brainwash | ali1234: what's broken in xfdesktop? | 00:12 |
---|---|---|
ali1234 | what *isn't* broken? :) | 00:12 |
ali1234 | the whole multiple desktop and workspace thing... doesn't work | 00:12 |
ali1234 | and when you relog it resets the wallpaper to default | 00:12 |
brainwash | only tested it once, set a different wallpaper for every workspace and it worked and was consistent | 00:14 |
=== xnox_ is now known as xnox | ||
ali1234 | yes, that's the *only* case where it works properly :) | 00:14 |
ali1234 | where you set a unique wallpaper for every monitor and workspace, without any spanning | 00:14 |
brainwash | oh | 00:15 |
ali1234 | if you have two monitors and you set a wallpaper to spanning it doesn't work properly, and if you set "same on all workspaces" it doesn't work properly | 00:15 |
ali1234 | when you start it up it tries to get the wallpaper for each workspace and monitor | 00:16 |
ali1234 | if that fails it looks for legacy settings | 00:16 |
ali1234 | the new version uses the monitor plug name, legacy uses the number | 00:17 |
ali1234 | that means that defaults have to be set using legacy, because we can't know the plug name in advance, it's something like "DVI-I-0" | 00:17 |
ali1234 | but that means that there is always a legacy setting | 00:17 |
brainwash | is the monitor naming consistent or does it change? | 00:17 |
ali1234 | it's consistent on a single machine if you don't even unplug anything | 00:17 |
ali1234 | so anyway, there are legacy settings in the defaults.xml, which means they can't be deleted | 00:18 |
ali1234 | that means when you log in and you didn't specify a wallpaper for some monitor or workspace, it tries to load the legacy setting and that resets your whole system to defaults | 00:19 |
ali1234 | also sometimes it loads up the default wallpaper and then quickly changes to what you asked for | 00:19 |
brainwash | wow, so much fun with a multi monitor setup | 00:19 |
brainwash | + xfdesktop :) | 00:19 |
ali1234 | yes | 00:19 |
ali1234 | and further more i can't even set the wallpaper on all my monitors because some of them overlap | 00:20 |
ali1234 | setting a wallpaper on those monitors wouldn't even make sense | 00:20 |
brainwash | so xfdesktop and lightdm-gtk-greeter really need some multi monitor rework | 00:20 |
ali1234 | looks like it | 00:20 |
ali1234 | i guess you've seen what gtk-greeter does when monitors overlap? | 00:20 |
brainwash | no, only running single monitor setups here | 00:21 |
ali1234 | oh... well it draws the wallpaper multiple times | 00:21 |
ali1234 | it's not pretty | 00:21 |
brainwash | did you also test unity-greeter? | 00:22 |
ali1234 | no | 00:22 |
brainwash | maybe we can do some copy&paste | 00:22 |
ali1234 | maybe | 00:22 |
ali1234 | probably not though | 00:22 |
ali1234 | hang on let me try it... | 00:23 |
brainwash | at least to get an idea on how to improve it (in case unity-greeter handles it better) | 00:24 |
ali1234 | heh | 00:24 |
ali1234 | brb | 00:24 |
brainwash | xfce4-session-4.11.0 has been released | 00:28 |
Noskcaj | How much of the new 4.11 stuff do we want for 14.04? | 00:32 |
Noskcaj | Everything but session is now packaged for debian | 00:33 |
brainwash | should we go full 4.11? :) | 00:34 |
Noskcaj | If the release team allows it, i want to | 00:37 |
Noskcaj | minimum is settings bugfix release, ui bugfix release, garcon | 00:37 |
brainwash | I see lots of bugfixes in the various changelogs | 00:38 |
brainwash | maybe it's worth to ship the full 4.11 setup | 00:40 |
ali1234 | how do i change the lightdm greeter? | 00:45 |
ali1234 | ochosi: accounts service background works in test mode but not in the real login window | 00:49 |
ali1234 | ok so unity-greeter is even more hilariously broken | 01:19 |
ali1234 | it doesn't honour the monitor layout at all | 01:19 |
ali1234 | so by default all the monitors do not overlap (it just uses whatever the xorg default is) | 01:19 |
ali1234 | that means you have to reconfigure your monitor layout after every login | 01:19 |
ali1234 | i suspect if it wasn't horribly broken it would show the same thing as gtk-greeter does, which is this: http://imagebin.org/295107 | 01:27 |
brainwash | ali1234: this would drive me nuts | 01:31 |
ali1234 | well it's better than what unity-greeter does | 01:31 |
ali1234 | which is this: https://bugs.launchpad.net/ubuntu/+source/unity-greeter/+bug/1283615/+attachment/3992319/+files/IMG_20140223_005551.jpg | 01:31 |
ubottu | Launchpad bug 1283615 in unity-greeter (Ubuntu) "unity-greeter ignores user monitor configuration" [Undecided,New] | 01:31 |
ali1234 | at least it is *functional* | 01:32 |
brainwash | mmh, makes me wonder.. was multi monitor support always that bad? | 01:33 |
brainwash | or do we see some sort of regression here | 01:33 |
brainwash | maybe hardware/driver related? | 01:34 |
ali1234 | unity greeter used to work properly with xorg.conf | 01:34 |
ali1234 | it isn't driver related | 01:34 |
ali1234 | with unity-greeter you get the default layout as if you had no xorg.conf | 01:34 |
ali1234 | they've obviously decided to ignore it | 01:34 |
ali1234 | which is pants-on-head retarded, so about right for unity | 01:34 |
ali1234 | but yeah, multimonitor support has always been terrible. nvidia is the only way to make it work ina reasonable way | 01:35 |
ali1234 | and they did it by rewriting half the xorg stack | 01:35 |
ali1234 | that bug does not affect unity-webkit-greeter btw | 01:36 |
ali1234 | and probably not lightdm-kde-greeter either | 01:36 |
brainwash | not at all? | 01:36 |
ali1234 | the webkit one does something very similar to what the gtk one does with multiple backgrounds | 01:37 |
ali1234 | but it honours the monitor layout like it should | 01:37 |
brainwash | maybe I should extend my laptop screen with another monitor and do some testing | 01:42 |
ali1234 | that would be cool | 01:45 |
=== eric_the_idiot is now known as Guest2522 | ||
=== eric_the_idiot_ is now known as Guest46634 | ||
=== yofel is now known as Guest77623 | ||
Unit193 | Oh right, how did the CC catch up go? | 06:44 |
=== jackson is now known as Guest16889 | ||
=== Guest16889 is now known as Noskcaj | ||
forestpiskie | it didn't | 07:58 |
=== OutOfControl is now known as benonsoftware | ||
elfy | lderan: you managed to get anywhere writing testcase for your app? | 09:12 |
=== Guest77623 is now known as yofel | ||
=== yofel_ is now known as Guest39928 | ||
lderan | elfy, is this on the right tracks http://paste.ubuntu.com/6981366/ ? | 12:10 |
elfy | lderan: yep that looks fine to me - get it in to bzr push lp:~YOURUSERID/ubuntu-manual-tests/YOURBRANCHNAME | 12:14 |
elfy | then I'll check/approve and sync it | 12:14 |
lderan | \o/ will do so now | 12:15 |
elfy | ok :) | 12:15 |
lderan | and i think i've just found a bug with thunar | 12:17 |
elfy | I bet it's a known issue :p | 12:19 |
lderan | bet so | 12:20 |
lderan | if you have 2 rename dialog boxes open the first stops working fully :P | 12:21 |
elfy | I can't even open 2 rename dialogues from the same instance of thunar - works fine with another thunar opened | 12:23 |
lderan | it partially works for me, the apply button on the first one opens doesn't close the box, but the apply button on the 2nd closes both | 12:24 |
elfy | can't confirm that as I can't even get that far :) | 12:24 |
lderan | :P | 12:24 |
=== Guest39928 is now known as yofel | ||
elfy | lderan: when you've done that testcase and I've synced the tracker - I'll do a call for testing - but I don't remember the PPA | 12:26 |
lderan | shall try to find it for you | 12:26 |
elfy | thanks | 12:26 |
elfy | biab | 12:26 |
lderan | elfy: https://launchpad.net/~light-locker-settings-team/+archive/stable , trust at the moment | 12:29 |
elfy | lderan: in future when you do a mr for testcases - don't number them :) | 12:41 |
elfy | someone might be syncing one and the number will be wrong ;) | 12:41 |
lderan | oh sorry | 12:41 |
elfy | lderan: you sure you've finished it? the first 2 tests look decidedly odd and the <dd><dd> won't work :) | 12:44 |
elfy | off for a few hours now | 12:45 |
lderan | shall check it :) | 12:45 |
elfy | can you check spelling too please :) | 12:46 |
elfy | cya later | 12:46 |
lderan | cheerio | 12:46 |
elfy | minuter - must have then copy pasted the same word elsewhere lol | 12:46 |
lderan | gah apologies | 12:47 |
elfy | ochosi: will get this testcase synced and then called for later today | 12:48 |
ochosi | elfy: which one? | 12:48 |
elfy | light-locker-settings | 12:48 |
ochosi | oh great | 12:49 |
ochosi | ty | 12:49 |
lderan | elfy, have redone it :P hopefully all okay now | 13:03 |
ochosi | elfy: shouldn't it be "sudo apt-get update" instead of "upgrade"? | 13:31 |
ochosi | (re: ML test-case announcement) | 13:34 |
jjfrv8 | ochosi, you still here? | 14:39 |
elfy | lderan: thanks - all done | 15:02 |
elfy | ochosi: yep ... | 15:02 |
jjfrv8 | ochosi, I can't stick around but wanted to ask something about light-locker... | 15:08 |
jjfrv8 | after testing the light-locker-settings ppa, I tried the light-locker testcase... | 15:08 |
jjfrv8 | when I click Suspend and go to resume, it brings up the old, flaming xScreensaver dialog. That's not right, is it? | 15:09 |
elfy | jjfrv8: purge xscreensaver | 15:09 |
jjfrv8 | alright, let me try that. | 15:10 |
jjfrv8 | elfy, that's better. Thanks. Can't keep all the stuff I read in the logs in my head :( | 15:11 |
elfy | I did have a note in the testcase to make sure that was done - was sure that the seed had been fixed to not include xscreensaver | 15:12 |
jjfrv8 | apparently not, this was on a fresh install from today's daily | 15:13 |
elfy | mmm | 15:13 |
Unit193 | Fix Committed, not Fix Released. | 15:16 |
elfy | thanks Unit193 | 15:16 |
jjfrv8 | ochosi, also wanted to make sure you had seen that xfesktop Prefs is ready for review. | 15:17 |
jjfrv8 | bbl | 15:17 |
ochosi | thanks jjfrv8 | 18:51 |
ochosi | will review tomorrow | 18:51 |
knome | http://open.knome.fi/2014/02/23/logging-in-with-ubuntu-one/ | 18:58 |
knome | ochosi, you feeling better today? | 18:59 |
Noskcaj | Anyone know where micahg went? He's usually online in some form | 19:01 |
* knome wonders how the liquid micah types | 19:03 | |
knome | or the gas-form micah | 19:03 |
elfy | very very carefully in both cases | 19:06 |
Noskcaj | :) | 19:07 |
knome | compressed air then i guess | 19:08 |
brainwash_ | Noskcaj: bug 1282509 has been fixed upstream, what is the next step now? normal merge request in ubuntu, debdiff or fix it in debian and sync? | 19:09 |
ubottu | bug 1282509 in xfdesktop4 (Ubuntu) "xfdesktop crashed with SIGSEGV in xfce_desktop_refresh()" [Medium,Triaged] https://launchpad.net/bugs/1282509 | 19:09 |
Noskcaj | I'll ask corsac. The sponsor queue is pretty big currently | 19:10 |
brainwash_ | ok, I'm a bit lost, not sure what the best move would be | 19:11 |
ali1234 | why did thunar get SRU'd in 13.10? | 19:13 |
Noskcaj | it did? | 19:13 |
ali1234 | the update is crashing all over the place | 19:14 |
ali1234 | apparently | 19:14 |
ali1234 | Noskcaj: can you see error reports on errors.ubuntu.com? | 19:15 |
ali1234 | https://errors.ubuntu.com/problem/cb0629c4765654432162e73197d145eea640134a | 19:15 |
Noskcaj | ali1234, No i can't | 19:16 |
Noskcaj | although it doesn't look SRUed | 19:16 |
ali1234 | hmm... maybe i'm reading it wrong | 19:16 |
ali1234 | https://launchpad.net/ubuntu/saucy/i386/thunar | 19:17 |
ali1234 | ah, uploaded two months before release | 19:17 |
ali1234 | anyway, there's 17000 reports of that problem ^ | 19:17 |
ali1234 | it's the old magazing_chain_pop_head nonsense again, crashing deep inside glib | 19:18 |
Noskcaj | I thought that was fixed in saucy | 19:23 |
ali1234 | still crashing in trusty | 19:24 |
ali1234 | it's on the top 100 most common bugs | 19:24 |
Noskcaj | wow. Is there an upstream bugfix? | 19:24 |
brainwash_ | I did not notice any thunar crashes in trusty so far | 19:25 |
ali1234 | there is a bug report on launchpad but i can't see that | 19:25 |
ali1234 | bug 1203296 | 19:25 |
ubottu | Error: Launchpad bug 1203296 could not be found | 19:25 |
ali1234 | 65th most common bug in ubuntu in the past day | 19:26 |
ali1234 | the xfdesktop bug was 48th | 19:27 |
brainwash_ | all marked as private? | 19:28 |
Noskcaj | public now | 19:30 |
Noskcaj | bug 1203296 | 19:30 |
ubottu | bug 1203296 in thunar (Ubuntu) "thunar crashed with SIGSEGV in magazine_chain_pop_head()" [Medium,Confirmed] https://launchpad.net/bugs/1203296 | 19:30 |
Noskcaj | the xfdesktop fix is done in debian. | 19:32 |
brainwash_ | that's great, so it will be ready for ubuntu B1, right :) | 19:39 |
ali1234 | hmm... so of the top 10 most frequent crashers in xubuntu, thunar is number 2, 3, 4, 5, 6, 7, 8, and 10 | 19:40 |
Noskcaj | brainwash_, Provided corsac uploads it | 19:40 |
ali1234 | and it looks like most of that is magazine_chain_pop_head - which means memory corruption | 19:40 |
Noskcaj | ali1234, wow | 19:41 |
brainwash_ | in trusty? | 19:41 |
ali1234 | in all versions, past 24 hours | 19:41 |
brainwash_ | how is that even possible?! =S | 19:42 |
brainwash_ | ah, I recall reading about this | 19:45 |
brainwash_ | something related to "Pane Side -> Tree is enabled." | 19:45 |
brainwash_ | the tree view | 19:45 |
ali1234 | it does look like the bug was introduced between 1.6.2. and 1.6.3 | 19:49 |
ali1234 | or maybe not | 19:50 |
ali1234 | there are like 3 reports of it with 1.6.2 and 20000 with 1.6.3 | 19:51 |
ali1234 | hmm i see some changes to g_list code... that's the same thing that crashed xfce-terminal | 19:55 |
ali1234 | argh... yeah, this is it | 19:55 |
ali1234 | http://git.xfce.org/xfce/thunar/commit/?id=d5b6768a104786b44542522768d6263cff07a886 | 19:55 |
ali1234 | bad commit... | 19:56 |
Noskcaj | it was terminal that we patched to stop magazine_pop_chain, maybe the patch can be modified to thunar | 19:56 |
ali1234 | magazine_chain_pop_head crashes are very generic | 19:57 |
ali1234 | it just means the program did a double free or something | 19:57 |
ali1234 | that commit is suspicious but it might not be it | 19:57 |
ali1234 | any g_list code would be a place to look though | 19:57 |
ali1234 | especially when using prepend | 19:57 |
Noskcaj | ok | 19:57 |
ali1234 | basically it means memory was corrupted... it could have happened anywhere in the program, backtraces are not useful :( | 19:58 |
ali1234 | some more g_list stuff here: http://git.xfce.org/xfce/thunar/commit/?id=58fa477fb20aa2ac8eaea8d81d7e4d8c00180600 | 20:01 |
ali1234 | of course it could even be in a plugin... | 20:03 |
brainwash_ | how do you trigger the memory corruption? with the tree view? | 20:03 |
Noskcaj | Do we want dev-tool or libxfce4util 4.11? both just got uploaded | 20:04 |
Noskcaj | util is probably too big a package to do this late, but dev-tools is probably safe to sync | 20:05 |
ali1234 | brainwash_: i don't know | 20:05 |
ali1234 | brainwash_: opening the tree view takes a suspiciously long time | 20:06 |
brainwash_ | ali1234: not here, 2sec | 20:09 |
ali1234 | 2 sec is too long | 20:09 |
brainwash_ | not for a very slow HDD | 20:09 |
brainwash_ | but still.. nothing crashes | 20:09 |
ali1234 | Noskcaj: can you see this page? https://errors.ubuntu.com/?user=a-j-buxton&period=day | 20:24 |
ali1234 | (or anyone else for that matter) | 20:25 |
Noskcaj | yep. Turns out i wasn't logged in | 20:27 |
ali1234 | that page should be visible to anyone i think | 20:27 |
Unit193 | It is. | 20:29 |
ali1234 | Noskcaj: watch out when making bug reports public, that thunar one has a torrent download filename in it | 21:12 |
Noskcaj | My one? | 21:13 |
ali1234 | 1203296 | 21:13 |
Noskcaj | Oops, must have missed that | 21:14 |
ali1234 | that guy likes his torrents, lots on the dupes too | 21:14 |
Noskcaj | Is it private info though? | 21:15 |
ali1234 | i dunno | 21:15 |
brainwash_ | not the usual linux distro torrents? :D | 21:15 |
ali1234 | it's evidence of illegal activity though | 21:15 |
bluesabre | quick, append "-parody" to each of them | 21:16 |
ali1234 | lots of the dupes were made public by apport | 21:16 |
ali1234 | this is on ProcCmdLine: field btw | 21:16 |
ali1234 | hmm | 21:17 |
ali1234 | "This incident occurred when closing Thunar after moving a file from my PC hard drive to my external drive. The cut/paste operation was complete, I clicked the "X" to close Thunar and got the "Ubuntu experienced an error" message." | 21:18 |
ali1234 | "Thunar crashed after pressing the close button" | 21:18 |
ali1234 | "I closed thunar, this error happened" | 21:19 |
ali1234 | well, that's a pattern if i ever saw one | 21:19 |
Noskcaj | thunar has a heap of different errors when it closes. I just made a bug for another one | 21:19 |
ali1234 | "I closed Thunar, and got this error. Pane Side -> Tree is enabled." | 21:19 |
ali1234 | it has all different errors because memory is getting randomly corrupted somewhere | 21:20 |
ali1234 | that can cause many different symptoms, but the root cause is almost certainly the same | 21:20 |
ali1234 | if you can find a way to reliably reproduce it... | 21:20 |
ali1234 | something is overwriting random bits of memory | 21:21 |
ali1234 | probably an uninitialized pointer somewhere | 21:21 |
=== slickyma4ter is now known as slickymaster | ||
ali1234 | it's particularly telling that all the bugs have roughly the same frequency except for the top one or two | 22:48 |
ali1234 | well, most happened only once | 22:49 |
brainwash_ | I'm not affected, because I don't move big torrent downloads from A to B :) | 23:01 |
brainwash_ | but thunar needs some cosmetic fixes, 1) use correct icon for "open with" menu entry 2) sort "open with..." sub menu 3) use theme color for selection rectangle | 23:04 |
brainwash_ | I'll put it on my todo list | 23:04 |
ali1234 | which versions of thunar have shipped in xubuntu releases? | 23:06 |
ali1234 | !info thunar precise | 23:07 |
ubottu | thunar (source: thunar): File Manager for Xfce. In component universe, is optional. Version 1.2.3-3ubuntu2 (precise), package size 311 kB, installed size 898 kB | 23:07 |
ali1234 | !info thunar quantal | 23:07 |
ubottu | thunar (source: thunar): File Manager for Xfce. In component universe, is optional. Version 1.4.0-1ubuntu2.1 (quantal), package size 315 kB, installed size 909 kB | 23:07 |
ali1234 | !thunar raring | 23:07 |
ali1234 | !info thunar raring | 23:07 |
ubottu | thunar (source: thunar): File Manager for Xfce. In component universe, is optional. Version 1.6.2-0ubuntu1 (raring), package size 320 kB, installed size 899 kB | 23:07 |
ali1234 | hmmmmmmmmmm | 23:07 |
brainwash_ | ?! | 23:08 |
ali1234 | this bug has never been reported with raring as far as i can tell | 23:08 |
ali1234 | so it may be a bug in some library | 23:08 |
brainwash_ | glib? | 23:08 |
ali1234 | no | 23:08 |
ali1234 | some xfce library that only thunar uses | 23:08 |
ali1234 | !info thunar saucy | 23:08 |
ubottu | thunar (source: thunar): File Manager for Xfce. In component universe, is optional. Version 1.6.3-1ubuntu1 (saucy), package size 314 kB, installed size 904 kB | 23:09 |
ali1234 | 1.6.2 showed the bug in saucy but not raring, and that was during development | 23:09 |
ali1234 | although actually, it *could* be due to glib changes... same thing happened with xfce-terminal | 23:10 |
brainwash_ | shouldn't there be reports outside of the ubuntu universe too? | 23:10 |
ali1234 | yes, but frankly... where? | 23:33 |
Noskcaj | arch, debian, suse, mint? | 23:33 |
ali1234 | the only reason i'm seeing these reports is because of automatic error reporting | 23:34 |
ali1234 | i think people tend not to report one-off unreproducable crashes | 23:34 |
ali1234 | mainly because triagers just close the bugs if they can't be reproduced | 23:34 |
ali1234 | i think the only option here is to go through the code and audit it by hand :( | 23:35 |
ali1234 | cppcheck didn't find anything obvious, valgrind isn't showing any corruption... i don't know what else to try | 23:36 |
ali1234 | i'm actually impressed we even got on the top 100 bugs... must be a lot of users out there | 23:37 |
ali1234 | especially since it seems pretty hard to reproduce | 23:38 |
knome | humm, i'm seeing a debian wallpaper on the installer... | 23:52 |
knome | :) | 23:52 |
Noskcaj | ali1234, yay? | 23:55 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!