=== jalcine_ is now known as jalcine === jalcine is now known as JackyAlcine === JackyAlcine is now known as jalcine [04:16] Good morning [06:02] good morning [06:03] bonjour didrocks, ca va? [06:06] pitti: guten morgen! I'm fine, thanks, and you? [06:07] didrocks: quite fine, thanks! just got up very early [06:07] didrocks: my brain started telling me all the things I need to do today before I get sedated and teeth operation :) [06:08] pitti: ah, teeth operation, round 2 today? :) [06:08] yep, the other two wisdom teeth [06:09] good luck, hope it will go as smootly as the first two of them :) [06:10] * didrocks catches up on Friday + week-end's email [06:10] skying was nice, but backlog when you come back is not ;) [06:11] * RAOF wonders what "skying" is. [06:12] didrocks: skying? [06:12] oh, skiing? [06:12] * pitti hopes didrocks didn't fall from the sky [06:13] "if at first you don't succeed, then maybe skydiving isn't for you" [06:13] ahah :) [06:13] Heh. [06:13] yeah, skiing [06:14] was clearly the last possible week-end for it seeing the temperature :) [06:14] but weather was splendid and snow was good [06:14] we had a marvellous spring Sunday here [06:14] we got out the garden table and seats again, and after Badminton we just sat outside for two hours and let the sun shine on our belly [06:15] ah nice, temperature here should be approximatly the same, I'm pondering working from my balcoon again :) [07:10] the window borders' grab area is 1pix wide after some recent update? [07:10] yes, here too, even on 3D [07:10] (it was never fixed on 2D, I believe) [07:10] what happened to the resize handles / bigger area? [07:11] yeah the lower right corner area is gone as well [07:12] new light-themes perhaps? [07:15] nope [07:17] but looks like light-themes 0.1.8.30-0u3 changes got lost [07:18] tjaalton: gtk resize grips patches reverted? [07:18] that happened, idk if that's the cause of what you're seeing though [07:18] tjaalton: where was that? [07:19] ah, gtk [07:19] "unity has resize borders".. doesn't look like it [07:25] so, off to doctor stuff, see you tomorrow! === tkamppeter_ is now known as tkamppeter === rickspencer3_ is now known as rickspencer3 [07:38] see you pitti, good luck :) [08:31] hm, after today's update, "git pull" (over ssh) is asking for my ssh password again, what's the trick to make ssh-agent do its job ? [08:33] G'Morning all! [08:34] good morning Sweetshark [08:34] didrocks: since pitti is at the doctor getting rid of his wisdom, how about you uploading libreoffice-3.5.1-1ubuntu1 [08:34] ? [08:35] Sweetshark: that can't wait for tomorrow? I'm not sure to be able to do that today, I have a lot of shutdown/restart to do while releasing -2d [08:37] didrocks: no, it was already postponed from friday and 3.5.1 fixes some nasty bugs that are spawning dupes on launchpad without end. [08:37] didrocks: besides our current LO is actually ftbfs because someone broke it with a kdelibs update [08:38] Sweetshark: I can do it, but only this evening I guess [08:38] Sweetshark: I'll have to restart my session, and the time to upload a libroffice release… [08:38] Sweetshark: I'm the other pilot today, but I won't be able to get to it for another 18.5 hours [08:38] ah, yeah, harass the patch pilot otherwise, it's done for that :) [08:39] bug 958781 [08:39] Launchpad bug 958781 in libreoffice "kde4libs updates breaks LibreOffice build (ftbfs)" [Undecided,New] https://launchpad.net/bugs/958781 [08:39] didrocks: cant you upload directly from chinstrap? [08:40] Sweetshark: there is not my gpg key there [08:40] Sweetshark: if the build is going to fail, it shouldn't be uploaded, have you fixed the failure that bryceh found yet? [08:40] micahg: yes, its fixed. [08:42] didrocks, micahg: keep in mind that when you upload now and it will still take some ~6 ours to show up in the repos on the fastest platforms ... [08:43] still can't upload it before this evening apart if you want to handle the unity 2d release, make some gnome-control-center patch as well :p [08:44] Sweetshark: I'm about to go to sleep, so no go here [08:44] Sweetshark: you might be able to get bryceh to do it in ~7-9 hours [08:45] micahg: k, thanks [08:57] hey [09:05] is somebody here getting the gtk "highlight of menus doesn't work"? [09:06] i.e the touchpad mouse enter bug [09:13] pitti, hi [09:21] tkamppeter, hey, not sure he's still around, he has a dentist appointement today for his remaining wisdom teeth, not sure at what time it is though [09:26] hello mvo [09:27] hey glatzor, how are you? [09:28] Servus seb128! Fine. How are you? [09:28] I'm good thanks! [09:30] hey glatzor [09:35] seb128, tkamppeter: pitti already waved off at 8:30. wish him luck. [09:38] seb128, desrt: mvo says I should show you this curious problem: "gsettings get com.ubuntu.update-notifier auto-launch" returns false, but gconf-editor shows it as true, and update-notifier is behaving as if it is true. [09:39] mpt, dconf-editor you mean? [09:39] seb128, yes [09:39] yes, known issue [09:39] the editor doesn't understand overrides [09:39] which is what we use to set distribution default values when they are different from upstream [09:40] the "update-notifier is behaving as if it is true" is a bug in update-notifier [09:40] the gsettings api will return the correct value [09:41] seb128, thanks. To return the favor, I'm getting the "highlight of menus doesn't work" bug. [09:41] Anything I can help you with? I know there's a bug report about it, at least [09:42] mpt, I would welcome you install the gtk version from https://launchpad.net/~ubuntu-desktop/+archive/ppa when it will be done building [09:42] i.e in half an hour [09:42] and let me know if that fixes the menu issue for you [09:42] ok [09:42] mpt, thanks [09:42] good morning everyone [09:42] hey chrisccoulson, how are you? [09:43] seb128, yeah, good thanks. got some peace and quiet here today (jo and ruby are going to see family) :-) [09:43] how are you? [09:44] chrisccoulson, I'm good thanks [09:44] seb128: hm, but mpt is getting "gsettings get com.ubuntu.update-notifier auto-launch" false so update-notifier is behaving "correct" (for this setting). but he said he never changed/touched this setting. any ideas about this? [09:45] mvo, "update-notifier is behaving as if it is true." he wrote? [09:45] mvo, seb128: The behavior changed as soon as I upgraded to Pangolin. [09:45] seb128, ah, sorry. update-notifier is behaving as if it is false. [09:45] So maybe not a bug in update-notifier but a bug in the upgrade [09:45] ok, it makes sense [09:46] mpt, what about "gconftool --get /apps/update-notifier/auto_launch" [09:47] seb128, "No value set for `/apps/update-notifier/auto_launch'" [09:47] mpt, [09:47] gsettings reset com.ubuntu.update-notifier auto-launch [09:47] gsettings get com.ubuntu.update-notifier auto-launch [09:48] seb128, "false" [09:48] so, that's not right [09:48] mpt, strace gsettings get com.ubuntu.update-notifier auto-launch 2>&1 | grep compiled [09:48] mpt, can you pastebin that ? [09:49] seb128, http://paste.ubuntu.com/890407/ === fenris is now known as Guest97336 [09:50] mpt, grep auto-launch /usr/share/glib-2.0/schemas/* [09:52] seb128, http://paste.ubuntu.com/890414/ [09:52] mpt, mvo: ^ there you go, xubuntu override the key [09:53] mpt, xubuntu-default set a different value [09:53] that sucks [09:53] is there not a way to do that on a per-session basis? [09:53] seb128, that didn't happen when I installed the Xubuntu stuff in 11.10 ... only after the upgrade [09:53] chrisccoulson, didrocks by then did something like that for gconf IIRC, not sure about gsettings [09:54] mpt, well, maybe they were not overriding the key in 11.10 or doing it in a smarter wy [09:54] way [09:54] maybe [09:54] but your issue comes from xubuntu [09:54] I see [09:54] it's unfortunatly not possible with gsettings [09:54] so talk to the xubuntu people I guess ;-) [10:25] Sweetshark, seb128, I remembered shortly after entering "pitti, hi" and went on with other stuff, but thanks anyway. [11:10] if somebody feels like doing some GNOME 3.3.92 updates I put a few on the pad [11:40] seb128: dude, can we please get the drag corner back...I can't resize my windows! [11:41] mdeslaur, yeah, need to talk to dx ;-) [11:42] but I'm sure didrocks watch those issue [11:42] seb128: it's getting fixed with a theme or with compiz? [11:42] yeah, this one is known, not sure what regressed though (if it's the theme or compiz) [11:43] mdeslaur: will surely be fixed for beta2 freeze anyway (not for 2d as it's a more complicated issue) but for 3d [11:43] didrocks: ok, thanks, I'll be patient for a while longer :) [11:44] I don't understand, it's not on every machine, which seems to indicate that's a compiz issue [11:44] seb128: do you have it as well? [11:44] didrocks, yes [11:44] didrocks, bug #953839 [11:44] Launchpad bug 953839 in compiz-core "[regression] Invisible resize border is now only 1px wide" [High,Confirmed] https://launchpad.net/bugs/953839 [11:45] seb128: I guess you would have noticed if it was at last compiz release? [11:45] (it's part of the test we are using as wellà [11:45] so I guess it's something else making it regressed [11:45] asac_: ^ [11:45] didrocks, no, I never resize stuff using the border [11:45] I use the handle in the corner [11:46] yeah, I got used to using the handle too, but now that's gone [11:46] anyway, will get fixed this week :) [11:46] didrocks, see dvv's comment on that bug, he points to the likely buggy commit [11:46] didrocks, is the "wm decorator buttons shouldn't have a tooltip" on your list as well? [11:47] seb128: right, we discussed about it this particular morning and he told sam will look at it [11:47] ok [11:47] seb128: you should upgrade :p [11:47] didrocks, ppa? [11:47] (see my ubuntu-artwork upload) [11:48] oh [11:48] compiz took my patch I made 2 years ago, but changed the key path and dropped my distro-patch… [11:48] I updated but I didn't restart my session [11:48] cool [11:48] :) [11:54] seb128: mdeslaur: for me the corner handle is very hard to use now ... and the borders are even worse [11:54] seb128: you'll need to rewrite 51_lock_screen_on_suspend.patch with the new gnome-settings-daemon, as the logic has changed there [11:54] seb128: FYI [11:54] mdeslaur, oh? how? [11:54] mdeslaur, I will check for it [11:55] another odd thing i noticed is that the top left corner handle is not at the top left corner of the window decorations [11:55] but rather right below the top window bar [11:56] seb128: their lock_screensaver function needs to use the old gsetting, and the upower_notify_sleep_cb function needs to use the new gsetting [11:57] mdeslaur, ok, thanks, I will probably ping you for review before uploading to make sure I got it right ;-) [11:57] seb128: sure [11:58] asac_, not sure to understand that description [12:01] seb128: put the mouse pointer in the left top corner of the window border -> nothing. .. put it below the top border on the left of the window -> resize grab appears === MacSlow is now known as MacSlow|lunch [12:07] oh, huh, I get both, top left corner, and edge between window title and window [12:24] asac_, works for me in the corner [12:25] same as mdeslaur, I get both [12:26] yeah... seems sometimes one of them is hard to get [12:48] good morning [12:48] happy beta2 week [12:48] hey desrt, had a good w.e? happy beta2 week! [12:48] ya. pretty good [12:48] desrt, happy GNOME hard freeze week as well ;-) [12:48] finally got my taxes done for 2010 and 2011 [12:53] mterry, hey [12:53] seb128, hello [12:53] mterry, we are discussing the gtk bug on #gtk+ on irc.gnome.org if you are interested [12:53] just read your comment [12:53] mterry, how are you btw? ;-) [12:54] mterry, it's weird that this patch which is only dealing with menus would impact on scrollbars? [12:54] seb128, I've been sick :-/ [12:54] :-( [12:55] seb128, I didn't double-confirm that the first-time scrollbar thing wasn't present in previous gtk [12:55] seb128, maybe they worked in previous GTK and I just never noticed... [12:55] likely, users reported the bugs as "only the first menu items shows highlighted then highlight stop working" [12:56] which seems to suggest that it was already behaving this way [12:56] get better! [12:56] christ [12:56] that's 3 major gtk issues you're tracking now [12:56] desrt, yeah, and you made fun of me when I told you I was not happy with gtk after 3.3.18 [12:57] i still reserve the right to make fun :) [12:57] ;-) [12:57] seb128: so this hashtable assertion [12:57] i think i pushed a fix for it [12:58] desrt, when? should it be in the current version? [12:58] over the weekend [12:58] basically: dbusmenu drops the reference on the old 'root' menuitem before it tells me that it did [12:59] so my weak reference to it is already invalid by the time i find out [12:59] and i was using that reference to remove items from the hashtable [12:59] solution: take a full reference [12:59] ok, I see r238 in your vcs [12:59] good ;-) [12:59] i'm not 100% sure it solves the problem since i would have actually expected a crash in this case [12:59] but maybe the old memory was still there or something [13:00] ok, I will do another upload today I think with whatever you want to get tested [13:00] we will get a round of tarball for beta2 on wednesday I guess [13:00] that should be all [13:00] desrt, you still aim to land xul support for wednesday? [13:00] i've generated a bunch of work for gord to sort through at this point [13:00] xul support is already done [13:01] desrt, it's not merged in though [13:01] oh [13:01] i should m-r that then :) [13:01] ;-) [13:01] hopefully it doesn't make too many new crashes :p [13:01] lol [13:01] yeah, I might backport it, so we can notice eventual issues before beta2 freez [13:01] i'm going to submit the request [13:01] maybe ted will do a release today [13:01] thanks [13:01] tedg: hi :) [13:02] and then i'm going to spend the day working on gnome stuff [13:02] dconf mostly [13:02] the #dx guys are bugging me for a fix and a release [13:02] and i have a couple of other patches to merge as well [13:02] desrt, Is there an app that uses GMenuModel for window menus? [13:02] tedg: yes. many. [13:03] i like to test with gnotravex since it was the first [13:03] desrt, name one :-) [13:03] :) === greyback is now known as greyback|lunch [13:03] * tedg is installing this program blindly [13:04] tedg: epiphany-browser is another [13:04] gnome-documents, gnome-contacts [13:04] probably boxes too, i guess [13:04] Cool, okay. I was able to find gnomine which has an application menu. [13:04] But it doesn't seem to have window menus. [13:05] ya [13:05] window menus are quite rare [13:05] it seems that all of the new apps are going for only having an app menu [13:05] and, incidentally, those are all of the ones that are using the new setup [13:05] (mostly since they need to use it to get the app menu) [13:05] i think it will be until next cycle before we see apps using it for their traditional menubars [13:06] there's one example, though: in the gtk source tree in examples/ we have 'bloatpad' [13:06] and there's also 'plugman' which is a copy of bloatpad made at one point and modified to support "plugins" [13:06] which basically dynamically add/remove menu items [13:07] Okay [13:08] desrt, Thanks! [13:08] * tedg will have defeated travex here soon and will be able to work again [13:10] i'm so confused about merge requests [13:10] it looks like there's already an active merge request on my branch [13:10] sooo, the quicklist patch to totem breaks mimetype registration for it ;) [13:12] chrisccoulson, is there another of those "whoever did it put the new section in the middle of the .desktop"? [13:12] chrisccoulson, I wonder why they did that, it should be at the end [13:13] seb128, oh, the patch adds it to the end. it's just that the totem build system adds the mimetypes to the end of the desktop file [13:13] oh [13:13] interesting [13:13] seb128, bug 956597 [13:13] Launchpad bug 956597 in totem "Firefox doesn't handle mms: URLs" [Low,Triaged] https://launchpad.net/bugs/956597 [13:14] seb128: the XUL stuff has been committed, merge-proposed and reviewed [13:14] tedg: merge!! [13:14] desrt, "it looks like there's already an active merge request on my branch" ... that's because you keep using the same vcs so you already have merge requests for it [13:15] seb128: i'm saying that it looks like there's already an active merge request on my branch for the XUL changes, specifically [13:15] charles reviewed and ACKed it [13:15] desrt, oh, good [13:15] wtf @ bug 955710? [13:15] Launchpad bug 955710 in firefox "tudou.com/playlist/p/a74514i120368879.html" [Undecided,New] https://launchpad.net/bugs/955710 [13:16] can people please stop reporting crappy bugs? :) [13:16] it appears not to contain the latest fix [13:16] which confuses me... i thought i update existing merge requests simply by pushing to them [13:18] desrt, you do [13:18] so this one didn't update :) [13:18] maybe it's because it was already 'approved' when i pushed [13:18] and it doesn't make sense to modify something already approved [13:18] desrt, https://code.launchpad.net/~desrt/indicator-appmenu/hud-rewrite-wip/+merge/97716 [13:18] you mean? [13:18] yes [13:18] it shows 238 in a blue rectangle [13:18] it has all but my most recent fix [13:19] at the bottom of the conversation [13:19] and in the revisions [13:19] desrt, it seems there to me [13:19] desrt, it's in the diff as well [13:19] oh. doh. [13:19] i was searching for the wrong string in the diff :) [13:19] ya. it's there [13:20] so the flip side of that is cool, then [13:20] i can push more changes into a merge request that's already been approved, after it's been approved :) [13:20] desrt, that's until #systems do something like didrocks' jenkins [13:20] sonny, let me tell you [13:20] desrt, the unity team merge reject those with a "new revision since the mp got approved" [13:21] in my day, we committed directly to master [13:21] and we liked it! [13:21] lol [13:22] anyway.... feel free to sort that out :p [13:22] hello [13:22] * desrt goes to fight with vala [13:22] is there anyway to interact with the HUD without launching unity-2d-shell ? [13:22] abondis: why yes, there is! [13:23] i want to use my favorite tiling manager [13:23] but i like the idea of hud [13:23] abondis: there is a hud-cli too, and latest versions of the hud have hud-gtk in the source [13:23] they're not really real interfaces though -- more like debugging tools [13:23] * didrocks waits for desrt telling "you can run unity-3d" :p [13:23] hey desrt ;) [13:23] didrocks: i'd never suggest that to -anybody- ;) [13:23] ah really ? in a ppa or in the precise repo ? [13:23] abondis: i think hud-gtk may end up in the precise repo today or tomorrow [13:24] i added it to the build system on a branch that'll be merged today and i think seb is planning to package it today as well [13:24] but it's really a debug tool.... [13:24] cause i tried xmonad with unity-2d-* [13:24] it's almost what i want [13:26] Sweetshark: still need sponsoring? === MacSlow|lunch is now known as MacSlow [13:26] * desrt wonders wtf header he needs to include to get S_IFREG [13:26] tkamppeter: hello [13:26] lsb says sys/stat.h. glibc disagrees. [13:28] pitti, hey, how are you? [13:28] seb128: feel like beaten up, but ok; merci! [13:28] arghghg. C99. [13:29] pitti, are you done now? i.e got the 2 remaining ones pulled out? [13:29] pitti: yes [13:29] seb128: yes [13:29] pitti: 3.5.1-1ubuntu1 as found on chinstrap [13:29] pitti, ok, at least that's something ;-) [13:33] seb128: btw, for the touchpad menu bug, I don't get that any more [13:33] I was going to test the upstream patch, but like that it won't tell us much [13:33] pitti, I still get it and the patch doesn't fix it [13:34] pitti, I put the ppa in the desktop ppa already [13:34] the patch in* [13:42] * pitti throws hands into air about bug 955147 [13:42] Launchpad bug 955147 in unity "[FFe, UIFe] Automatically add launchers for newly installed applications" [Undecided,New] https://launchpad.net/bugs/955147 [13:44] pitti: at least it doesn't change the UI of the default install? [13:44] that should give the documentation folks some breathing room compared to last time, at least [13:44] that yes, but that's not what I'm concerned about [13:45] I have no problem with the timing, it's the change itself that drives me nuts [13:45] so at first we got rid of the bfb and used up a launcher space [13:45] then we put the 12th link to the control center there, using another [13:45] which caused it to overflow in a default install on a netbook screen [13:45] this one will now cause it to grow beyond limit [13:46] pitti: but how will users ever get to experience the awesome accordian effect? [13:46] then we disable autohide on the launcher because users don't know how to reveal it after removing the bfb from the panel ;) [13:46] we should be proud of our technologies [13:46] someone who cannot figure out that right-clicking allows to make it sticky won't find out that right-clicking allows you to remove it again [13:46] chrisccoulson: and that, too [13:46] * desrt wants accordian by default on a 2560x1600 monitor. make it happen, pitti! [13:46] pitti, you can dnd to the bin icon [13:46] natty was really good [13:46] pitti: yeah, that's a good point [13:46] anyway, I've been through discussions like this several times [13:47] i dislike the idea of adding icons for newly installed apps to the launcher by default [13:47] I'd just appreciate if that anonymous "it has been found that..." had some more concrete data [13:47] 10 minutes of user testing is not what would give you a good balance on decisions like that [13:47] pitti: it's an option on software-center, ask mvo/tremolux, but I'm pretty sure you can put it off [13:47] didrocks: yes; you can put off everything [13:47] but nevertheless, the kidn of user who doesn't know how anything in the desktop works won't find that either [13:48] it basically turns the launcher into a really bad start menu [13:48] I agree [13:48] is the dash really that bad? [13:48] i mean.. if user's can't find stuff in the dash... [13:48] no it's not [13:48] maybe instead of shoving more and more things into the launcher, maybe it should be easier to find stuff in the dash? [13:48] the bfb, the control center, and this bug is basically an outright capitulation that the dash is too hard to use by that ominous "it has been determined that..." [13:48] maybe have the newly installed app show up in the front page of the dash as soon as you open it or something [13:48] desrt, well, the dash has the same issue than any menu or dash like feature, how do you find an icon for something new in the middle of 100 icons [13:49] so why don't we make that easier to discover? [13:49] with the search string pre-filled [13:49] seb128: so now we turn the launcher into exactly that [13:49] indeed [13:49] and the launcher is even harder since you have to scroll it [13:49] I like the idea to have new stuff showing at the end of the "active" list [13:49] you need to do a better job of training users to use the dash to search [13:49] that's what i.e iOS do [13:49] didrocks, pitti, chrisccoulson: you can shut the auto-add to launcher behavior off in Software Center by unchecking "View"->"New Applications in Launcher" [13:50] tremolux: yes, I know [13:50] when you install something on an iOS device it adds an icon on your workspace [13:50] my recommend: when you install a new app, the dash icon gets a pulsing blue glow or so [13:50] and when you open it, the search field is pre-filled and the icon for the new app is there [13:50] seb128: the iOS or android app launchers have more space than just a single vertical row [13:50] its' much more comparable to the dash layout [13:50] pitti, well the principle is the same, give you an obvious way to start what you installed [13:51] * desrt contributes that piece of feedback to the bug [13:51] but yeah, I'm not sure it should "stick" there, maybe just show after install until you run it once [13:51] pitti: you need to spend more time with an iPhone [13:51] once you used it once it will show in "most recently used" [13:52] well, the ipod nano I've seen uses a full-screen multi-page app selector [13:52] well it collects everything installed === greyback|lunch is now known as greyback [13:52] but anyway, even if iOS does it exactly like proposed in bug 955147 that doesn't make it any better for me :) [13:52] Launchpad bug 955147 in unity "[FFe, UIFe] Automatically add launchers for newly installed applications" [Undecided,New] https://launchpad.net/bugs/955147 [13:54] pitti, desrt: btw I think having the launcher "accordian" effect by default was a "feature", not a "bug", i.e something they wanted to show by default [13:54] seb128: that's hilarious [13:54] desrt, why? [13:54] because that launcher works so much better with 30 instead of 9 icons.. [13:54] are you kidding me why? [13:54] with 30 it entirely loses its functionality as a task switcher replacement [13:54] ya. what he said. [13:54] pitti: there is no way that iOS does that [13:55] not to mention your muscle memory goes to crap because the scrolling effect moves the icons around all the time so you have no way to get used to "this is where the firefox icon is" [13:55] dobey, they do add everything you install on your background [13:55] but I guess I give up; I argued about the single control center icon for over three days back in oneiric [13:55] seb128: yes, but not the dock [13:55] dobey, is "dock" their 4 icons at the bottom? [13:55] seb128: and we do that already. everything is in the dash [13:55] seb128: yes [13:55] well maybe we should have the dash always open on the background... [13:56] right now a default configuration already takes much more space than GNOME 2 ever did [13:56] seb128: think of processor caches [13:56] seb128: practically by the laws of physics, the bigger they get, the slower it is to access them [13:56] it's really the same concept [13:56] ok, not with the menu bars [13:56] seb128: we should just have one big button that does whatever it is you want [13:56] dobey, ;-) [13:56] Sweetshark: throwing libreoffice_3.5.1-1ubuntu1_source.changes launchpadwards [13:57] desrt, pitti: note that I didn't say stuff show "stick" on the launcher, but I like having them with maybe a special look until run once [13:57] seb128: that's even worse, in my opinion [13:57] then they are in "most recently used" [13:57] I guess another option is to autostart them for you after install [13:58] because it was there and now it's gone [13:58] so they would be added in "recently used" for you as well :p [13:58] which is just annoying/frustrating... "where did it go?" [13:58] desrt, well same difference as anything else you run [13:58] autostart after install is not a bad idea [13:58] desrt, when you close it the icon is gone [13:58] same argument like for killing the super-useful intellihide, I guess [13:58] with confirmation, of course [13:58] I will not comment on that, I hate *hide ;-) [13:59] it makes the launcher useless [13:59] you clearly have a too big monitor :-) [13:59] I clearly don't see launcher progress bars or count when they are hidden [13:59] and I've difficulties to remember what is running when it's not on screen [14:00] so I often look to the launcher as an indicator of what is open [14:00] but I guess different people work differently [14:00] anyway, I understand the desire to see it all the time [14:00] but I don't understand how autohide is any better than intellihide [14:00] the sabdfl did a detailed post about that on the ayatana design list [14:00] pitti: autohide is just generally annoying [14:01] it basically comes down to "some people don't understand the intel behind intellihide and gets very confused trying to understand what happens" [14:01] but yeah [14:01] yes, I know [14:02] so your user testing is telling you something quite loudly [14:02] hard to conciliate everybody's need in one desktop :-( [14:02] well, one of two things [14:02] 1) the dash sucks [14:02] 2) people really don't understand the dash at all [14:02] desrt: hang on [14:02] i suspect it's #2 [14:02] desrt: _some_ people don't understand it at all [14:02] searching to launch application is a bit of a new concept [14:03] there's demonstrably enough who do [14:03] yes. of course. [14:03] pitti, hi [14:03] even my sister figured it out, and heck, you need to talk to her about computers like you talk with a dog [14:03] but a substantial enough portion of users are getting stuck that you keep getting asked to add more stuff to the launcher [14:03] as far as i can tell, "add more stuff to the launcher" is a way to prevent people from ever having to open the dash [14:03] but that kills itself [14:04] because then we'll have hordes of users who don't figure out how to make the launcher smaller, and we need a new launcher with "active tasks" [14:04] pitti: your sister figured it out, i bet, because you told her [14:04] it's not a hard concept to understand [14:04] desrt: no, I didn't; in fact she didn't even warn/told me that she was upgrading to natty [14:04] "click here". "search" [14:05] she told me the other day "that's a cool new thing you have there" [14:05] ah. okay [14:05] of course she did not know about the WIndows key [14:05] that's what I told her [14:05] she opened it with the mouse [14:05] so she's one of the lucky ones :) [14:05] i still maintain that the user testing is showing that there is a good chunk of people who can't find stuff using the dash [14:06] and i think the only thing that is missing there is a short explanation [14:06] or some hints [14:06] well, I still think that our user testing has a rather large bias towards "first-time" users [14:06] my kids figured out the search stuff, without even skipping a beat [14:06] they just need to be told "click here, start typing" [14:06] but i did have to explain to them how to show the launcher [14:06] if you keep killing long-term efficienty for first-time ease, you inevitably ruin the net sum of user happiness [14:07] we need one of those "Welcome to Unity 98! Would you like to take a tour?" things that pops up the first time you login :) [14:07] everyone loves those [14:07] yes; at that point that would make me much happier than stuff like that [14:07] because it fixes the problem that we have -- the first five minutes, not the years after [14:08] indeed [14:08] can we add a startup music as well? please please :) [14:08] i hate those boxes, though [14:08] because they're not the first five minutes of using unity [14:08] they're the first five minutes after every new install :) [14:08] although i guess that's only a problem that bothers the ultra-nerdy [14:08] since most people are not reinstalling every few months [14:09] fwiw, android has some awesomeness here [14:09] and those who do reinstall every month keep their $HOME [14:09] out of a factory install (at least on the nexus series) you have this little dude on your home screen [14:09] pitti: thanks [14:09] and he's like "click me" [14:09] but still, even as a nerd it's faster to click away the intro than finding where to disable all those stupid settings again [14:09] and each time you click him, he presents a new tip -- just a sentence or two. there are about 6 tips [14:10] telling you how to add stuff to the home screen, how to delete things from it, etc [14:10] and the last tip is "okay. now delete me by dragging me to the trash" [14:10] i. e. 'don't fill up my launcher', remove the unnecessary ones, finding how to disable the "add automatically", enable autohide, etc. [14:10] ricotz: pitti just threw 3.5.1-1ubuntu1 launchpadwards, so you have green light for backports ;) [14:10] having an icon like that on the launcher by default would be cool [14:12] the dash may be a bit undiscoverble, compared to gnome activities for example. you have an "Activities" label up there, and i think it's pretty obvious that that's where the action is. and once you click it, everything is pretty clear [14:14] s [14:16] pitti: even better.... we should bring back the first-time dialog from the GNOME 1 days [14:16] Are you a [x] Beginner [ ] Intermediate [ ] Advanced user? [14:17] [x] Here only for 15 minutes of user testing [14:18] jbicha wanted to add a launcher icon for the ubuntu user guide previous cycle [14:18] but I guess people wouldn't like that either :p [14:18] desrt, yeah, maybe the "small guy walking around the screen and giving you some hints" would be cool [14:19] < your apps are there [14:19] ^ your menus are there [14:19] ^ your im contact are there [14:19] type that key for accessing the dash [14:19] the help is added by default in the gnome-shell dash [14:20] i'm a bit surprised you guys didn't tackle this issue already [14:20] gnome-shell and unity are similar enough.... [14:20] desrt, what issue? [14:20] but like... sort of unlike most other desktop interfaces [14:21] the issue of people not being used to searching for apps on their desktops [14:21] well it's hard to "tackle" [14:21] i think we have this phobia of first-run screens/features [14:21] probably it's well-justified [14:21] also one issue is that the search is,was not smart enough [14:22] but it would go a long way to treat this issue (which clearly _is_ an issue) [14:22] that's partially being fixed with the new keywords support [14:22] not sure that's enough [14:22] like people have an hard time finding stuff because they type the wrong work [14:22] like typing "jpg" will not list eog [14:22] or photo will not list "image viewer" [14:22] that's tough :) [14:22] or "picture" [14:23] so "photo" brings up shotwell for me [14:23] (in g-s) [14:23] and "picture" eog [14:23] presumably due to some keywords in the desktop file or something [14:23] desrt, yeah, I'm trying to make up example but I suck at it [14:23] seb128: well, it's sort of a good point, actually [14:23] desrt, well as said we added keywords this cycle [14:23] because "photo" and "picture" are practically the same word [14:23] eog is one example where we did it [14:23] and i get two completely different apps for them [14:24] desrt, "jpg" does list anything? [14:24] not in the shell [14:24] neither in unity [14:24] it's not a keyword [14:24] i'm not sure it should be [14:24] desrt, same, how do you get "excel" [14:24] people have an hard time finding their stuff, they open the dash and type for what they know [14:24] do the new keywords include stuff like "watch", "view", "edit", "write" and so on? [14:24] i.e excel :p [14:24] seb128: to some extent, the ability to browse is important [14:25] topyli, Keywords=Picture;Slideshow;Graphics; [14:25] that's the eog ones [14:25] * desrt almost thinks that we should take the other approach [14:25] seb128: ok, no verbs or "tasks" [14:25] topyli, the specs doesn't define the content [14:25] find a way to collect usage data of what people are searching for in the dash [14:25] take the top 1000 items [14:25] it's just a ";" separated list of "words" [14:25] figure out what they should ideally do [14:26] seb128, should locate index /dev ? [14:26] make a big list..... [14:26] mhr3, no [14:26] desrt, yeah, we would need a way to let people opt in to participate for providing those datas [14:27] anyway... [14:27] i think the problem is more fundamental than that, even [14:27] i think it simply doesn't occur to people to click on the dash icon and type [14:27] seb128, yet there are people for whom it's indexed [14:27] so who cares what the results are? they aren't seeing them :p [14:27] seb128 (dunno if it's you to tell): FYI gnome-online-accounts 3.4 has facebook support if you build with --enable-facebook. it is not (yet) enabled by default because GNOME board wants lawyer to read the terms of service first. [14:27] xclaesse, talk to kenvandine rather [14:28] kenvandine, ^ [14:28] desrt, right... [14:28] * kenvandine reads back [14:28] seb128, any idea what could cause that? [14:28] mhr3, maybe they have a symlink to /dev/something ? [14:28] xclaesse, i'll look at it [14:28] kenvandine, will be in next release [14:29] mhr3, /dev is type devtmpfs which is in the updatedb.conf PRUNEFS list here [14:29] mhr3, maybe ask them for "mount" output to see if they have a different fs type for it or something? [14:30] xclaesse, ok [14:30] seb128: btw: something that may interest you slightly (but nothing has changed) [14:30] kenvandine, dunno what legal procedure ubuntu has for this, but I would personally love if that could be enabled in ubuntu's package :) [14:30] seb128: it's now official that with gsettings overrides, files that come "later" have more power [14:30] ie: 20_foo will override 10_bar [14:30] kenvandine, at least now we have the 3 most popupar web services in one place [14:30] xclaesse, not sure i want to enable that for the lts... [14:31] (google, windows live and facebook) [14:31] desrt, excellent, thanks [14:31] :) [14:31] seb128: also: all files should be called like nn_* now [14:31] you have one that is not [14:31] desrt, btw related topic that came earlier, is there a way to have "default settings by session type"? [14:31] kenvandine, does ubuntu-unity use GOA? [14:31] (again -- no changes in code -- just updated in the docs) [14:31] seb128: yes. there are two ways and i told them both you before [14:31] *to you [14:31] you didn't like either one :) [14:32] xclaesse, nope, but if you login with GNOME you get it [14:32] pitti, ping [14:32] kenvandine, what does it change that it is LTS? [14:32] seb128: one is to have different gsettings schema directories per session and use GSETTINGS_SCHEMA_DIR (and either install twice or symlink all schemas) [14:33] at least all schemas that may have overrides... [14:33] xclaesse, since it isn't enabled by default, and facebook being a web service that isn't always easy to maintain over a long period of time... [14:33] tkamppeter: hello [14:33] seb128: the other is using a dconf database to do the same and have a 'unity' and 'gnome-shell' dconf profile [14:34] xclaesse, just a big commitment if we enable it and GNOME ends up not really supporting it down the road [14:34] (gnome does not support anything 6months later anyway) [14:34] desrt, hum, right, I don't really like either of those :p I think I will delay that to UDS, next cycle, no hurry to do it [14:35] kenvandine, anyway, up to you ;) [14:35] seb128: so an interesting option could be to make a modification to the schema compiler to deal with this [14:35] desrt, it came because xubuntu overrides the update-notifier "auto launch" key and mpt was wondering which that key had a wrong default value for him [14:35] kenvandine, what is implemented in GOA is the officially supported and documented way from facebook ;) [14:35] pitti, can you upload CUPS to Debian and Ubuntu? I have fixed a crash. [14:35] seb128: it could look at the override files in a particular directory and source the schemas (only the ones referenced from those overrides) from another directory [14:35] xclaesse, yeah, i'll think about it [14:36] tkamppeter: please upload to ubuntu as -1bzr1 or so (don't commit that release to bzr) [14:36] xclaesse, thx for pointing it out [14:36] desrt, I like that one better ;-) [14:36] tkamppeter: I'd like the current version to go into testing again, the previous migration was a while ago [14:36] seb128: then you could set the GSETTINGS_SCHEMA_DIR to the location of the overrides [14:36] and it would check there, then the normal system directory [14:36] desrt, that's basically what we were doing for gconf [14:36] pitti, I already committed to BZR today in the morning. [14:37] so the unity session could set GSETTINGS_SCHEMA_DIR=/usr/share/unity/gsettings-overrides or so [14:37] then you could just do nothing for gnome-shell (ie: use pure upstream) [14:37] desrt, I like that one ;-) [14:37] seb128: it's interesting. i'll think about it some more. [14:37] desrt, I will remind you at UDS ;-) [14:37] it would only require some rather minor tweaks in the compiler for it to work [14:38] you could do it yourself already with a few symlinks :) [14:38] ie: have your overrides package install the overrides files plus symlinks to the appropriate schemas in the normal system directory [14:38] you'd just need to make sure the glib trigger updated both directories [14:39] pitti, I have also a new cups-filters package today or tomorrow (most important: bug 955553). Should I commit and ask you to put in both distros or better upload Ubuntu-only. [14:39] Launchpad bug 955553 in cups-filters "12.04 Beta 1: Brother MFC8680DN cannot print text file" [High,In progress] https://launchpad.net/bugs/955553 [14:39] i should really spend next cycle doing dconf/gsettings work :P [14:39] ;-) [14:39] olli has been asking us all what we want to do... [14:39] * desrt writes that down [14:40] desrt, pitti: btw about that s-c unity launcher bug, one option would be to have the icon in the launcher during install, so you get the "installing" with the progress bar info and then an animation moving the icon to the dash icon? [14:40] desrt: get rich or die trying? [14:41] seb128: that's an interesting proposal. you should write it on the bug. [14:41] hi guys, I'm trying to debug a weird problem with dash search results - when I type 'f' in the home lens, it gives 'full' and 'fuse' entries, which are useless. with help from mhr3 we already confirmed it's because locate has indexed 92 entires from my /dev, including those two [14:41] i'd say the only negative point against that is that if you look away from the screen (or go to get some tea because it's taking a long time) then you will miss the animation [14:41] desrt, just did [14:42] desrt, well if you miss the animation it's not any worth that the solutions where there is no clue at all :p [14:42] seb128: so if we mixed our two ideas... [14:42] ie: the app animates into the dash and then the dash starts glowing [14:42] i think that would be pretty cool [14:42] right [14:42] I've a new install with today's daily iso, standard parition layout. One thing worth mentioning is I'm using sshfs on that box [14:43] anyway [14:43] i should be doing dconf work :) [14:43] desrt, have fun! [14:43] pstolowski_, what does "mount | grep dev" gives you? [14:44] seb128: if you do one thing today, please get my fixed hud into the distro :p [14:44] pstolowski_, "mount | grep "on /dev"" [14:44] desrt, hud? [14:44] desrt, I plan to get xul support into distro ;-) [14:44] seb128: and the hashtable assertion fix [14:45] (Oh, it doesn't do XUL yet? I was wondering why it didn't work in Firefox...) [14:45] desrt, right [14:45] seb128, /dev/sda1 (ext4), udev on /dev and devpts on /dev/pts [14:45] mpt: it will still require a unity fix after this in order to fully work [14:45] pstolowski_, that's missing the info I need [14:46] pstolowski_, what is written after "type" for /dev [14:46] seb128, https://pastebin.canonical.com/62547/ [14:46] seb128, btw, same for me as for the other testers in bug 949414, the new package has the same bug. Let me know if I can do any more. [14:46] Launchpad bug 949414 in gtk+3.0 "Some GTK+3 events are not emitted when using a touchpad (but are with a mouse)" [High,In progress] https://launchpad.net/bugs/949414 [14:47] mpt, that's all for now but I will ping you back later if needed, thanks! [14:48] pstolowski_, dunno then [14:50] seb128, ok, thanks. they shouldn't be indexed by locate, right? [14:51] Hmmmm, right-clicking doesn't work any more [14:53] pstolowski_, no, /etc/updatedb.conf has devtmpfs in the PRUNEFS= list [14:53] pstolowski_, but maybe you get something symlinking back to / or something [14:59] seb128, I see, thanks [15:04] seb128, empathy-call segfault with that message: empathy-ERROR **: Missing cluttersink [15:04] seb128, and indeed gst-inspect has no cluttersink [15:04] libclutter-gst is installed of course [15:04] something is broken in its packaging? :/ [15:05] jbicha, ^ [15:05] ricotz, ^ [15:05] thanks for the redirect [15:05] xclaesse: what architecture are you using? [15:05] 64bits [15:05] jbicha, the build log has [15:05] # list-missing files result: [15:05] -./usr/lib/gstreamer-0.10/libgstclutter.la [15:05] -./usr/lib/gstreamer-0.10/libgstclutter.so [15:06] -./usr/lib/libclutter-gst-1.0.la [15:06] jbicha, I guess the .so is important :p [15:06] https://launchpadlibrarian.net/96829417/buildlog_ubuntu-precise-i386.clutter-gst_1.5.4-0ubuntu1_BUILDING.txt.gz [15:06] oh, maybe that explains why empathy-call didn't seem to do anything here ;) [15:07] :D [15:10] jbicha, btw gnome-contacts and gnome-shell should depend on empathy-call [15:10] those 2 does not support old empathy-av [15:11] or recommend maybe [15:13] yeah I think gnome-contacts should recommend empathy-call, of course then Contacts couldn't be on the CD but Contacts doesn't work too well any way [15:14] afaik empathy depends on gnome-contacts nowadays [15:14] * didrocks shows again his dh_install --fail-missing to jbicha :) [15:15] jbicha, in empathy, right-click a contact -> information, that will launch gnome-contacts [15:15] didrocks, how do you deal with stuff like .la that you don't want to install? need to do rules hack to clean those? [15:15] xclaesse: Contacts isn't installed by default, but Empathy will use it if it's available [15:15] jbicha, xclaesse: I think it does install on demand if needed [15:15] kenvandine said it tried and it was working [15:16] seb128: right, you have to add override_dh_install: anyway, so then, I rm; dh_install --fail-missing [15:16] didrocks, would be call if you had a .install.override like the lintian override [15:17] where to list stuff you don't install [15:18] seb128: hum, interesting way to dealing with those. you will still need to override dh_install to use --fail-missing, but we can maybe integrate that to dh8 directly with an env variable rather [15:18] something to look at for next cycle ;-) [15:19] i.e after release rush, at UDS or somethingh [15:20] seb128: right! [15:21] seb128, jbicha, oh this looks like a greater fallout , sorry :\ [15:21] * didrocks notes that somewhere [15:21] seb128, yeah empathy will use pkgkit, but I don't know if that works on ubuntu [15:21] xclaesse, it does [15:21] it will use aptdaemon just fine [15:21] cool :) [15:23] seb128: nothing against moving unity-gtk-greeter ownership to another team (people from xubuntu, lubuntu), making that team part the current lightdm upstream team part of it (to still have commit access)? [15:23] xubuntu/lubuntu want to hack on the code, upload pot files… [15:23] didrocks, lightdm-gtk-greeter you mean? [15:23] oupss [15:23] yeah [15:23] autotypo :) [15:24] didrocks, I've difficulties to parse what you asked for, do you want to move the ownership and give access to the lightdm team? or keep the ownership and those team to the lightdm hackers? [15:25] seb128: create a lightdm-gtk-greeter-hackers team, give ownership of this project/move trunk to them [15:25] and have lightdm-hackers part of lightdm-gtk-greeter-hackers (to still have commit access) [15:26] didrocks, works for me [15:26] didrocks, robert_ancell hacked the ownership change on https://bugs.launchpad.net/ubuntu/+source/lightdm/+bug/908926 [15:26] Launchpad bug 908926 in lightdm ""Large Font" style option does not work" [Undecided,Confirmed] [15:26] ups, "acked" rather [15:27] didrocks, anyway please do ;-) thanks [15:27] seb128: great, will do the launchpaderies* now then! (hoping I have access to change ownership) [15:28] didrocks, if you don't let me know, I've access [15:29] Sweetshark, thanks, will try to get to it later [15:31] seb128: everything went smootly, no need! :) [15:31] smoothly* [15:31] didrocks, excellent [15:32] pitti, did you see my last messages? [15:35] hiya [15:35] I had a couple of questions about lightdm before I wander off and file bugs that have already been filed, etc :) [15:35] how do I "enable" the unity-greeter background switcher? for some reason it doesn't seem to work for me - is there a bug about it? [15:36] dholbach: do you use an encrypted disk? [15:36] also sometimes it seems after waking up from suspend the session session doesn't come back up - I have to stop/start lightdm in the console [15:36] didrocks, no [15:37] pitti, please do not upload current BZR snapshot from CUPS. I have to modify it as Mike has posted an upstream fix which is different. [15:37] tkamppeter: ack; FTR, I'm not really "here" today, I'm on a sick day [15:37] * dholbach hugs pitti [15:37] pitti, I hope you're better soon again [15:37] pitti, and shut down that computer! [15:37] should be much better on Wednesday or so [15:37] dholbach: need to do a pygobject upstream release :) [15:38] * dholbach rolls eyes [15:38] dholbach: did you see that dedication? :) [15:38] hard code freeze and tarballs due midnight today [15:38] didrocks, you should have asked jasoncwarner_ ;-) [15:38] heh, indeed :) [15:40] seb128: hey [15:40] desrt, hey [15:40] seb128: if i give the 'dconf compile' capability do you still plan to use it this cycle for the guest session? [15:41] dholbach, what background to you use? a personal image? it works only for system backgrounds [15:42] seb128, ah, it was a personal image [15:42] desrt, I might yes (not at the top of my list but I don't rely much relying on the gconf->gsettings migration to have screen locking stopped) [15:42] nevermind then - see it's good I asked :) [15:43] seb128: just wondering if i should bother tucking it into this dconf release [15:43] or wait for next cycle to do the work [15:43] desrt, it can wait next cycle I think [15:44] desrt, too much locking is not a security issue and the current version works fine [15:44] sure [15:44] desrt, worth thing we can maybe sru the compile stuff [15:44] punted [15:44] but let's see next cycle [15:44] cool [15:44] i'm starting to think dconf could use some substantial refactoring in the next cycle anyway [15:45] it's not code that i am particularly proud of :p [15:45] it was written with too much of an eye toward efficiency [15:45] so it does really bad things from a software engineering standpoint for some very minimal savings [15:45] result: ugly hard-to-maintain code [15:45] right [15:45] didrocks, seb128, did you hear anything about the unlock dialog not coming up after waking up from suspend? [15:46] dholbach, no, what do you mean? it doesn't lock your screen when it should, or you need to press a key to get the dialog to display or ...? [15:46] I need to Ctrl-Alt-F1 stop lightdm; start lightdm [15:46] not every single time, but maybe every 2nd time [15:47] dholbach, lightdm has nothing to do with locking [15:47] hum, well I get a black screen and a moving mouse pointer, but no unlock dialog [15:47] dholbach, can you kill gnome-screensaver if you go on a vt? does it solve it? [15:48] I'll try next time [15:48] thanks [15:48] thanks for the tip seb128 [15:48] but no, not knows [15:48] dholbach, yw ;-) [15:48] ok [15:48] hrmph. [15:48] Background=/home/desrt/.cache/gnome-control-center/backgrounds/6b61801e8ab18a61264418bfaa9dca68cff4e3c89933fc5e8641dd9ae274cd6a [15:49] looks like the accountservice background patch could use a bit more tweaking... [15:49] even with a public image file used, this happens [15:49] desrt, what's wrong? [15:49] seb128: that file is not readable by lightdm [15:49] even though the actual file i set for my background is [15:49] desrt, right, but that's the accountservice patch [15:49] that's *not* [15:50] that's the file used for your background no? [15:50] no [15:50] i.e we cache the backport to avoid rescaling [15:50] background [15:50] doh [15:50] ya. that's the problem [15:50] the account service should know about the original file [15:50] but instead it gets told about the rescaled cache file in a private directory in my ~ [15:50] ie: where it can't possibly access it [15:51] it's an .svg file -- that might be why this happens [15:51] not svg specific I think no, there is a bug open about it [15:51] ah. good. [15:52] as long as it's a known issue [15:52] where's the bug? [15:52] desrt, I'm not sure it's a bug, nobody took a stand on whether that's one or not so far [15:52] desrt, I think part of it is to not leak the path and filename infos [15:52] https://bugs.launchpad.net/bugs/928553 [15:52] seb128: it prevents me from seeing my public background image at the login screen [15:52] Launchpad bug 928553 in unity-greeter "Non-default wallpapers are not readable by Unity-Greeter" [Low,Triaged] [15:53] seb128: this is not the same issue [15:53] desrt, right, I guess somebody needs to come with a way to solve that without leaking infos that shouldn't go public [15:53] again: the original image file is in a _public_ folder (/, of all places) [15:53] it's world-readable [15:54] desrt, it is the same issue? [15:54] desrt, hum, we used the same bug for several similar issues in fact [15:54] so interesting to note -- this may in fact be a bug in the control centre [15:55] desrt, https://bugs.launchpad.net/ubuntu/+source/unity-greeter/+bug/918934 is your issue I guess [15:55] Launchpad bug 918934 in unity-greeter "Doesn't show manually added wallpapers from "Picture Folders" category (dup-of: 928553)" [Low,Confirmed] [15:55] Launchpad bug 928553 in unity-greeter "Non-default wallpapers are not readable by Unity-Greeter" [Low,Triaged] [15:55] or rather some misunderstanding between the control centre and the accountservice patch [15:55] desrt, we dupped it because I think it's basically the same bug [15:55] seb128: i don't think so. there's something much more complicated going on here [15:55] but maybe they should be undupped [15:55] gsettings has the private copy stored as the background image [15:56] right, that's what we want [15:56] the problem is that the accounts service assumes that the key in gsettings is appropriate for publishing to the accounts service [15:57] the .cache 700 permission was overlooked there I think yes [15:57] desrt, not sure what would be the right way, copy your image in /var/run/lightdm? [15:58] store it in ~ rather than ~/.cache (though that feels wrong) [15:58] seb128: use the original image file [15:58] desrt, then you have to redo the scalling at every boot and loose 1 second of boot time [15:58] if i manually set it to that it works fine [15:59] that's the reason we got that caching to start with [15:59] seb128: in my case that has to happen anyway since lightdm picks a different screen resolution than i have at my desktop [15:59] and it's an SVG, so... for all we know rendering the SVG at any resolution is faster than decoding a full-screen jpeg [16:00] hum, k [16:00] also: it doesn't really add a second to the boot because (a) if the login screen shows then the slowest thing happening there is waiting for the user [16:00] and (b) the background image loading happens in a thread and doesn't block the loading of the login screen [16:01] right [16:01] mterry, ^ what do you think? [16:01] desrt, I guess we only need a "check if the selected image is publicly available and set the filename only in this case" to not leak private paths [16:02] indeed [16:02] really, the checking should be very very simple [16:02] 1) no cache logic on the before-gsettings side of thing [16:02] 2) only push paths into the account service if they are not in my homedir [16:03] is the control-center caching thing upstream or us? [16:04] because it causes some flaky behaviour... [16:04] * mterry reads [16:04] ie: if i set a background image and then come back into the panel i find that my image is not selected in the list [16:05] seb128: can you binNEW unity-2d btw? I added a -common package [16:05] and if i select it again, i don't see the original filename but rather the checksum blob [16:05] didrocks, looking [16:05] desrt, 2) is wrong [16:05] desrt, like ~/Images is public readable I think [16:06] seb128: right. so maybe some better logic is required. [16:06] desrt, the caching is done by gnome-desktop I think [16:06] seb128: like iterate paths all the way from / to the file and make sure each step is public readable [16:06] seb128: thanks! [16:06] didrocks, yw! [16:06] seb128: and written back into gsettings from gnome-settings-daemon? i hope not :( [16:07] desrt, hum, maybe not, didrocks knows that part better, he did it ;-) [16:08] so there is gnome_bg_save_to_preferences but gnome-settings-daemon never calls it [16:08] this is a job for dconf blame :) [16:09] How would g-s-d get access to the original path instead of the cached path? [16:09] mterry, does gsd needs to? isn't g-c-c writig to accountsservice? [16:09] mterry: why not just have the control center be the one that pokes the account service? [16:09] at the time you elect the image [16:09] select [16:10] seb128, no, g-s-d is (to handle cases where non-g-c-c apps are writing to gsettings) [16:10] i think the first step is to make g-c-c not write the cached filename [16:10] because that's causing problems for g-c-c itself [16:10] There are other apps that let the user set the wallpaper [16:10] yeah, like eog [16:11] Firefox [16:11] meanwhile g-s-d has its own second copy of caching logic [16:11] so... i don't know why it would ever even do that in the first place [16:12] guilty: 04_new_appearance_settings.patch [16:12] or perhaps i misunderstand patch? [16:12] s/patch/quilt/ [16:13] desrt, in g-c-c? /me looks [16:14] i misunderstood. quilt stores an entire copy of the original file using the name of the patch -- that file contains the caching logic [16:14] cc-background-panel.c [16:15] looks like an upstream problem, really [16:18] it looks like logic to deal with downloading background images from http urls... [16:18] except it happens for local files too [16:22] yay, my first GNOME tarball release [16:22] pitti, congrats! [16:23] \o/ pitti [16:30] * desrt goes insane [16:31] * pitti AFK [16:31] phew, just worked off some of the unhealthy food i ate at the weekend [16:33] mterry, seb128: so interesting discovery [16:33] ~/.config/gnome-control-center/backgrounds contains a last-edited.xml [16:33] it has containing the cache filename [16:34] and containing the original filename [16:34] I'm totally bringing up an old conversation, but my wife yesterday tried typing "crop pictures" into the Dash yesterday which doesn't work [16:35] desrt, hmm, so g-s-d could check that file when the gsettings changes and use the original... [16:35] mterry: i still think we have a g-c-c bug on our hands, actually [16:35] source_url should only be set for http:// files [16:35] somehow it is set for local files as well in some cases [16:35] so I guess we'll need keyword verbs next cycle then [16:36] jbicha, well nothing stops you to put "crop" into the keywords? [16:37] I told her she could fix it too but she didn't seem too interested :) [16:39] okay. found the issue [16:39] the copying happens when you 'add single file' [16:39] cjwatson, stgraber: FYI, uploading new pygobject 3.1.92 now; this needs a small test fix in ubiquity, https://code.launchpad.net/~pitti/ubiquity/indicator-soname/+merge/98230 [16:40] ack [16:41] cjwatson: are you already merging that change? (I was just about to) [16:41] pitti: can we make this work both before and after somehow, at least for a while [16:41] ? [16:41] pitti: we're having an installer sprint at the moment and a new build-dependency for the tests would be inconvenient right at this poit [16:41] *point [16:42] cjwatson: I guess we can store the return value to a variable, then depending on len() do the assert or not (and assign the variables properly)? [16:42] mterry: so it looks like this happens in order to avoid the file that you add being deleted later or something [16:42] hello all! just now I made an automatic dist-upgrade, and I noticed (too late) that unity-2d was removed. Now I'm trying to re install it, and I'm getting: [16:42] The following packages have unmet dependencies: [16:42] unity-2d-panel : Depends: unity-2d-common (= 5.7.0-0ubuntu1) but it is not installable [16:42] ... [16:42] E: Package 'unity-2d-common' has no installation candidate [16:43] is this known? can I workaroung somehow? [16:43] stgraber: yeah, that was my thought - I could take care of that [16:43] workaround* [16:43] desrt, sure, makes some sense... or moved or whatever [16:44] cjwatson: sure, have fun :) I'm digging into more persistent usb stick weirdness [16:44] And the cached bg is under ~/.cache which is 0700... [16:44] mterry: i'm going to file a bug about this upstream [16:44] mterry: meanwhile, it's a pretty simple patch to fix it, i think [16:44] desrt, by fix it, you mean stop caching? [16:44] it's not a cache -- it's a straight copy [16:45] desrt, k. But problems-it's-causing-us aside, seems like a feature not a bug? [16:45] pitti, ping [16:45] mterry: maybe? [16:45] nessita: that's arch skew, it happens when amd64 builds before i386 where there are dependent arch:all packages needed [16:45] if it's a feature, it's a rather poorly thought out one, and implemented in a buggy way, in my opinion [16:45] fair [16:45] because it ends up dumping the cache filename (checksum blob) into the user interface [16:45] which is obviously wrong [16:45] eww yeah [16:46] pitti: I'd prefer something like http://paste.ubuntu.com/890906/ - OK? [16:46] micahg: ack, thanks, so shall I just wait for the rest of the builds to finish? [16:47] nessita: yeah [16:48] pitti: if so, I'll apply to a local branch and merge [16:54] pitti: ... done [17:04] mterry: okay. got a patch. [17:04] https://bugzilla.gnome.org/show_bug.cgi?id=672405 [17:04] Gnome bug 672405 in Background "'source-url' set on single image files causes problems" [Normal,Unconfirmed] [17:04] desrt, neat. want me to apply it to our package? [17:05] i want you to test it first :) [17:05] i tested it and it works here [17:05] but try to duplicate the problem first: [17:05] go into control center -> background. click the [+]. navigate to and select a file from /usr or something [17:05] desrt, yar sure :) [17:05] verify that ~/.cache/ is in the account service [17:05] apply patch, try again [17:20] mdeslaur, there? [17:20] seb128: yeah [17:21] mdeslaur, http://pastebin.ubuntu.com/890942/ [17:21] mdeslaur, does that seems right? [17:21] seb128: no [17:21] mdeslaur, why not? [17:23] seb128: in lock_screensaver(), we are checking "ubuntu-lock-on-suspend", but in that location (when the lid is closed), we actually want to check "lock-enabled" [17:23] mdeslaur, right, I did undo that change [17:23] seb128: oh, wait a sec... [17:24] mdeslaur, the patch I suggest replaces yours [17:25] mdeslaur, hum, right that patch is wrong... [17:25] seb128: so, in upower_notify_sleep_cb, we need to check "ubuntu-lock-on-suspend", but in do_lid_closed_action, we need to check either one or the other depending on what the user setting is [17:26] desrt, can confirm original bug and that your patch fixes it [17:26] seb128: I can look at if you want, just check it in and I'll poke at it [17:27] mterry, desrt: i'm about to update g-c-c, I can sneak that patch in [17:28] seb128: please do [17:28] ok [17:28] seb128: the upstream fix is wrong anyway, since it will lock all screens even if an external monitor is plugged in I think [17:29] seb128: i should mention that this patch has a side-effect [17:29] mdeslaur, I'm getting very confused by all the screen locking discussion, I'm not sure what we are supposed to do and when now ;-) [17:29] seb128: the control center adds the pictures from the cache directory to the "Pictures Folder" section [17:29] which ... is dumb [17:30] in any case, having it not copied to that directory means that it won't appear there [17:30] seb128: hey there! I'm currently working on the ubiquity code dealing with keyboard layouts. I'm setting org.gnome.libgnomekbd.keyboard layouts to an array of valid layouts (maximum 4 of them) [17:30] but it also prevents the issue with it having appear there with a checksum blob for the filename [17:30] seb128: but apparently I need to kill gnome-settings-daemon for it to actually get applied, what am I missing [17:30] stgraber: could be a g-s-d bug -- it's not properly listening for the change? [17:31] stgraber, dunno, I've 5 people highlightning me and I didn't make any progress in the 15 updates on my list due to constant ping, I've no clue about xjb [17:31] xkb [17:31] seb128: leave the original patch as-is, and I'll fix it for you :) [17:31] mdeslaur, I will commit a broken version now, if you could have a look I will appreciate, I will get going on g-c-c meanwhile [17:31] mdeslaur, just commit your update to the vcs if you want, that would be great [17:32] seb128: sure, be glad to [17:32] mdeslaur, pushed [17:33] desrt: it looks like it gets confused when you set a list that doesn't contain the current layout, it doesn't switch to the first one in that case and sometimes the list in the indicator is wrong too [17:36] apparently going outside was a very bad idea :( [17:37] dobey: were there other humans out there? :( [17:40] and pollen [17:41] lots of pollen [17:43] yuck [17:46] seb128: ok, commited untested patch update, that should be good [17:46] mdeslaur, thanks [17:47] seb128: np [17:51] mdeslaur, I'm not sure to understand the logic, why not just calling lock_screensaver() if the key is set? [17:55] seb128: because then lock_screensaver checks "lock-enabled" and bails out if that's not set [17:55] mdeslaur, oh ok [17:56] seb128: if someone suspends, we want to check "ubuntu-lock-on-suspend" only [17:57] seb128: repeating the few code in upower_notify_sleep_cb isn't the most elegant way to do this, but I figure it will make the patch a lot less intrusive for future updates [17:58] mdeslaur, right [18:12] * didrocks waves good evening [18:12] didrocks: goodnight [18:12] desrt: thanks, you too! :) [18:14] kenvandine, indicator-messages shows new IM's as my name (not the user who left the IM) my guess is it would be reported against indicator-messages but im not sure if thats empathy causing the issue or indicator-messages [18:14] hey bcurtiswx [18:14] it is telepathy-indicator [18:14] already reported [18:14] kenvandine, OK, thanks. I'll make sure to look there now. [18:15] bcurtiswx, it is on my todo list for today [18:34] desrt, is that gcc patch enough to fix any of the unity greeter issues? [18:34] seb128: no. [18:34] seb128: it's enough only to fix the issue of setting a world-readable file [18:34] the issue of what to do if the user sets a private background is still open for debate :) [18:35] desrt, well that's "an" issue, so does it fix any of the issues -> yes [18:35] desrt, thanks [18:35] oh [18:35] i read 'all'. sorry :) [19:41] kenvandine: pitti seb128 got a nice bug for you [19:41] smehow i updated 30 minutes ago [19:41] and now the program i was hacking on cant run anymore [19:41] ERROR:/build/buildd/pygobject-3.1.92/gi/pygi-marshal-from-py.c:1071:_pygi_arg_to_hash_pointer: code should not be reached [19:44] ugh [19:45] seiflotfy, it's a bug for pitti, but I think he's out of the evening; maybe downgrade to be able to resume hacking until it's fixed [19:46] seiflotfy, opening a bug with a testcase might be useful [19:46] which package should be downgraded [19:46] and why is it using pygobject 3.1 [19:47] seiflotfy, tbh I would assume it is upstream pygobject bug [19:47] xclaesse: i am not sure [19:47] it just worked like 30 minutes ago [19:48] yeah was working for me too [19:48] oh its the version number [19:48] * xclaesse don't remember what package got update [19:48] seiflotfy, what did you upgrade? python-gi [19:48] seiflotfy, yeah pitti synced from debian [19:49] oh [19:49] seiflotfy, that got updated to .92 today [19:49] yeah [19:49] where can i find the old package ? [19:49] seiflotfy, what arch? [19:50] 386 [19:51] seiflotfy, https://launchpad.net/ubuntu/+source/pygobject/3.1.1-0ubuntu1/+build/3227491 [19:51] the debs are on the left [19:51] * desrt indulges in debfoster [19:52] thanks [19:52] :D [19:52] hi desrt :D [19:52] seiflotfy: hey :) [19:55] * desrt wonders if his system is still working [19:56] desrt, not it's not and we can't read you [19:56] wow. working + faster. [19:57] desrt, no, can't read you , still [19:57] somehow apt-get autoremove was supposed to make it so i didn't need debfoster [19:57] but it doesn't really do a great job... [19:59] desrt, well, autoremove let you clean stuff that were installed by a package you uninstalled [19:59] seb128: it lets me clean stuff that was installed when i installed a package that i lated removed, rather [19:59] *later [19:59] ie: not working for stuff that came on the CD [19:59] desrt, that doesn't handle stuff like gedit was depending on gtk2 and is now depending on gtk3, gedit is still installed but I don't need gtk2 [20:00] yes. that case as well. [20:00] it's an hard problem to solve [20:00] it just removed a huge amount of stuff -- like 1.5GB installed size [20:00] waouh [20:00] and i didn't even run it with -i [20:00] that's when it gets really fun :) [20:00] yeah, I've quite some old libs here [20:01] like I noticed recently I had like still 3 so versions of libindicat* [20:01] ya... the apt-get support really has some issues :/ [20:03] then there are some things on the list that are like ... uh... why? [20:04] acpi-support really seems like it ought to be in ubuntu-standard, as an example [21:31] seb128: meh, suspend on lid close is broken for me now, and downgrading gnome-settings-daemon doesn't fix it :( [21:31] * mdeslaur will investigate tomorrow [21:31] mdeslaur, :-( [21:47] cyphermox, I remember that network-indicator being part of unity-greeter got dropped for 12.04. Was that just lack of time or were there hard technical problems? [22:04] jasoncwarner_, hey following up on your horizontal scrolling mouse bug - I see it's been fixed upstream now, and we have pulled that fix into ubuntu. When you get a chance can you update to latest precise and verify that the issue is now solved? [22:04] bryceh: sure...will do and let you know [22:07] mdeslaur: are you using a proprietary driver? [22:08] suspend on lid close is broken with the nvidia driver, it thinks its an external screen so it doesn't suspend [22:39] seb128: poke [22:40] desrt, hey [22:40] seb128: any word on having the control centre set gsettings keybindings yet? [22:40] desrt, no, nobody seems interested by that [22:40] hum [22:41] or interested enough to do the work [22:41] * desrt is interested in the work being done :) [22:41] I tried to ping jbicha and ricotz several times about it [22:41] maybe i can take a look now that XUL is finished [22:42] desrt, I planned to have a look this week but beta2 freeze is thursday so it will be tight [22:42] I count tomorrow to still be busy with GNOME 3.3.92 [22:42] it's on my list [22:42] the g-s-d patch for reverting using gsettings for custom key bindings is ruining my day :) [22:42] desrt, I think it might be easy, like the g-s-d revert can probably be dropped for sure [22:43] desrt, oh, how so? [22:43] I though g-s-d was not linked to wms, etc [22:43] seb128: trying to build gnome-control-center out of upstream git -> crashes when it can't find the keys in the gsettings schema for g-s-d [22:43] I though g-s-d had the multimedia keys and actions only [22:43] ah right [22:43] because your patch reverted them [22:43] yeah, I hit that recently as well :p [22:44] okay. stepping out for the evening [22:44] will take a look tomorrow [22:44] good night [22:44] desrt, so I think that an easy step would be to drop the g-s-d patch and update the g-c-c patch to drop the multimedia keys part [22:44] i.e just keep writting the wm keys to gconf [22:45] that would help, indeed [22:45] better would be to change that code to write to both gsettings and gconf [22:45] desrt, 'night [22:45] ciao