| tjaalton | bryceh: the fix for bug 778758 is already in oneiric | 08:52 |
|---|---|---|
| ubot4 | Launchpad bug 778758 in xorg (Ubuntu Oneiric) (and 2 other projects) "The apport hook discourages new bug reports (affects: 3) (dups: 1) (heat: 88)" [Undecided,Fix released] https://launchpad.net/bugs/778758 | 08:52 |
| tjaalton | or did you mean natty | 08:52 |
| bryceh | tjaalton, yes, natty | 08:54 |
| tjaalton | ah, yes it's fine | 08:54 |
| tjaalton | hrm, so reverting to the broken crtc confinement patch made the build pass again | 09:01 |
| tjaalton | ..and it really doesn't make sense comparing build logs generated with -j2 | 09:08 |
| tjaalton | something wrong with the xserver rules file, need to run debuild clean twice | 09:21 |
| tjaalton | or oneiric, the error: | 09:21 |
| tjaalton | No patches applied | 09:21 |
| tjaalton | rm -rf .pc stampdir/patch | 09:21 |
| tjaalton | debuild: fatal error at line 1322: | 09:21 |
| tjaalton | couldn't exec fakeroot debian/rules: | 09:21 |
| tjaalton | at the end of the first run | 09:22 |
| tjaalton | RAOF: think I found the bug in the patch | 09:55 |
| RAOF | tjaalton: Cool. My build failed with a missing list.c, but I hadn't actually looked back in at it to make sense of it. | 10:05 |
| tjaalton | yeah | 10:05 |
| bryceh | heya RAOF | 10:05 |
| tjaalton | RAOF: it's just a test, I've disabled building it and checking if it's enough to "fix" it | 10:06 |
| tjaalton | nope :) | 10:07 |
| tjaalton | fixes.c missing as well | 10:07 |
| tjaalton | duh | 10:07 |
| RAOF | bryceh: Good $TIME_OF_DAY :) | 10:08 |
| tjaalton | hmm, maybe just pull what got upstream | 10:11 |
| RAOF | I started doing that, but I'm not sure that all of it actually made it upstream. | 10:12 |
| tjaalton | part of it was merged in february | 10:14 |
| RAOF | Yeah, but that was some setup work. | 10:15 |
| RAOF | It didn't include (a) actually calling ConfineCursorHarder, or (b) any ConfineCursorHarder implementations ;) | 10:15 |
| tjaalton | details.. | 10:16 |
| RAOF | I'm going to backport the Xi 2.1 raw-events-delivery patches so that DX can implement proximity animations sanely… | 10:16 |
| RAOF | Hm. Have they not made it off the list yet? | 10:18 |
| tjaalton | seems like it | 10:21 |
| tjaalton | at least no reply from keithp about merging them | 10:22 |
| tjaalton | I'll probably split the pointer barrier patch in two, the v3 patch in its own file and the rest in the other | 10:22 |
| tjaalton | it needs some changes, to not build missing tests | 10:23 |
| tjaalton | hmm or just update the current one, meh | 10:27 |
| Amaranth | RAOF: By sanely do you mean without polling for the mouse position? | 10:46 |
| RAOF | Amaranth: Indeed. | 10:46 |
| RAOF | tjaalton: Do you want to finish fixing the crtc patch, or do you want to hand over to me? | 10:47 |
| tjaalton | RAOF: I'm done, test-building next | 10:48 |
| tjaalton | I just diff'ed the changes to cursor.c in v3 and ported the changes over | 10:49 |
| tjaalton | and dropped the change to test/Makefile.am | 10:49 |
| tjaalton | built | 11:07 |
| tjaalton | ok, maybe should test this sucker then | 11:22 |
| RAOF | Woot! | 11:23 |
| tjaalton | ..right after these updates :) | 11:24 |
| RAOF | Messages from our sponsor! | 11:27 |
| tjaalton | bah, cking took the stage :) | 11:28 |
| tjaalton | the projector is still reserved, need to find another one | 12:20 |
| tjaalton | works! | 12:38 |
| tjaalton | fixing bug 554984 | 14:43 |
| ubot4 | Launchpad bug 554984 in xserver-xorg-input-evdev (Ubuntu) "[lucid] enable trackpoint scroll emulation by default (affects: 17) (dups: 6) (heat: 68)" [Wishlist,In progress] https://launchpad.net/bugs/554984 | 14:43 |
| tjaalton | and enabling middlemouse emulation for those, to allow dragging with the emulation | 14:44 |
| tjaalton | i guess it's a compromise I'm willing to live with :) | 14:44 |
| lilstevie | tjaalton: you good with finding what is causing bugs in evdev? :p | 14:54 |
| tjaalton | lilstevie: there are none, silly you.. | 14:54 |
| lilstevie | tjaalton: hah | 14:55 |
| lilstevie | tjaalton: I am having a weird issue with a touchscreen and either evdev or xorg tripping on it | 14:56 |
| lilstevie | touch frame is correctly implemented :) | 14:56 |
| lilstevie | but I need to bash 2 fingers to some random tune to get it to detect the click event | 14:57 |
| tjaalton | hmm not sure, those are usually kernel bugs if evdev is used | 14:58 |
| lilstevie | kernel is not the bug here :) | 14:58 |
| lilstevie | kernel is about 3 months worth of hard work to make perfect :) | 14:58 |
| lilstevie | and it is fine to maverick | 14:58 |
| tjaalton | maybe something to do with the multitouch patches | 14:59 |
| lilstevie | hmm | 15:00 |
| lilstevie | i see | 15:00 |
| tjaalton | just guessing | 15:00 |
| lilstevie | it is weird cause the mouse cursor doesn't follow my finger but x is tracking the location the pointer should be at | 15:00 |
| lilstevie | and the weird tappy pattern is just a strange quirk, the touch frames are identical for each tap | 15:01 |
| tjaalton | hmm, cnd ^ ? | 15:01 |
| lilstevie | I have been logging the kernel output to make sure it wasnt that | 15:01 |
| lilstevie | heh cnd said it was time to start looking at evdev | 15:01 |
| cnd | lilstevie, it could be something where to get a tap you need to physically press down and up within a certain amount of time, but you must also move one pixel for it to register | 15:03 |
| cnd | maybe that's the bug? | 15:03 |
| lilstevie | cnd: hmm that would be difficult to confirm | 15:03 |
| cnd | yeah :) | 15:03 |
| lilstevie | cnd: it is also hard to reliably trigger a full click | 15:05 |
| lilstevie | cnd: but that at least 1px movement is something that could certainly be an issue | 15:05 |
| lilstevie | in none of the successful cases has there not been +px | 15:06 |
| cnd | you could use evemu-record to record good and bad clicks to see what is different | 15:06 |
| lilstevie | oh dear god :p | 15:06 |
| cnd | :) | 15:06 |
| lilstevie | going to do that now | 15:08 |
| lilstevie | cnd: I forgot the syntax for evemu, what did I need to do again | 15:15 |
| cnd | lilstevie, https://wiki.ubuntu.com/Multitouch/Testing/uTouchEvEmu#Debugging | 15:15 |
| lilstevie | ty | 15:15 |
| lilstevie | ok got then | 15:17 |
| lilstevie | them :D | 15:17 |
| lilstevie | cnd: http://lilstevie.geek.nz/touch/device.prop and http://lilstevie.geek.nz/touch/device.record | 15:26 |
| cnd | lilstevie, I won't be able to look at them for a while :( | 15:34 |
| cnd | you can read more legible output if you run evtest on the device node and then play the recording back | 15:35 |
| cnd | comparing the good and bad taps may give you an idea | 15:35 |
| lilstevie | hm ok | 15:36 |
| lilstevie | cnd: ok I am seeing a theme here | 15:41 |
| lilstevie | cnd: only one of X/Y are reported on touches that work | 15:43 |
| cnd | on touches that work? | 15:43 |
| cnd | I'd expect touches not to work if X and Y aren't both provided | 15:43 |
| lilstevie | the touches that register, as in those that follow the click action report only 2 of the 3 params | 15:45 |
| lilstevie | either X+Y, X+BTN_TOUCH, or Y+BTN_TOUCH | 15:45 |
| lilstevie | cnd: http://pastie.org/2134438 for instance is a touch that worked | 15:47 |
| cnd | lilstevie, that's weird | 15:51 |
| cnd | it shouldn't handle a touch unless both X and Y are provided | 15:51 |
| cnd | oh wait | 15:51 |
| cnd | evdev takes care of that | 15:51 |
| lilstevie | hm | 15:52 |
| cnd | so the problem is when BTN_TOUCH goes down but X and Y are not provided? | 15:52 |
| cnd | either X and Y are not provided, that is | 15:52 |
| cnd | shoot, that's not right | 15:52 |
| cnd | when both X and Y are not provided | 15:52 |
| cnd | if (!X && !Y) | 15:52 |
| cnd | :) | 15:52 |
| lilstevie | no the problem occurs when both X and Y are provided | 15:53 |
| lilstevie | actually nvm I phail that is the second fingers dummy data | 16:01 |
| lilstevie | cnd: the touch that works is showing btn_touch, X, and Y on finger being lifted as well | 16:01 |
| lilstevie | but the kernel driver only reports if movement has occured | 16:02 |
| cnd | lilstevie, so the kernel doesn't report BTN_TOUCH unless there's motion? | 16:03 |
| lilstevie | no btn_touch is reported when it changes | 16:03 |
| lilstevie | the kernel driver doesn't report ABS_X if ABS_X has not changed since last report | 16:03 |
| lilstevie | so when I lift my finger BTN_REPORT changes to 0, but X/Y haven't changed so no report | 16:04 |
| cnd | so it sounds like a kernel driver bug | 16:04 |
| cnd | do you understand how to fix it up at this point? | 16:05 |
| lilstevie | I think I have an idea, | 16:05 |
| cnd | cool | 16:05 |
| lilstevie | but it means using those cheap tricks I was hoping to not use, part of my cleaning up the driver for release is probably the one thing that will fux | 16:06 |
| lilstevie | fix this | 16:06 |
| tjaalton | lilstevie: is bug 637874 related? | 16:51 |
| ubot4 | Launchpad bug 637874 in xserver-xorg-input-evdev (Ubuntu) "Ubuntu Patches in evdev cause regressions on touchscreen (affects: 8) (dups: 1) (heat: 46)" [Undecided,Confirmed] https://launchpad.net/bugs/637874 | 16:51 |
| lilstevie | tjaalton: nah it is different | 16:52 |
| tjaalton | ok | 16:52 |
| lilstevie | tjaalton: this one has touch register in xorg core, but the mouse pointer doesn't move | 16:53 |
| lilstevie | I still get stuff like tooltips for where I touched | 16:53 |
| lilstevie | the "pointer" ends up where I touch | 16:53 |
| tjaalton | hmm, was the "hide a cursor after a few seconds" a change in the xserver or elsewhere? | 17:18 |
| bryceh | tjaalton, I recall the Dx team talking about something like that last sprint | 17:24 |
| bryceh | (think it was for removing 'clutter' from the screen) | 17:25 |
| tjaalton | i remember tseliot pointing out the change at some point | 17:26 |
| tjaalton | isn't the first time I'm asking :) | 17:26 |
| tseliot | tjaalton, bryceh: unclutter does it but I guess gtk applications do it now too | 17:29 |
| tseliot | there's no such thing in qt apps AFAIK | 17:30 |
| tjaalton | ok, so where do I throw bug 774434? | 17:32 |
| ubot4 | Launchpad bug 774434 in xserver-xorg-input-evdev (Ubuntu) "mouse pointer disappears in ubuntu 11.04 (affects: 12) (dups: 2) (heat: 124)" [Undecided,Confirmed] https://launchpad.net/bugs/774434 | 17:32 |
| bryceh | RAOF, what's the low down from that kernel graphics bug process meeting? | 17:32 |
| tjaalton | I could just close it | 17:32 |
| tjaalton | nah, tossed it over | 17:33 |
| RAOF | bryceh: Oh! Basically that JFo's on a leave of absence and so we shouldn't be subscribing/assigning things to him and expecting that to do anything; we should just be tagging kernel-handoff-graphics to get things on their radar. | 17:35 |
| tjaalton | bug 796325 is fun | 17:35 |
| ubot4 | Launchpad bug 796325 in xserver-xorg-input-evdev (Ubuntu) "USB HD used as pointer input dev (affects: 1) (heat: 254)" [Undecided,New] https://launchpad.net/bugs/796325 | 17:35 |
| tseliot | tjaalton: but we don't install unclutter by default. I guess it's gnome/gtk/etc. that does it | 17:35 |
| tseliot | I can reproduce it here without unclutter | 17:36 |
| tjaalton | tseliot: ah ok | 17:36 |
| tjaalton | I have it installed in oneiric | 17:36 |
| tseliot | tjaalton: it turns out that I was right: https://bugzilla.gnome.org/show_bug.cgi?id=328516#c6 | 17:39 |
| ubot4 | Gnome bug 328516 in general "Hide mouse cursor when keyboard is in use and mouse isn't" [Enhancement,Unconfirmed] | 17:39 |
| bryceh | RAOF, ok | 17:39 |
| tjaalton | tseliot: yeah, thanks | 17:40 |
| tseliot | np | 17:40 |
| bryceh | RAOF, I re-tagged all the 'xorg-needs-kernel-fix' tagged bugs to 'kernel-handoff-graphics' yesterday | 17:40 |
| RAOF | bryceh: Great; we're therefore doing what we should be doing :) | 17:41 |
| bryceh | RAOF, did they indicate who on their end will be reviewing that list? | 17:42 |
| tjaalton | uh, cursor up key not working if xserver is not restarted (kubuntu)?? | 17:42 |
| tjaalton | on logout | 17:42 |
| RAOF | bryceh: No; they no longer have a dedicated triager so I assume they're all scanning that list. | 17:43 |
| bryceh | ok, touched base with ogasawara | 17:50 |
| bryceh | sounds like they'll be scanning that list weekly, but for critical items we should ping them on #ubuntu-kernel | 17:51 |
| RAOF | Ok. That sounds easy enough. | 17:59 |
| === yofel_ is now known as yofel | ||
| horstle | does anybody know what this means "NVRM: Xid (0000:03:00): 6, PE0001"? | 23:05 |
| horstle | seen at dmesg | 23:05 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!