| RAOF | Man, writing upstart rules is super-easy. | 00:32 |
|---|---|---|
| RAOF | Now I just need to figure out the proper dbus-monitor incantation I want. | 00:32 |
| broder | RAOF: http://dbus.freedesktop.org/doc/dbus-specification.html#message-bus-routing-match-rules :) | 00:39 |
| jasoncwarner_ | RAOF and robert_ancell :0 glad to see both of you ! I have a lightdm-that-might-be-an-x question | 00:52 |
| jasoncwarner_ | RAOF and robert_ancell , so I got my fancy new x220 (intel, of course) and everything is working great now...after some tweaking | 00:55 |
| jasoncwarner_ | but on the unity-greeter screen, my mouse has HUGE amounts of lag on it...I can't really move it around... | 00:56 |
| jasoncwarner_ | not sure how to debug it and get you both some useful information | 00:56 |
| robert_ancell | jasoncwarner_, anything in the -greeter.log file in /var/log/lightdm? | 00:58 |
| * jasoncwarner_ looking | 00:58 | |
| robert_ancell | jasoncwarner_, also, once logged in can you run 'dm-tool add-nested-seat' and see if the same problem occurs? | 00:58 |
| jasoncwarner_ | robert_ancell: what should I be looking for in the greeter log? | 01:00 |
| robert_ancell | anything looping around? | 01:00 |
| robert_ancell | i.e. hundreds of lines of something? | 01:00 |
| jasoncwarner_ | robert_ancell: no, nothing like that... | 01:01 |
| jasoncwarner_ | dm-tool, should I just run that and then reboot? | 01:01 |
| robert_ancell | no, it will pop up a nested X window with a login screen | 01:02 |
| robert_ancell | which is running everything the same as normal, except it uses the Xephyr X server instead of the normal one | 01:02 |
| jasoncwarner_ | do you know package name ot install for xephyr? | 01:03 |
| robert_ancell | jasoncwarner_, apt-get install xserver-xephyr | 01:03 |
| jasoncwarner_ | thanks, installing... | 01:03 |
| jasoncwarner_ | thanks,,, robert_ancell | 01:04 |
| jasoncwarner_ | it does not happen in the dm-tool | 01:04 |
| jasoncwarner_ | mouse moves fine | 01:04 |
| jasoncwarner_ | robert_ancell: guessing X then... | 01:07 |
| robert_ancell | that suggests an X problem. It could be gnome-settings-daemon or something like that though | 01:07 |
| robert_ancell | the next step is to log out, and then go to a text terminal and run top and see if anything is acting up | 01:08 |
| robert_ancell | (I'm assuming it always occurs?) | 01:08 |
| jasoncwarner_ | robert_ancell: yeah, it has always happened ... "always" being, 'since I got this computer yesterday...every time on this comp" | 01:09 |
| jasoncwarner_ | I'll log out now and check top...be back in a few | 01:09 |
| robert_ancell | at least it's consistent | 01:09 |
| jasoncwarner_ | hey robert_ancell I just tried to shutdown/restart from the greeter | 01:13 |
| jasoncwarner_ | #1 - the box wasn't themed with a GTK style...it was base GTK | 01:13 |
| jasoncwarner_ | #2 - nothing worked? couldn't restart and couldn't power down... | 01:13 |
| robert_ancell | do you have a /etc/lightdm/unity-greeter.conf? | 01:13 |
| jasoncwarner_ | robert_ancell: rebooting again...will check in a sec | 01:14 |
| jasoncwarner_ | robert_ancell: I hate writing this, and I don't know how you'll feel reading it (RAOF either) | 01:15 |
| jasoncwarner_ | but when I went to a tty and then back to unity-greeter | 01:15 |
| jasoncwarner_ | the mouse just started working fine | 01:16 |
| jasoncwarner_ | and then when I rebooted...it seemed to work fine | 01:16 |
| jasoncwarner_ | that is my def of "magic" :) | 01:16 |
| robert_ancell | oh yay | 01:16 |
| robert_ancell | um, so problem solved right? | 01:16 |
| robert_ancell | :) | 01:16 |
| jasoncwarner_ | :) that would be an awesome note in the release notes: if laggy mouse, do this ;) | 01:17 |
| jasoncwarner_ | btw | 01:17 |
| jasoncwarner_ | i do have a unity-greeter.conf file | 01:17 |
| robert_ancell | jasoncwarner_, does it sound like bug #828112? | 01:18 |
| ubot2` | Launchpad bug 828112 in unity-greeter "Password field feedback slow at times" [Medium,Incomplete] https://launchpad.net/bugs/828112 | 01:18 |
| jasoncwarner_ | robert_ancell: I think it resembles, that but not exactly. | 01:21 |
| jasoncwarner_ | though, when I am having the problem, yes, the password field does have lag as well | 01:21 |
| robert_ancell | could you put your notes on that bug and note the mouse was also a sympton | 01:22 |
| jasoncwarner_ | robert_ancell: will do | 01:24 |
| jasoncwarner_ | robert_ancell: lag is back after reboot | 01:29 |
| jasoncwarner_ | and, yes, it was solved simply by this | 01:29 |
| jasoncwarner_ | CTRL+ALT F2 | 01:29 |
| jasoncwarner_ | CTRL+ALT F7 | 01:29 |
| jasoncwarner_ | no lag after that | 01:29 |
| jasoncwarner_ | rebooting again... | 01:29 |
| jasoncwarner_ | robert_ancell: could it be gnome-settings-daemon related? | 01:31 |
| jasoncwarner_ | I am getting pretty consistent crashes with g-s-s | 01:32 |
| jasoncwarner_ | and that might be why I don't have a GTK theme on the shutdown dialog? | 01:32 |
| robert_ancell | I'm just speculating, but both the indicators and g-s-d could run all sorts of things that unity-greeter has no control over | 01:32 |
| TheMuso | The sound indicator runs pulse for one... | 01:33 |
| jasoncwarner_ | I imagine if g-s-s fails to load I would have quite a few problems...and g-s-s crashes each and every boot for me. | 01:35 |
| jasoncwarner_ | so perhaps... | 01:35 |
| jasoncwarner_ | https://bugs.launchpad.net/ubuntu/+source/gnome-settings-daemon/+bug/838502 | 01:35 |
| ubot2` | jasoncwarner_: Error: <Bugtracker.plugin.Launchpad instance at 0x8d65a2c> bug 838502 not found | 01:35 |
| jasoncwarner_ | robert_ancell: after rebooting quite a few times,I think that bug you linked before and mine are probably the same | 02:20 |
| jasoncwarner_ | the op prob just didn't move his mouse to see it not moving | 02:21 |
| robert_ancell | jasoncwarner_, oh good. Is it coming back again? | 02:21 |
| jasoncwarner_ | yeah | 02:21 |
| robert_ancell | yeah, that was my guess too | 02:21 |
| jasoncwarner_ | robert_ancell: did you see my above? I can get it to go away with switching to TTY and back again | 02:23 |
| jasoncwarner_ | and when I lock my screen and go to switch user (which brings me back to unity-greeter) | 02:23 |
| jasoncwarner_ | I get it as well | 02:23 |
| robert_ancell | RAOF, ^ does that sound remotely X related? | 02:23 |
| robert_ancell | My first guess is video drivers with VT switching, or some sort of power management issue (i.e. idle checking) | 02:24 |
| jasoncwarner_ | robert_ancell: I'm on intel on this machine... | 02:25 |
| === qwebirc74340 is now known as cyphermox_ | ||
| RAOF | robert_ancell: That could be X related; VT switching drops drm master, and switching back regains it, which might mean that if it failed the first time then VT switching could re-enable acceleration? | 02:35 |
| robert_ancell | RAOF, anythink we could check to confirm that? | 02:36 |
| RAOF | robert_ancell: /var/log/Xorg.0.log would contain information to confirm or deny that hypothesis. | 02:36 |
| robert_ancell | jasoncwarner_, ^^ | 02:36 |
| jasoncwarner_ | RAOF: what should I be looking for int aht log? | 02:37 |
| RAOF | jasoncwarner_: Well, pastebinning it would be good. The drmOpenDevice lines would be part of the bit. | 02:38 |
| jasoncwarner_ | RAOF: http://pastebin.com/HwtT0JZn | 02:40 |
| RAOF | Well, that doesn't contain what I'd expect. | 02:41 |
| RAOF | Oh! Compiz is being really slow? Can you check your top output? I've had a problem where something's spamming the dbus, and dbus-daemon has high CPU usage, and because the panel is in the compositor that means that _everything's_ slow. | 02:42 |
| RAOF | (I think in my case it was because indicator-datetime was endlessly trying to enumerate all my calendars) | 02:42 |
| RAOF | VT switching could possibly halt compiz for long enough to unwedge things. | 02:43 |
| jasoncwarner_ | RAOF: my top didn't show compiz being slow...I'll reboot and look again...but remember, switching to a VT seems to clear it all up...I'm not sure I'll be able to see it... | 02:44 |
| jasoncwarner_ | oh...I'll log in remotely to see if it is there | 02:44 |
| jasoncwarner_ | RAOF: ok...I'm actually goign to finish these three emails from my TODO first...been all day and I haven't gotten 'em done...be back in a bit | 02:48 |
| RAOF | jasoncwarner_: No problem :) | 02:49 |
| RAOF | robert_ancell: Have you played around with dbus-monitor? I can't seem to get it to display anything but signals, and I want to trace method calls instead. | 02:49 |
| robert_ancell | RAOF, no, sorry | 02:50 |
| stgraber | RAOF: system or session bus? and do you need the content of the call of just interface, path and function name? | 02:51 |
| broder | RAOF: you can't use dbus-monitor to get method calls by default | 02:51 |
| RAOF | stgraber: System bus, and I really just want to know _when_ it gets called. | 02:51 |
| RAOF | broder: Ah, well now my results make sense :) | 02:52 |
| stgraber | RAOF: sorry can't help much with system bus unless you can tell your client that it's at another location than /run/dbus/system_bus_socket | 02:52 |
| stgraber | (I wrote a dbus proxy as part of my work on arkose which logs everything that goes through it) | 02:52 |
| broder | RAOF: if you put http://pastebin.ubuntu.com/679351/ at /etc/dbus-1/system-local.conf and then do pkill -HUP dbus-daemon, you'll be able to see them | 02:53 |
| RAOF | broder: Can I mess with the system bus config to make that work? | 02:53 |
| broder | but that completely compromises the integrity of anybody that's depending on dbus for security :) | 02:53 |
| broder | e.g. NM | 02:53 |
| RAOF | Well, it's just for local debugging, so that's fine. | 02:54 |
| robert_ancell | RAOF, does it make sense to disable TCP connections in lightdm by default for security reasons (Bug #838152) | 02:54 |
| ubot2` | Launchpad bug 838152 in lightdm "hard-coded "-nolisten tcp"" [Low,Triaged] https://launchpad.net/bugs/838152 | 02:54 |
| RAOF | robert_ancell: Not listening on TCP is certainly the correct default. | 02:55 |
| robert_ancell | RAOF, ta | 02:55 |
| RAOF | It's something that you might want to offer as an option, though. | 02:56 |
| broder | RAOF: fwiw, i think there are less heavyhanded ways of enabling limited eavesdropping, but i've never been able to get them to work | 03:13 |
| RAOF | broder: Nyargh! Is there a *more* heavyhanded way of enabling eavesdropping? I've dumped that in, and can see _other_ method calls, but not the ones that I'm interested in! | 03:14 |
| broder | hmm..that's not a problem i've had before | 03:15 |
| broder | are you sure the calls are actually getting made? | 03:15 |
| RAOF | Ah, there we go. | 03:15 |
| RAOF | Apparently you need to restart d-feet before that'll work. | 03:15 |
| RAOF | Oh, yeah! The dbus reload probably needs that, doesn't it. | 03:16 |
| broder | huh, weird | 03:16 |
| RAOF | I wasn't expecting getting dbus-monitor to work to be the most difficult part of this :) | 03:18 |
| RAOF | Writing upstart rules is plesantly easy. | 03:20 |
| randrums | https://plus.google.com/115895054477394715961/posts/84KCPX2xgUh | 03:21 |
| DBO | ping RAOF | 03:22 |
| DBO | so good news bad news time | 03:23 |
| DBO | good news: the signal for figuring out when a resume happens works | 03:23 |
| DBO | bad news: recreating the FBO's doesn't fix the corruption issue | 03:23 |
| RAOF | DBO: Hm. That's non-good. | 03:44 |
| DBO | so looking at it more carefully | 03:44 |
| DBO | what I observe is this | 03:44 |
| RAOF | DBO: I thought you'd previously tested and found that to work? | 03:44 |
| DBO | it did in nux | 03:45 |
| DBO | I was assuming that would transfer over... | 03:45 |
| DBO | (we had the inverse problem last cycle, nux was corrupted and compiz wasn't) | 03:45 |
| DBO | so my observations seem to be as follows: | 03:45 |
| DBO | 1) the FBO, without being reloaded, actually pants fine (contents are uncorrupted when painted into) | 03:46 |
| DBO | 2) painting the texture the fbo is bound to onto another fbo works fine | 03:46 |
| DBO | 3) painting the compiz FBO onto the backbuffer somehow fails terribly | 03:46 |
| DBO | which really makes me wonder, why do the nux fbo's paint just fine | 03:47 |
| RAOF | Because magic. | 03:47 |
| RAOF | Presumably :) | 03:47 |
| DBO | I have a screenshot | 03:48 |
| DBO | RAOF, http://i.imgur.com/TFOXc.png | 03:50 |
| DBO | see how the corruption is a predictable pattern | 03:50 |
| DBO | I suspect somehow the driver is blending the FBO onto the backbuffer with an incorrect format | 03:50 |
| RAOF | That looks like it's likely to be tiling-related. | 03:50 |
| DBO | tiling? | 03:51 |
| RAOF | Given that it occurs on what looks like tiling-sized blocks. | 03:51 |
| RAOF | Textures aren't linear in memory, at least not if you want it to be fast. | 03:52 |
| DBO | oh | 03:52 |
| DBO | is that done automatically by the driver? | 03:52 |
| RAOF | They're tiled into crazy patterns so that the multi-bar memory controllers are used maximally. | 03:52 |
| RAOF | Yeah, that's done by the driver. | 03:52 |
| DBO | maybe an error in the fixed function pipe | 03:53 |
| DBO | (sam, he's silly like that) | 03:53 |
| DBO | obviously unity blends fine | 03:54 |
| RAOF | But it's a fertile source of bugs in the kernel/DDX/mesa layers, at least for the free drivers. | 03:54 |
| DBO | and functionally, it too is painting FBO's | 03:54 |
| RAOF | Yeah, if Unity's doing shadery things and compiz is using fixed-function that could be different, I guess. | 03:55 |
| DBO | well I mean painting the contents of an FBO shouldn't be functionally different than painting a texture | 03:56 |
| DBO | which still works fine | 03:56 |
| RAOF | DBO: Will you still be around in ~20 minutes? I'd like to relocate back from this cafe to home. | 03:56 |
| DBO | yes | 03:57 |
| RAOF | Ta. | 03:57 |
| RAOF | Catch you in 20. | 03:57 |
| RAOF | DBO: Back. | 04:23 |
| DBO | welcome back RAOF | 04:23 |
| DBO | RAOF, if you got any ideas that would be great... | 04:26 |
| DBO | RAOF, is the backbuffer tiled? | 04:26 |
| RAOF | *Everything* is likely to be tiled; back buffer, front buffer, textures, the whole enchilada. | 04:26 |
| RAOF | Because it's like a 300% performance win. | 04:26 |
| DBO | hmmm | 04:27 |
| RAOF | (Well, not exactly, but it's a huge performance win) | 04:27 |
| DBO | I hope we dont have to recreate the GL context to get around this | 04:27 |
| RAOF | In fact, from what I gather from nouveau the nv5x+ cards don't actually handle *untiled* surfaces. | 04:27 |
| DBO | that would kind of suck | 04:27 |
| RAOF | That would rather suck, yes. | 04:28 |
| RAOF | Can you recreate the backbuffer or something? Like, swap out the composite overlay window? | 04:28 |
| DBO | I dont think so | 04:34 |
| DBO | maybe? | 04:34 |
| RAOF | That might be something to try, if some tiling bit isn't getting reset properly across suspend. | 04:35 |
| DBO | wow | 04:36 |
| DBO | I just realized that the opengl plugin in compiz creates a shader in the most crazy ass fashion | 04:37 |
| DBO | im betting if we change how we paint this thing | 04:37 |
| DBO | that will fix it | 04:37 |
| RAOF | Hah! | 04:38 |
| RAOF | Always it is with the craaaazy compiz! | 04:38 |
| jbicha | robert_ancell: good work on unity-greeter today, fixed 2 of my bugs! | 04:47 |
| jbicha | my indicators are in a crash-loop, it's "fun" to see the appmenu appearing and reappearing | 04:49 |
| jbicha | are there any logs I can look at it to see what's breaking with my indicators? | 04:57 |
| robert_ancell | jbicha, :) | 05:23 |
| pitti | Good morning | 05:23 |
| jbicha | pitti: good morning | 05:25 |
| kenvandine | good morning pitti | 05:25 |
| pitti | hey guys, how are you? | 05:26 |
| TheMuso | Morning pitti. | 05:30 |
| RAOF | Hey pitti | 05:30 |
| pitti | so, interesting release night.. | 05:31 |
| kenvandine | pitti, how is the release looking? | 05:34 |
| pitti | was respun during the night due to a pretty serious bug with GPT partitions | 05:34 |
| didrocks | good morning | 05:46 |
| pitti | bonjour didrocks | 05:47 |
| RAOF | Oh, no! It's didrocks! Quick, hide! :) | 05:48 |
| didrocks | guten morgen pitti | 05:52 |
| didrocks | RAOF: not fast enough! :-) | 05:52 |
| didrocks | the more I see bug report, the more I will be in favor to revert Ctrl + Del to simply Del. The trash is there for fat fingers… | 05:57 |
| jasoncwarner_ | didrocks: I agree with that.. | 05:57 |
| pitti | and now crtl+del is harder to do than shift-del, which immediately removes it.. | 05:58 |
| pitti | this was one of these rather pointless changes which destroy the habits learned in years *sigh* | 05:59 |
| jasoncwarner_ | pitti didrocks change it to just DEL? makes most sense to me. | 05:59 |
| didrocks | can do it if you both agree :) | 06:00 |
| didrocks | and +1 with pitti's remark, it's harder that shift + del, which is some kind of hilarious :) | 06:00 |
| didrocks | than* | 06:00 |
| pitti | didrocks: we'd need to update the GNOME documentation for that, too, though | 06:00 |
| didrocks | pitti: indeed, so, opening/find a bug and subscribe the doc team to it? (I think there is no patch to the doc for all the indicator/systray removal anyway) | 06:01 |
| pitti | not sure whether it's in gnome-user-guide or nautilus help or both | 06:01 |
| didrocks | I can have a look, will do that after the cjk landing | 06:02 |
| didrocks | pitti: so, a bunch of package will land with real cjk support (there was a first attempt one ok ago while I was on vacation if i understood correctly), but this time, the correct patches will land | 06:03 |
| pitti | just saw the confirmation in the bug, nice to hear! | 06:03 |
| didrocks | this touches apt-xapian-index, zg-extension, sotware-center and unity | 06:03 |
| didrocks | jasoncwarner_: FYI ^ | 06:04 |
| didrocks | pitti: ok, will land today if all go well :) | 06:04 |
| jasoncwarner_ | didrocks: saw that....big change...lots of testing needed | 06:04 |
| didrocks | jasoncwarner_: indeed | 06:04 |
| didrocks | jasoncwarner_: what was missing from the first round was the rebuild of the index once cjk is there | 06:05 |
| jasoncwarner_ | didrocks: ah...cool...thanks | 06:05 |
| jasoncwarner_ | once this gets in then, back to oneconf? ;) | 06:05 |
| didrocks | jasoncwarner_: I finished the first round of oneconf in parrallel, it's now working with software-center gtk3 :) | 06:08 |
| didrocks | still missing some speedup improvment (caching and stuff) that I made for the gtk2 version, but it's usable (the whole software-center gtk3 needs some speedups anyways) | 06:09 |
| jbicha | didrocks: I'll take care of making sure the Ubuntu docs say just Delete | 06:58 |
| didrocks | jbicha: I'm writing a script for that, will be way easier :) | 06:58 |
| jbicha | it's not in that many places, is it? | 06:58 |
| didrocks | jbicha: it's in gnome-user-guide FYI, I just run a script to do the sed and such | 06:59 |
| didrocks | jbicha: yeah, but there are the translations ;) | 06:59 |
| jbicha | didrocks: the real docs are in ubuntu-docs, gnome-user-docs is currently just a copy from Debian, it's unclear whether we want to maintain a diff there | 06:59 |
| didrocks | jbicha: if I press F1 on nautilus, it's gnome-user-docs which is used, isn't it? | 07:00 |
| jbicha | didrocks: yes that will need to be patched too, thanks :-) | 07:01 |
| didrocks | jbicha: I'm patching this one (and so, add the tool for regenerating the patch for easier use), can you take care of ubuntu-docs, please? | 07:01 |
| jbicha | didrocks: actually can you correct the nautilus help link to point to ghelp:ubuntu-docs instead of gnome-user-docs? | 07:02 |
| jbicha | gnome-user-docs isn't installed by default | 07:02 |
| jbicha | I thought Nautilus shipped its own help but it just piggybacks on the system help since it's Core | 07:03 |
| didrocks | jbicha: you're right, it's not installed by default | 07:06 |
| didrocks | weird, it's in main, maybe for the dvd? | 07:06 |
| jbicha | didrocks: upgraders will still have it & it's pulled in for users who install gnome shell | 07:06 |
| jbicha | for Natty, we heavily modified gnome-user-docs instead of just using ubuntu-docs so it was installed in Natty | 07:07 |
| RAOF | pitti: You're our resident PolicyKit maestro, right? In bug #837851 it seems that policykit hates colord running as not root; is there a way to tell PolicyKit that it's ok for the colord user to check whether $USER has authorisation on its interfaces? | 07:09 |
| ubot2` | Launchpad bug 837851 in colord ""Color" tool in GNOME System Settings only recognizes my webcam as color-manageable device" [High,Confirmed] https://launchpad.net/bugs/837851 | 07:09 |
| didrocks | jbicha: ah, makes sense then, now everything is in ubuntu-docs? I should not touch gnome-user-docs then | 07:09 |
| jbicha | didrocks: yes, it's just easier to keep gnome-user-docs in sync with Debian for now | 07:09 |
| didrocks | jbicha: ok, I remove my script then :) | 07:10 |
| didrocks | jbicha: adding an ubuntu-docs task and assign to you, ok? | 07:10 |
| jbicha | didrocks: that's fine, we were thinking about pushing a new ubuntu-docs update this weekend | 07:11 |
| RAOF | pitti: Specifically, colord asks polkit, as the colord user, whether $USER is authorised to perform the actions, and polkit only allows root to query whether someone else is authorised to perform an action (presumably for security purposes). | 07:11 |
| didrocks | jbicha: but that means that we will have to patch every application to reference ubuntu-docs rather than gnome-user-docs? | 07:11 |
| jbicha | didrocks: there aren't that many apps that use g-u-d, but gnome-control-center probably does | 07:13 |
| didrocks | jbicha: ok, let's try to change it for nautilus first | 07:13 |
| chrisccoulson | good morning everyone | 07:14 |
| pitti | hey chrisccoulson | 07:15 |
| chrisccoulson | hi pitti, how are you? | 07:15 |
| didrocks | jbicha: so, basically, now it's ghelp:gnome-help#… should it be ghelp:ubuntu-help#… ? | 07:15 |
| didrocks | good morning chrisccoulson | 07:15 |
| chrisccoulson | hi didrocks | 07:15 |
| pitti | chrisccoulson: pretty well, thanks! | 07:15 |
| pitti | RAOF: (will get to you, still stuck in release juggling and CD testing in parallel) | 07:15 |
| RAOF | pitti: No problem! | 07:15 |
| chrisccoulson | hmmmm, i just did apport-bug compiz, and apport seems to have frozen :/ | 07:18 |
| jbicha | didrocks: yes, when we push the new ubuntu-docs we'll have to change it again though because the new url will be help: | 07:18 |
| jbicha | instead of ghelp: | 07:18 |
| didrocks | jbicha: ok, let's do that change once you pushed the new ubuntu-docs package, ok? | 07:19 |
| didrocks | jbicha: I'll just push the Ctrl + Del -> Del | 07:19 |
| jbicha | didrocks: yes the other isn't ready yet | 07:20 |
| pitti | RAOF: so, polkit | 07:33 |
| pitti | RAOF: right, I'm afraid the 'ownuser || uid==0" check is hardcoded, so that you can't spy on other uers | 07:33 |
| pitti | users | 07:33 |
| RAOF | That seems like a misfeature. | 07:34 |
| pitti | so for this we'd need a provision to mark a particular system user as being able to do the same | 07:34 |
| RAOF | Right, that was what I was thinking. | 07:34 |
| pitti | or a particular program | 07:34 |
| pitti | like, this privilege could become a polkit privilege | 07:34 |
| pitti | (recursivity FTW) | 07:35 |
| RAOF | org.freedesktop.polkit.check_authorisation :) | 07:35 |
| RAOF | Hm, no. That wouldn't work; it needs to be a per-action thing. | 07:35 |
| RAOF | You want the colord user to be able to check whether $USER is authorised for org.freedesktop.colord.*, but not anything else. | 07:36 |
| RAOF | So, an extension to the action semantics could do this. | 07:36 |
| pitti | RAOF: what does colord do if it cannot check this? | 07:36 |
| RAOF | Refuses to add the device, so the cups and g-s-d integration fail. | 07:36 |
| pitti | asked the other way round, what are the default permissions on org.freedesktop.colord.* ? | 07:37 |
| RAOF | at_console | 07:37 |
| pitti | I'm not entirely sure why it asks a particular user's permissions to add devices to cups, as cupsd is a system service | 07:37 |
| pitti | RAOF: eww | 07:37 |
| RAOF | Yeah, I'm not sure if cups is actually going to work *at all* | 07:37 |
| pitti | colord uses polkit _and_ at_console? | 07:38 |
| pitti | at_console is an ancient hack for programs which don't use polkit | 07:38 |
| RAOF | Sorry, I couldn't remember the polkit name for at_console. It's default-deny, allow for active user. | 07:38 |
| pitti | ah, ok | 07:38 |
| RAOF | Ah, allow_active. | 07:38 |
| RAOF | Will allow_active allow cupsd access? | 07:40 |
| pitti | no | 07:40 |
| pitti | cupsd is not on any console | 07:40 |
| Sweetshark | Morning all! | 07:40 |
| pitti | the consolekit stuff only works for actual sessions, not for system daemons | 07:40 |
| pitti | hey Sweetshark | 07:40 |
| Sweetshark | pitti: libreoffice_3.4.2-2ubuntu3 for upload on chinstrap | 07:40 |
| pitti | RAOF: another option would be to actually stop using polkit, and control this with d-bus policy | 07:41 |
| RAOF | pitti: Yeah, thought so. So, cupsd isn't going to work even once I fix the session stuff. | 07:41 |
| pitti | RAOF: i. e. allow calling the API from root (that'll cover cupsd) and at_console | 07:41 |
| pitti | cf. my comment about at_console being deprecated, though | 07:42 |
| pitti | RAOF: yet another alternative is to let colord run as root, and lock it down with an apparmor profile | 07:42 |
| RAOF | pitti: That would be reasonably easy, and a transitional plan while fixing polkit. | 07:42 |
| pitti | RAOF: depends on how much patching is involved to remove all the polkit checks from colord | 07:43 |
| RAOF | Although running as root _still_ won't fix cupsd, right? | 07:43 |
| RAOF | pitti: colord has a --no-polkit option. | 07:43 |
| pitti | RAOF: but I guess it could be factored out, and the functino would only ask polkit if geteuid() == 0 | 07:43 |
| pitti | and otherwise just allow it | 07:43 |
| pitti | then we put a d-bus policy in front of it | 07:43 |
| pitti | that patch should even be upstreamable | 07:44 |
| pitti | RAOF: ah, that sounds promising, too | 07:44 |
| RAOF | It wouldn't need to be a patch at all; we can just build colord without policykit support. | 07:44 |
| pitti | you'd need to patch the d-bus policy, but that's relatively simple | 07:45 |
| RAOF | Yeah, I'm pretty sure I know how to do that. | 07:45 |
| pitti | some of the interfaces might be save for everyone to call | 07:45 |
| pitti | and the ones that aren't should be removed from context="default" and added to <policy at_console="true"> | 07:46 |
| pitti | and again to <policy user="root"> | 07:46 |
| pitti | to cover g-s-d (first) and cupsd (second) | 07:46 |
| pitti | RAOF: /etc/dbus-1/system.d/NetworkManager.conf does that, for example | 07:47 |
| RAOF | Great. | 07:47 |
| RAOF | And then, a patch to polkit to allow doing this there. | 07:48 |
| pitti | and a patch to its policy to allow access to root users, not just allow_active | 07:52 |
| pitti | RAOF: ^ that's easy to do, I can help you with that once it becomes relevant | 07:52 |
| pitti | (man pklocalauthority) | 07:52 |
| pitti | something like | 07:52 |
| pitti | [allow access to cups] | 07:53 |
| pitti | Identity=unix-user:root | 07:53 |
| pitti | Action=org.fd.colord.whatever | 07:53 |
| pitti | ResultAny=yes | 07:53 |
| didrocks | pitti: when you get a second: bug #838623. Do you need a bug/FFe for activating the CJK support? (it was supposed to already be activated, it wasn't the case though) | 07:54 |
| ubot2` | Launchpad bug 838623 in software-center "FFe, UIFe: bring back OneConf to Software-Center" [Medium,In progress] https://launchpad.net/bugs/838623 | 07:54 |
| pitti | didrocks: no, bug fix | 07:54 |
| didrocks | ok, thanks :) | 07:55 |
| pitti | everything is supposed to work in CJK :) | 07:55 |
| RAOF | Oh, there isn't an <allow_root>yes</allow_root> specifier in the actions. That's kinda strange. | 07:56 |
| didrocks | pitti: indeed :) | 07:56 |
| RAOF | Let's play with dbus policy! | 07:56 |
| pitti | RAOF: not in the .policy files, no; that needs to go into pklocalauthority | 07:57 |
| RAOF | The strange corners you find in polkit! | 07:59 |
| kamstrup | didrocks: don't sweat over dee - we don't need it really, all required things are distropatched by ken, it's just so that other distros can ship gwibber without cherry picking extra api from trunk :-) | 08:00 |
| didrocks | kamstrup: I know, was just kidding :-) | 08:01 |
| kamstrup | didrocks: phew, i wouldn't wanna upset you already after my second day back ;-) | 08:01 |
| didrocks | kamstrup: it's never too late to upset me, seems that's dx paradigm ;-) | 08:02 |
| rodrigo_ | morning | 08:10 |
| rodrigo_ | kamstrup, did you get someone to upload jhbuild? | 08:23 |
| kamstrup | rodrigo_: I poked lool (as he was the original maintainer afaik), but he was not overly keen on jumping in :-) | 08:24 |
| rodrigo_ | kamstrup, ok, I can do a merge proposal then | 08:24 |
| rodrigo_ | what's the jhbuild version you need? | 08:25 |
| rodrigo_ | hmm, there's only 2.32.4 from last December in download.gnome.org | 08:25 |
| kamstrup | rodrigo_: that's also the one :-) | 08:26 |
| rodrigo_ | that's the one you need? | 08:26 |
| chrisccoulson | jasoncwarner_, enjoying your new laptop? ;) | 08:50 |
| jasoncwarner_ | chrisccoulson: yes, yes I am | 08:51 |
| chrisccoulson | excellent :) | 08:52 |
| jasoncwarner_ | and, chrisccoulson , glad you are here! :) I had a question for you. I just installed aurora build of FF | 08:52 |
| jasoncwarner_ | and the menu and ubuntu theme went away | 08:52 |
| jasoncwarner_ | anyway to get those back? | 08:52 |
| chrisccoulson | i really want to buy a new laptop, but i fear that jo might end my life if i spend money on one, when we're meant to be saving for our own house ;) | 08:52 |
| chrisccoulson | hmmm, the menu disappeared entirely, or appears in it's own window? | 08:52 |
| jasoncwarner_ | chrisccoulson: it is just not in the global menu position. I would like to renable that because it takes up precious screen realestate on the smaller monitor! :) | 08:53 |
| jasoncwarner_ | chrisccoulson: just tell her for hte good of your relationship you need a new laptop | 08:53 |
| chrisccoulson | jasoncwarner_, you could try reporting a bug with apport-bug, as the new apport hook i wrote attaches lots of useful information which might save me asking lots of quetions :) | 08:53 |
| jasoncwarner_ | and then watch how quickly a good relationship can go bad ;) | 08:53 |
| chrisccoulson | heh | 08:53 |
| jasoncwarner_ | chrisccoulson: I think the menu is the new firefox "we don't enable third party libraries anymore" stuff | 08:54 |
| jasoncwarner_ | it says it will be renabled when it is "compatible" | 08:54 |
| chrisccoulson | jasoncwarner_, this is a build from the firefox-aurora PPA isn't it? | 08:54 |
| jasoncwarner_ | whatever that means | 08:54 |
| jasoncwarner_ | yes | 08:54 |
| chrisccoulson | jasoncwarner_, yeah, apport-bug firefox might help me figure out what's wrong :) | 08:54 |
| jasoncwarner_ | ok.. | 08:55 |
| chrisccoulson | it should let you report a bug from the PPA | 08:55 |
| chrisccoulson | and it will be useful to find out if the new hook works from there :) | 08:55 |
| jasoncwarner_ | what about the ubuntu theme? can that be reenabled as well? | 08:55 |
| jasoncwarner_ | stock firefox isn't exactly pretty :/ | 08:55 |
| chrisccoulson | jasoncwarner_, we only shipped with the stock firefox theme before (it just inherits colours from our theme) | 08:56 |
| jasoncwarner_ | this "every new release of firefox breaks all the plugins" thing is getting old... | 08:56 |
| jasoncwarner_ | I highly doubt all those plugins are broken | 08:56 |
| jasoncwarner_ | would like to be able to override and try it out anyway | 08:57 |
| chrisccoulson | jasoncwarner_, did you manage to report a bug? | 09:00 |
| jasoncwarner_ | on a call...do it in a bit...I'll let you know... | 09:00 |
| chrisccoulson | thanks | 09:00 |
| jbicha | jasoncwarner_: don't you use https://addons.mozilla.org/addon/add-on-compatibility-reporter/ ? | 09:01 |
| chrisccoulson | jbicha, that won't ever work for the globalmenu-extension | 09:11 |
| chrisccoulson | if it's incompatible, forcing it on will likely just make it crash | 09:11 |
| jasoncwarner_ | chrisccoulson: https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/838671 | 09:13 |
| ubot2` | Ubuntu bug 838671 in firefox "Auroa PPA does not work with global menu" [Undecided,New] | 09:13 |
| chrisccoulson | jasoncwarner_, ah, silly me. i need to refresh the aurora branch with the latest version of the extension ;) | 09:14 |
| pitti | Sweetshark: oh, you apparently built with -sa; I'll edit the sources.changes in-place to drop the tarballs | 09:14 |
| chrisccoulson | i should have checked that already | 09:14 |
| chrisccoulson | https://launchpadlibrarian.net/78687096/Extensions.txt | 09:14 |
| zen_monkey | nights, how can i report an issue with gnome-settings-daemon from the gnome3-team? | 09:14 |
| pitti | Sweetshark: oh, or was this done for libreoffice_3.4.2.orig-ext-human-updates.tar.gz? i. e. is that one new? | 09:15 |
| zen_monkey | it's seriously hogging cpu... lappy will melt :D | 09:15 |
| pitti | Sweetshark: I'll try whether uploading only this will work | 09:15 |
| pitti | Sweetshark: uploaded | 09:16 |
| chrisccoulson | jasoncwarner_, i'll push the latest version and respin the build | 09:16 |
| chrisccoulson | jasoncwarner_, oh, were you actually using a custom theme before btw? | 09:18 |
| chrisccoulson | https://launchpadlibrarian.net/78687110/Themes.txt shows that the default is enabled, but you have an incompatible theme installed in your profile | 09:18 |
| jasoncwarner_ | yeah...I was... | 09:18 |
| Sweetshark | pitti: Yes, I had too build with -sa: there is one new tarball *-ext-human-updates (I hope you didnt drop that one ...) | 09:18 |
| pitti | Sweetshark: no, that's the one I kept :) | 09:19 |
| Sweetshark | *relieved* | 09:19 |
| chrisccoulson | jasoncwarner_, the theme seems to work on a current nightly with addon compatibility checking disabled | 09:21 |
| chrisccoulson | jasoncwarner_, https://addons.mozilla.org/en-US/firefox/addon/add-on-compatibility-reporter/ | 09:21 |
| chrisccoulson | jasoncwarner_, i'm respinning the aurora builds now, so you should get an update in a few hours | 10:11 |
| chrisccoulson | wow, i've reported 3 bugs this morning. i think that's the most i've reported in ages ;) | 10:11 |
| didrocks | pitti: small pygi question, how do you handle out parameter for basic types, like GtkMenuPositionFunc (x and y) ? | 10:21 |
| pitti | didrocks: normally, when you just simply call them, they are part of the return tuple | 10:22 |
| pitti | but in this case these are out values of a callback | 10:22 |
| pitti | didrocks: in theory, the callback should just return them, but I don't know whether anyone did that before | 10:23 |
| didrocks | pitti: ok, let's have a try then ;) | 10:23 |
| didrocks | pitti: interestingly, it seems to work for y, but not for x… I try adding random parameters, it doesn't complain… | 10:31 |
| ronoc | pitti, hey ! do you know a way to query if the guest session is disabled on a system ? I could not see anything obvious on d-feet | 10:31 |
| pitti | didrocks: with somethign as complex and unusual as correctly marshalling out arguments for callbacks I'd not be surprised if it's buggy.. | 10:31 |
| pitti | ronoc: hm, that's a question for robert_ancell | 10:32 |
| pitti | ronoc: I'm not familiar with how lightdm handles that, I'm afraid | 10:32 |
| ronoc | pitti, sure no probs | 10:32 |
| didrocks | pitti: yeah, confirm, only y is working, where should I open a bug, can get a minimal reproduceable testcase | 10:33 |
| pitti | didrocks: https://bugzilla.gnome.org/enter_bug.cgi?product=pygobject | 10:33 |
| didrocks | pitti: thanks! | 10:34 |
| pitti | didrocks: does that block you? | 10:34 |
| pitti | didrocks: I used popup_for_menu() in computer-janitor, but with a None callback, so I didn't notice that | 10:34 |
| didrocks | pitti: not, it's just a little bit uglier (the popup menu will appear where the cursor is) | 10:34 |
| pitti | yeah, that seems to be a reasonable default (menu where you click) | 10:34 |
| didrocks | so, not a blocker, just a polish | 10:35 |
| pitti | right | 10:35 |
| didrocks | that's not the behavior of what I was using: gtkMenuToolButton | 10:35 |
| didrocks | it position the button above or below | 10:35 |
| didrocks | but this button has two part: the button itself and the dropdown to show the menu | 10:35 |
| didrocks | in mpt's spec, it's only one button (https://wiki.ubuntu.com/SoftwareCenter#oneconf) | 10:35 |
| didrocks | so trying to polish this with a regular button which shows the popupmenu | 10:36 |
| didrocks | ok, it works on the minimum "not reproducer" :/ | 10:46 |
| didrocks | pitti: oh I get it, it's because I was on my secondary screen, and so x was outside my monitor and so, it sticks on the monitor border | 10:47 |
| didrocks | so in a nutshell, it works, I just have the get the absolute coordinate | 10:48 |
| pitti | nice! | 10:48 |
| pitti | didrocks: so the callback just returns the actual return value and all the out values in order? | 10:48 |
| didrocks | pitti: indeed, and you just have to not list them as args of the method of course :) | 10:49 |
| pitti | right | 10:49 |
| pitti | nice | 10:49 |
| jasoncwarner_ | pitti: just sent you my bootcharts for today...now I'm going to bed! have to get up around 5 for a call :/ | 11:21 |
| rodrigo_ | hmm, I don't get any mail when packages in the queue are accepted (or rejected, don't know), is that normal? | 11:34 |
| Sweetshark | pitti, jasoncwarner_: tommorrow morning I wont be on IRC as I will fly to http://wiki.documentfoundation.org/Hackfest2011 | 11:57 |
| jasoncwarner_ | Sweetshark: ok, thanks. have fun and crack some heads ;) | 11:57 |
| jasoncwarner_ | Sweetshark: see if you can get any traction on simplifying the UI! :) | 11:58 |
| Sweetshark | jasoncwarner_: yep | 11:58 |
| * jasoncwarner_ notes that he sounds like a broken record sometimes | 11:58 | |
| Sweetshark | jasoncwarner_: Christoph will be there too ... | 11:59 |
| jasoncwarner_ | nice...when are you back? | 11:59 |
| Sweetshark | jasoncwarner_: I fly back on sunday. | 11:59 |
| jasoncwarner_ | Sweetshark: ok...we'll catch up next week. cheers. have fun... | 12:00 |
| * jasoncwarner_ going to bed for real this time... | 12:00 | |
| === MacSlow is now known as MacSlow|lunch | ||
| pitti | jasoncwarner_: ok, we need some further mods for that | 12:05 |
| * rodrigo_ -> lunch | 12:10 | |
| * didrocks has a super small testcase working, but returning the same coordinate in pressing enter or clicking on the mouse doesn't do the same thing in software-center O_o | 13:47 | |
| ronoc | pitti, is mvo on holidays ? | 13:50 |
| pitti | ronoc: yes | 13:51 |
| stgraber | didrocks: hey, for Edubuntu we're customizing the big-friendly-button of unity. Apparently diverting /usr/share/unity/4/launcher_bfb.png does the trick for the 3d version but not for the 2d one. Do you happen to know the equivalent? | 13:55 |
| didrocks | stgraber: I guess that agateau used the vendor-logo or something like that, but he should fix it for using the same | 14:00 |
| didrocks | agateau: FYI ^^ | 14:00 |
| stgraber | agateau: bug 838829 | 14:00 |
| ubot2` | Launchpad bug 838829 in edubuntu-artwork "[UIFe] Big friendly button not customized in Unity 2d" [Medium,New] https://launchpad.net/bugs/838829 | 14:01 |
| stgraber | agateau: I'm fine diverting another file if needed but I guess using the same icon for both versions would be better :) | 14:01 |
| agateau | stgraber: is it for Natty or Oneiric | 14:01 |
| agateau | ? | 14:01 |
| stgraber | agateau: Oneiric | 14:01 |
| agateau | stgraber: : I sent an email to design and dx asking if everyone agreed on using distributor-logo... it seems I was ignored :/ | 14:01 |
| didrocks | agateau: stop using standard icons! :) | 14:02 |
| === MacSlow|lunch is now known as MacSlow | ||
| agateau | didrocks: heh, my mistake, won't happen again :) | 14:03 |
| didrocks | hope so! :-) | 14:03 |
| agateau | stgraber: I think bug 838829 should be reassigned to unity-2d (or at least marked as "also affect") | 14:05 |
| ubot2` | Launchpad bug 838829 in edubuntu-artwork "[UIFe] Big friendly button not customized in Unity 2d" [Medium,New] https://launchpad.net/bugs/838829 | 14:05 |
| stgraber | agateau: Sure, I can change that to "Make unity-2d use the same bfb icon as unity". Will I have that done by beta2? | 14:06 |
| agateau | stgraber: I think so | 14:06 |
| * agateau starts a branch | 14:06 | |
| === smstudy is now known as smspillaz | ||
| stgraber | agateau: should I assign that bug to you? | 14:11 |
| agateau | stgraber: please do | 14:11 |
| Daviey | GRRRRR.. who bound Alt+a ? :( | 14:14 |
| Daviey | How do you change keybindings in unity? | 14:22 |
| pitti | Daviey: bound to what? | 14:28 |
| pitti | Daviey: these days, in ccsm I think | 14:28 |
| pitti | but Alt+A doesn't do anything here, it gets into the application here | 14:29 |
| pitti | it's not supposed to be bound to anything | 14:29 |
| === davidcalle_ is now known as davidcalle | ||
| oier | I have seen that there is some kind of alt+' switcher but in my ccsm unity plugin I don't see any switcher tab to activate or configure it | 15:03 |
| oier | am I missing an update or a package? | 15:04 |
| tedg | pitti, Hey, what is the machine that has the archives on it that we can grep through? I can't remember its name. | 15:24 |
| pitti | tedg: people.canonical.com? that's lillypilly | 15:24 |
| pitti | tedg: it's not about "Label Empty", is it? | 15:24 |
| tedg | pitti, I don't think so... | 15:25 |
| pitti | tedg: ok; I subscribed you to a bug, but it's debugged nwo | 15:25 |
| tedg | Ah, okay. I forgot to start up my mail this morning, so it's still downloading :-) | 15:25 |
| === tkamppeter_ is now known as tkamppeter | ||
| didrocks | rebooting on unity-2d, brb | 15:41 |
| chrisccoulson | g'ah, trying to build the lightning source package is painful in oneiric | 15:42 |
| chrisccoulson | just the source | 15:42 |
| chrisccoulson | it takes nearly 15 minutes with bzip compression | 15:42 |
| chrisccoulson | and i have to do it 4 times :( | 15:42 |
| kenvandine | chrisccoulson, that sounds painful | 15:45 |
| didrocks | chrisccoulson: thunderbird-bin doesn't integrate to the session isn't it? (I always have it running and so, the gnome-session dialog) | 15:45 |
| chrisccoulson | didrocks, i'm not too sure what you're asking | 15:47 |
| didrocks | chrisccoulson: when I quit my session, I almost always have the "thunderbird-bin doesn't respond" dialog | 15:47 |
| didrocks | chrisccoulson: even if I closed the thunderbird window before | 15:47 |
| chrisccoulson | hmmm, i've not seen that before. in any case, there is no integration at all on the default install (for tbird or firefox), because libgnome isn't on the CD | 15:49 |
| chrisccoulson | so you won't see that dialog normally ;) | 15:49 |
| didrocks | chrisccoulson: I get it almost everytime | 15:50 |
| chrisccoulson | does thunderbird actually appear to close? | 15:50 |
| didrocks | chrisccoulson: the -bin is still their, but there is no more window | 15:53 |
| chrisccoulson | ah, ok | 15:54 |
| chrisccoulson | any chance of attaching gdb to the process? | 15:54 |
| chrisccoulson | i've seen a shutdown hang before | 15:54 |
| didrocks | rebooting with new gnome-session | 15:57 |
| didrocks | chrisccoulson: sure, will do that (probably tomorrow) | 15:57 |
| chrisccoulson | thanks | 15:57 |
| Daviey | pitti: sorry, missed your response. Seems that Alt+a is now doing 'similar' behaviour to alt+tab.. showing a preview rather than icons tho. | 16:09 |
| pitti | Daviey: hm, that's supposed to be Alt+` | 16:09 |
| pitti | njpatel, didrocks ^ | 16:09 |
| Daviey | pitti: Hmm, yes - showing the same thing on both shortcuts for me | 16:09 |
| Daviey | a real pain the butt for irssi users :) | 16:10 |
| pitti | ok, that sounds like a bug | 16:10 |
| didrocks | Daviey: do you have unity trunk? | 16:10 |
| Daviey | didrocks: no, oneiric up-to-date | 16:10 |
| pitti | Daviey: at least it's not meant to behave like this, so it's fixable :) | 16:10 |
| Daviey | pitti: hurray! | 16:10 |
| didrocks | Daviey: ok, so, there is a change right now about it | 16:10 |
| Daviey | I am such a noob, i'm not quite sure what to raise it against. | 16:10 |
| didrocks | Daviey: normally, now, it tries to detect the key above the tab | 16:10 |
| pitti | Daviey: raising it against #ubuntu-desktop is not a bad start :) | 16:11 |
| didrocks | Daviey: I just pushed unity, it will be available in a few hours | 16:11 |
| Daviey | lol | 16:11 |
| pitti | didrocks: how is it supposed to detect where a key is located? | 16:11 |
| pitti | didrocks: my ` key is in the lower left corner of the keyboard.. | 16:11 |
| Daviey | webcam + OCR? | 16:11 |
| pitti | above the Tab key is +/= for me | 16:11 |
| didrocks | yeah, webcam is required now :) | 16:11 |
| Daviey | heh | 16:11 |
| didrocks | pitti: some low level xkeyboard API | 16:12 |
| didrocks | to have the ranges, and such | 16:12 |
| didrocks | mutter is doing the same btw | 16:12 |
| pitti | didrocks: that sounds crazy and bug prone | 16:12 |
| Daviey | didrocks: For the impatient, is there a PPA with trunk? | 16:12 |
| pitti | I've heard several people now who have that on totally unexpected keys | 16:12 |
| didrocks | pitti: there is a shortcut anyway, it seems the keycode is 45 | 16:12 |
| didrocks | Daviey: no, there isn't | 16:12 |
| didrocks | pitti: for gnome-shell | 16:12 |
| didrocks | or unity? | 16:12 |
| pitti | unity, I suppose | 16:12 |
| didrocks | pitti: it was an hardcoded key before | 16:13 |
| pitti | didrocks: that sounds better | 16:13 |
| didrocks | worked only on US config | 16:13 |
| pitti | didrocks: how on earth is software supposed to be able to figure out how my keyboard is physically arranged? | 16:13 |
| didrocks | pitti: is it better to get the current state which is Alt + Altgr + è on a french keyboard, for instance? | 16:14 |
| pitti | Can't it be Alt+Tab and Ctrl+Tab or so? | 16:14 |
| didrocks | pitti: mutter is doing the same for a release already, I just saw that the unity guys went the same path (avoid random keys) | 16:14 |
| pitti | hardcoded is the opposite of random? | 16:14 |
| didrocks | and as they ignored modifier, it resulted in Daviey's bug, where A is mapped | 16:14 |
| didrocks | pitti: random in term of layout on the keyboard? | 16:15 |
| didrocks | pitti: Alt + Tab is already used, isn't it? | 16:15 |
| pitti | didrocks: Alt+Tab is hardcoded as well, isn't it? | 16:15 |
| didrocks | pitti: yeah, but on all keyboards, it's approximately at the same place | 16:16 |
| didrocks | anyway, it's not mentionned anywhere, not in the changelog, in any bug | 16:16 |
| didrocks | I just saw the change because I'm looking at unity trunk | 16:16 |
| pitti | didrocks: so using ctrl+tab for the "other" switcher would be a lot more predictable? | 16:16 |
| didrocks | pitti: that's a question for DBO and design I guess | 16:17 |
| pitti | how do you want to document/explain to people how to find the other switcher? | 16:17 |
| pitti | "Alt+some key on your keyboard, we can't tell you what it is" :) | 16:17 |
| DBO | you mean "alt + key above tab" | 16:18 |
| DBO | ? | 16:18 |
| didrocks | pitti: again, it's not a question for me :) I'm not even supposed to be aware about it… | 16:18 |
| pitti | DBO: except that it's not :) | 16:18 |
| DBO | pitti, how do you figure? | 16:18 |
| pitti | DBO: the key above tab is =/+ for me, and it doesn't do anythign with compiz | 16:18 |
| didrocks | pitti: it's in the incoming release | 16:19 |
| DBO | pitti, are you running trunk? | 16:19 |
| pitti | DBO: no, current oneiric | 16:19 |
| DBO | yeah you need trunk, the key detection works | 16:19 |
| DBO | I had several people do testing | 16:19 |
| DBO | you can too if you want | 16:19 |
| DBO | there is a test program in unity trunk you can run | 16:19 |
| DBO | it will tell you what the key above tab is on your layout | 16:19 |
| * pitti branches lp:unity | 16:20 | |
| pitti | DBO: so no need to report all the bugs from e. g. Daviey? | 16:20 |
| pitti | we'll test that program from trunk instead? | 16:20 |
| DBO | yes just run the program | 16:20 |
| DBO | test-keyutil | 16:21 |
| DBO | in the tests directory | 16:21 |
| DBO | if it prints out a key that is not your key above tab | 16:21 |
| DBO | it failed | 16:21 |
| DBO | so far, it has not failed | 16:21 |
| pitti | ah, I guess I need to configure/make etc. first? | 16:21 |
| pitti | ah, no configure | 16:21 |
| pitti | didrocks: is there a trick how to only build tests/test-keyutil from a trunk checkout? | 16:22 |
| didrocks | pitti: mkdir build | 16:22 |
| didrocks | cd build | 16:22 |
| didrocks | cmake .. | 16:22 |
| didrocks | then, I guess cd tests && make | 16:22 |
| didrocks | (instead of make in the root dir) | 16:22 |
| desrt | arghghgh | 16:24 |
| * desrt spends the last two days chasing a ghost | 16:24 | |
| DBO | pitti, sorry, didn't mean to leave you hanging | 16:24 |
| didrocks | DBO: did you saw that the keycode can always be 49 btw? | 16:24 |
| didrocks | DBO: that's why I read on some other code ;) | 16:24 |
| DBO | desrt, in the end, it's revealed that you are the ghost | 16:24 |
| didrocks | desrt: ghost window? :-) | 16:25 |
| desrt | DBO: i see kernel bugs? | 16:25 |
| DBO | desrt, I think you do | 16:25 |
| DBO | didrocks, what? | 16:25 |
| desrt | can everyone download and run a test program for me? | 16:25 |
| desrt | http://bugzilla-attachments.gnome.org/attachment.cgi?id=195399 | 16:25 |
| Daviey | DBO: If you need helping testing something, please do let me know :) | 16:26 |
| desrt | it should print out a bunch of dots on your screen. at some point, the dots should stop. | 16:26 |
| DBO | Daviey, what keyboard layout do you use? | 16:27 |
| didrocks | DBO: see the comment on http://git.gnome.org/browse/mutter/commit/?id=4ea00e102b6afe25e2b84f9def2f44da1b4953c6 | 16:27 |
| DBO | desrt, it does this in a window, right? | 16:27 |
| desrt | DBO: it's a console program... | 16:27 |
| desrt | its interaction with you is limited to putchar('.'); fflush(); | 16:28 |
| Daviey | DBO: Oddly, English (US) atm | 16:28 |
| Daviey | DBO: i do have an Xmodmap rebinding # to the button below Esc tho. | 16:28 |
| desrt | depending on your luck, the crash could happen immediately or after a thousand or so iterations | 16:28 |
| desrt | i've had to run it for as much as 5 minutes before | 16:28 |
| DBO | Daviey, do what pitti is doing | 16:29 |
| pitti | still building here | 16:29 |
| pitti | apparently cmake generates broken Makefiles which don't work with parallel building :( | 16:29 |
| Daviey | pitti: Are you building trunk unity for amd64? | 16:29 |
| pitti | Daviey: yes | 16:29 |
| DBO | I do -j4 constantly pitti | 16:29 |
| didrocks | pitti: it works with parallell building here (thanksfully) | 16:30 |
| Daviey | pitti: for the lazy, are you able to scp them somewhere? | 16:30 |
| pitti | it still seems to build the world, even when running make in tests/ | 16:30 |
| pitti | Daviey: sure | 16:30 |
| didrocks | but still takes 30 minutes just for unity | 16:30 |
| Daviey | thanks. | 16:30 |
| DBO | didrocks, I am making it longer by adding more test applications | 16:30 |
| DBO | sorry! | 16:30 |
| didrocks | DBO: what, you do test? where is the snow? :) | 16:30 |
| pitti | Daviey: http://people.canonical.com/~pitti/tmp/test-keyutil-amd64 | 16:31 |
| pitti | Key above Tab is grave | 16:31 |
| pitti | well, what I thought | 16:31 |
| pitti | I suppose that's based on the standard physical layout of a keyboard, not my real keyboard :) | 16:32 |
| pitti | but I guess I'm a bit of a special case here | 16:32 |
| pitti | DBO: & | 16:33 |
| pitti | DBO: and when I switch to German layout, it still says the same | 16:33 |
| DBO | pitti, what? | 16:33 |
| pitti | sorry, that was meant to be a ^ | 16:33 |
| DBO | pitti, I honestly have no idea if that is right for you or not | 16:34 |
| pitti | DBO: anyway, Alt+` was workign for me, so I guess Daviey's output is more interesting | 16:34 |
| pitti | DBO: http://www.kinesis-ergo.com/images/kb_adv-pro_met720x471.jpg is my keyboard | 16:34 |
| DBO | pitti, okay screw that | 16:34 |
| didrocks | pitti: what is the keycode for reference, | 16:35 |
| pitti | but as there is no way to find out about the physical layout of a keyboard, I guess that's the next best thing | 16:35 |
| pitti | DBO: I'm not worried about my case | 16:35 |
| pitti | DBO: I'm worried about Daviey's | 16:35 |
| pitti | DBO: but it seems there is no point in reporting Daviey's bug now, when that got changed in trunk after all, right? | 16:35 |
| pitti | didrocks: for what? | 16:35 |
| didrocks | += | 16:35 |
| DBO | pitti, what layout does your keyboard report itself as? | 16:36 |
| pitti | DBO: US | 16:36 |
| Daviey | pitti: sorry, before i exec this - what is it? :) | 16:36 |
| pitti | Daviey: don't worry, just an alias for rm -r | 16:36 |
| Daviey | oh good | 16:36 |
| pitti | Daviey: no, it just prints out a list to stdout, nothing more | 16:36 |
| didrocks | pitti: I meant, the keycode of the key += (the one above the Tab) | 16:37 |
| DBO | didrocks, his keyboard is lying to the xserver | 16:37 |
| DBO | clearly that wont work | 16:37 |
| * Daviey trusts pitti :) | 16:37 | |
| pitti | KeyPress event, serial 36, synthetic NO, window 0x4200001, | 16:37 |
| pitti | root 0xbe, subw 0x0, time 40499628, (150,16), root:(766,853), | 16:37 |
| pitti | state 0x0, keycode 21 (keysym 0x3d, equal), same_screen YES, | 16:37 |
| didrocks | yeah, it won't work then :) | 16:37 |
| Daviey | pitti: http://pb.daviey.com/0QFc/ | 16:37 |
| pitti | "Key above Tab is numbersign" | 16:37 |
| pitti | Daviey: does that sound true? | 16:38 |
| Daviey | pitti: blame me for that, Xmodmap. | 16:38 |
| DBO | so its true for you Daviey? | 16:38 |
| Daviey | $ cat ~/.Xmodmap | 16:38 |
| Daviey | keycode 49=numbersign | 16:38 |
| Daviey | keycode 94=dead_tilde | 16:38 |
| Daviey | remove Lock = Caps_Lock | 16:38 |
| desrt | DBO: did it crash yet? | 16:39 |
| Daviey | I can retest without that, if it helps test this? | 16:39 |
| didrocks | pitti: and for your question about discoverability, this is another way than the arrow to show the instances FYI | 16:39 |
| DBO | desrt, oh sorry I forgot to run it with everything else | 16:39 |
| desrt | s'ok :) | 16:39 |
| desrt | didrocks: can you test too, maybe? | 16:39 |
| didrocks | desrt: sure | 16:40 |
| DBO | desrt, didn't crash | 16:40 |
| DBO | but | 16:40 |
| DBO | stopped printing dots | 16:40 |
| desrt | DBO: the stopping is the crashing | 16:40 |
| desrt | can you give me your uname -r? | 16:41 |
| DBO | 3.0.0-9-generic | 16:41 |
| desrt | thanks | 16:41 |
| desrt | apt-cache policy for libc, too? | 16:41 |
| didrocks | desrt: same here, same kernel | 16:41 |
| didrocks | 2.13-17ubuntu2 | 16:42 |
| desrt | thanks | 16:42 |
| DBO | desrt, same as didrocks | 16:42 |
| DBO | desrt, it prints a different number of dots each time if that helps | 16:43 |
| pitti | Daviey: would be interesting to confirm, anyway, I guess | 16:43 |
| desrt | DBO: it's a race... it happens by random chance | 16:43 |
| desrt | we're on eglibc these days, right? | 16:43 |
| DBO | pitti, you jelly of my key above tab code? | 16:43 |
| pitti | DBO: "jelly"? | 16:44 |
| DBO | sorry... too much reddit lately | 16:44 |
| DBO | you dont like the key above tab stuff? | 16:44 |
| pitti | DBO: my original question was debugging Daviey's issue of Alt+A, which is certainly unexpected | 16:44 |
| pitti | and then we found that trunk uses a different approach than what's in oneiric right now | 16:45 |
| pitti | so I guess the original bug is moot now | 16:45 |
| DBO | more or less | 16:45 |
| pitti | I still think Ctrl+Tab might be a little easier to document/explain/be more robust, but as long as it doesn't grab letters any more, I'm fin e:) | 16:45 |
| DBO | pitti, ctrl+tab is used by other applications | 16:46 |
| DBO | and you cant switch from alt-tab to ctrl-tab without the switcher going away | 16:46 |
| Daviey | pitti: http://pb.daviey.com/1Cbb/ | 17:05 |
| Daviey | unmodified layout. | 17:05 |
| Daviey | (incidently, without my modified layout alt+a works again) | 17:09 |
| desrt | DBO, didrocks; can you try again with -static? | 17:18 |
| DBO | desrt, sure | 17:18 |
| desrt | annoyingly, static linking appears to solve the problem for me | 17:19 |
| DBO | same | 17:19 |
| desrt | which prevents me from checking old libc versions with new kernels | 17:19 |
| desrt | puts more weight on chances of it being a libc bug, though | 17:19 |
| desrt | time for a libc bug, methinks | 17:22 |
| kenvandine | pitti, are you still around? | 18:13 |
| pitti | kenvandine: again, yes (was at dinner) | 18:25 |
| kenvandine | pitti, nm... unping :) | 18:25 |
| pitti | heh | 18:26 |
| dbarth | Amaranth: see https://bugs.launchpad.net/nux/+bug/838693 | 18:28 |
| ubot2` | Ubuntu bug 838693 in compiz "Add support for GL-ES" [Undecided,New] | 18:28 |
| Amaranth | dbarth: so the compiz changes are 59 files changed, 2330 insertions(+), 2934 deletions(-) | 18:29 |
| Amaranth | and that's just compiz, not even including compiz-plugins-main | 18:29 |
| Amaranth | yay for removing more code than added, I guess :) | 18:30 |
| dbarth | Amaranth: woops | 18:30 |
| dbarth | oh right, smspillaz mentioned the plugin rewrite you had to do | 18:30 |
| Amaranth | haven't rewritten any plugins yet, just ported them all | 18:30 |
| dbarth | well, yeah, in that case it's a significant distro-patch | 18:31 |
| * didrocks waves goodnight | 18:38 | |
| DBO | its 3PM and all is quiet in the desktop team | 19:00 |
| pitti | it's 9 pm, and my wife gets angry :) | 19:00 |
| pitti | FTR, pygobject 2.90 uploaded \o/ | 19:00 |
| === zyga is now known as zyga-afk | ||
| pitti | good night everyone! | 19:31 |
| bryceh | pitti, night | 19:33 |
| cyphermox | night pitti | 19:33 |
| === eeejay_is_afk is now known as eeejay | ||
| === eeejay is now known as eeejay_is_afk | ||
| === eeejay_is_afk is now known as eeejay | ||
| dupondje | Seems like network-manager-openvpn is broken | 22:33 |
| dupondje | it does not change default route to the vpn tunnel, even if wanted | 22:33 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
