[00:06] robert_ancell: Aha! Looks like the tests were right, but my xserver implementation was relying on uninitialised memory. [00:59] RAOF: oh hai [01:00] RAOF: pointer needed to whomever does battery stuff [01:00] lifeless: What particularly about battery stuff? Colin King would be our resident "please don't chug electrons" expert. [01:00] I'm on battery [01:00] been on it for an hour [01:01] applet shows 'Charged' and no timer. [01:01] Aha. [01:01] So many ways this could be wrong :) [01:01] power control panel says [01:01] 'using battery power - 1h33m remaining' [01:01] Is anything else on your panel updating? [01:01] I want the timer back in the applet :) [01:02] if I change volume the volume icon in the panel changes [01:02] The problem would either be in indicator-power or unity panel; probably therefore in indicator-power. [01:03] The changelog suggests that kenvandine would be your man! [01:03] kenvandine: yo ^ [01:10] tedg would also be indicated for indicators, but he's not here :) [06:07] Good morning [06:08] Riddell: btw, I bought a DisplayPort/HDMI -> DVI adapter to use my new laptop with my old monitor (unless you want to replace your monitor anyway) [06:10] Unless you need an active connector, the DP→DVI cable should be of negligable cost :) [06:10] yes, it was some 20 EUR [06:24] pitti: What's the likelyhood of a MIR or two going through at this stage? I presume it's on the low side. [06:25] RAOF: pretty much depends on what it is for and the kind of package [06:25] we can't avoid some, but we of course want to reduce them to a minimum [06:25] It'd be for google-test and libgtest-xorg; if we want some extra build-time tests in the Xserver. [06:26] that seems fine [06:26] more tests are always good [06:28] * RAOF will have to get onto that, then. [06:35] good morning [06:36] pitti, I'm writing a regression test for bug 877766 - it appears that the root user can have chdir rejected. a) WTF?! b) do you know of other cases I should test for [06:36] Launchpad bug 877766 in lightdm "lightdm login fails with NFS home and strict (mode 0700) permissions" [Medium,Triaged] https://launchpad.net/bugs/877766 [06:36] does that mean you literally can't cd from bash on a system like this? [06:41] robert_ancell: yes, NFS usually uses root_squash [06:41] so root can't do anything [06:41] which absolutely makes sense for a remote fs :) [06:42] pitti, so the kernel knows everything about the filesystem, but it doesn't let the root user know? [06:42] robert_ancell: I think having a test case with a 0700 home dir, and perhaps an ecryptfs one, sounds good; it's a relatively strong case and checks that you don't make too many assumptions about the home dir, especially before PAM unlocking [06:43] robert_ancell: I think it's more like that the NFS server doesn't accept "root" as a remote user [06:43] pitti, oh, I see [06:43] robert_ancell: it's not unlike ssh -- as root on the client you don't automatically have access to all user accounts on a remote server [06:44] so it should only be chdir/read/write that are affected? [06:44] now, NFS allows root more than ssh does [06:44] robert_ancell: as root you can still handle world-readable/writable files [06:44] and dirs [06:44] but as root you are effectively "nobody" on NFS [06:44] aha [06:44] (that's literally the account it uses, I think) [06:45] I was reading the docs on this but it was a little jargony [06:45] I used NFS in my university, but it's been a while === tkamppeter_ is now known as tkamppeter [08:36] pitti: DisplayPort/HDMI -> DVI connector does me no good for a VGA monitor [08:37] Riddell: ah, sure :) I had assumed you used DVI [08:42] You *can* get DP→VGA connectors, but they're active and so cost quite close to a new (crappy) monitor. [08:42] hey [08:43] hey seb128 [08:43] hey pitti, how are you? [08:43] seb128: quite fine, thanks! [08:44] seb128: FYI, retracer crashed again today, on the dupe confusion [08:44] I have some time now, so I'm dissecting this more thoroughly instead of just restarting it [08:44] pitti, ok, I untagged a bug yesterday evening and restarted them [08:44] ah, that's fine [08:44] well them -> the amd64 one [08:46] bug 942891 is a broken one it stumbled over [08:46] and it's quite similar to the other ones [08:47] Hmmm am I the only one that noticed a weird behavior with changing /etc/hosts and dnsmasq? [08:59] pitti, is apport known to be broken btw? [08:59] seb128: broken how? [08:59] pitti, it writes 0 byte root owned files there [08:59] not very useful... [08:59] seb128: if you don't have whoopsie installed, yes [08:59] shouldn't it depends on it? [09:00] seb128: bug 942326 [09:00] Launchpad bug 942326 in apport "Apport creating 0 byte crash files" [High,Fix committed] https://launchpad.net/bugs/942326 [09:00] already fixed in trunk [09:00] seb128: no, just get along with not having it [09:00] pitti, thanks ;-) [09:00] I need this freeze to end :) [09:00] * seb128 installs whoopsie [09:00] seb128: oh, you are one bug ahead of me! [09:01] hehe! [09:01] hm, 6 fixes staged in apport trunk [09:01] pitti, weird that none of your rebuilt from yesterday are listed there [09:01] it still amazes me how much time I spend on fixing bugs in a bug reporting program.. [09:02] is the list clever enough to filter stuff that actual have a receiving side on launchpad ? [09:02] seb128: so I guess it's clever enough to only count bugs which have a task for that source? [09:02] rather than just .changes numbers? [09:02] pitti, seems so [09:02] phew, alt-tap fix now in unity/ppa :) [09:02] I don't know how that works, it's bdmurray's baby [09:02] seb128: anyway, for this morning's rebuild round I changed the syntax, in any case [09:02] pitti, well in any case I've a few fixes in unapproved but I don't think it goes to 6 [09:03] I need to keep working ;-) [09:03] didrocks, lut, nice! [09:03] I'm running staging here today [09:03] I didn't notice any difference with 5.4 yet ;-) [09:03] seb128: salut ;) [09:03] seb128: please update, there are some cases where alt-tab was stuck [09:04] ok [09:04] ok, we won a lot of keywords against the ubuntu: branch, as planned :/ [09:04] the hud still doesn't give back the focus to the app :-( [09:04] * didrocks dig and move them to the right target [09:04] seb128: yeah, I know… [09:04] still on the priority list [09:05] and the notify-osd color stuff is still not fixed [09:05] didrocks, keywords -> I plan to send the GNOME one to GNOME, to get them commited and wait .91 [09:05] didrocks, not sure they are worth adding a patch before .91 [09:06] seb128: hum, no distro-patch? [09:06] didrocks, you can let the GNOME ones to me if you want [09:06] seb128: well, it's fine, I can handle them [09:06] I can stage some for the defaults, and report upstream [09:06] didrocks, well, distro patch if you want, but it seems work over what's useful, at least no upload for those please [09:06] didrocks, ok, your call, thanks ;-) [09:06] seb128: oh, I'll just stage them, no worry :) [09:07] pitti, btw either you misread or count changed by we are tie it seems ;-) [09:10] seb128@ubuntu.com has 191 fixes [09:10] martin.pitt@ubuntu.com has 190 fixes [09:10] seb128: that's on http://reports.qa.ubuntu.com/reports/bug-fixing/precise-fixes-report.html [09:10] oh, fun -- http://reports.qa.ubuntu.com/reports/bug-fixing/canonical-desktop-team-precise-fixes-report.html has 195 for both of us [09:11] right, I look at the desktop list [09:11] desktop team for the win in any case, we secured the first 3 places I think ;-) [09:14] it's also interesting that kirkland has more fixes than most canonical employees; community FTW: ) [09:14] ;-) [09:19] waouh [09:19] new webkit built in the ppa \o/ [09:19] the ldflag seems to have helped [09:20] seb128: which flag did you use? [09:21] micahg, -Wl,--no-keep-memory [09:21] ah, ok [09:21] nice [09:41] good morning everyone [09:45] chrisccoulson, hey, how are you? [09:48] Sweetshark, I reopen bug 916291 [09:48] Launchpad bug 916291 in libreoffice "failed to upgrade from Oneiric to Precise: ERROR: Cannot determine language! - exit status 134" [High,Confirmed] https://launchpad.net/bugs/916291 [09:48] Sweetshark, it is still a problem with daily upgrade testing [09:48] hi seb128. i'm good thanks [09:48] how are you? [09:49] chrisccoulson, I'm good, a bit tired though [09:49] seb128, yeah, i know that feeling :) [09:49] ;-) [09:50] wow, precise-fixes-report makes it look like i don't do anything useful ;) [09:50] updated to the current unity from the ppa, glad that they fixed the alt tap stuff, my go to ws1 ws2 ws3 are alt based and I kept opening the hud before [09:51] chrisccoulson, heh, we have quite some desktop bugs waiting for you if you want to pick it up ;-) [09:51] chrisccoulson, but don't worry, I'm sure everybody appreciate having a working uptodate firefox and tb ;-) [09:51] seb128, i still have quite some firefox bugs too, like dropping the libgnome requirement :( [09:51] * seb128 hugs chrisccoulson [09:51] i'm going to do that one today though [09:51] seb128: these retracer crashes are indeed quite interesting, I did an analysis in bug 943117 (you don't need to read if you aren't interested, of course) [09:51] Launchpad bug 943117 in apport "retracer crashes on wrong duplication" [Medium,In progress] https://launchpad.net/bugs/943117 [09:51] * chrisccoulson hugs seb128 [09:52] pitti, oh, I'm interested, looking ;-) [09:52] seb128: just look at the last comment, the previous ones are boring analysis [09:53] hmmm, i'm really liking mercurial patch queues :) [09:54] interesting [09:54] chrisccoulson, how do they work? [09:55] seb128: so, we currently assert on "new bug is a dupe of A with address sigs and a dupe of B with symbolic sigs" [09:55] seb128, it's basically just quilt [09:55] seb128: (the cautious approach) [09:55] it's designed to allow you to break up work in to multiple patches [09:55] seb128: but so far all of the bugs I looked at actually turned out that A was a dupe of B which we didn't originally detect as such" [09:55] seb128, https://developer.mozilla.org/en/Mercurial_Queues [09:56] chrisccoulson, thanks [09:56] although, you don't need an extension now. it's built in to newer versions of mercurial [09:56] pitti, why do you need to stop on those? worth cases we have a dup not marked as such and 2 master bugs for the same issue? [09:56] pitti, doesn't seem worth stopping the retracers on? [09:56] seb128: well, that's the bug [09:56] ok [09:57] seb128: I originally added this when we introduced address sigs, to ensure that we don't process a thousand bugs with a flawed algorithm [09:57] seb128: but now that we have some experience/trust with it, we can be more bold, I think [09:57] right [09:57] and consider either capable of identifying dupes [09:57] s/either/both/ [09:58] do you know why they didn't get detected as dup from the start if their signature match? [09:58] or is that only the symbol sig matching? [09:58] seb128: yes, I wrote in the bug -- neither of the sigs matched [09:58] i.e not the address sig? [09:58] seb128: it's the third bug which matches the address sig of A and symbolic sig of B [09:58] oh ok, right, I see [09:58] seb128: but both address and symbolic sig of A != B [09:59] seb128: that's what I called "interesting" [09:59] it's a bit of a special case, I'm surprised it happens so often in practice [09:59] mvo, around? [09:59] glatzor, hey [09:59] seb128: well, most cases actually were because I changed the symbolic sigs in the middle -- still remember the __add_sse24() stuff? [09:59] seb128: that doesn't seem to be the case in this synaptic crash, though [10:00] pitti, yeah I remember those [10:00] seb128: so most of the fallout was that the dupe DBs got inconsistent because we changed the symbolic algorithm [10:00] but still I think it's reasonably safe to just make all three duplicates [10:00] +1 [10:01] which will cancel out the noise from each other's approach [10:01] seb128: effectively that's what we did manually anyway -- we looked at the log's assertion message and manually duped [10:01] (at least that's what I did) [10:02] I didn't because I was unsure what to do with those, I just called you for you help ;-) [10:02] pitti, thanks for the explanations [10:02] and +1 from me for just dupping things in those cases [10:03] what a brain tangler [10:05] chrisccoulson: yes, but in a call [10:05] pitti: what synaptic crash? I fixed one this morning [10:05] mvo, sure, no problem. i can wait until you're finished :) [10:05] mvo: bug 936677 [10:05] Launchpad bug 936677 in synaptic "synaptic crashed with SIGSEGV" [Medium,In progress] https://launchpad.net/bugs/936677 [10:06] mvo: but I'm mostly talking about the apport retracer duplicate identification [10:08] chrisccoulson: if you write the question here I can think about it during the call in a background thread [10:09] mvo - is there a "correct" way to enable extra software channels (such as partner). say, for example, i want to offer the adobe flashplugin from the partner repo to users the first time they go to youtube [10:09] i looked at what apturl does, but that seems a little bit hacky [10:09] and it doesn't look like software-center does it either [10:10] chrisccoulson: um, you might have a policy issue with that unless you're only doing it if partner is enabled [10:10] micahg, we would enable partner [10:10] chrisccoulson: you can't do that AIUI without explicit consent [10:11] micahg, yes, i'm well aware of that already [10:11] chrisccoulson: you could call sofware-center it has code for this, I don't think the code is extracted to something else [10:12] mvo, hmm, i tried looking in software-center already. mind pointing me in the right direction? :) [10:12] i'd like to try and do this without running software-center if possible, and it's pretty easy for me to display the eula in firefox already [10:12] chrisccoulson: also, why not just encourage people to enable the HTML5 trial? [10:13] micahg, how? [10:13] that is if youtube is the use case [10:13] youtube.com/html5 [10:13] micahg, how do you encourage users who have gone to youtube to do that? [10:14] I'll have to test that next time I have a VM up, I think youtube might actually do that already [10:14] it didn't the last time i tried it [10:14] you could just show a link to enable the trial the same way you'd ask to enable partner [10:14] in addition to the fact that a lot of video's still don't work with it [10:15] ah, if some videos don't work, I guess that's not a universal solution [10:15] micahg, adding website specific hacks to the plugin finder doesn't seem like a great solution [10:15] in fact, that would probably be more work [10:17] dobey, et al. is there documentation somewhere for how I should import desktopcouch into a python app with goi these days? [10:17] chrisccoulson: sorry, that's what I thought you were doing ;) [10:19] rickspencer3: are there significant API differences when you move to gir1.2-desktopcouch-1.0? [10:19] pitti, I doubt it [10:19] pitti, the problem is I can't figure out how to write the import statements :( [10:20] from desktopcouch.records.server import CouchDatabase [10:20] from desktopcouch.records.record import Record [10:22] this is the last thing I need to do before quickly.widgets is ported to GOI, which I need to finish in order to get the package onto seb128's computer [10:22] >>> from gi.repository import Couchdb [10:22] >>> Couchdb.Database [10:22] [10:22] that seems to be the first one [10:22] thanks pitti [10:23] rickspencer3, salut ;-) [10:23] rickspencer3: I can't find anything resembling a Record class, though [10:23] pitti, I can stumble around some more iPython [10:23] rickspencer3: "Couchdb" has {get,set}_record_type() methods in the Document class [10:24] rickspencer3: good luck! [10:24] ooh, that sounds ugly [10:24] maybe I should just dump couchgrid and write sqlitegrid instead [10:25] given that we don't love couchdb any more, that might be nice indeed :) [10:26] let me see how hard this is to port [10:26] I've been meaning to do a sqlite implementation anyway [10:28] "we don't love couchdb anymore"? [10:28] (not trolling, just asking) [10:30] the U1 team stopped using it for their services, and evolution-couchdb was dropped a while ago [10:30] as apparently it was too hard to make it do what we need to [10:35] jibel: k, thx [10:37] Sweetshark: your bug yesterday re removing open suse from a bug was done for you, had to get webops to remove it [10:38] czajkowski: thx, saw it. [10:38] pitti, hi [10:38] hi guys, latest precise, gnomeshell power-indicator isnt updating values [10:39] hello tkamppeter [10:39] upower --dump shows the battery is at 94% right now, but the indicator says 100% [10:39] Amoz, did you open a bug? [10:39] where does the indicator get it values from, or how can I debug it? [10:40] seb128, potential duplicate #934233 [10:40] pitt, you have got my mails from yesterday? I have found a solution for the trigger problem and applied it to CUPS, ready for upload after beta1. [10:40] reporter says the indicator displays it as charing although he undocked it [10:40] Amoz, I guess you need http://git.gnome.org/browse/gnome-shell/commit/?id=8bcbf3030f1c6c3088d4aa2b921d5d7031828ce9 [10:40] pitti ^^ [10:41] seb128, that could be something [10:41] seb128, restarting gnomeshell updates the values it seems [10:41] tkamppeter: yes, I got it, thanks! [10:42] Amoz, well try that patch [10:42] pitti, have you any idea on bug 936629? Ghostscript accessing /proc/XXX/auxv and being stopped by AA? [10:42] Launchpad bug 936629 in cups "Printing fails after printing first document " [Undecided,Incomplete] https://launchpad.net/bugs/936629 [10:42] no, not yet [10:43] pitti, should I assign it to you? [10:44] tkamppeter: I have the mail in my mbox, but sure [10:45] pitti, OK. [10:45] hi desktop folks! bamfdaemon has just crashed on me and I'm getting a bunch of duplicate entries in the launcher. Is there a way to restart it, other than logging out and back into the session? [10:46] dpm, run "unity" [10:46] or compiz --replace [10:46] dpm, hey btw ;-) [10:48] [...] [10:48] unity-panel-service: no process found [10:48] Initializing unityshell options...done [10:48] Segmentation fault (core dumped) [10:48] bummer [10:49] hey seb128, how's it going? [10:49] dpm, well, you can run it again, run "unity" from a vt if needed [10:49] it seems the second time around it worked [10:49] dpm, busy but good ;-) how are you? [10:49] seb128, it seems we're all the same :) [10:51] jbicha, hey [10:53] seb128, looks like the patch won't apply, the code it's replacing doesn't look the same in the patch and in my power.js [10:53] Amoz, ok, well I guess gnome-shell's code needs updating in some way for g-s-d 3.4 anyway [10:53] same reason as described on this bug [10:53] the signals changed [10:54] you should open a bug on launchpad if there is not already one [10:54] seb128, depends if you think 934233 is a dup [10:54] no idea, could be, I don't use gnome-shell [10:54] nor work on it [10:54] I'll open a new one then [10:54] what I know is that it needs to be updated for gsd changes [10:54] thank you [10:54] yw [10:56] seb128: good morning [10:56] jbicha, how are you? [10:56] jbicha, Amoz was reporting that the gnome-shell battery indicator doesn't update as it should, do you know if that's a known issue? [10:57] jbicha, it probably needs an update for the g-s-d signal changes (I pointed a commit from trunk but it doesn't apply to 3.2 it seems) [10:58] seb128: I've been running gnome-shell 3.3.90 this week, maybe new gnome-shell will land in precise next week? [10:59] ooh, new gnome-shell yes please [10:59] jbicha, you tell me, you should know better [11:00] jbicha, I'm neither using it nor working on it ;-) [11:00] jbicha, wasn't you saying the we only should update some packages and stay on 3.2? [11:00] the new gnome-shell will clutter up precise ;) [11:00] ok, no way then [11:00] lol [11:00] brb, need to restart my session [11:01] speaking of batteries, I powered my laptop at 0915, not it's 1201 and I still have more than an hour of batteries left. [11:01] Amoz: I was just reporting what the current status was, since seb128 did the gnome-control-center/g-s-d update, things have changed a bit [11:03] ok [11:05] re [11:05] jbicha, hey, no joking, the new clutter is lot of changes we can't do that at the end of the cycle [11:05] seb128: someone in #ubuntu+1 yesterday said they had problems with indicator-power's battery indicator too [11:06] jbicha, that was fixed a week ago or so [11:06] jbicha, maybe their mirror is not uptodate or they didn't restart since the update? [11:07] maybe [11:07] seb128: if you don't think clutter should be updated, you should comment at bug 941617 [11:07] Launchpad bug 941617 in clutter-1.0 "FFe: Update clutter/cogl to 1.9" [Wishlist,In progress] https://launchpad.net/bugs/941617 [11:08] jbicha, I will [11:08] jbicha, sorry but it would be crazy, they did ton of changes and refactoring this cycle and ricotz said another cogl soname change was still coming at least [11:08] is there anything I can do? [11:09] Amoz, sleep? walk? eat? ;-) [11:09] Amoz, or what was the question? ;-) [11:09] seb128: that's why it was waiting until next week when the cogl soname bump would land [11:09] seb128, about the power-indicator =) [11:09] jbicha, well I'm not in the r-t but that seems crazy [11:09] should i file a bug? [11:09] Amoz, yes, I told you that before ;-) [11:10] oh sry then [11:11] Amoz, no worry [11:12] ;) [11:15] help me out here, a good summary, "power-indicator not updating due to faulty signal connectors" ? [11:26] seb128, I can set it to confirmed as well? [11:26] it's reported at #943183 [11:26] Amoz, if you reported it no, the reporters shouldn't confirm their own bug [11:26] ah [11:26] that defeat the purpose of having 2 status otherwise [11:26] true [11:55] hi, I am improving kubuntu's software center to submit information to the rnr service [11:55] now when I send a request I get an error saying: "Authorization Required" [11:55] can somebody give me a hint about what's going on? [12:11] apol_, hi, you might want to ask on #software-center === MacSlow is now known as MacSlow|lunch [12:13] seb128: I asked there but nobody answered... anyway, I'll be patient :P [12:22] pitti: hey, when you get a chance, can you try pressing alt on your unity 5.4 with a gtk3 (or firefox/thunderbird) app? [12:22] pitti: tell me if you see the menus in the tope panel :) [12:22] (trying to see if it's a 5.6 regression) [12:28] didrocks, i can't [12:28] :) [12:28] but that's 5.6 [12:28] didrocks, i don't see a menu with 5.4 [12:28] rye: ah, I need someone with 5.4 :) [12:28] and i have 5.4 [12:29] chrisccoulson: ok, you are not using the unity-team ppa, right? [12:29] 5.4.0-0ubuntu3 to be precise -> no menu [12:29] ok :) [12:29] didrocks, i am. but i haven't updated yet [12:29] thanks for confirming [12:29] i'm still on 5.4 ;) [12:29] didrocks, is there a bug i can +1 ? [12:29] * didrocks put on the priority list [12:29] https://bugs.launchpad.net/unity/+bug/943194 [12:29] Launchpad bug 943194 in unity "[regression] Pressing alt doesn't show the menu title bar in top panel" [Critical,Triaged] [12:30] do any of you have a gtk2 app installed by any chance? [12:30] hm, nautilus and xterm works, xchat/firefox/thunderbird, i guess that's gtk2 [12:31] firefox, thunderbird don't for me [12:31] xchat/firefox/thunderbird do not [12:31] * rye needs to wake up, it's 2PM [12:31] :) [12:32] rye: xchat doesn't work then? [12:32] (this one is gtk2) [12:32] didrocks, xchat-gnome is gtk2, right [12:32] ok, great :) [12:32] thanks again [12:33] gtk3 apps have the menu on alt, gtk2 don't [12:33] hum [12:33] gnome-terminal doesn't here [12:33] chrisccoulson: same for you? gtk3 apps work? [12:33] * didrocks is getting confused [12:33] didrocks, yeah, seem to [12:33] ah [12:33] so a regression [12:38] not sure why it's broken in some apps and not others [12:38] chrisccoulson, gtk2 against gtk3? [12:39] seb128, i don't think that should be relevant. we have our own code in firefox / thunderbird for handling this: http://bazaar.launchpad.net/~extension-hackers/globalmenu-extension/trunk/view/head:/extensions/globalmenu/components/src/uGlobalMenuBar.cpp#L494 [12:39] ok [12:39] yeah, i just tested it, and we never even see the alt keypress [12:39] so, definitely not my bug :) [12:40] *phew* [12:47] chrisccoulson: and alt + F in gnome-terminal? [12:47] chrisccoulson: does it open the menu? [12:48] didrocks, yeah. and it works for firefox too [12:48] ok, seems a regression only in gnome-terminal [12:48] it's just showing the menubar with the alt key which is broken in firefox [12:49] didrocks, and, as expected, it works if i turn off the shortcut for the HUD [12:49] chrisccoulson: care to comment on the bug? ;) [12:52] didrocks, done [12:52] thanks :) [13:16] lol [13:16] * ritz wrong window, sorry [13:26] hey ritz [13:26]