ControlsfreekI'm looking at Bug #1099815, which seems to be a bitesize bug. I'm completely new to unity/nux. Trying to figure out the lifecycle of LauncherOptions. Is there documentation on how the options get populated (aside from their default values in the constructor)?02:32
ubot5bug 1099815 in unity (Ubuntu) "Launcher starts with default size then jumps to configured one on login" [Undecided,New] https://launchpad.net/bugs/109981502:32
ControlsfreekWas expecting to see a function that loads values from a config file of some sort...02:33
ControlsfreekA nudge in the right direction would be much appreciated! Thanks02:34
jrrhow might I make the top menu bar only display on my primary screen?02:54
popeyseb128: seems unity in ubuntu-desktop/daily-build is broken today?09:37
popeyi get no launcher or panel09:37
popey(morning btw) ☺09:37
popeysegfault in libunityshell.so09:38
popey[   96.792983] compiz[2973]: segfault at 70 ip 00007f86ac6c291a sp 00007fff37a906a0 error 4 in libunityshell.so[7f86ac4ea000+409000]09:38
davidcallepopey, I can confirm09:39
popeythanks davidcalle09:39
duflupopey: CLI; way of the future [tm]09:39
* popey gets his VT101 out09:39
seb128popey, hey, I'm not running the daily ppa but seems you got people to confirm09:40
* davidcalle was running trunk via unity-team/staging when it happened yesterday afternoon09:42
popeydavidcalle: do we have a bug filed for it?09:42
davidcallepopey, I wasn't sure if it was a driver or Unity issue, since the move to SNA was done at the same time. No bug yet AFAIK.09:43
popeyI'm on nVidia09:44
davidcalleIntel for me09:44
sil2100So maybe the latest commit?09:45
popeysil2100: can you take a look pls?09:46
sil2100popey: will do09:46
sil2100davidcalle: you're using staging, right?09:48
davidcallesil2100, not anymore, but I was indeed09:48
* popey purges the ppa until fixed09:49
sil2100popey: what version did you use that was segfaulting?09:54
sil2100Was it 6.12.0daily13.01.21-0ubuntu1 for unity?09:54
popeyone mo09:57
popeylooks like from http://paste.ubuntu.com/1558044 that I installed 13.01.22 atr ~09:30 today09:58
sil2100Oh, ok, so what versions did you use that were crashing? from daily0build?09:58
popeyoh, no, hang on09:58
popeyyes, 22 at 9:3009:59
popeyif I'm reading that grep of /var/log/dpkg right09:59
popeyso i updated at 9:30 (my time in the logs) and saw the issue at 09:37, yes looks like 13.01.2209:59
sil2100Ok, so hm, there's no compiz update this day10:01
sil2100Thanks, looking further10:01
seb128sil2100, the segfault is in unity10:01
seb128so doesn't need to be a compiz update10:01
sil2100popey: is it segfaulting every time? Since the difference between 21 and 22 is just one build-related commit10:02
sil2100popey: did you try 21 yesterday without any problems?10:03
popeyyes, 21 was fine10:03
sil2100Actually, nux might also be involved10:03
popeyi had 21 installed and running when 22 was installing10:03
popeyran all day with 21 running fine yesterday10:03
sil2100If there are some new commits of course...10:03
sil2100Ok, testing something10:05
sil2100popey: ok, so, try updating nux as well10:10
popeyto what?10:10
sil2100popey: it seems that there was probably some inside ABI break in the last big commit and unity deps didn't get updated10:10
sil2100popey: to the one in daily10:10
sil2100popey: since from your logs I didn't see nux getting updated10:10
popeyi specifically grepped for unity10:11
sil2100So, nux was 22 as well?10:11
sil2100Since I experienced a similar segfault when having upgraded unity and compiz without nux10:12
popeyone mo..10:12
sil2100But this might be a different case10:12
popeysil2100: http://paste.ubuntu.com/155812410:12
popeythats the full log from 20th to now10:13
sil2100Ok, so it was there, interesting10:14
sil2100popey: is it possible for you to use latest unity but revert nux to the previous daily? (18)10:18
rperierwhat is the module used to ask the Launcher what is the running app ? BamfApplicationManager ?10:18
rperierhey btw10:18
popeylets see10:18
popeysil2100: http://paste.ubuntu.com/1558193  - that works10:24
sil2100popey: ok, so as I thought, nux is the problem - and the reason why I don't get the crash is probably that I'm using staging right now10:25
sil2100popey: and it seems that there might have been a difference in what was built before what10:26
sil2100hah, indeed!10:27
sil2100libnux-4.0-common amd64 4.0.0daily13.01.18-0ubuntu110:28
sil2100So, the problem is:10:28
sil2100unity in the daily PPA has been build with nux 4.0.0daily13.01.18-0ubuntu1, which is the old one, after which the 22nd nux has been built - and in nux there seems to be some ABI break that requires unity to be built with the latest nux to work10:29
sil2100So, because unity was built using old nux, and the daily PPA has the new nux in it, which is not compatible with unity built with the old nux, there is a crash10:29
sil2100In staging there is no such problem because nux has been built there 18 hours ago and unity 5 hours ago, so unity was built using latest nux10:30
sil2100How to fix that? Nux probably needs an ABI bump first of all10:30
MCR1Hi :) I can confirm staging is still working on raring and the Dash flies with the new blur 8-)10:31
MCR1Quality of the blur has somewhat regressed, but functionality and speed are much more important for the Dash10:32
MCR1it was quite unusable with free gallium-radeon driver at least...10:33
sil2100popey: now, I don't really have a clue how to fix the daily PPA...10:39
sil2100popey: since the broken packages are already there, not sure if I can push anything to that place10:40
seb128sil2100, you should be able to land an unity rebuild no?10:40
sil2100seb128: I don't think I have any permissions for that PPA10:41
seb128sil2100, should it autoland rebuilds if there are new commits?10:44
seb128where is didrocks today?10:44
sil2100seb128: not entirely sure how daily works, but I think it's only rebuilding per-day10:44
sil2100seb128: so, since the last build was 6 hours ago, I think it will rebuild tomorrow at soonest10:45
sil2100seb128: not sure, I was hoping he'll be around ;p10:45
seb128he might still be jetlagged10:45
MCR1speaking of the devil ;)10:54
sil2100didrocks: my man!10:55
didrockshey sil210010:55
sil2100didrocks: how are you?10:55
didrocksjetlag apparently10:56
didrocksawaken from 1am to 4am10:56
sil2100For everyone interested: https://code.launchpad.net/~sil2100/nux/bump_abi/+merge/14427010:57
MCR1didrocks: Hi :) Do you know why Unity is recommending the package "hud", there is no such package anywhere ? Is this some kind of joke ?10:58
didrocksMCR1: this package will come soon10:58
didrockswhy would we put joke in the packaging?10:59
MCR1ah, ok10:59
MCR1I do not know - I was just wondering10:59
sil2100It's an easter egg10:59
sil2100We're waiting for easter to release it10:59
sil2100It hatches then10:59
MCR1It sounds good: Unity recommends using the hud ;)10:59
rperierthe bug 692444 is still valid and need to be fixed ? I see  "confirmed->Fix committed" and "fix committed -> confirmed"  a lot of times11:00
ubot5bug 692444 in nautilus (Ubuntu) "Launcher - clicking trash or device icons multiple times opens multiple instances" [Undecided,In progress] https://launchpad.net/bugs/69244411:00
sil2100rperier: let me check, but from what I remember it's still an issue11:00
MCR1yes it is11:00
MCR1rperier: Still valid11:00
luvI am planning to alter my unity to show list of windows when i right click an icon in the launcher - that is i can switch to the window I want based on it's windows title without going through the spread11:04
luvsaves time and all11:04
rperierok, as the trash is "hardcoded" into launcher/TrashLauncherIcon.cpp and executes "xdg-open trash://" directly,  we could ask bamf to focus the window with the title "trash" (or ask the launcher to focus the window), what do you thing ? There is a module to do that easily in unity ? or I need to use bamf directly ?11:04
luvanyone would be interested in the patch or should i not bother and just keep it to myself?11:04
sil2100rperier: hm, I think the best person to ask is Trevinho ^11:05
rperierTrevinho: ping11:05
luvor have a look at BamfLauncherIcon.cpp how it is done there11:05
sil2100luv: I would be interested in such a patch - what I would recommend is to send this to the design team for comment11:05
sil2100luv: for instance, you could file in a bug that's targetting unity as Wishlist and ayatana-design, attach the branch there and wait for comment from the design11:06
luvsil2100: well, i'm going to get it done anyone for myself so i dont really need design team's comments i know what i want ;-)11:06
rperierlunch time, bbl11:06
luvi will put the bug on my github and see you guys link when im done and then you can take it from there11:07
sil2100luv: it's not about modifying your approach, but getting to know if the unity design team would find it useful11:07
luvsil2100: right, i will do that when im done11:07
luva bit of sarcasm here ... do i have to be subtle about the idea and make it seem they actually came up with it to make them accept it ;-)11:08
luvanyway, im not comfortable just filing a bug (and wishlist for that matter) because in my experience it always leads nowhere (no hard feelings, that's just how it is ... and I have been told bunch of times to use ML or IRC)11:11
=== _salem is now known as salem_
=== mmrazik is now known as mmrazik|lunch
=== mmrazik|lunch is now known as mmrazik
rperierbefore trying to solve the above bug, I am trying to install unity-trunk from 12.10 in $HOME/staging (so I need bamf 0.4) . bamf from trunk tries to install Bamf-3.gir into /usr/share/gir-1.0 and does not respect the $PREFIX variable12:14
Trevinhorperier: I've already done some nautilus work to get that working properly12:25
rperierThe problem is the trash launcher does not call nautilus directly but xdg-open12:26
rperierso as long as xdg-open is used from the trash launcher, the solution should work for all files managers compatibles with xdg-open and should not be specific to nautilus . Except if you plan to call "nautilus trash://" directly12:30
rperierhowever, I am still a beginner, I could be wrong12:30
Trevinhorperier: well, caliing that is ok to me... Relying on windows name is not a proper solution, I'd prefer to fix nautilus since it's what we use12:34
rperierok, so that's probably better for me to find another bug in this case, no ? (if the fix is in nautilus and not in unity...)12:36
seb128Trevinho, what you mean "fix nautilus"?12:37
=== MacSlow is now known as MacSlow|lunch
didrocksseb128: I think he meant "exposing the tabs through a dbus API" or something like that12:50
didrockslike location path…12:50
seb128didrocks, oh, ok12:51
rperierTrevinho: If the problem is mostly on the nautilus side, does unity need to be fixed or not ?13:03
=== dandrader is now known as dandrader|lunch
rperierwell, I will find another bug, because apparently the previous one is still under discussion and I am probably not the good person to work on it ;)13:56
ryeuhm, i can reproduce flickering in dash on my acer aspire one device14:00
ryeah, sorry, ubuntu-x, not here - https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/110097014:01
ubot5Launchpad bug 1100970 in xserver-xorg-video-intel (Ubuntu) "transparent dash background flickers on intel graphics" [Undecided,Incomplete]14:01
=== MacSlow|lunch is now known as MacSlow
ryehowever, it looks like dash blur "overlay" is rendered too low and too to the left here - https://bugs.launchpad.net/ubuntu/+source/unity/+bug/1100970 - added unity task too14:15
ubot5Launchpad bug 1100970 in xserver-xorg-video-intel (Ubuntu) "Dash background flickers and blur is misplaced on intel graphics" [Undecided,Confirmed]14:15
didrockssil2100: if you bump abi, you need to have unity deps on the new nux version14:35
didrocksotherwise, in the ppa, unity will maybe still build against the previous nux14:36
didrocksmterry: nux is always breaking its ABI, we have something similar to compiz to handle that14:36
didrocksmterry: basically, this file is stripped down and generate a virtual provides:14:36
didrockssee Provides: libnux-abiversion-${nuxabiversion}14:36
didrocksand it's using that string in debian/rules to generate the right version14:37
didrocksthen, when building against nux14:37
mterrydidrocks, ah, OK thanks14:37
didrockswe have a similar thing in debian/rules to see against which nux we built against14:37
didrocksand dep against that14:37
didrocksnot stellar, but no good other way when upstream keeps breaking the ABI14:37
didrocksmterry: however, as told to sil2100, to ensure next daily-build is working, we need to ensure unity build-dep until nux is built14:38
didrocksmterry: can you ensure we get that dummy merge bumping the version in?14:38
mterrydidrocks, sure.  sil2100 , have you started working on that?14:39
sil2100didrocks: ok!14:45
sil2100mterry: I'm working on it right now14:47
mterrysil2100, alright, I'd be happy to do the review when you're done14:47
sil2100didrocks: so a dummy empty merge is enough?14:56
didrockssil2100: hum? no, you need to bump the build-dep for the next day ppa build14:57
sil2100didrocks: ah, yes yes ;) I wasn't thinking when I was asking that it seems14:58
didrocksbetter to ask before anyway ;)14:58
didrocksping me if you have issues to identify what versionning you should put here14:59
didrocksmterry: maybe you already know the rule btw ^14:59
mterrydidrocks, not off the top off my head.  I imagine SONAME gets bumped and the day will be incremented15:00
mterrydidrocks, so libnux-4.0-1 with a version of the next daily build?15:01
didrocksmterry: no, the binary packages statys the same (as we don't version ABI changes thanks to this provides:)15:04
didrockslet's see with which version numbers comes in in his MP and then we can discuss :)15:04
sil2100My branch is open to discussion ;p15:06
mterrydidrocks, oh I misunderstood how the magic happened then.  You're saying the magic is on the consumer (unity) side?15:07
didrocksmterry: yeah, we need to tell unity to build after nux15:08
didrocksas both are built at the same time, uploaded to the ppa15:08
didrockssil2100: ok, that was a nice try :)15:08
didrockssil2100: but in fact not!15:08
didrockstheorically it's fine, for the daily release :)15:09
didrockswe need first to pass the jenkins merger15:09
didrockswhich doesn't have this version15:09
didrocksso we'll get a rejection15:09
sil2100uh, right15:09
didrocksso we need to use the version which corresponds to that change in the staging ppa15:09
sil2100Well, I was thinking we could get this merge in tomorrow, but this would mean a one-day lag that's probably not necessary...15:09
sil2100Oh, ok15:10
sil2100Ah, indeed, since it's using daily + bzr revision15:10
didrockssil2100: so, the version is not that important, bu you can put: 4.0.0daily13.01.18bzr75215:10
didrockseven if it's not exact by one commit :)15:10
didrocksanyway, distro won't have that one15:11
didrocksit will get 4.0.0daily13.01.23 as you invocated15:11
mterrydidrocks, I know why we need to bump unity, I meant the magic of calculating which version is which ABI is happening on the unity side?  I guess I'm confused by what you said happened automatically when the ABI was bumped15:11
didrocksand 4.0.0daily13.01.18 < 4.0.0daily13.01.18bzr752 < 4.0.0daily13.01.2315:11
didrocksso we are good15:11
sil2100Yes, but I see compiz anyway has something like that, since they're compatible15:11
didrocksmterry: oh, that's separate, yeah, let's finish then and I'll reexplain :)15:11
didrockssil2100: yeah, for compiz, it was a day when I just triggered a rebuild15:12
didrocks(of unity)15:12
didrockssil2100: mterry: I'm still opened to the discussion on the fact we can avoid that btw15:12
didrockslike autobumping automatically everyday the build-dep on what's part of the stack15:12
mterrydidrocks, but the daily build has 01.2215:12
mterrydidrocks, not .1815:13
didrocksmterry: ah, daily ppa you mean15:13
mterrydidrocks, yeah15:13
didrocksmterry: yeah, but this one was never published15:13
mterrydidrocks, oh...15:13
didrocksit failed because of the ABI breakage15:13
didrocksmterry: https://jenkins.qa.ubuntu.com/view/cu2d/view/Unity%20Head/15:13
didrockssee this beautiful red :)15:13
mterrydidrocks, OK15:14
mterrydidrocks, and if it had gotten released, the staging would have based its bzr suffixes on .22 anyway15:14
didrocksmterry: yep15:15
didrocksmterry: kenvandine: btw, I was thinking that we can have a google hangout on how to monitor those dailies now that they are working15:15
kenvandinedidrocks, hey yeah we should15:15
didrocksI was thinking like in 45 minutes from now, but on Thursday15:15
didrocksdoes this work for you?15:15
didrockslet me try to schedule this on our calendars15:16
didrocksrobru and cyphermox ok'ed that time15:16
didrocksmterry: kenvandine: I won't be around next week, so you will have to look closely to the releases ;)15:17
didrockssil2100: btw, tell me once you did the change :)15:18
didrockssil2100: approved15:25
didrocksmterry: FYI ^15:25
didrocksmterry: so on the nux ABI thingy15:25
didrocksmterry: this string is used to generate a virtual package15:25
didrockswe extract the date from the file in debian/rules from nux15:26
mterryyup like libnux-abiversion-2012120415:26
didrocksand when generating with dh_gencontrol the control file, we inject it15:26
mterrybut unity doesn't use it?15:26
didrocksyep :)15:26
didrocksunity uses it15:26
didrocksso this string ends up in a header15:26
didrocksthat libnux-dev installs15:26
sil2100didrocks: thank you!15:26
didrockssil2100: yw, thank to you :)15:27
didrocksin unity debian/rules we do something similar15:27
didrocksNUX_ABIVERSION := $(shell sed -rn 's/^\#define[[:space:]]+NUX_ABIVERSION[[:space:]]+//p' /usr/include/Nux-4.0/Nux/ABI.h )15:27
didrocksand in debian/control we dep on libnux-core-<date>15:27
didrocksthat we inject with     dh_gencontrol -- -Vcoreabiversion=$(CORE_ABIVERSION) -Vnuxabiversion=$(NUX_ABIVERSION)15:27
didrocks(coreabiversion is the misnamed compiz one)15:28
MCR1\o/ seems we got a working Compiz Firepaint GLES port !15:28
didrocksmterry: making more sense?15:28
mterrydidrocks, so all consumers of nux have to do that?15:29
didrocksmterry: right, "all" being one… unity15:30
mterrydidrocks, (and that would normally mean we don't have to manually bump unity's dep version like we just did -- but we had to because we used the new blur API?)15:30
didrocksmterry: all compiz consumers have the same thing too (it was the plugins before)15:30
didrocksmterry: yeah, theorically we don't need to bump it, but the new blur API forced us to15:30
didrocksmterry: I'm still open to a discussion on autobumping the build-dep automatically everyday for what is part of the stack15:31
didrocksit's theorically wrong, but TBH, should we really care?15:31
mterrydidrocks, not sure.  It's not like we make it possible to backport to previous releases anyway (i.e. we don't need the deps to be accurate to help other people doing a backport)15:32
didrocksmterry: yeah, the question is in practice, do we really backport only part of the stack?15:33
mterrydidrocks, but that means that if part of the stack has a problem (like nux did), it will stop other parts of the stack15:33
didrocksmterry: most of the time, we need to backport everything within the stack15:33
didrocksmterry: I have no strong opinion TBH, if we have this kind of issues too often, I guess this is a possible workaround15:35
=== dandrader|lunch is now known as dandrader
=== mmrazik is now known as mmrazik|afk
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader
Trevinhoseb128: any idea why my compiz vsize/hsize parameters are set back to old default values when I change them from ccsm? bschaefer has the smae20:56
seb128Trevinho, no idea no21:23
=== salem_ is now known as _salem
=== lborda_ is now known as lborda

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!