/srv/irclogs.ubuntu.com/2011/09/16/#ubuntu-desktop.txt

jasoncwarner_hey robert_ancell did you see andy's comment about the white flicker in LightDM in the latest FFB email thread? Curious if others are seeing that. You hear anything?00:42
robert_ancelljasoncwarner_, my flicker has returned...00:42
jasoncwarner_robert_ancell: :( ah, ok...is it related to the debugging you and I were doing a few days ago? that is when I saw it00:43
robert_ancellRAOF, bryceh, perhaps you can help me with this..  When lightdm starts, there is a chance of the background colour being drawn when the window is show if the draw routine doesn't execute fast enough.  Is there a safe way of ensuring the first frame is drawn in time?00:44
robert_ancelljasoncwarner_, so, anything that slows the system down can trigger this00:44
jasoncwarner_robert_ancell: ah, ok...so it might (or seemingly) be related to the lag I am seeing?00:45
robert_ancellyes00:45
robert_ancellI'm hoping there's some X magic that can make this work properly.  But the real answer might be "wayland" :)00:46
robert_ancellthe difficultly is it's a big window, with a number of elements and it's being drawn when the system is under load00:47
robert_ancelland none of the content looks like the default background color so it's really obvious00:47
chrisccoulsonhi jasoncwarner_00:49
jasoncwarner_hey chrisccoulson , how is the new laptop00:49
chrisccoulsonjasoncwarner_, fast :)00:49
jasoncwarner_robert_ancell: instead of white flicker, can we make it a purple flicker ;) so it looks like grub (talk about a bandaid!)?00:50
jasoncwarner_robert_ancell, bryceh RAOF obviously fixing the lag would work much better if we can figure it out... !00:50
robert_ancelljasoncwarner_, it's a last second option, but I'd like to fix it properly if we can00:50
jasoncwarner_robert_ancell: ageed...totally00:51
=== eeejay is now known as eeejay_is_afk
smspilla|zjasoncwarner_: compiz tarballs with the stacking fix will be in a ppa today01:16
smspilla|zbut I'm off today01:16
smspilla|ztoodleloo!01:16
smspilla|zrobert_ancell: protip, set the backing_pixel to (thepurplecolor)01:17
smspilla|zwhen you create the window01:17
smspilla|zI think you can do that with get01:17
smspilla|zyou can do it with X at least01:17
robert_ancellsmspilla|z, yeah, but I *really* dont want X to draw the first window until I have the frame ready01:18
smspilla|zrobert_ancell: the next option is compositing I suppose01:18
smspilla|zdo XCompositeRedirectSubwindows with CompositeRedirectManual01:19
smspilla|zcreate your window but don't draw it on-screen till you know its read01:19
smspilla|z(use XCompositeNameWindowPixmap to get a pixmap for the offscreen window and you can draw it on screen fairly easily using cairo)01:20
smspilla|zstackinspector is a good example of that01:20
smspilla|zok01:20
smspilla|zbye!01:20
mdeslaurjasoncwarner_: I'm curious...where is FFB being discussed?01:23
mdeslauris there a list for it?01:23
tremoluxpitti: hello! I just wanted to let you know that I've prepared the tzdata updates for 2011j, bug 80277805:17
ubot2Launchpad bug 802778 in tzdata "tzdata 2011j-1 available" [Undecided,In progress] https://launchpad.net/bugs/80277805:17
tremoluxpitti: and thanks!!05:18
didrocksgood morning05:36
rickspencer3good morning didrocks05:48
rickspencer3robert_ancell, hey, am I ok to upgrade this morning?05:49
didrockshey rickspencer305:49
rickspencer3I saw something about a bug that you fixed05:49
rickspencer3didrocks, fwiw, I don't think compiz/unity crashed once for me yesterday05:49
robert_ancellrickspencer3, yes, it's all updated05:49
rickspencer3thanks robert_ancell05:49
rickspencer3robert_ancell, it's looking sweeeet05:50
didrocksrickspencer3: ah, nice to know! I got unity crashes with the new music lens using ubuntuone (but it's only in the desktop team ppa). We'll debug it today, was too late yesterday05:50
rickspencer3hmmm05:50
rickspencer3music lens is quite the problem, I guess05:50
rickspencer3I see more new compiz stuff05:51
* rickspencer3 crosses fingers05:51
didrocksrickspencer3: the compiz stuff coming isn't dangerous05:51
didrocksrickspencer3: new ws switcher settings, with what's needed to upgrade them05:51
rickspencer3thanks didrocks05:51
didrocksthe music lens is in a ppa anyway, won't be for beta2 and it only crashes when trying to show data from u105:52
didrocks(and only for french apparently ;))05:52
rickspencer3heh05:52
rickspencer3didrocks, isn't beta 2 next week?05:53
didrocksrickspencer3: indeed, but the freeze was yesterday evening05:53
rickspencer3seems like it might be worth it for Dx to try to fix the music lens and get it in for beta 205:54
didrocksrickspencer3: the production server side will be only ready on Tuesday… Not sure it's wise for us to push it before the server is ready05:56
rickspencer3oh05:58
rickspencer3does anyone know why aptd takes up 100% of one of my cores each start up?06:12
glatzorrickspencer3, how long does this high load phase take?06:14
rickspencer3glatzor, seems a couple of minutes06:15
rickspencer3I haven't timed it06:15
rickspencer3I usually just go to a vt and renice it06:15
glatzorrickspencer3, that is far too long indeed06:15
htorquerobert_ancell: hi! before filing a bug report: what's supposed to happen when i log out a second user?06:15
htorquerobert_ancell: right now this sends me to the unity-greeter (on tty8) with the first user selected (and marked as logged in), but it doesn't accept my password.06:15
htorquerobert_ancell: however, if i switch to tty7, i'm presented with the unlock screen for the first user and can log in just fine.06:16
robert_ancellhtorque, it should return you to the greeter and entering the password for the first user should return you to the first session06:16
htorquerobert_ancell: could this be connected to the auth issue from yesterday (i haven't updated yet)?06:17
robert_ancellhtorque, the lightdm.log will be the most useful one as it will say what PAM said after you provided your password06:17
robert_ancellhtorque, it shouldn't be related06:17
glatzorrickspencer3, could you add the -d parameter to the Exec statment in /usr/share/dbus-1/system-services/org.debian.apt.service and append the syslog output of aptd to a bug report?06:18
htorquerobert_ancell: ok, thanks. will reboot to get fresh logs and file a bug report.06:19
robert_ancellhtorque, thanks06:19
rickspencer3glatzor, sure, logging another bug right now, but I'll do that next06:19
rickspencer3robert_ancell, htorque I think I just logged a related bug: bug #85161206:28
ubot2Launchpad bug 851612 in lightdm "Logging out from a FUSA session does not reliably return to VT7" [Undecided,New] https://launchpad.net/bugs/85161206:28
rickspencer3glatzor, of course, aptd was perfectly well behaved this time06:33
robert_ancelldoes anyone know if vt switches are logged by the kernel?  I'd like to check the requests are getting there from lightdm and nothing else is overriding it06:39
chrisccoulson_hi didrocks07:40
chrisccoulson_shouldn't we be tracking bug 805087 in oneiric?07:40
didrockshey chrisccoulson_, how are you?07:40
ubot2Launchpad bug 805087 in unity "Dash and launcher appear underneath windows" [Critical,In progress] https://launchpad.net/bugs/80508707:40
chrisccoulson_seems pretty serious to me ;)07:40
chrisccoulson_yeah, i'm good thanks07:40
chrisccoulson_how are you?07:40
=== chrisccoulson_ is now known as chrisccoulson
rodrigo_morning07:42
chrisccoulsonhi rodrigo_07:42
didrockschrisccoulson: still the stacking issue, sam apparently just got a fix yesterday, but it needs testing07:42
didrockschrisccoulson: and he rolled the wrong tarballs07:42
didrocksso not sure where the fix is, he's off today07:42
didrockshey rodrigo_07:42
didrockschrisccoulson: we'll certainly push on Monday the fix in a ppa, and you can be a goog guinea pig07:43
chrisccoulsonheh :)07:43
didrockschrisccoulson: see the "didrocks-oneiric-list"07:43
didrockstag07:43
didrocksIt means I track it :p07:43
chrisccoulsonah :)07:43
chrisccoulsonrodrigo_, were you still looking at bug 832603 btw, or do you want me to?07:43
ubot2Launchpad bug 832603 in gnome-settings-daemon "gnome-settings-daemon crashed with SIGSEGV in g_simple_async_result_complete()" [High,Triaged] https://launchpad.net/bugs/83260307:43
rodrigo_hi chrisccoulson, didrocks07:44
rodrigo_chrisccoulson, hmm, no07:44
rodrigo_chrisccoulson, it disappeared from my list as it's assigned to you07:45
rodrigo_chrisccoulson, assign it back to me if you want07:45
chrisccoulsonrodrigo_, ah, ok. i can still look at it if you don't have time07:45
rodrigo_chrisccoulson, I have time, so assign it back if you want07:45
chrisccoulsonrodrigo_, can i assign you all my bugs?07:45
chrisccoulsonj/k ;)07:45
rodrigo_chrisccoulson, no, only this one :)07:46
tjaaltondidrocks: I'm a good guinea pig too, have had that bug for a loong time07:46
didrockstjaalton: ok, it will be in the ubuntu-desktop ppa by Monday I hope :)07:46
tjaaltondidrocks: thanks07:47
* didrocks weirdly doesn't have a lot of stacking issue, compared to last cycle07:47
chrisccoulsontjaalton, yeah, that's the reason i've been using unity-2d for most ot the cycle07:47
didrockschrisccoulson: seems firefox is making that worse, a lot07:47
chrisccoulsonheh07:47
didrocksmaybe that's why I don't get it07:47
didrockschrisccoulson: not a joke! that's what sam told me :)07:47
chrisccoulsondidrocks, it's strange that i've never had a stacking issue in mutter or metacity ;)07:48
rodrigo_talking about firefox, chrisccoulson what package do you have installed for flash plugin?07:48
chrisccoulsonrodrigo_, i don't have the flash plugin installed at all atm07:48
didrockschrisccoulson: that's my answer for the past year as well :)07:48
chrisccoulsonfresh install ;)07:48
tjaaltonchrisccoulson: right, like didrocks said firefox seems to trigger it here. full-screening windows may also un-trigger it07:48
didrockschrisccoulson: "it's because of the compositing (but metacity has compositing enabled now)"07:48
rodrigo_chrisccoulson, oh, because of it being broken in oneiric?07:48
didrocksoh well, let's just hope it fixes :)07:48
rodrigo_flashplugin-installer that is07:49
chrisccoulsondidrocks, yeah, i'm using metacity + compositing without issues too, and i don't get it in gnome-shell either :/07:49
rodrigo_I've installed gnash and it doesn't work well on some pages07:49
chrisccoulsonrodrigo_, what's wrong with flashplugin-installer?07:49
chrisccoulsoni try to put off installing flash for as long as insanely possible ;)07:49
rodrigo_The following information may help to resolve the situation:07:49
rodrigo_The following packages have unmet dependencies:07:49
rodrigo_ flashplugin-installer : Depends: flashplugin-downloader but it is not installable07:49
rodrigo_                         Depends: nspluginwrapper but it is not going to be installed07:49
rodrigo_E: Unable to correct problems, you have held broken packages.07:49
chrisccoulsonrodrigo_, oh, you need to enable ,multiarch ;)07:50
chrisccoulsondo you not read mailing lists? :P07:50
rodrigo_chrisccoulson, only sometimes :)07:50
rodrigo_so what do I need to enable?07:50
chrisccoulsonrodrigo_, 1 second, i'll try and find the mail07:50
rodrigo_ok thanks07:50
mvorodrigo_: echo "foreign-architecture i386" > /etc/dpkg/dpkg.cfg.d/multiarch07:51
mvoand then apt-get update07:51
chrisccoulsonrodrigo_, https://lists.ubuntu.com/archives/ubuntu-devel-announce/2011-August/000886.html07:51
rodrigo_i386 even if I'm using 64bit?07:51
chrisccoulsonyeah, what mvo said ;)07:51
rodrigo_ok07:51
tjaaltonhmm so if sam pushed a fix for compiz I'll just test it right away07:51
mvo(the rleease upgrader will do that automatically btw ;)07:51
didrocksrodrigo_: you should subscribe to ubuntu-devel-announce07:51
rodrigo_yeah, I'm not on that list afaics07:52
didrockstjaalton: we don't really know where the fix is and he rolled wrong tarballs apparently (he's off today)07:52
didrocksrodrigo_: you should, really low traffic (3/4 emails a month as an average?)07:52
tjaaltondidrocks: i mean if it's in bzr "somewhere"07:52
didrockstjaalton: it's a merge of multiple branches AFAIK, hence the "no definite answer" :)07:53
tjaaltonahah, ok :)07:53
tjaaltonI'll sit tight and wait then07:53
didrocksI would have pushed it otherwise ;)07:53
rodrigo_didrocks, ok07:53
tjaaltondidrocks: sure07:53
didrockschrisccoulson: btw, any chance playing with nautilus?07:54
chrisccoulsondidrocks, yeah, i was looking at it last night, but ended up having to sleep when i felt like my head was going to explode ;)07:54
chrisccoulsonand i ended up dropping it to investigate bug 851055 too07:54
ubot2Launchpad bug 851055 in lightdm "[Oneiric] 'Not Authorized' error given when mounting device" [Critical,Fix released] https://launchpad.net/bugs/85105507:55
didrockschrisccoulson: come on, head exploding just for a tiny tiny 5 px width window?  :)07:55
chrisccoulsonheh07:55
didrocksyeah, bug 851055 seems to have made a "strike" on the french forum07:55
didrocks(as in bowling, not the french strike :))07:56
chrisccoulsonyeah, it broke basically everything (software-center, update-manager, audio, restart/shutdown etc) ;)07:56
chrisccoulsoneverything using policykit or relying on device ACL's ;)07:57
mvoweeh07:57
rodrigo_need to reboot08:00
chrisccoulsoni'm not the only one who thinks that the icons in the dash are broken am i?08:10
chrisccoulsoni can't believe that's the design :/08:10
chrisccoulsonhttps://bugs.launchpad.net/unity/+bug/851259/+attachment/2409198/+files/Ghost.png08:10
ubot2Ubuntu bug 851259 in unity "Icon ghosting/colour bleed in dash (dup-of: 838925)" [Undecided,Confirmed]08:10
ubot2Ubuntu bug 838925 in ayatana-design "Dash - reflections should be fixed on the centre of the viewport, *NOT* the centre of the entire page including occluded areas" [High,Fix committed]08:10
didrockschrisccoulson: it's by design08:10
didrockschrisccoulson: so, the story is:08:11
didrocks- the dash is a glass08:11
chrisccoulsonthe icons look absolutely terrible in my screenshot :(08:11
didrocks- what is under the dash (the windows) are blurred by the glass (it's an opaque one)08:11
didrocks- the icons above the dash are reflected by the glass08:11
didrocksthen, for anything more, talk to JohnLea :)08:12
chrisccoulsonright, he already closed my bug though ;)08:12
didrockstalk again? :-)08:12
didrocksI find the ones at the bottom are really away from the original icon08:13
chrisccoulsoni get a headache just looking at the dash. it looks how i would expect it to look after having a few beers08:13
chrisccoulsondouble vision ;)08:13
didrockschrisccoulson: so, even no need to drink!08:13
didrockssee*08:13
chrisccoulsonlol08:13
didrocksunity is definitively for you :)08:13
chrisccoulsonwhen the stacking issues are fixed ;)08:13
chrisccoulsonin the meantime, i think i will continue to drink actual beer :)08:14
didrockschrisccoulson: come on! the stacking issue enables you to not have any headache!08:14
didrocksit hides the dash :)08:14
chrisccoulsonlol08:14
didrocksI think you really need better arguments :p08:14
tjaaltonand steals the input -> hilarity08:14
* didrocks stops kidding :-)08:14
rodrigo_didrocks, :)08:19
rodrigo_hmm, now no flash plugin shows in about:plugins in firefox08:21
rodrigo_I guess it doesn't find the 32bit plugin?08:21
rodrigo_chromium does find it08:25
tjaaltonhmm, wonder when the fix for bug 807950 enters oneiric, if it's already fixed upstream08:30
ubot2Launchpad bug 807950 in zeitgeist "zeitgeist-daemon crashed with LookupError in remove_from_connection(): <_zeitgeist.engine.remote.RemoteInterface at /org/gnome/zeitgeist/log/activity at 0xb74ee2cc> is not exported at a location matching (None,None)" [Undecided,Fix released] https://launchpad.net/bugs/80795008:30
seiftjaalton, , we did not release it yet08:45
seifits fixed in trunk08:46
TommehAck08:51
TommehGnome3 PPA reverted to baby fonts *again*08:51
TommehWhat's the default font/size meant to be?08:52
tjaaltonseif: two weeks ago?08:53
seiftjaalton, we fixed it08:54
seifbut we did not release it yet08:54
tjaaltonright, but that commit should be pushed to oneiric08:54
seiftjaalton, i will look into getting a release done this weekend08:54
tjaaltonseif: ok, thanks08:54
rodrigo_everytime there is a unity release, I get lots of spam from a spammer called didrocks :)09:18
didrocksrodrigo_: oh really? sorry about it! :-)09:18
rodrigo_didrocks, no problem, I'm used to it :)09:18
rodrigo_didrocks, that's your unify script, right?09:20
didrocksrodrigo_: right, it opens/assign/close/retarget all the needed tasks :)09:22
rodrigo_ok, so I won't add you to the list of spammers then :)09:23
didrocksrodrigo_: heh :-)09:36
chrisccoulsondidrocks, so i confirmed now that the nautilus issue is the grab handle separating 2 panes (the desktop icon view is in a GtkPaned container)10:03
didrockschrisccoulson: ahah! I wasn't *that* wrong then :-)10:04
chrisccoulson:-)10:04
didrockschrisccoulson: is there a way from the master icon view class to know that we are in desktop and so, hiding it?10:04
chrisccoulsonit is GtkPaned which creates the offset GdkWindow, and the offset is the width of the grab handle10:04
chrisccoulsoni'm just looking at a way to fix it now ;)10:04
didrockschrisccoulson: I had to reparent my oneconf view in software-center to hide the Pane…10:08
chrisccoulsondidrocks, could even be a gtk bug. perhaps gtk shouldn't draw the handle if there is no first child (which is the case for the desktop window, as the sidebar isn't created)10:09
didrockschrisccoulson: yeah, it's drawing the handler, that's what I saw in oneconf10:09
chrisccoulsondidrocks, yeah, from the docs - "No separator is drawn if one of the children is missing"10:15
chrisccoulsonso, i'm pretty sure that's a gtk bug10:15
chrisccoulsoni'll try and write a small test case for that10:16
didrockschrisccoulson: indeed :)10:16
chrisccoulsonand then send it upstream10:16
didrockschrisccoulson: good work!10:16
rodrigo_ugh, monitor-lost-signal bug again, twice in a row10:39
rodrigo_iirc, it might have been caused by overheating, so switching to my laptop for a bit10:39
Shred00why is libproxy not being updated to the currently maintained (i.e. by upstream) release?11:34
Shred00the release that's in Oneiric is plain and simply broken.11:34
Laneyoh man, c# bindings11:42
LaneyShred00: can you explain in a bug why the current version is non-functional? We're quite close to release so it's difficult to squeeze new versions of libraries in but it may be doable in the case of important bugs.11:47
Laneyall I can see are comments saying it is broken but not why11:48
chrisccoulsondidrocks, oh, it's difficult for me to get it to happen in a test case, as everything draws so quickly. i shall just write a gtk patch and be done with it ;)11:50
didrockschrisccoulson: take a slower machine then! :-)11:50
Shred00it crashes any application that links to it when there is a PAC proxy configuration11:50
Shred00Laney: it crashes any application that links to it when there is a PAC proxy configuration11:50
didrockschrisccoulson: yeah, if you found the difference in gtk where it does the unwanted effect, sure :)11:50
chrisccoulsoni reckon this will do it - http://paste.ubuntu.com/690735/11:57
chrisccoulsonjust need to rebuild and test :)11:57
Shred00Laney: i.e. https://bugs.launchpad.net/ubuntu/+source/libproxy/+bug/660455 and for upstream reference http://code.google.com/p/libproxy/issues/detail?id=153&can=1&q=segfault wherein the completely admit that they don't maintain 0.3 any more.11:59
ubot2Ubuntu bug 660455 in libproxy "clock-applet continually crashing" [Medium,Triaged]11:59
=== MacSlow is now known as MacSlow|lunch
cyphermoxgood morning!12:16
cyphermoxpitti: could you please review https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/849994 ?12:16
ubot2Ubuntu bug 849994 in network-manager "FFE: add NetworkManager DUID support" [Undecided,New]12:16
chrisccoulsoncyphermox, pitti isn't around today12:26
cyphermoxd´ oh12:26
chrisccoulsondidrocks, fixed \o/12:42
chrisccoulsonhmmm, need to remember my bugzilla password now ;)12:43
didrockschrisccoulson: you rock so much!12:44
mterryvictorp, if my machine froze up in the middle of a "System Testing" run, what's the best thing to do now?13:14
victorprestart13:16
victorpif you then restart the test it will just continue from when it left13:17
victorpmterry, what test were you running?13:17
mterryvictorp, ok.  I had just clicked "next" after the "how much memory" question13:19
victorpah, it run a memory stress test after that13:19
victorpI guess your computer wasnt too happy with that13:20
chrisccoulsonhmm, i might take a look at this metacity crasher now13:20
victorpmterry, if we get similar reports from other people we can always remove that test13:20
chrisccoulsondidrocks, https://bugzilla.gnome.org/show_bug.cgi?id=65924113:20
ubot2Gnome bug 659241 in gtk "GtkPaned initially allocates space for separator, even if there is only one child" [Normal,Unconfirmed]13:20
didrockschrisccoulson: it ook more time for you to find your password than getting the fix? :)13:21
chrisccoulsonheh :)13:21
didrockschrisccoulson: the patch looks good, can you stage it in the gtk branch?13:21
chrisccoulsondidrocks, yeah, sure, i'll do that shortly13:22
didrockschrisccoulson: the metacity crash, you mean bug #797078 ?13:22
ubot2Launchpad bug 797078 in metacity "metacity crashed with SIGABRT in raise()" [High,Triaged] https://launchpad.net/bugs/79707813:22
chrisccoulsondidrocks, yeah, was going to take a look at that if nobody else is already13:22
didrocksI was about to jump on it, but I won't remove you this pleasure to work on non xul-related stuff :)13:23
mterryvictorp, froze me again.  Is it possible to submit what I have without finishing?13:23
mterryvictorp, or skip that test?13:23
chrisccoulsonheh :)13:23
roadmrhello!13:24
victorproadmr,  mterry is getting a system freeze when the automated memory test run, can he skip that test somehow?13:25
roadmrmterry: how are you running checkbox? can you deselect it on the initial test selection screen?13:26
mterryvictorp, (FYI also, one of the sound tests refers to "Sound Preferences" in the sound indicator -- but in Oneiric that's "Sound Settings...")13:26
mterryroadmr, I don't remember an initial test screen.  Also, at this point, I'm in the middle, so if I resume the test, I don't believe I get such a screen for the rest of the tests13:26
roadmrmterry: you're right, then let me come up with a workaround13:27
roadmrmterry: memory/check is the troublesome test, right?13:28
didrockscyphermox: hey, are you around?13:28
cyphermoxhey didrocks, yeah, I´ m there13:28
cyphermoxwhat´ s up13:28
mterryroadmr, not sure.  The one right after seeing how much memory you have13:29
roadmrmterry: please run this command to temporarily disable the threaded_memtest binary13:29
didrockscyphermox: just a small question on the network indicator, when clicking on it, I get the menu, then it hides, then it shows again (very clickly)13:29
roadmrmterry: chmod 000 /usr/share/checkbox/scripts/threaded_memtest13:29
cyphermoxdidrocks: yeah :/13:29
mterrydidrocks, same here13:29
didrockscyphermox: or sometimes, I get a first menu (without the wifi access point, and then, get them 0.5s later (even if I'm connected for a long time already)13:29
didrockshum, interesting13:29
mterryroadmr, ok, will re-run13:29
cyphermoxit´s on purpose, kinda. I need to update the menu from the initial thing there was13:29
didrocksseems to only be on this indicator, isn't it?13:29
cyphermoxand indicators can´t not have a menu13:29
cyphermoxright13:30
didrockscyphermox: yeah, and you don't do this in the backend?13:30
roadmrmterry: let me know how it goes, also, once your test run is finished, I'd appreciate if you could run the memory test by itself and see if it has the same result13:30
cyphermoxdidrocks: I do update the menu, but it needs to be replaced at some point, and only when you click on it to display it to get the most up to date information13:30
mterryroadmr, maybe checkbox could detect such scenarios and offer to skip the next test when resuming?13:30
didrockscyphermox: maybe, there is a rendering workaround that is possible, can  you check with tedg and njpatel about them?13:30
mterryroadmr, k13:30
didrockscyphermox: even when the menu didn't change?13:31
roadmrmterry: that would be possible, not sure if we can get that into oneiric at this point but we'll look at implementing a crash detection system, so to speak13:31
mterryroadmr, quite a few tests are internally crashing too.  I'm clicking "yes" to report them as bugs.  Is that info being bundled up with the rest of the report?13:35
cyphermoxdidrocks: it always gets a new menu, because so many things can have changed13:35
roadmrmterry: nope, it should send you to launchpad to file the report, I think our apport integration is a bit wonky still :(13:35
roadmrmterry: the report will have information about the failed tests anyway, so it might be a starting point13:36
didrockscyphermox: and you can't detect if the menu changed or not? That would remove a lot of flickering already, wdyt?13:36
didrockscyphermox: we need a md5 of the menu! :p13:36
didrockscyphermox: (I meant, updating the menu after the click)13:37
mterryroadmr, submitted!  :)13:37
cyphermoxdidrocks: for wifi it makes no difference; this works quite well with gtkstatusicons13:37
didrockscyphermox: but with this "redraw my whole menu" glitches (disappear and appear againà13:38
cyphermoxdidrocks: not with the gtkstatusicon13:38
cyphermoxso this is really just specific to the indicator13:38
cyphermoxwhere it does make sense to poke tedg or njpatel about it13:38
didrocksah ok, so they have to cache it13:39
didrocksand make the diff locally13:39
roadmrmterry: yay!13:39
didrocksanyway, this will make sense with a lot of indicators I guess, thanks cyphermox :)13:39
mterryroadmr, about to manually test the mem, I may not respond for a bit if I freeze  :)13:40
cyphermoxdidrocks: yeah13:40
roadmrmterry: ok, just change the executable's permissions to 755, cd /usr/share/checkbox/scripts/ and run ./memory_test (it's a Python wrapper)13:40
mterryroadmr, seems to be further than it got before13:41
roadmrmterry: hmm, if it only crashes when running through checkbox it'll be a lot of fun to debug heh13:42
mterryroadmr, I'm willing to remote-debug with you if it helps13:43
roadmrmterry: thanks, if we can't reproduce on any of our systems I'll take you up on the offer13:43
mterryroadmr, if this console-test finishes, is there anything useful I should do with the results, or just note that I didn't freeze up?13:44
roadmrmterry: if you could pastebin the output, just for reference, that'd be ok, but if it finishes I guess the test will say all went well13:45
tedgcyphermox, You really shouldn't be removing the menu and starting over.  You should update the items that changed, including adding or deleting.13:47
cyphermoxtedg: that´ s not really going to be possible, tracking down which ap disappeared that was in the menu and which was now added, then making sure it shows up in the right place13:49
cyphermoxtedg: like i said, this is working just fine with gtkstatusicons, so there has to be something that can be done13:49
tedgcyphermox, Yes, put everything in the same process.13:50
tedgcyphermox, With status icon it blocks on the creation.  Which we're explicitly avoiding.13:51
tedgBad architecture does solve some problems, but we've moved beyond that.  We need to embrace that.13:51
roadmrmterry: is your installation amd64? could you let me know which version of checkbox you're running?13:53
cyphermoxtedg: and I did ask about blocking the display of the menu until the menu is updated in about-to-show, shouldn´ t that work?13:54
mterryroadmr, x86 and 0.12.713:54
roadmrmterry: oh ok.... but the system has 4 GB RAM I see.. that might give a clue, but let's wait until it's finished13:55
mterryroadmr, ok, test just finished, no freeze13:55
tedgcyphermox, No, it doesn't work.  There's some space in the protocol for it, but it's not been implemented.  I think that in general, it's a bad idea.13:56
roadmrmterry: hm, ok, I'm looking at your report right now13:56
tedgcyphermox, And in that case it won't take a destruction of the menu, only postpone display.13:56
cyphermoxtedg: destroying the menu has nothing to do with it, all I´m doing is updating, i destroy after13:57
tedgcyphermox, Never destroy!  Keep updated.13:57
cyphermoxwell, *right now* it´s building a new menu13:57
tedgYes, and that's why it flickers.13:58
cyphermoxbecause tracking down all that may have changed (esp. APs) is hard if not impossible13:58
tedgcyphermox, Why don't you just mark the entries that are APs and remove them, and then update just that list.14:01
tedgcyphermox, g_object_set_data(menuitem, "is-ap", GINT_TO_POINTER(TRUE)) on each one.14:02
cyphermoxdon´ t think this is any better than what´s being done now, but I have another idea that might work14:02
dbarthtedg: seems chrisccoulson is doing the same with xul, as i'm seeing something similar14:05
chrisccoulson??14:05
dbarthie, a small 0-entry menu embryo, and then an update with the content14:05
chrisccoulsonthat's because menus are generated dynamically14:05
roadmrmterry: ok, what the memory test does is run a command twice, so could you run this on two terminals at the same time?14:05
dbarthchrisccoulson: i noticed that recently in thunderbird, maybe my system was slowed down by the ZG bug at that time14:05
roadmrmterry: /usr/share/checkbox/scripts/threaded_memtest -qpv -m1024m -t6014:06
cyphermoxdbarth: the menu is never 0-entry, it just doesn´t have time to show before it gets updated14:06
dbarthchrisccoulson: i understand, but that creates a display flicker14:06
cyphermox(at least for nm-applet)14:06
chrisccoulsondbarth, that's why we need a mechanism for delaying the menu opening whilst we build it14:06
mterryroadmr, ok, hold on one sec14:08
dbarthchrisccoulson: hmm, i thought that about-to-show would give apps enough time to update and avoid the need for that14:09
chrisccoulsondbarth, not really. the panel still displays the menu as soon as it has sent that14:10
chrisccoulsonso it's already open before the application does anything with it14:10
dbarthwhich maybe the problem in the first place14:10
mterryroadmr, froze me14:11
mterryroadmr, seemingly due to the second terminal running it too14:11
roadmrmterry: sorry :( that's a slightly destructive test14:12
roadmrmterry: i'm trying to replicate your results on a system I have here14:15
=== chris|| is now known as chris|
roadmrmterry: 4 of your test errors (bluetooth, wireless) were due to bad permissions in the package, we're working on that14:22
roadmrmterry: usb/storage_transfer might fail if you didn't have usb storage plugged in14:23
mterryI believe I did...14:23
roadmrmterry: and the two memory-related tests fail obviously - other than these your report looks good :)14:23
mterryyay-ish!14:23
nessitapitti: hi there! would you please remind me where I can file bugs for ubuntu-dev-tools?14:23
roadmrmterry: oh ok, let me check the test definition for usb storage14:23
roadmrmterry: ok, could you please plug in a usb thumbdrive, make sure it gets detected/mounted, and then /usr/share/checkbox/scripts/usb_test -t14:27
roadmrmterry: btw, if you're busy this all can wait, I don't want to impose or take up your time or anything14:28
mterryroadmr, permission denied14:28
roadmrmterry: argh, for the usb test too? that's why it failed then :)14:28
roadmrmterry: oy yes, my bad, I just checked the ubuntu branch of the code and a bunch of scripts lack the correct permissions14:29
=== MacSlow|lunch is now known as MacSlow
rickspencer3dobey, I don't seem to have the U1 store in my Banshee15:44
Laneyyou're supposed to use the whizzy new installer to turn that on now15:45
rickspencer3Laney, ?15:46
rickspencer3the store?15:46
rickspencer3I thoguht it was a banshee plugin15:46
Laneyyou use the ubuntuone-installer to install it. or something.15:47
chrisccoulsondidrocks, oh, the metacity crash is easily reproducible :)15:47
chrisccoulsonjust starting xchat crashes it15:47
Laneyit's part of their new delivery mechanism15:47
roadmrmterry: whenever you get a minute (and don't mind the system possibly freezing) could you run this command: /usr/share/checkbox/scripts/threaded_memtest -qpv -m2048m -t6015:47
didrockschrisccoulson: oh really?15:47
didrockschrisccoulson: I'm suprised (that some people are using xchat ;))15:48
chrisccoulsonyay to screen, i managed to get a decent backtrace with --sync, and i caught it all with xtrace too15:48
didrockschrisccoulson: good that's easy to get! would be easier to fix :)15:48
chrisccoulsonnow to figure out what's going on :)15:48
didrockschrisccoulson: oh great!15:48
Laneyrickspencer3: it sucks that people who had it automatically installed might get it removed and have to re-enable it manually though15:48
Laneyis that what's happening to you?15:48
chrisccoulsondidrocks, oh, actually, i would have caught it all in xtrace if i started metacity on the fake display. oops ;)15:49
Laneyor did you just install the system?15:49
chrisccoulsonbrb, need to try again15:49
rickspencer3hey Laney15:53
rickspencer3tbh, I never used Banshee before15:53
rickspencer3I got it when I upgraded15:53
rickspencer3I just looked for the music store, and it wasn't there15:53
Laneyah15:53
Laneywell the U1 people have moved to this new delivery model15:54
Laneyyou have to run their installer to get the pieces you want15:54
Laneyor I suppose you can install the individual packages yourself15:55
didrocksLaney: hey, is the amazon store on by default?15:57
Laneyshould be15:57
didrocksok, that sounds weird then :)15:57
Laneythat's what they want15:58
Laneyno U1 on the CD basically15:59
mterryroadmr, running16:02
chrisccoulsondidrocks, i think I solved the metacity crash16:04
mterry_roadmr, froze16:04
chrisccoulsonwill test a patch in a bit, but i need to disappear for an hour first16:04
didrockschrisccoulson: you're too fast, even not funny :)16:05
chrisccoulsonheh :)16:05
didrockschrisccoulson: sure, I'll probably go on week-end in a hour, if I don't see you, enjoy your week-end!16:05
dobeyrickspencer3: yeah, it's not there by default and the installer installs it.16:06
roadmrmterry_: oops, sorry again :-/16:07
dobeychrisccoulson: is tb-couchdb uploaded?16:23
roadmrmterry_: sorry for the delay. Here's one more thing to try: try running that command in a vt (ctrl-alt-f1), and see if you get any error messages, kernel panics, anything whatsoever when it freezes16:27
mterry_roadmr, k16:28
mterry_roadmr, now it's not freezing me in either vt or X?  ...  slows me down a bunch though16:30
mterry_roadmr, I have to go for an errand16:30
rodrigo_ok, time to go out, have a good weekend all!16:31
roadmrmterry_: hm, it may be dependent on memory load and all16:31
roadmrmterry_: what a coincidence, me too :) We'll continue this in a bit then, let me know when you have some time to try other stuff16:31
mdeslaurAm I the only one who's window decorations disappear every time the workspace switcher is used?16:36
mterry_roadmr, I may also have mistaken a very slow computer as a frozen computer on the previous freeze report with that command line16:49
mterry_roadmr, when running in checkbox, it for sure froze though16:49
didrockshave a nice week-end everyone!16:55
roadmrmterry_: so when running through checkbox, it outright froze and you had to reboot? was the system responsive at all? (i.e. if you press capslock, does the LED respond?), how long did you wait?18:09
mterry_roadmr, back, sorry.  It outright froze.  Music I had playing in the background got stuck in a 1-second loop.  I couldn't switch VTs.  I didn't try caps lock, but I don't have any of those LEDs anyway.  I waited around 15s?18:26
roadmrmterry_: ok, thanks... did it freeze instantly, or did it work for a short while, then crashed?18:28
mterry_roadmr, it was very quick.18:28
mterry_roadmr, somewhere between instant and short while  :)18:29
roadmrmterry_: ok :)18:29
roadmrmterry_: could you try running, on two VTs, the threaded_memtest command? /usr/share/checkbox/scripts/threaded_memtest -qpv -m1024m -t6018:30
roadmrrun this on vt 1, then switch to vt2 and run another instance, so we get two running in parallel18:30
roadmrit seems weird to me that things would crash that hard without any messages at all18:31
mterry_roadmr, yeah, in a sec.  doing something I don't want forzen right now18:31
roadmrmterry_: and I assume the system doesn't crash under normal workloads?18:31
mterry_roadmr, right18:31
roadmrmterry_: yes, no rush, whenever you get a minute :) thanks for your help debugging this18:31
=== kklimonda_ is now known as kklimonda
dobeychrisccoulson: upload tb-couchdb already! :)19:02
chrisccoulsondobey, i'll do it before i sleep tonight ;)19:03
dobeychrisccoulson: btw, where is thunderbird-couchdb upstream source? i guess it would be nice for us to have it in our nightlies ppa19:03
dobeyalthough, i suppose that might be complicated and require new thunderbird too19:04
chrisccoulsondobey, there will be once i've pushed it. but it's only around 10 lines of code anyway (if you don't count the support code i wrote for using ctypes)19:04
dobeysure19:04
dobeyactually this presents a new issue we need to deal with for our nightlies builds :-/19:05
mterry_pitti, heyo -- i still have crazy compiz stacking problems, that I've heard have been fixed upstream.  Do you know the status of that?  I'd gladly test patches, I'm going crazy here19:17
cyphermoxmterry_: just curious, stacking related to minimizing/switching apps?19:18
DBOmterry_, lp:~compiz-team/compiz-core/oneiric.stack_sync_fix19:19
DBOthat branch + latest animation plugin fixes *everything*19:19
mterry_cyphermox, um, like, dash being behind all windows, not the right window appearing when you switch apps19:19
cyphermoxmterry_: yeah, guess it's what I'm seeing too19:19
mterry_DBO, is that in a release or just that branch?19:20
DBOwe have a tarball with some minor issues19:20
DBOsorry we suck at dropping tarballs19:21
DBOand the ABI was bumped for no reason (that needs to be undone)19:21
DBOactually no19:21
mterry_DBO, so sounds like Desktop would want to wait for another tarball?19:21
DBOracarr has a proper tarball19:21
DBOyeah let me get it from him19:21
mterry_DBO, yeah, roll me some of the good stuff19:23
* mterry_ smoketests it19:23
DBOmterry_, https://chinstrap.canonical.com/~racarr/tarballs-oneiric-0.9.5.94ub1.tar.bz219:24
DBOmterry_, I am being told didrocks knows of this and will be testing it on monday or something19:25
mterry_DBO, looks official  :)19:25
DBOmterry_, so enjoy the tarball for you own happy loving window stacking goodness19:25
* DBO goes back to fuming about stacking19:25
mterry_didrocks is the best.  DBO, OK, I'll check in on him on Monday then19:25
DBOdidrocks is a national hero19:25
mterry_DBO, stacking is fixed, you say!  Nothing to fume about, pending our ability to get it into the beta19:26
mterry_or right after19:26
* mterry_ makes no promises19:26
micahghungry, hungry panel: 19250 micah     20   0 1389m 908m 7164 S    0  5.7   3:00.77 unity-2d-panel20:27
chrisccoulsonthere's already a bug for those leaks somewhere20:27
chrisccoulsonright, metacity crasher fixed :)20:28
micahgok, I guess I just have to play the waiting game...20:28
Nafallohmm. lightdm didn't show up after reboot, and I could't figure out how to fix it myself.20:44
NafalloI had just a black screen with a white pointer.20:44
Nafalloanyone seen it?20:44
Nafalloand yeah, apt-btrfs-snapshot saved my ass ;-)20:45
Nafallomvo+++20:45
htorquemicahg: bug 850320 - fixed in trunk (for the most part) :-)20:48
ubot2Launchpad bug 850320 in unity-2d "bad memory leak in unity-2d-panel" [Undecided,Confirmed] https://launchpad.net/bugs/85032020:48
micahghtorque: thanks, any upload planned soon?21:16
desrtrodrigo_: word.21:21
htorquemicahg: no idea, i'm just a curious tester ;-)21:25
=== seif is now known as seiflotfy
dupondjeLibreoffice is like a hell to work with atm23:34
dupondjetons of artifacts?23:34

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