bamboo user/dev [21:48] ie jason gerecke, the wacom dude :) [21:48] tjaalton: ah at the login screen it is 'absolute mode' if that means the cursor follows my finger [21:48] bdmurray: yess, and that's what it should always have [21:50] a lazy person would just reopen gnome bug 670655 and let them know that touchscreens are broken, but why not try to create a patch in the meantime [21:50] Gnome bug 670655 in plugins "Wacom 'touch' devices are initialized to absolute mode by default" [Normal,Resolved: fixed] http://bugzilla.gnome.org/show_bug.cgi?id=670655 [21:54] bryceh, bdmurray: if you have a moment, can you add Touch=false to the bottom of /usr/share/libwacom/serial-wacf004.tablet and try logging in again? [21:54] that would just disable the touch part? [21:55] sure [21:55] there is no touch part [21:55] huh? [21:55] its a friggin touch screen not a touchpad [21:55] yes, touch screen with finger + pen touch [21:55] the touch detection i see in g-s-d is only for wacom touchpads built into tablets to put them in relative mode.. [21:56] Sarvatt, alright done. what am I looking for? [21:56] just if its absolute after logging in [21:57] nope, same misbehavior as before [21:57] ok cool, thanks for checking it wasn't something that easy [21:58] that was added to the [Features] section fwiw [21:58] bryceh: wait, one more? Pad=false ? [21:58] okie [21:58] i'm just thinking maybe its assuming things if its not explicitly defined [21:58] doesn't that only change the configurator? [22:00] tjaalton: what a headache :) [22:00] nope, [22:00] hmm [22:01] btw in that file I see a DeviceMatch for serial:0000:0000 [22:01] are there pci ids to be checked? [22:01] nope its serial [22:01] gsd_wacom_device_is_screen_tablet() does seem to ask stuff from libwacom [22:01] so looks like both need an update [22:01] libwacom_is_builtin [22:02] /usr/share/libwacom/isdv4-e3.tablet is what blows my mind [22:02] Touch=true for touchscreen, its only wacom for the touchscreen.. [22:02] libwacom is a ton of bug reports waiting to happen [22:03] what do you mean "its only wacom for the touchscreen"? [22:03] tjaalton: oh wait a second here [22:03] surely isn't, there's Stylus etc there [22:03] Builtin means it doesn't parse the external files doesnt it.. [22:04] ah nope [22:04] so maybe you can make g-s-d not apply relative if BuiltIn=true [22:04] since those are all touchscreens [22:05] tjaalton: BuiltIn=false + Touch=true is ok for relative, just people using the pad surface with touch as a touchpad [22:06] yes, that's bamboo [22:06] should be able to use the BuiltIn definition to separate out touchscreens where you'd want absolute from touchpads where you'd want relative [22:09] ah yea libwacom_is_builtin like you found earlier [22:10] right [22:10] and it should work.. [22:11] I mean, libwacom should be ok [22:11] yeah [22:11] its already used elsewhere to determine what graphics are shown in the ui [22:11] yep [22:27] nifty, fedoras actually generating wacom udev rules from libwacom [22:27] thats a great idea [22:27] yep [22:28] bryce is going to try http://kernel.ubuntu.com/~sarvatt/wacom-touch-fix.patch out [22:32] alright :) [22:32] also got some debuggery folded in with that in case it doesn't work [22:37] btw, I see BuiltIn=true in the .tablet file if that matters... [22:38] yes, as should be [22:38] bingo sarvatt [22:39] http://paste.ubuntu.com/916692/ [22:39] (gnome-settings-daemon:7808): wacom-plugin-WARNING **: gsd_wacom_device_is_screen_tablet = 0 [22:39] (gnome-settings-daemon:7808): wacom-plugin-WARNING **: libwacom_is_builtin = 1 [22:40] wth? [22:40] my thoughts exactly... [22:42] oh [22:42] gsd_wacom_device_is_screen_tablet [22:42] that might just check if its assigned to a screen tjaalton [22:43] since you can bind the tablet to specific screens [22:44] hum, ok [22:44] ftr this is the patch I used - http://paste.ubuntu.com/916697/ [22:48] Sarvatt: or maybe device->priv->is_screen_tablet isn't initialized yet? [22:53] oh well, I don't care :) bryceh thanks for checking that out! [22:55] btw, downgrading to lightdm 1.1.9 didn't help my laptop [22:56] still can't login to the X session [22:56] the greeter seems stuck [22:56] doesn't bail out [22:59] tjaalton, sure thing [22:59] so, do we have a pushable patch? [22:59] bryceh: uploaded an update to the ppa [22:59] great [22:59] likely so, but I'm not sure if seb pushed something already [23:00] so maybe it's best to let people test the ppa version and upload the fix to the distro after easter [23:00] * bryceh nods [23:00] bdmurray, ^^ [23:02] what ppa and package? [23:02] ppa:tjaalton/ppa [23:02] gnome-settings-daemon [23:02] 3.4.0-0ubuntu3.2 [23:03] takes some time to build [23:31] * bryceh uploads xdiagnose 2.4 [23:34] I added a script xpci [23:34] humber:~/src/xdiagnose/xdiagnose$ xpci [23:34] RV770 (1002:9442) xserver-xorg-video-ati [23:34] humber:~/src/xdiagnose/xdiagnose$ xpci 10de:0604 [23:34] G92 (10de:0604) GeForce 9800 GX2 xserver-xorg-video-nouveau,nvidia-graphics-drivers [23:35] humber:~/src/xdiagnose/xdiagnose$ grep 'VGA ' /tmp/lspci_vvnn.txt | xargs xpci [23:35] gm45 (8086:2a42) cantiga xserver-xorg-video-intel [23:38] * skaet not feeling that comfortable about libwacom though.... [23:43] bah :) [23:44] bryceh: the libwacom change is so straight forward it isnt funny, i'm sure it'll pass review :) [23:44] me mentioning the "merge" probably scared her off [23:44] then again there is this nasty change: +DM-Upload-Allowed: yes [23:45] doesn't affect ubuntu at all [23:45] yeah was a joke [23:45] hah, ok