[00:45] Aaah, good. That's not a new crash, just an old one :/ [00:59] I committed a fix for bug 855100 but which series should I push it to? [00:59] Launchpad bug 855100 in totem "python plugins not working" [Undecided,Fix committed] https://launchpad.net/bugs/855100 === eeejay is now known as eeejay_is_afk [01:39] jbicha, what was the reason for holding back totem 3.2? [01:44] it depends on clutter-gst and at least in the past clutter hasn't worked on quite a bit of hardware [01:45] we should get totem 3.2 in the GNOME3 PPA though [01:46] seems to work on nvidia [01:46] so far [01:48] maybe just ARM, I don't really know [01:48] it would be good to figure out before UDS where it works & doesn't work though [01:51] want the scripts so you can build/install it on yours? that would eliminate intel [01:57] bjsnider: yeah you can send me what you've got; I started on totem 3.2 but I didn't get all the way yet [01:58] and you should ask tomorrow what the criteria for joining the GNOME3 team is, as I don't know [02:00] I've heard reports of clutter having problems on ATI in the past, that may be resolved now though. [02:03] I think GNOME Shell fglrx users are still waiting for next month's version to finally get gshell more usable [02:03] that's ridiculous [02:04] but unsurprising [02:17] It of course also means that unless you've got 3D you can't play videos. [02:17] Which is perhaps just a *little* bit aggressive :) [02:25] ah, then there's quite a bit of hardware where 3D isn't working [02:27] Indeed. [02:38] RAOF, they can use something else to play videos [02:38] That's true. [02:38] But the thing we'll pop up by default, and the only video player we install by default, won't work. [02:39] is it going to be substantially different in 6 months? [02:40] We have six months to fix it? :) [02:41] 6 months to get everybody on 3d [02:41] RAOF: llvmpipe ? === smspillaz is now known as smspillaz|o [02:43] bjsnider: Or 6 months to add a non-3d fallback to totem. Or to fix clutter on swrast. [02:43] maybe there's already a fallback [02:43] i haven't tested it === _Tommeh is now known as Tommeh [04:14] Good morning [04:16] Hey pitti. [04:17] Hey pitti! [04:21] hey TheMuso, hey RAOF, how's downunder? [04:22] Damp and cold, apparently. [04:22] More like spring these days thanks. [04:22] Spring is here! [04:22] Yep, warmer weather being the spring I am referring to. :) [04:23] ah, we just turned from "nice summer" to "cold and windy", brrrr [04:24] * ajmitch would like summer to come around a bit faster [04:26] * RAOF would like to be proved wrong about his understanding of gnome-desktop's crtc assignment algorithm [04:29] pitti: what series should I push the totem plugin fix to? [04:29] jbicha: just saw your upstream commit pointer, nice! [04:29] jbicha: oneiric-proposed, please [04:29] jbicha: I can accept it already, so that people can test it [05:04] STACKING!!!!1111 [05:04] smspillaz|o: You're fixing all those problems in an SRU, right? :) [05:09] -.- [05:10] RAOF: its not possible to fix them 'all' unless I can reproduce them 'all' first [05:10] RAOF: though I did give didrocks some patches which fixed a few cases, not sure if that's been uploaded yet [05:11] see here's what sucks about writing window managers [05:11] when you write a normal application and a user reports a bug about how they did something crazy and then it did something crazy in return [05:11] you can just say [05:11] 'dont do that' [05:12] when you write a display server or window manager, what on earth are you supposed to tell the user [05:12] 'dont use your desktop? [05:12] ' === smspillaz|o is now known as smspillaz [05:12] Inconsistency detected by ld.so: ../sysdeps/x86_64/dl-machine.h: 466: elf_machine_rela_relative: Assertion `((reloc->r_info) & 0xffffffff) == 8' failed! [05:12] wonder what's up with that [05:12] * smspillaz reboots [05:19] RAOF: also, whenever someone says the word "stacking" I just think of this http://www.youtube.com/watch?v=U9SEVop9sc0&feature=related [06:15] good morning [06:18] Hey didrocks [06:18] hey RAOF! [06:19] RAOF: I again had a X stopping on too much write [06:19] RAOF: in addition to Xorg.log, I have dmesg output, interested? [06:19] Yes please! [06:20] bonjour didrocks [06:20] good morning pitti! Seems you didn't sleep much, how are you? [06:21] didrocks: I'm fairly well, cold is almost gone [06:21] didrocks: I got 8 hours, from 10 to 6, that's quite enough :) [06:21] pitti: great! you should take some rest though in addition to just answering bugs from 9:30 and starting again at 6:20 :-) [06:22] wow, I'm being watched :) [06:22] We have your best interests in mind :) [06:23] pitti: heh! ;) [06:26] RAOF: http://people.canonical.com/~didrocks/xcrash/ [06:28] didrocks: And that dmesg includes the crash? It's depressingly silent. [06:28] RAOF: well, the "stop" at least, not sure if it's a real crash :) [06:28] but yeah, it contains this "zomg, no more Xorg" [06:30] Bah [06:31] * RAOF hates on Xorg's backtrace handler that we don't quite manage to stop messing up crash dump production. [06:34] At least it seems to be consistently crashing in evdev. I'll think about how best to grab useful state. [06:35] RAOF: keep me posted :) === tkamppeter__ is now known as tkamppeter [07:01] Greetings. [07:11] BigWhale: Greetings, fellow Human! [07:12] Funny thing saying this to a whale :> [07:17] pitti: compiz and compiz-plugins-main into -proposed FYI [07:17] \o/ [07:18] lot of fixes from smspillaz :) [07:19] yay, thanks [07:28] smspillaz: great work, thanks for all these fixes [07:28] :) [07:28] more coming next week :) [07:42] chrisccoulson: You've prodded the bowels of gnome-desktop's RANDR code recently - am I correct in my analysis that it can fail to accept a valid configuration because it's stupid about the restrictions on CRTCs? [08:09] hey [08:09] bonjour seb128 [08:09] hey pitti, how are you? [08:10] seb128: quite well, thanks! [08:10] salut seb128 [08:11] pitti, did you get over your cold? [08:11] lut didrocks [08:11] pitti, it's meeting reminder day! [08:11] seb128: mostly, yes [08:11] seb128: ooh, right [08:11] great ;-) [08:13] ok, proposed is getting quite some activity, I'm torned between testing what will be on the CD or switching to proposed ;-) [08:14] * pitti is on -proposed [08:14] I have my mini10 on a fresh install from this morning without -proposed [08:15] I've my 10v on a fresh install for yesterday afternoon [08:15] and I'm finally enjoying some quiet hours for apport hacking :) [08:15] need to rsync and redo an install today [08:15] pitti, yeah, apport hacking \o/ [08:15] seb128: the bcmwl installation is really great these days, did you notice? [08:15] just bug 872119 which stands between "now" and "perfect" [08:15] Launchpad bug 872119 in ubiquity ""Error occurred while copying the network settings" on bcmwl machine" [Undecided,Confirmed] https://launchpad.net/bugs/872119 [08:15] ev is on it [08:16] pitti, well, I had wifi working on the liveCD for most of the cycle but yesterday's install had no wifi listed in ubiquity [08:16] need to try again [08:16] hm, worked here [08:16] it automagically installed bcmwl at the start of ubiquity [08:17] pitti, right, that's what I mentioned to you at the rally iirc, wifi was working for me on the liveCD this cycle, you didn't believe me by then :p [08:17] well wifi was up but I had no network listed in ubiquity [08:17] seb128: yeah, I didn't realize that we shipped gcc, dkms, toolchain, bcmwl-kernel-source on the desktop CDs [08:17] it didn't pick my ap for some reason [08:18] by maybe it was just being slow to pick it up and I didn't wait or something, I will try again in a few [08:18] I also didn't bother much because I figured most people who install online so I wanted to test an offline install ;-) [08:19] pitti, oh, you improved your sru script to pick names? ;-) [08:20] seb128: yes, it hardcodes "Dear seb, plz test everything, love pitti" [08:20] :p [08:21] morning everyone [08:21] it was done to more clearly point out that we actually expect and appreciate feedback from the original reporters, not (just) QA team [08:21] hey jasoncwarner_ [08:21] oh, a jasoncwarner_ [08:21] let's pretend we are working! [08:21] jasoncwarner_: are you in London now? [08:21] hey seb128 and pitti, how are things? [08:21] jasoncwarner_, working hard!!! ;-) [08:21] pitti: yup! finally made it. boy, I love traveling ;) [08:21] *clickedicklackediclick* [08:21] lol [08:22] * pitti ignores the blatantly huge irony tag there [08:22] jasoncwarner_: had to rebuild during the night, and today's testing again uncovered some warts, but nothing earthshattering so far [08:22] seb128 pitti , how are things looking? [08:22] pitti: oh, ok... [08:23] either of you know how lightdm multi-=monitor is looking? I didn't catch robert this morning [08:23] jasoncwarner_: us having the time and nerve to do bad jokes is usually a good sign :) [08:23] jasoncwarner_, no real surprise desktop wise, we are good, not great but good [08:23] jasoncwarner_, some SRUs start flowing in proposed already [08:23] jasoncwarner_: bryceh did some analysis on http://people.canonical.com/~bryce/ExternalMonitor/results.html [08:23] seb128 pitti cool, I'm going to be putting in some blueprints this week then...get ready for P! [08:23] RAOF, what sort of restrictions? [08:23] hello btw :) [08:23] jasoncwarner_: seems bearable on the lightdm side [08:23] and good morning everyone [08:23] jasoncwarner_, great [08:24] hey chrisccoulson [08:24] hey chrisccoulson, good morning; how are you? [08:24] hi seb128, how are you? [08:24] morning chrisccoulson ! [08:24] hi pitti - i'm good thanks, how are you? [08:24] hi jasoncwarner_, how's london? [08:24] pretty well, thanks [08:26] chrisccoulson, I'm fine, you? [08:26] chrisccoulson: not rainy? ;) seems nice so far. never been here. [08:26] seb128, yeah, good thanks [08:27] jasoncwarner_, oh, it's not raining? [08:27] chrisccoulson: saw the river, london bridge...all good so far [08:27] i guess that's because you're 3 hours away from birmingham! [08:27] chrisccoulson: I know, right? I thought it always rained in london [08:27] bryceh, pitti, jasoncwarner_: just for the record it's likely that login screen and the Unity (GNOME) session should behave differently with multimonitor, lightdm has this point let xorg deal with screens where Unity (GNOME) has gnome-settings-daemon reacting to xorg events [08:27] should->will [08:27] i haven't been to london for years. in fact, i don't think i've been since i joined canonical [08:28] which is a bit strange ;) [08:29] pitti - i fixed ubufox being disabled on upgrade to the firefox beta yesterday, although you'll need to manually reenable it if it's already disabled [08:29] but the new version won't be disabled on upgrade now [08:29] i really don't see the point in this new feature in firefox [08:29] when it's so easy to bypass [08:30] i've already fixed all of our extensions to work around it ;) [08:32] You people are chatty this morning. :> [08:33] not only this morning ;-) [08:33] BigWhale: it's not Internet Relay Silence :) [08:33] pitti, what do you think about bug #840858 [08:33] ? [08:33] Launchpad bug 840858 in gnome-control-center "No obvious way to disable Ubuntu startup sound" [Low,Fix committed] https://launchpad.net/bugs/840858 [08:34] seb128: OK for me for an SRU, although it really feels like a workaround [08:34] it's the last place someone woudl look for sound events [08:35] but at least it's some way, so fine for an SRU [08:35] it is a workaround but Jeremy and some other feels like it's a frequent request and the workaround is better than not having the option [08:35] not a solution for precise, though [08:35] right [08:35] we will [08:35] - turn it off by default in precise [08:35] - change the sound configuration dialog to have a checkbox for turn it on, off [08:36] sounds great [08:36] seb128: do you know if GNOME has a startup sound? [08:36] as there's no option for it, I figure "no"? [08:36] define "GNOME", "GNOME OS"? ;-) [08:36] shell [08:36] upstream gnome 3.2 install, I mean [08:36] well, the autostart is part of libcanberra [08:37] so it's really a matter of how distribution package and ship libcanberra binaries [08:37] perhaps they'll take at least that checkbox then [08:37] there is not a reference GNOME desktop that would allow to answer about what upstream recommends [08:37] yeah, let's see ;-) [08:38] to be honest if we turn it off I don't care much about the checkbox [08:38] not sure how many users actually want a login sound [08:38] same here [08:38] it feels like a thing of the past century [08:39] without mobile devices to annoy your fellows [08:40] heck, we took the systray away, we can certainly get away with killing that effing login sound :) [08:41] ;-) [08:42] I was wondering yesterday if screensavers are from the previous century as well :p [08:42] it's weird, windows8 as well removed the screensaver [08:42] seb128 pitti john and connor said we could remove startup sound [08:43] well, screensavers are useful to not damage physically crt screens [08:43] that's an issue we don't have nowadays [08:43] jasoncwarner_, yeah, we ack-ed that decision and are happy about it, we just missed to do it before uif in Oneiric so we will do in next cycle [08:44] seb128: indeed, and as it's quick to shut down one, contrary to monitor which have to be warmed [08:44] seb128: :) it does feel like "polish" anyway...cleaning it up ;) [08:44] seb128: actually, the indicator icons managed to burn into my TFT; I was quite surprised [08:45] didrocks, well, it's rather than crt screens can get marked if a static image stay for too long, should be the case with tft ones [08:45] though pitti is making me lie apparently now :p [08:46] well, I bought this screen in 2003, it's admittedly not the shiniest one around [08:46] I'm not really using screen savers, if I'm away from the computer for longer than 20 mins or so I just switch it off [08:46] seb128: yeah, the other solution would have been to shutting down tft which takes time as I told, that's what is used for crt now as the screen is disabled quite quickly as it can be brought back quickly [08:47] no reason to waste 30 W [08:47] didrocks: I thought that already happened? [08:47] it does work here (when I forget to turn it off before lunch break or so) [08:47] well anyway it's not a rational thing, some users just like screensavers it seems ;-) [08:48] I'm really unsure we should bring it back though [08:48] well I just don't see it being enough of a priority to get done [08:48] (on this point, its simple enough for me to technically do this, would take me about 15h to get working fully) [08:49] G'morning all. [08:49] smspillaz, well, we are speaking about "hacks", i.e screensaver animations [08:49] yeah, I know [08:49] smspillaz, we still want compiz to do the locking if you can get that done so we can use lightdm as a frontend for lock screen and switch users [08:50] +1 [08:50] well [08:50] * Sweetshark is a bit scared now that he sees on the weekly summary how much he has to do at the Libreoffice conference ... [08:50] now that I come to think of it, I'm not a fan of compiz doing the locking [08:50] hey Sweetshark [08:50] I'm more a fan of [08:50] having the lock screen run a compositor and window manager inside of itself [08:50] Sweetshark: heh, but don't plan to stop sleeping [08:50] which replaces compiz [08:51] but that would be ... difficult to do architecturally [08:51] smspillaz, how would that work? [08:51] right... [08:51] seb128: well, the idea would be that it takes the compositor selection away from compiz so compiz stops doing compositing [08:51] or compiz keeps compositing into an offscreen window [08:51] while the onscreen one is the lockscreen [08:51] or something [08:51] smspillaz, could one process composite different xorg sessions? [08:52] no [08:52] this would be an external process [08:52] maybe I'd need to write it all down one day and nut it out [08:52] but the idea is to have compiz' compositing redirected into an offscreen window [08:52] doesn't seem like a lts cycle thing ;-) [08:52] no, it isnt [08:52] well all we want is to make sure you can't go back from lightdm to open sessions [08:52] and then have the "lock screen" take over the "real" compositing from X [08:52] pitti: I work 24 hours a day and if that isnt enough I'll work nights ... [08:52] one other way would be to make sure you can't vt switch [08:53] seb128: oh, I was thinking more stability [08:53] seb128: like, as in, if you can make compiz crash, you won't have a lock screen [08:53] with this, you are minimizing what can crash [08:53] right [08:53] but its complicated [08:53] let's see what we can get done ;-) [08:53] requires a bit more thought, but I sort of came up with it today in light of the handoff stuff I did a few days ago [08:54] what we don't need is complicated things to add on the compiz list for next cycle :p [08:54] seb128: yeah, it's definitely not for next cycle that for sure [08:54] seb128: however, getting compiz to do the screenlocking is something we can do this cycle [08:54] smspillaz, btw in compiz gsettings in a ready to land state for UDS? [08:54] seb128: since I've already done it ;-) [08:54] would be nice to start by landing that early [08:54] seb128: should be. [08:55] smspillaz, noted for the locking, let's see at UDS what is the compiz list and make sure we don't overcommit [08:55] like if I had my say it would be gsettings and then bug fixing and stabilization for the cycle: p [08:56] hello [09:00] pitti, hi [09:00] hey tkamppeter [09:01] hey rodrigo_ [09:01] hi seb128 [09:09] pitti, seems that the reporter of bug 653132 has disappeared. Can you test it to verify whether it is fixed? On the desktop of a second account try the actions add printer, change options, delete printer and check whether the right things happen when canceling, entering nothing, wronmg, correct credentials, and whether, once correct credentials (the ones of the first user) are entered, the user is not asked for credentials any more un [09:09] til closing s-c-p. [09:09] Launchpad bug 653132 in system-config-printer ""Add Printer" dialog requests root password if user is not in Configure Printers group" [High,Fix committed] https://launchpad.net/bugs/653132 [09:11] tkamppeter: added to my todo list [09:12] pitti, thanks in advance. [09:22] * pitti fixes four year old bug 132904 [09:22] Launchpad bug 132904 in apport "Ignore crashes for programs which got updated in between the crash and reporting" [Medium,Fix committed] https://launchpad.net/bugs/132904 [09:23] pitti: that remidns me [09:24] pitti: is it possible to allow apport to still give you the backtrace (eg, install dbgsym, retrace locally and give it to you) even if you have some outdated packages ? [09:24] smspillaz: apport-retrace doesn't care about outdated packages, you can call that locally [09:24] ah awesome [09:24] smspillaz: oneiric's version even doesn't make a mess out of your workstation, it just works in a temp dir without any root rights [09:25] cool [09:25] yeah, I figured there must have been [09:25] smspillaz: http://www.piware.de/2011/08/apport-retrace-made-useful/ [09:25] since there's nothing more than "hey X crashed" "ooh, can I have a backtrace?" "nope, you have an outdated libnobodycaresabout" "FFFFUUUUU" [09:26] *more annoying than :) [09:26] pitti: <3 [09:27] smspillaz: if you actually want to report that, you can also run APPORT_IGNORE_OBSOLETE_PACKAGES=1 apport-bug /var/crash/... [09:27] (see man apport-bug) [09:27] :) [09:27] smspillaz: bug 75901 is on my precise agenda, too [09:27] Launchpad bug 75901 in apport "Integrate apport-retrace into GUI" [Wishlist,Triaged] https://launchpad.net/bugs/75901 [09:27] that should make it a lot more enjoyable for developers [09:28] pitti: in other news, I fixed the bug in compiz where the apport dialogs would come to the front and then not get focus and then you wouldn't be able to stack your other window on top of them unless you focused the apport dialog first [09:30] nice! I didn't see that for a while, though [09:31] seems current oneiric's version at least made this a whole lot less frequent [09:31] I get that bug often with update-manager [09:32] so yeah for fixing \o/ [09:32] i.e I often run it on an different workspace, click upgrade, go back to where I was and since it's a bit slow I get the "download" dialog on that ws [09:33] which often lands on front and not focussed [09:35] it was a one line typo [09:36] smspillaz, is the fixed in the sru uploaded today? [09:37] should be [09:45] pitti, hey could you please take a look at https://code.launchpad.net/~loneowais/python-distutils-extra/gobject-introspection [09:45] OwaisL: hello [09:45] Hi :) [09:45] OwaisL: that's on my radar, many thanks for this [09:46] Cool, thanks! [09:59] re [09:59] yeah, 'launcher not revealing" issues are fixed with the compiz sru ;-) [10:01] rodrigo_, btw if you didn't notice I'm filling your bug list regularly ;-) [10:02] seb128, yeah, I've seen it :) [10:02] rodrigo_, there are some bug fixes for oneiric but it's mostly next cycle work on g-c-c [10:03] seb128, which one do you refer to? [10:03] rodrigo_, bug fixing or next cycle? [10:03] for next cycle [10:04] rodrigo_, bug #869080 bug #845915 bug #742544 [10:04] Launchpad bug 869080 in ayatana-design "Launcher - Add launcher configuration option to System Settings Appearance dialogue" [Medium,Fix committed] https://launchpad.net/bugs/869080 [10:04] Launchpad bug 845915 in gnome-control-center ""All Settings" is confusing" [Wishlist,Confirmed] https://launchpad.net/bugs/845915 [10:04] Launchpad bug 742544 in gnome-control-center "Launcher is shown on the wrong screen in some multi-monitor setups" [High,Confirmed] https://launchpad.net/bugs/742544 [10:04] ok [10:05] rodrigo_, well it's basically the other items from the google doc we didn't have time for this cycle [10:05] ok [10:05] rodrigo_, i.e having the all settings in the g-c-c shell changed to be location aware, i.e all settings> region [10:05] the xrandr panel has well to be a bit more unity like visually [10:07] * Sweetshark just got the ultimate nerdshirt: GSoC2011 [10:17] rodrigo_, bug #845915 we could perhaps start by opening a bug on bugzilla about it? [10:17] Launchpad bug 845915 in gnome-control-center ""All Settings" is confusing" [Wishlist,Confirmed] https://launchpad.net/bugs/845915 [10:18] seb128, yes, I just want john to do the convincing, he knows better what he wants [10:18] but yes, I'll open the bug myself with his info [10:18] rodrigo_, good luck, design tend to be busy and not know "upstream" well [10:18] i.e I'm not sure that's going to happen [10:19] yeah, I know [10:19] but it's worth trying, you should probably discuss it directly on IRC with John though [10:19] it's a try to not have to be discussing with 2 design teams :) [10:19] not sure that the bug comment will be enough [10:19] seb128, yes, already waiting for him [10:19] ok [10:19] thanks [10:22] rodrigo_, oh, I've also dropped some bugs from your list now [10:22] rodrigo_, i.e g-s-d natty issues that I don't think we will work on now since natty is not a LTS and Oneiric is the current focus [10:22] cool [10:36] smspillaz, the fix for bug #748840 doesn't work or is not enough for the update-manager case [10:36] Launchpad bug 748840 in unity "Windows should not automatically be focused when opened if the focus is on another application" [Critical,Fix committed] https://launchpad.net/bugs/748840 [10:40] seb128: the update-manager case is probably different then === smspillaz is now known as smspillaz|d === MacSlow is now known as MacSlow|lunch === Riddelll is now known as Riddell [11:00] smspillaz|d, ok === smspillaz|d is now known as smspillaz [11:13] did we drop tsclient? [11:14] https://launchpad.net/ubuntu/oneiric/+source/tsclient [11:16] yes [11:17] https://launchpad.net/ubuntu/+source/tsclient/+publishinghistory [11:17] see the first line it has the reason [11:17] Debian bug #547314 [11:17] Debian bug 547314 in ftp.debian.org "RM: tsclient -- ROM: unmaintained, better alternatives" [Normal,Open] http://bugs.debian.org/547314 [11:18] remmina is the recommended option nowadays [11:18] * AlanBell installs [11:20] looks good [11:21] great ;-) [11:34] RAOF: around ? [11:34] actually [11:34] tseliot: around ? [11:34] smspillaz: sure, what's up? [11:35] tseliot: is there an nvidia-glx package thats supposed to exist ? I installed the binary driver to test something but it didn't seem to install nvidia's libgl [11:35] (I've got nvidia-current installed now) [11:35] (at least, I've noticed that I've still got mesa's libgl installed) [11:36] smspillaz: nvidia-current is the only package you need. what's the output of update-alternatives --display x86_64-linux-gnu_gl_conf ? [11:36] hang on [11:37] http://paste.ubuntu.com/705991/ [11:37] (don't see a libgl) [11:37] tseliot: I'm happy to install it from nvidia's website, though I understand that's a slightly more involved process these days [11:38] smspillaz: that would mean breaking your system [11:38] I figured [11:38] * smspillaz remembers the good old days where that actually worked [11:38] tseliot: any ideas then ? [11:38] smspillaz: what does "ldconfig -p | grep GL" say? [11:39] libGL.so.1 (libc6,x86-64) => /usr/lib/nvidia-current/libGL.so.1 [11:39] hm, interesting [11:39] Error: couldn't find RGB GLX visual or fbconfig [11:39] <- at least that's what I get on a simple glxinfo [11:40] so I figure that maybe it's not loading that [11:40] * smspillaz tries LD_PRELOAD [11:40] ooook so it works with LD_PRELOAD [11:40] *shrug* good enough for me [11:41] smspillaz@XPS-UNITY:~$ ldd /usr/bin/glxinfo | grep GL libGL.so.1 => /usr/lib/libGL.so.1 (0x00007f99863db000) [11:41] or I could jump symlink [11:41] *just [11:41] don't do that [11:42] tseliot: don't worry, I used arch linux, I'll be fine [tm] :) [11:42] smspillaz: I'm wondering what the X log says [11:42] tseliot: nothing particularly of interest there [11:42] (no WW or EEs [11:43] tseliot: I'll probably just export the LD_PRELOAD for now. used to switch graphics drvers around tons back in the day, should be fine [11:43] smspillaz: ok, it's still weird though [11:44] * tseliot -> lunch [11:45] tseliot: I compile tons of stuff from source, didn't expect it to all work :) [12:00] * rodrigo_ lunch === MacSlow|lunch is now known as MacSlow [12:33] popey: hey, how are you? did you see my question in bug 795475? [12:33] Launchpad bug 795475 in libimobiledevice "[iOS5 devices do not work] Unhandled lockdown error (-4)" [Undecided,Confirmed] https://launchpad.net/bugs/795475 [12:33] no pitti I hadn't [12:33] * popey looks [12:44] smspillaz: ok, that's understandable then ;) [12:45] :p [12:46] pitti: left a comment, let me know if there's anything else you need [13:07] Sweetshark: ping? === m_conley_away is now known as m_conley [13:27] popey: hm, weird; I used the very same patch you pointed to [13:46] rodrigo_, there? [13:51] jasoncwarner_: pong. [13:51] Sweetshark: see my dm to you [13:51] tremolux, am I supposed to see "for purchase" apps in oneiric yet? [13:52] pitti, I'm looking at bug 871895. pkgstriptranslations seems to correctly make a _static_translations.tar.gz file (as well as a _i386_translations.tar.gz for the mo files), but I don't see the files in language-pack-gnome-fr-base (for example) [13:52] Launchpad bug 871895 in deja-dup "Deja-dup help translations are not packaged in Ubuntu" [Undecided,Confirmed] https://launchpad.net/bugs/871895 [13:53] mterry: ah, that would be bug 869815 [13:53] Launchpad bug 869815 in langpack-o-matic "does not properly install mallard help files" [High,Fix released] https://launchpad.net/bugs/869815 [13:54] mterry: I suppose the next oneiric langpack update that we do will need to be a full -base refresh, to pick up the lost translations [13:54] sorry about that [13:54] I hacked around this for ubuntu-docs, but didn't pick up the other files any more [13:55] pitti, ah, ok [13:58] mterry: heyo, for purchase apps are not yet deployed for Oneiric, but will be very soon I expect [13:59] tremolux, ah, ok. I read a pre-review that talked about apps for purchase, but they must have just been anticipating based on natty. I thought something might be wrong on my end. Guess not! [14:00] mterry: I'll try to get an ETA, I'm curious myself [14:02] seb128, yes [14:03] rodrigo_, unping, I was having some bug in g-c-c with the search but seems like LANG impacts on the ui but it was still searching on french keywords [14:03] jasoncwarner_: sorry for the late reply, I madly creating presentation slides ;) [14:03] rodrigo_, with a correct setup it works ;-) [14:03] seb128, ah ok :) [14:08] Sweetshark: no worries [14:09] Sweetshark: just confirmed on robbie's machine [14:09] jasoncwarner_: yikes [14:10] jasoncwarner_: what kind of X drivers do you guys have? [14:10] intel [14:10] * popey wonders if there's likely to be any progress on broken tomboy sync by release time [14:10] Sweetshark: I see two issues [14:10] nvidia here. [14:11] bug 853098 if anyone has time to look [14:11] Launchpad bug 853098 in tomboy "Oneiric Tomboy sync auth doesn't complete (dup-of: 845321)" [Medium,Confirmed] https://launchpad.net/bugs/853098 [14:11] Launchpad bug 845321 in tomboy "Unable to setup Ubuntu One sync with Tomboy" [Undecided,Confirmed] https://launchpad.net/bugs/845321 [14:12] jasoncwarner_: intel sucks in other ways too https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/816075 [14:12] Launchpad bug 816075 in xserver-xorg-video-intel "X crashes when Impress antialiasing on Lenovo X220 Intel drivers" [Undecided,Incomplete] [14:13] Sweetshark: this only happens with libreoffice app, not writer, calc or others but just the main libreoffice splash app [14:18] Sweetshark: I'm having a bunch of problems with that app, fyi....main libreoffice apps (calc, writer etc) are working, but the generic libreoffice launcher app is having launcher, menu and alt+tab issues as well. Something is going on... [14:24] Sweetshark: can you test this? open libreoffice and click on new spreadsheet. Do you have a double menu? one is nautilus. Do you have it in the launcher (I don't) and do you have it in alt+tab? [14:28] jasoncwarner_: urgh. yes, I can reproduce it now. [14:28] * Sweetshark bets it is a bamf issue again, not being able to handle the startcenter. [14:29] Sweetshark: can you run this down? [14:29] Sweetshark: not sure what needs to get done, but grab jay or njpatel and see what you can find...thanks! [14:29] jasoncwarner_, you can count it on the list of "known unity issues assigned to dx for 2 cycles" [14:29] hello [14:30] seb128: this one feels new to me, though? [14:30] seb128, feeling grumpy today? [14:30] njpatel, lol, not at all ;-) [14:30] seb128: I could swear this was working for me just like two weeks ago...? [14:30] :) [14:30] * Sweetshark backs with seb128 ;) [14:30] njpatel, but we had a workaround for the startcenter in natty that we dropped in Oneiric because you guys were supposed to make it work ;-) [14:30] startcenter? [14:31] njpatel: libreoffice startcenter app...not calc or wrtier, but the app called 'libreoffice' [14:31] njpatel: right now I'm seeing a couple of pretty icky probs with it [14:32] jasoncwarner_, what problems? I can ask DBO to put them on the SRU1 list [14:32] njpatel: search dash for libreoffice, open it. you should see some graphics glitches, maybe a menu issue and no window controls (if fullscreened) [14:32] njpatel: then, click on an new spreadsheet [14:32] (if they are fixable in time, of course) [14:33] jasoncwarner_, you have found waldo [14:33] njpatel: you should then see that there is no launcher, no alt+tab and a possibly a double menu (one is nautilus) [14:33] the WORST possible application in the world [14:33] the libre-office launcher thing [14:33] you cant work with it, it's just evil... [14:33] we need to remove its .desktop file if you ask me [14:33] bug #732412 [14:33] Launchpad bug 732412 in unity-2d "LibreOffice Writer has no icon in Unity and is labeled "LibreOffice Small Device Format Importer"" [Medium,Confirmed] https://launchpad.net/bugs/732412 [14:33] jasoncwarner_, sorry, I keep getting document recovery that keeps failing, so I can't reproduce it here [14:33] DBO and njpatel are you guys seeing those? [14:33] jasoncwarner_, ^ is that this bug? [14:34] unity-2d is the only thing I have seen with the no-menu controls bug [14:34] * njpatel deletes some libreoffice folders [14:34] seb128: this is in 3d for me (looks closely at bug) [14:34] jasoncwarner_, well, the bot picked one component but it affacts both [14:34] seb128, if we got rid of the libre-office start launcher thing? [14:34] would that be bad? [14:34] jasoncwarner_, ah, got the renderering errors [14:34] seb128: I think _part_ of it ..yeah. [14:34] DBO, we did and put it back in Oneiric because it was a workaround and you said you would fix bamf [14:35] njpatel seb128 Sweetshark DBO I'll record my desktop once I'm out of thise meeting so you can see what I'm seeing [14:35] seb128, I never said I could fix that [14:35] DBO, but I've no opinion, talk to Sweetshark he maintains libreoffice ;-) [14:35] Sweetshark, what graphics chip do you have? (intel/nvidia/?) [14:35] Sweetshark, you see the same issue, right? (Renderering)? [14:35] seb128, there is literally no way to fix this short of me making bamf just ignore that .desktop file [14:36] njpatel: nvidia [14:36] yes, I see it now. === kancerman_ is now known as kancerman [14:40] * Sweetshark guesses killing the .desktop file in the install yet again would be quite simple, but rebuilding libreoffice takes three days on armel [14:40] thats definitely a no-go for release ... but you could do an SRU [14:40] pitti: apart from a small glitch in the bookmarks (not sure why I have "foo" with quotes), all seems very great for the french respin, keyboard, live or installer correct, first screen ok, radios… [14:42] didrocks: seb128: do you know if there's a bug # off hand for this "window decorations don't appear on the spread" bug ? [14:42] I just fixed [14:42] it [14:42] DBO njpatel seb128 Sweetshark video emailed... [14:42] smspillaz, bug #865696 [14:42] ? [14:42] Launchpad bug 865696 in compiz "Windows from other workspaces missing decorations in window spread" [Low,Fix committed] https://launchpad.net/bugs/865696 [14:43] ok, that is not fix committed [14:43] * smspillaz reverts the status :) [14:43] seb128: thanks!@ [14:43] smspillaz: you told it was fixed, hence why it's fix committed :) [14:43] yw [14:43] didrocks: I think I confused it for another bug [14:44] didrocks: there was a similar one with expo [14:44] jasoncwarner_, do you have libreoffice-filter-mobiledev installed? [14:44] didrocks, oh hey, btw, thanks for fixing glade2 and quickly. A long time ago, I knew that would cause a problem if not taken care of, but forgot to take care of it [14:44] smspillaz: it's not reproduceable with the -proposed version? [14:44] didrocks: I'm running up to date trunk and reproduced and fixed it :) [14:44] smspillaz: the bug is about windows in other ws missing decorations [14:44] didrocks: right, so the expo one and this one are kind of similar but fundamentally different in one way [14:44] mterry: no worry, was interesting to deal with that :) [14:45] seb128: no [14:45] smspillaz: ok [14:45] the expo one was about windows that were visible in expo losing their decorations sometiems when you switched workspaces. this one is about windwos losing their decorations when transformed but not on a transformed screen [14:45] (subtle difference, which is why I confused it) [14:45] didrocks, smspillaz: the title is misleading, it happens on the same workspace as well, they seem to be there sometime and flicker a bit, it's still happening with the proposed version [14:46] seb128: yeah, that's why I got confused [14:46] I thought you meant "spread" as in "expo view" not "scale view" [14:46] seb128: seem to be hw/driver dependant, I really don't get it there [14:46] didrocks: it should happen on everything, it was a software thing :) [14:46] anyways I've fixed it now [14:46] jasoncwarner_, hum ok, so probably a different bug from the one I pointed, the "buggy" .desktop is coming from that binary [14:46] smspillaz: hum, can do a screenshot [14:46] but I swear I have them :) [14:47] or maybe you just mean on maximized apps? [14:47] didrocks: move a window partly offscreen and press super-w [14:47] the decoration will be clipped incorrectly [14:47] didrocks, open 2 gedits on a workspace and click on the gedit icon in the launcher and move your mouse around the screen while in the spread [14:48] smspillaz, is that bug #870925 [14:48] Launchpad bug 870925 in unity "Windows dragged to another workspace in expose don't have decoration" [Undecided,Confirmed] https://launchpad.net/bugs/870925 [14:48] seb128: nope [14:48] see, this is why I said it's confusing :) [14:48] smspillaz, i.e when you dnd in workspace overview between 2 workspaces it "cut" the decoration at the limit [14:48] seb128: right, I've fixed that in -proposed [14:48] or at least it should be [14:48] smspillaz, no you didn't [14:48] I didn't ? [14:48] I'm running proposed, it's still cut while the dnd happen [14:49] I get seb128's bug about a window between 2 ws [14:49] it comes back when youn unpress [14:49] huh [14:49] I wonder if we missed a patch then [14:50] * smspillaz looks at whats in -proposed [14:50] smspillaz: it's what you told me to put in :) [14:50] smspillaz: + the additional mess cleanup you are aware about [14:50] smspillaz: so, if you are not speaking about that one, which bug are you talking about? (test case?) [14:51] https://bugs.launchpad.net/unity/+bug/865696 [14:51] Launchpad bug 865696 in compiz "Windows from other workspaces missing decorations in window spread" [Low,Fix committed] [14:51] also, please add jittery windows fix [14:51] njpatel: it needs to be reviewed first! [14:51] is it known that nautilus crash a lot in oneiric? [14:51] smspillaz: ok, in spread, not expo, got it! [14:51] didrocks: s/spread/scale/ [14:51] :) [14:51] I get confused otherwise [14:51] yeah ;) [14:51] it keeps crashing when I open some folders here [14:51] * kenvandine waves [14:51] xclaesse, Hey kenvandine [14:51] smspillaz: can you retitle to remove the confusion, please? [14:52] hey xclaesse [14:52] hum, the ubuntuone fix hasn't been pushed? [14:52] kenvandine: ? ^ [14:52] didrocks: so it looks like we're missing it in the package, but that doesn't matter since the fix I did for this bug will add the same fix for the other one in the patch anyways [14:52] smspillaz: ok, so for next SRU? [14:52] +1 [14:52] jasoncwarner_, is your screen corrupted this way or is the video corrupted? that seems pretty broken indeed [14:52] jasoncwarner_, njpatel: works mostly fine for me [14:52] video [14:53] smspillaz, good point, please see if DBO can do it [14:53] Im right here [14:54] DBO: my active reviews [14:54] on it [14:54] and compiz-team/s active reviews [14:54] except the big_fbo one [14:54] jasoncwarner_, njpatel: well in any case I'm not getting that issue on my fresh install from yesterdya [14:54] there's a driver bug on nvidia that I have to hack around *sigh* [14:54] didrocks, the ubuntuone bug that was crashing nautilus was fixed and uploaded [14:54] this is why you should use nouveau [14:54] must be another bug [14:55] kenvandine: ok, thanks, so xclaesse issue should be something else [14:55] kenvandine, when, where? [14:55] xclaesse: stacktrace? [14:55] seb128: :) for you :( for me. I'd take :| for everyone ! [14:55] seb128, last week, remember? [14:55] kenvandine, the bug I pointed to you before? it didn't get backported [14:55] oh... i haven't looked at that one [14:55] kenvandine, no, that's the double unref one he's talking about [14:55] the other one was fixed [14:55] ok [14:55] ah [14:55] didrocks, xclaesse: it's pending upload [14:55] so, yeah, this one needs in -proposed [14:55] * didrocks didn't do all the debug for dobey for nothing :) [14:56] bug 865115 [14:56] Launchpad bug 865115 in ubuntuone-client-gnome "nautilus crashed with SIGSEGV in gconf_client_get(), if both nautilus-open-terminal and ubuntuone-client-gnome are installed" [High,Confirmed] https://launchpad.net/bugs/865115 [14:56] xclaesse, ^ [14:56] what? [14:56] I get a long list of (nautilus:18976): GLib-GObject-CRITICAL **: g_value_get_object: assertion `G_VALUE_HOLDS_OBJECT (value)' failed [14:56] then crash [14:56] dobey, that bug keeps taking nautilus down for users :p [14:56] dobey: we are talking about the patch you fixed, kenvandine need to sponsor it :) [14:56] xclaesse, the list of warning is tedg's fault [14:56] dobey, eta on that? [14:56] oh, it is ready to be sponsored? [14:57] kenvandine, the fix got merged upstream so I guess so [14:57] kenvandine: soon. we have some other fixes we need to put in SRU also [14:57] and yes, the crash is in gconf_client_get [14:57] xclaesse, remove one of the 2 nautilus-... [14:57] seb128, ok thanks for the workaround [14:57] yw [14:58] seb128: do we need to get it in release, or SRU is still good enough? [14:58] cool, that fixed it :) [14:58] dobey, too late for release [14:59] ok [14:59] dobey, but get a SRU out today, if your other fixes are not ready you will do another SRU next week [14:59] dobey: SRU is fine as nautilus-open-terminal isn't installed by default [14:59] so upgrade can be safely dealt with [14:59] seb128: the other fixes are ready in upstream branch as well. [15:00] seb128: so yes, should have it ready later today [15:00] will upload it to -proposed and bug kenvandine when ready :) [15:01] dobey, well if you can upload no need to bug kenvandine ;-) [15:01] you need to bug pitti for approvals :p [15:01] actually, that might be the only one for ubuntuone-client gnome. we have some ubuntuone-client fixes also [15:01] njpatel, jasoncwarner_: weird, I get lo to act weird on my laptop install, not on my netbook though [15:01] sorry, blanked for a minute. the -gnome fix will need sponsored :) [15:03] dobey, just shout when it is ready [15:04] hum, with a11y on on the live, orca isn't launched there [15:04] jibel: is that known ? ^ [15:06] *munch* *munch* yay icecream [15:06] way to reward yourself for the way from the supermarket :) [15:07] and our Italian ice cafe closes on the 23rd [15:08] pitti, him, icecream ;-) weather is winter like here today though :-( [15:08] oh, uh; after some days of rain, 18 degrees and sunny again here [15:09] pitti: seb128: really? it's summer there :) [15:09] sunny and 25 °C ;) [15:09] seb128: we maybe don't have extra 2 slacking days stolen from Germany there, but at least, we have sun! [15:10] nice trade [15:10] didrocks, oh, sunny and 25°C, that's what I call winter weather [15:10] summer is 35° and lot of sun ;-) [15:10] :p [15:10] seb128: ahah, try to fake it now ;) [15:11] 35 sounds pretty hot [15:11] indeed, 25°C is way better ;-) [15:12] seb128 Sweetshark DBO njpatel at the manager's sprint, just confirmed that 3/5 people have issues I have. [15:12] jasoncwarner_, yeah, I get it too, will see if it's something compiz side or unity side [15:13] jasoncwarner_, I've it as well on my laptop [15:13] jasoncwarner_, for SRU1, though, we're at the limit for SRU0 [15:13] MORE BUGS FOR ME WOOOO [15:13] thanks, njpatel [15:13] (sorry couldn't help it, I'm addicted to bugs these days) [15:13] smspillaz: come on, you fix stuff on plugins we don't use by default, like the "command" one! you need more challenging things ;-) [15:14] didrocks: that was actually a patch someone gave me [15:14] smspillaz: you can have some bugs from me, I trade them against finished slides for the Libreoffice conference ... [15:14] smspillaz: I know, just kidding :-) [15:14] jasoncwarner, Sweetshark, bryceh, chrisccoulson, didrocks, tremolux, Riddell, kenvandine, cyphermox, mterry, rodrigo_, seb128, tkamppeter, pedro_: reminder, meeting in 15 [15:14] didrocks: and come on, I just fixed like 3 severe geometry issues today :p [15:14] and a graphical corruption bug [15:14] (saw the "ooppssss I didn't merge that sooner, sorry ;)") [15:14] smspillaz: great value! [15:14] pitti: o/ [15:15] didrocks: :p [15:15] Sweetshark: that's the webcam behind you [15:19] didrocks: dont talk about webcams to me, Im still shocked by oneiric showing my ugly face without asking on install and proposing to make it my avatar! [15:20] njpatel: gnome-settings-daemon was the #1 package uploaded in oneiric, it's not unity, what have you done dude??? :-) [15:20] noooooo [15:20] Sweetshark: heh, I have on my tablet a win8 preview to play with, my avatar is quite weird on it ;) [15:24] how does one propose a branch to oneiric-proposed, if there is no oneiric-proposed branch? [15:24] propose it to the oneiric branch [15:24] but with oneiric-proposed in changelog, the sponsor will deal with it [15:25] ok [15:25] kenvandine: https://code.launchpad.net/~dobey/ubuntu/oneiric/ubuntuone-client-gnome/fix-crash/+merge/79000 deal with it :) [15:26] thx [15:26] We need to deal with lunch :) [15:26] * kenvandine restarts before meeting, brb [15:26] dobey, "We" again :) [15:28] didrocks, g-s-d was the first package uploaded to oneiric? [15:29] seb128: in number of uploads, see ubuntu-devel ML [15:29] oh [15:30] jasoncwarner, Sweetshark, bryceh, chrisccoulson, didrocks, tremolux, Riddell, kenvandine, cyphermox, mterry, rodrigo_, seb128, tkamppeter, pedro_: meeting o'clock [15:30] * pedro_ waves [15:30] hey [15:30] * kenvandine waves [15:30] o/ [15:30] njpatel, you lost to the spanish mafia, shame on you :p [15:30] hey pitti [15:30] https://wiki.ubuntu.com/DesktopTeam/Meeting/2011-10-11 [15:31] and wow... we have jasoncwarner_ for our meeting! [15:31] maybe :) [15:31] so first, congrats everyone for oneiric [15:31] kenvandine, not sure it's a wow, he keeps complaining about bugs this week :p [15:31] so it became a reasonably usable and stable release after all :) [15:31] haha [15:31] yesterday banshee, today libreoffice [15:31] ;-) [15:31] pitti, \o/ [15:31] works pretty fine for me as well [15:31] install and runtime ;-) [15:31] hi [15:31] still needs to pass the WAF test [15:32] (here at least) [15:32] seb128, I only recognise the French mafia, they are the ones that rule my life [15:32] njpatel, ;-) [15:32] kenvandine: anything to discuss for partner this week? [15:32] njpatel, and don't you forget it [15:32] hi [15:32] * pitti hopes kenvandine doesn't have a list of "ten things utterly broken in oneiric" [15:32] pitti, it's on the wiki, tedg just got back from a week sprinting then summit over the weekend [15:32] chrisccoulson, your multi monitor bug is fixed courtesy of Trevinho [15:33] ah, reloading FTW [15:33] njpatel, excellent, thanks [15:33] he will get me a list of planned SRUs later today and i'll link them on the wiki [15:33] nice [15:33] and i have dobey's u1 fix i am about to sponsor [15:33] that is all i have right now [15:33] didrocks: thanks for the unity update [15:33] new SRU looking good, lots of +1 already [15:33] yeah, I leverage the french forum :-) [15:33] so that njpatel knows more that he is really ruled by the french mafia :) [15:34] will do the same tomorrow with unity [15:34] 14 left on http://status.ubuntu.com/ubuntu-oneiric/canonical-desktop-team.html, but I was monitoring that, nothign serious there [15:34] will just postpone to P [15:34] nooo [15:34] so we got 389 work items done, [15:35] I don't mess with the French forum. As gord says, there are a lot of of you. [15:35] I'll do some stats over the past cycles to have an estimate what we can go for in precise [15:35] so, nothing from me this week except for "go out and test images and SRUs" :) [15:35] does anyone have something we should discuss? [15:36] pitti: no more updates, not even security ones for OpenOffice.org available, even if Apache OOo would come around one day it would be completely different. [15:37] pitti: so maybe consider backporting Libreoffice to lucid, maverick? [15:37] yuck [15:37] heh [15:37] Sweetshark: you don't think that the odd security update can be backported? [15:37] there haven't been terribly many USNs for OO.o/LibO so far [15:38] pitti: Im not sure, but LO has gone through some major code cleanups since OOo times. [15:38] if the affected part of LibO didn't change fundamentally, then the patches ought to be backportable; and if it did, it might not even apply [15:39] Sweetshark: do we have any currently outstanding CVEs which we could look at? [15:40] uhm, security isn't a valid reason to get LO into old releases [15:40] pitti: http://blog.documentfoundation.org/2011/10/05/the-document-foundation-publishes-details-of-libreoffice-3-4-3-security-fixes/ CVE-2011-2713 would likely apply also to OOo [15:40] Sweetshark: ** RESERVED ** This candidate has been reserved by an organization or individual that will use it when announcing a new security problem. When the candidate has been publicized, the details for this candidate will be provided. (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2011-2713) [15:40] chrisccoulson: for "Review update plan for extensions with security team", has this been discussed? [15:41] pitti - no, but there isn't much to discuss anymore. i've got that all pretty much figured out [15:41] chrisccoulson: ah, good; setting to "done" then [15:41] thanks [15:41] Sweetshark: is that such a huge patch? [15:42] Sweetshark: this sounds like the kind of issue which is usually an one-liner with an additional boundary check [15:43] pitti: no, IIRC those patches are rather small. [15:43] pitti: however, WMF exploits might not even work on linux anyway. [15:43] Sweetshark: due to our fortified toolchain? [15:44] pitti: But I am bringing the topic more in a general sense -- maybe needs a more careful discussion on UDS. [15:45] Sweetshark: can do; my gut feeling is that we should just apply the security patches and otherwise leave it alone; we could consider a PPA for people who really want it, of course [15:46] pitti: well, or because we dont do anything with the WMFs anyway, while on windows the exploit might wreck havoc with windows services ... [15:46] @all: any other topics we need to discuss? otherwise let's adjourn [15:47] Sweetshark: perhaps; but in principle such attacks work under linux, too, just with different code, of course, and the fortified toolchain makes stuff a lot harder in many cases [15:49] ok, so thanks everyone! [15:49] let's release that [15:49] let's release that sleepy big cat [15:49] thanks pitti [15:49] thanks pitti :) [15:51] Sweetshark: enjoy the conference next week! should be a relatively short ride, I guess? [15:51] Sweetshark: how long does that take to Paris, 4 hours on the train or so? [15:52] pitti: one hour with air france (unless the ATC strike) [15:52] oh, good luck then, crossing fingers [15:57] pitti, my fixes on s-c-p concerning authentication of admin users were all accepted upstream by Tim Waugh. [16:17] good night everyone [16:19] 'night pitti [16:20] good night pitti [16:24] good night pitti [16:37] mvo, could you check on https://bugs.launchpad.net/ubuntu/+source/software-properties/+bug/853231 [16:37] Launchpad bug 853231 in software-properties "Make the software-properties .policy file dynamically translatable" [Undecided,Confirmed] [16:37] ? [16:37] https://code.launchpad.net/~kelemeng/software-properties/bug853231/+merge/75903 [16:37] mvo, seems trivial and worth including in the next upload ;-) === eeejay_is_afk is now known as eeejay [16:41] seb128: indeed [16:41] mvo, thanks [16:41] seb128: I uploda that to proposed now [16:41] mvo, bug #854818 seems to get some duplicates [16:41] Launchpad bug 854818 in software-properties "software-properties-gtk crashed with AttributeError in toggle_source_use(): 'NoneType' object has no attribute 'disabled'" [Undecided,Confirmed] https://launchpad.net/bugs/854818 [16:41] mvo, when disabling ppas they say [16:42] if it's an easy one you might want to squeeze that in the same upload ;-) [16:43] though I can't confirm it [16:45] seb128: this may actually be a dupe of #820028 [16:45] bug #820028 [16:45] Launchpad bug 820028 in software-properties "software-properties-gtk crashed with UnicodeEncodeError in RemoveSource(): 'ascii' codec can't encode characters in position 99-100: ordinal not in range(128)" [High,Fix committed] https://launchpad.net/bugs/820028 [16:47] mvo, oh ok [16:47] mvo, sorry, reviewing the component bugs while I'm at it, I can confirm bug #838507 [16:47] Launchpad bug 838507 in software-properties "apt-add-repository --remove doesn't remove deb-src entry" [Undecided,New] https://launchpad.net/bugs/838507 [16:47] thanks seb128 [16:47] I check it out [16:48] seb128: I go for dinner now, if you find more, please mail it to me or tell me tomorrow morning :) [16:48] mvo, ok [16:48] * mvo waves [16:48] mvo, should I just assigned them to you you can review the list and unassign if you have no interest? [16:48] mvo, enjoy! [16:55] dobey, i uploaded ubuntuone-client-gnome [16:56] kenvandine: thanks [16:56] let me just say wow... amazing number of dupes on that one :) [16:56] although i couldn't reproduce the crash... [16:57] We could reproduce it easily on 11.04, actually, without nautilus-open-terminal [16:57] different backtrace, but same issue [16:57] yeah [16:57] but that's what happens when you corrupt memory :) [16:57] the fix was obviously correct [16:58] clearly a bug :) [16:58] yes [16:58] just weird it isn't reliably reproduce-able on 11.10 [16:58] yeah, really weird that it "works" without nautilus-open-terminal installed [16:59] wfm with it installed [16:59] kenvandine: it is with nautilus-open-terminal installed [16:59] kenvandine: did you killall nautilus? [16:59] yup [16:59] even restarted my session [16:59] no crash.. [16:59] tried 3 times... oh well [16:59] kenvandine: I got a 70% crash [16:59] it's a heisenbug [16:59] it was clearly a bug though [16:59] hehe [16:59] like most mem corruption bugs are [17:00] don't know why it was so easy to crash on 11.04 though [17:00] probably gtk3 tries to be "smarter" about memory allocation, and does the weird stuff with GSlice [17:01] and nautilus 2.x didn't or something [17:02] good night everyone [17:04] oh right, someone has to approve it going into -proposed === eeejay is now known as eeejay_is_afk === eeejay_is_afk is now known as eeejay [17:05] dobey, hopefully nobody complains about the 2 fixes being together [17:05] depends on who reviews it i guess [17:05] * kenvandine is happy to get both fixes in [17:07] kenvandine: well, technically they both fix the bug in slightly different ways :) [17:07] but they are also separate issues that needed to be dealt with [17:07] ok, we can use that argument if it comes up :) [17:08] kenvandine: so the missing config.h was causing that file to be built using gconf API instead of gsettings, while the rest of the code was using gsettings [17:08] oh... ugly [17:08] kenvandine: and the double-unref was in the gconf-using code :) [17:08] yes [17:09] really, it could have been a much worse problem, but luckily it seems to only happen, probably when another gconf-using extension is installed [17:10] since we were corrupting the default gconf client instance, but if all the extensions use gsettings, they'll never hit the invalid memory :) [17:11] but since nautilus-open-terminal apparently uses gconf still, it hit it [17:11] guess that's why it was easier to debug on 11.04 [17:12] anyway, the patch fixes all those issues :) === eeejay is now known as eeejay_is_afk [19:35] who here knows some internals of Unity or something like jumplists? [19:36] !ask | federico1 [19:36] federico1: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience [19:39] federico1, check in #ayatana too [19:39] federico1, check in #ayatana too [19:40] federico1, quite a few unity developers hang out there [19:43] federico1, i would point you at DBO, but he isn't online right now [19:44] kenvandine: cool, thanks [19:45] federico1, and if you don't find anyone... email ayatana-dev@lists.ubuntu.com [19:45] tedg is probably a good person to ask as well [19:45] federico1, yeah, if your question is really about dbusmenu which is what the quicklists use [19:45] tedg is your man [19:45] kenvandine: (who's dbo?) [19:46] jason [19:46] * kenvandine tries to remember his last name... [19:46] lol [19:46] he did a lot of the launcher work [19:46] smith [19:46] seb128, hehe.. i knew it was something common :) [19:46] ;-) [19:46] i've called him DBO for years though :) [19:47] federico1, he's from the unity team, he has been working on compiz and gnome-do before [19:47] ah, okay :) [19:49] kenvandine: smith :) [19:49] hey jbicha [19:49] jbicha, not sure that the totem patch you dropped and the upstream one are the same thing [19:49] at least it's not something crazy like my last name ;) [19:51] seb128: they aren't the same, upstream used a different bbc plugin (named iplayer) [19:51] I can't really test the functionality as I don't have a UK IP address :( [19:51] jbicha, I think those are different bbc services [19:52] chrisccoulson or other uk people might know better [19:52] ok, it was disabled anyway but we can easily readd it if it's something different & useful [19:53] well I doubt it's get ranked high enough in our priorities to be worked [19:53] but yeah, if we need it it's easy enough to get the patch back ;-) === eeejay_is_afk is now known as eeejay === SanbarCo1puting is now known as SanbarComputing [20:31] re [20:32] hey seb128 [20:32] i just noticed the patch adding the static quicklists to gedit got dropped [20:32] looking at the changelog it was probably accidental :( [20:33] kenvandine, hum [20:33] kenvandine, seems like upstream reverted the commit because it caused error in the validator [20:33] ugh [20:33] that explains it [20:34] kenvandine, http://git.gnome.org/browse/gedit/commit/?id=d5855ef158714fb574183c29aa9af03d16ef2dc3 === federico1 is now known as f_food [20:35] https://bugzilla.gnome.org/show_bug.cgi?id=653470 [20:35] Gnome bug 653470 in general "Please add Unity Launcher quicklist to gedit desktop file" [Enhancement,Resolved: invalid] [20:35] "error: file contains group "NewDocument Shortcut Group", but groups extending [20:35] the format should start with "X-" [20:35] " [20:36] well, I told tedg by then that we should prefix non standard groups with X- :p [20:36] he argued that was not required for groups [20:36] hehe [20:38] chrisccoulson_, there? [20:38] he was wrong [20:38] "Alternatively, fields can be placed in their own group, where they may then have arbitrary key names. If this is the case, the group should follow the scheme outlined above, i.e. [X-PRODUCT GROUPNAME] or something similar. These steps will avoid namespace clashes between different yet similar environments." [20:38] hi seb128 === chrisccoulson_ is now known as chrisccoulson [20:38] hey chrisccoulson [20:38] chrisccoulson, did you get a reply from federico about the docking, fn-f7 thing? [20:38] chrisccoulson, if not he's there, take the opportunity :p [20:39] g'ah, i still haven't e-mailed him [20:39] stupid vacation ;) [20:39] i'll do that now [20:39] chrisccoulson, ok, or just ping him when he's back from being f_food [20:39] you might have a chance to get a reply ;-) [20:39] thanks [20:40] yw [20:43] chrisccoulson: Still here? [20:43] RAOF, yep :) [20:43] hmmm, seems i've really broken the firefox nightlies :/ [20:43] seb128, dobey, you can just make "X-Ayatana-Desktop-Shortcuts: XAyatana Foo;XAyatana Bar;" [20:44] So, I want to check whether my understanding of gnome-desktop's RANDR algorithm is correct. [20:44] tedg: the implementation doesn't follow the desktop-entry-spec correctly :) [20:44] seb128: i'm curious - what docking/fn-f7 thing is this? it sounds like it might be the same as something i'm running into [20:45] dobey, I blame the users of it ;-) [20:45] broder, dell laptops generate a fake fn-f7 keypress when you dock them [20:45] broder, chrisccoulson has been debugging screen setups not being correct on latitude e series after docking [20:45] chrisccoulson: ha! it is the same thing i was running into :) [20:45] heh [20:45] tedg: you are the user of the desktop-entry-spec :) [20:45] well other models could have the same issue [20:45] no, we're seeing it on dells [20:46] anyway [20:46] chrisccoulson, seb128: is there an ubuntu bug i can cross reference in our internal db? [20:47] broder, not yet. feel free to create one though :) [20:47] broder, chrisccoulson: or probably several [20:47] but not triaged correctly [20:47] or not connected to the docking signal [20:47] chrisccoulson: Oh, thinking of which - the gnome-desktop in oneiric-proposed fixes your hang, right? [20:48] RAOF, i've not upgraded yet. i'm still catching up on post-vacation stuff [20:48] will do that tonight though :) [20:48] yeah, sure. my favorite part is that the fn-f7 gets emited on the first post-dock keypress [20:49] chrisccoulson: anyway, if you find anything exciting, i'd be curious to know more. also, i have lots of hardware with the issue if you need testing [20:49] mterry, hey, could you check on bug #872462 and see if we need to do something? [20:49] Launchpad bug 872462 in gedit-developer-plugins "gedit crashes on launch, fails import of _gtk" [Undecided,Fix released] https://launchpad.net/bugs/872462 [20:50] mterry, i.e add a conflicts, or drop the gedit plugin from bzr-gtk, or update to gtk3 (seems not for a sru though) [20:50] chrisccoulson: My understanding is basically: foreach output where output is on { foreach crtc { if crtc can drive output and (crtc is unassigned or output is clone of crtc) assign crtc } }. Does that match your understanding? [20:51] RAOF, yeah, i think so [20:51] RAOF, No, but that matches my compiler's understanding ;-) [20:51] heh :) [20:51] :) [20:51] that seems to make sense, and matches my understanding of how it works too [20:52] seb128, I think perhaps my favorite would just be to stop distributing the gedit plugin. That would pass SRU [20:52] i'll assign to me [20:52] mterry, ok, that's my favorite as well, thanks! [20:52] Unfortunately, it's broken in the case where there are some outputs which must be driven by specific CRTCs. That it seems to work now is a happy accident, I think. [20:53] RAOF, you might be able to help me actually. is there any way i can access the X server that the lightdm greeter is using (before i've logged in)? [20:53] i want to look at the properties on the root window [20:53] but i don't seem to be allowed to do that if i'm not the lightdm user [20:54] You need some X credentials to be able to connect. [20:54] Where does lightdm store them? [20:54] i keep getting an annoying bug where pulseaudio fails to start on login, but when i look at the root window properties in my session, i see a bunch of stale PULSE_* properties [20:54] i'm wondering if those are coming from the greeter [20:54] it uses the same server doesn't it? [20:54] It does indeed, yes. [20:55] I suspect that /var/lib/lightdm/.Xauthority is your winning ticket. [20:55] thanks [20:55] i'll try again next time i log in :) [20:55] :) [20:56] chrisccoulson, you config is a bugs magnet it seems [20:56] if i try to start pulse with those properties still on the root window, i just get "pa_context_connect() failed: Connection refused" [20:56] you are the only one there who seems to have got the gnome-desktop locking issues RAOF's change created as well [20:56] but if i delete the properties, then it works [20:56] heh ;) [20:56] it's good that i catch these bugs ;) [20:57] RAOF: what happens in the screw case? gnome-desktop refuses to apply the configuration? [20:57] err, s/gnome-desktop/RandR/, i guess [20:58] broder: Right. If, say, LVDS can only be driven by CRTC 0 and LVDS appears after VGA in outputs, then VGA will get assigned CRTC 0 and LVDS will be undrivable, and gnome-rr will say "whoops, you can't do that". [20:59] broder: The first part of that is extremely common; I think *all* my laptops' panels can only be driven by CRTC 0, and it's a happy accident that they're also the first output listed. I'm not aware of any protocol reason for that. [20:59] yeah, deleting the PULSE_SERVER property results in pulseaudio immediately being spawned by gnome-settings-daemon [21:00] so i wonder if this is cruft from the greeter [21:00] chrisccoulson, switch to the gtk greeter and see if it happens still? [21:00] chrisccoulson, or uninstall indicator-sound and see if it happens, I guess if something starts pulse in the greeter it's it [21:00] seb128, yeah, i could try that. but i wanted to get the root window properties from the greeter session to see if the properties match [21:01] chrisccoulson, how do you unset the property btw? [21:01] that will be a good sign that it's from the greeter :) [21:01] seb128, xprop -root -remove PULSE_SERVER [21:01] ok, thanks, you never know that could be useful for me one day as well ;-) [21:01] heh [21:02] RAOF: huh. let me take a look at my ATI laptop. my recollection is that VGA is listed before LVDS, but i don't know what the allowed CRTC mappings are [21:03] jbicha, just for the record I still don't agree with you that the login sound is an application :p but we discussed it with pitti today and he agreed that showing the .desktop in a sru is the easiest workaround for oneiric [21:03] jbicha, we will add a checkbox to the sound panel next cycle [21:04] RAOF: looks like both CRTCs can drive any output [21:05] broder: Ah, so hurray for ati I guess. [21:06] seb128: well we can get a better fix next cycle [21:06] how do you determine what CRTC can drive what output? [21:06] seb128: xrandr --verbose [21:06] jbicha, right, that's what I was trying to say, let's get the workaround for Oneiric but do it properly with a checkbox next cycle [21:07] xrandr --verbose will list them, if you don't particularly want to read raw xtrace output :) [21:07] broder, thanks [21:07] seb128: it abstracts the actual protocol-level CRTC numbers, which is annoying, but *shrug* [21:09] right [21:10] Hm. On further inspection, it looks like the nvidia card in my netbook is happy to drive anything with either CRTC, too. So, perhaps I should amend that statement to "My *Intel* laptops all want to drive LVDS with CRTC0 exclusively" :) [21:10] RAOF, chrisccoulson: what would be the correct login then? first deal with output which have specific crtc requirements? [21:10] RAOF, same on my intel dell laptop ;-) [21:11] RAOF: i think it's an issue with old intel. my sandybridge thinkpad can do either with either [21:12] broder: Excellent. So this problem will probably disappear in time. Although? got an eyefinity card? They've got 6 CRTCs and it's entirely possible some of _those_ can only drive DP outputs :) [21:13] ...huh. may have spoke too soon. just switched LVDS to use the 2nd CRTC and my screen went black [21:14] broder: There are some tricks there; did you just run "xrandr --ouptut LVDS1 --crtc $CRTC"? Because that'll set a mode of 0x0 on the crtc as well. You need to additionally specify the mode. [21:15] aha, yeah. that would be my problem [21:15] works if i add a --auto [21:15] Also, in my experience, xrandr would kindly like you to specify the XID of the crtc you're trying to set even though it doesn't actually expose that info. [21:15] yeah, i xtraced to get them [21:16] it'll take the 0/1 indices it prints out, but i don't think the mapping is stable :) [21:16] In my testing it actually *didn't* take the 0/1 indicies. Well, it'd take 0, but not 1. [21:16] It silently accepted 1, but didn't actually do anything. [21:16] that may be a more accurate interpretation of what i saw [21:17] To check, I suggest setting a different gamma ramp on each CRTC; that way you can easily tell when you've actually switched CRTC. [21:19] RAOF: with only one display, i think --crtc 1 always means "the other one" [21:20] Hm. It didn't seem to do that for me, but maybe I didn't try it enough :) === m_conley is now known as m_conley_away [22:06] RAOF, bryceh, you guys around for what I suspect will be a quick meeting? [22:07] TheMuso, I am [22:07] Yup. [22:07] Do we have a jasoncwarner_ , or is he in London sleeping off some pints? ? [22:07] He is in London, and afaicr from the email RAOF was asked to chair... [22:08] Hm. Wher [22:08] * RAOF rifles through his inbox to try and find said email. [22:08] Actually, heh. Let's just chair. Ensuring I actually get all the email people send me can wait! [22:09] bryceh: Want to start with the quiet X front? [22:09] sure [22:09] posted new xdiagnose last week with some fixes [22:09] probably going to split branch at this point, just fixes in stable, working on adding more functionality to release in precise [22:10] X bugs are accumulating, nothing *super* serious but some that look like they deserve some attention. Maybe some worth SRUing [22:10] otherwise seems fairly quiet on my end. [22:11] I'll be looking more into multi monitor issues going forward. [22:11] RAOF, you? [22:11] Likewise on my end of X. [22:11] I've uploaded a new gnome-desktop to oneiric-proposed, fixing my lack of paranoia in the initial check_gl_texture_size work. chrisccouslon's laptop should no longer hang when he docks it :). [22:13] While going through gnome-desktop, it seems that the algorithm it uses to handle configuration changes can wrongly reject configurations that could be set. I'm looking at that in conjunction with bug #861426 [22:13] Launchpad bug 861426 in xserver-xorg-video-intel "[Oneiric] [Regression] When disabling onboard LVDS display and just using external VGA screen corruption occurs" [High,Incomplete] https://launchpad.net/bugs/861426 [22:13] There is always something with Intel around release time, every release. :) [22:14] it troubles me that we have algorithms in xrandr and algorithms in gnome-desktop, that aren't the same, and the ones we rely on the most appear to be the less cogent of the two... [22:14] This has a cheap work-around. I can't quite reproduce the problem without using gnome-display-preferences/g-s-d, but I'm not yet convinced that there's not an X/intel/kernel bug in there. [22:14] TheMuso, heh, it's not just around release time! [22:17] heh ok. [22:17] oh, forgot, I looked into that Vino bug for seb last week. Seems to just be an fglrx bug; asked Alberto to send it up to AMD, but haven't heard back yet (maybe he's on vacation?)\ [22:17] So, going forward I'm going to be diving ever further into the morass that is RANDR-in-gnome. [22:17] Not much on my side of things, audio bug triaging, tracking an a11y bug related to legacy at-spi which is in universe, but still needs fixing. [22:17] And gearing up for P. [22:18] Is there anything that doesn't work with at-spi2? [22:18] Gok, which is in universe for one. [22:18] And dasher, although at-spi support was removed on the latest upload. [22:18] Oh and gnome-mag. [22:18] But thats being deprecated. [22:19] In summary: a bunch of things :) [22:19] yeah [22:19] But gok will likely go away soon enough as well/ [22:20] No at-spi for P? :) [22:20] No. [22:20] only at-spi2. [22:21] We don't have a Robert, but chrisccoulson was having a problem with pulseaudio that might be lightdm/unity-greeter related. [22:22] Did he give a bug number? [22:22] or file a bug? [22:22] not yet [22:22] Ok. [22:22] i'm trying to figure out where the issue is at the moment, but it's not particularly reproducible [22:23] chrisccoulson: What is the issue? [22:23] TheMuso, occasionally, pulseaudio fails to spawn at all in my session [22:23] Interesting. I don't get that at all here. [22:23] and when this happens, there are stale PULSE_* properties attached to the root window [22:23] SOunds like a race... [22:23] removing them results in pulseaudio starting immediately [22:24] * TheMuso hates the way the session loads. [22:24] and i'm wondering if this is cruft left over from the greeter [22:24] oh wow ok. [22:24] hrm... [22:24] Anything else? [22:24] AOB? [22:25] thanks [22:26] Nope. [22:26] Ok. [End Meeting]. Thanks one and all! ;) [22:26] RAOF: You looked into flights for Budapest yet? [22:26] TheMuso: No, I haven't. [22:26] Want to coordinate? [22:27] Ok that makes me feel better. :) [22:27] May as well. [22:31] RAOF: ha! i got pulled away before i could ask, but i think bug #861426 is actually the same as a different multihead bug we're seeing here [22:31] Launchpad bug 861426 in xserver-xorg-video-intel "[Oneiric] [Regression] When disabling onboard LVDS display and just using external VGA screen corruption occurs" [High,Incomplete] https://launchpad.net/bugs/861426 [22:31] we have an issue with older intel cards where just driving the VGA results in wonky timing-related issues [22:31] it seems like the monitor is detecting that it's being driven at a smaller resolution than X thinks [22:32] i had theorized it was because CRTC 0 was getting assigned to drive VGA, but i can't trigger the issue just by setting that mapping myself [22:33] Heh, yes. My problem as well :) [22:33] RAOF: any thoughts on useful debugging info if i can repro the issue? [22:34] xtrace-ing gnome-settings-daemon might be interesting? If nothing else, that should allow someone to write a simple binary that'll reproduce the problem. [22:34] ooh, hadn't thought of that. ok [22:42] RAOF: ok, got a trace. let me see if i can find anything exciting [22:45] RAOF: does http://pastebin.ubuntu.com/706368/ look interesting to you? [22:45] * RAOF rescues hundreds of mails from his spam folder. [22:46] ugh. apologies for the wrapping [22:46] Nothing looks obviously crazy. [22:47] I mean, apart from the fact that gnome-desktop first disables everything then sets up the new configuration from scratch. [22:47] That sounds kinda messy. [22:47] heh, right [22:48] it is definitely changing which CRTC it uses [22:48] the startup config is output 0x41 -> crtc 0x40; output 0x42 -> crtc 0x3f [22:52] I think you're right in that this is likely to be a similar problem to 861426. [22:54] oh, just remembered that i was testing on natty. let me make sure it still happens for me on oneiric [23:00] i guess there's no way to pass arguments to apport hooks, other than by setting environment variables, is there? [23:01] What do you want to do? I guess the question is - how is the hook triggered? [23:01] RAOF, for the firefox apport hook, i want to attach the contents of all the memory reporters in firefox [23:02] i can gather than information from inside firefox (when activating the Report a bug menuitem in the Help menu) [23:02] but i need to pass a filename to the apport hook (assuming i write the output to a file) [23:03] i basically want everything in about:memory [23:04] I think you could do something if it's being called like apport-gpu-error-intel.py; we call that from a udev rule, and could pass parameters to it if we wanted to. [23:06] ah, i see [23:06] thanks, i'll take a look at that [23:08] RAOF: i can't do the same operation in a single xrandr(1) invoke because it's unwilling to give LVDS1's CRTC to VGA1 [23:10] but if i let it pick the CRTC, it does unmap both of them before remapping VGA1 to CRTC 0x3f [23:10] xrandr does? What's the full call? [23:11] chrisccoulson, is this for reports that are being generated from a crash, or just for when a user wishes to intentionally file a bug? [23:11] bryceh, it would just be for people who want to report a bug [23:12] RAOF: http://pastebin.ubuntu.com/706376/ [23:12] that's from xrandr --output LVDS1 --off --output VGA1 --auto [23:12] we'll probably ask if the bug they are reporting is related to memory usage, and then attach the memory reporters to the bug report [23:12] as we'd ask for that straight away anyway [23:13] chrisccoulson, what I might do is instead of generating the dump inside firefox and passing it, to instead provide a mechanism to request the memory dump generally, and then have the apport hook request it [23:13] i guess that's another way we could do it [23:13] of course that's probably not feasible from a security point of view... [23:14] but that's the analog to what we're doing with the intel gpu hook, just that we're snagging it via sysfs [23:14] bryceh, well, it would be fairly trivial. extensions can add command line handlers, so i could add a handler to ubufox (ie, --dump-memory, or something like that), which outputs the information we want [23:14] chrisccoulson, you could also consider just writing it out to the filesystem and have the hook scan for files at that location [23:15] RAOF: also, after running that, if i then do xrandr --output VGA1 --crtc 0x3f --auto (which should be the problematic one), it's fine [23:16] broder: So both those xrandr calls work correctly? [23:16] they seem to [23:19] Hm. I wonder if it's timestamps being weird... [23:19] also weird: if i put it into the broken state, then run xrandr --output VGA1 --crtc 0x3f --auto (which should be a no-op), it fixes things [23:21] wait...dangit. i might have my crtcs backwards [23:22] the crtc numbers are stable for the lifetime of the X server, right? [23:23] Yes. [23:25] Ok. I got my CRTCs confused. If I break things with gnome-display-properties, I can't unbreak it with xrandr --output VGA1 --crtc 0x40 --auto, which is unsurprising, given that it *should* be a no-op [23:27] Hm. [23:29] So, crtc_assignment_apply gets passed a RANDR timestamp. It then proceeds to turn off all the CRTCs, then turn the appropriate ones back on, passing the original timestamp. Doesn't the configuration change at the "disable everything" step break that timestamp? [23:30] xrandr(1) does the same thing [23:31] Yes, but xrandr passes ServerTime as the timestamp. [23:31] Which is *always* current. [23:31] oh timestamp, not config timestamp [23:32] if that was happening..then turning off LVDS1 would be the only change taking effect [23:32] and VGA1 would still be running at x=(width of LVDS1) y=0 [23:33] but it would also be equivalent to "xrandr --output LVDS1 --off", because all that does is disconnect the CRTC [23:33] oh no, because xrandr manages to get the SetScreenSize call in [23:33] hmm... [23:35] And my g-s-d traces all show Success as the reply to SetCrtcConfig. [23:36] hmm, also true [23:42] Hm... Unless what happens is - g-s-d sets crtc 0 to off, invalidating timestamp for crtc 0, then sets crtc 1 to off, invalidating timestamp for crtc 1, then tries to set crtc 0 to drive VGA, whereupon the server silently refuses to set the mode with invalid timestamp? Seems unlikely. [23:43] This sounds like a job for Empirical Testing?. [23:45] working on some test code now [23:52] ok, if i just disable CRTC 0x3f, it...just disables that display [23:53] Bah! Why does my laptop no longer reproduce this?! [23:57] ...fascinating [23:58] i think you might be right about it being timestamp-related