=== mimico is now known as mimico_afk === chaoticuk_ is now known as chaoticuk [04:28] Good morning [05:03] I've just done a fresh install of precise on my server [05:03] (hard disk failure) [05:03] but I want to set a static IP address [05:03] yes... still IPv4 [05:04] what is the minimal bits I need to change? [05:04] thumper: hey [05:04] thumper: ah, you've come to the right place for UI-less servers ;-) [05:05] thumper: more seriously, do you want to edit conffiles or do you have GNOME/network-manager on this box? [05:05] pitti: hey... I have it running unity :) [05:05] oh, heh: ) [05:05] pitti: it is a desktop [05:05] but I treat it as a server [05:05] :) [05:05] my media server [05:05] so, network indicator -> edit connections -> cable based -> add [05:06] add, not edit? [05:06] create a new connection, siwtch to the "IPv4 settings" tab, change from "DHCP" to "Manual", and enter teh data [05:06] thumper: if you have an existing connection, sure, editing will do [05:06] when I just tried this, it wouldn't allow me to save [05:06] I don't have any (using Wifi) [05:07] ah... hang on, that time save enabled... [05:07] lets see if it works [05:07] thumper: the save button only becomes active if you add an IP [05:07] (or switch to "Auto") [05:08] ideally I didn't want to specify dns [05:08] but it seems without it... [05:08] it won't find anything :) [05:08] for more flexibility you can also just edit /etc/network/interfaces [05:09] something like [05:09] auto eth0 [05:09] iface eth0 inet dhcp [05:09] that should suffice [05:09] err, no, not "dhcp", but "static" [05:09] and add "address" and "netmask", see man interfaces [05:13] Morning. [05:15] way... adding the DNS servers made it happy :) [05:15] now to install sshd [05:30] Morning everyone! [05:31] hey Sweetshark [05:45] RAOF: hey Chris, how are you? [05:45] Good! [05:46] I think I may be closing in on why two-finger scrolling sometimes stops working until you reload psmouse. [05:46] RAOF: are you up for some SRU training wrt. handling kernel uploads? There's something I want to try [05:46] (not necessarily right now, in a few hours is fine as well) [05:46] Yeah, I can do that. [05:47] You know I can't copy to -updates due to LP bugs, though, right? [05:47] RAOF: why not? [05:47] permission issues? [05:47] Because the copy times out. [05:47] RAOF: ah, not any more [05:47] Ooooh! [05:47] there is never a reason to go to cocoplum any more to release SRUs [05:48] RAOF: sru-release was changed from the synchronous copyPackage (which often timed out) to the async syncPackage() [05:49] pitti: That's not quite true though, right? There's the handful of packages which still need manual tweaking with shell access to do (debian-installer & update-manager are the ones I remember offhand)? [05:49] RAOF: so you can release kernels just as any other SRU, except that most kernels also go to -security and thus you need the -s flag [05:49] RAOF: right, that's correct [05:49] Good, just keeping up :) [05:49] RAOF: releaseing kernels is easy now, though [05:49] RAOF: copying kernels to -proposed which do NOT bump API (i. e. don't require a new linux-meta) is also easy [05:50] just run the copy-package script on pending-sru.html, and you are good [05:50] RAOF: the tricky part is copying kernel abi bumps, as the new packages routinely land in universe and need to be promoted [05:50] Yay bugs. [05:50] I have the permissions to fix that up though, don't I? [05:50] RAOF: I usually run copy-proposed-kernel, then ssh cocoplum and move them to main [05:50] but I wanted to try something [05:51] RAOF: do you have a bookmark for "pending kernel SRU tasks"? [05:51] https://bugs.launchpad.net/~ubuntu-sru/+assignedbugs?field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE [05:52] RAOF: we just have one task right now [05:52] I do not, but there's the "open bugs" button on the pending-sru page, rigth? [05:52] RAOF: that doesn't help for the kernel tasks [05:52] the kernel SRU workflow is a bit different; we don't actually run sru-accept for kernels [05:52] RAOF: so I would like to run that one task with you, and try something [05:53] Ok. [05:53] let's switch to /msg to not spam the channel [06:31] pitti: what's the proper upstream name for gsettings-desktop-schemas (Capitalization wise) [06:33] micahg: hm, I think it's just that [06:33] * micahg was going to use GSettings desktop schemas (not sure about the last 2 words, but the first is like that in the README [06:36] micahg: looks fine [07:14] pitti, hi [07:14] hey tkamppeter [07:14] pitti, did you get mail from OdyX? [07:15] tkamppeter: yes, I added him to the project last night and told him, he just pinged bck [07:15] pitti, OdyX has uploaded a new cups-filters package to Debian as it seems that Debian has no libtiff5 any more. [07:16] pitti, up to nowthere is no new commit by OdyX. [07:16] tkamppeter: yes, he said he'd push "later tonight" [07:45] hey [07:46] bonjour seb128 [07:46] pitti, guten tag! wie geht's? [07:46] pitti, happy friday ;-) [07:46] je suis bien, merci! [07:46] and to you [07:46] I need to file the release team report, stuck in bug triage.. [07:47] I'm good thanks [07:47] well, now is a good time for bug triage [07:47] I've been triaged mostly this week, iso testing as well today I guess [07:48] right [07:48] I spent all day on hackign pygobject yesterday, that was a nice diversion [07:51] * pitti sighs at http://status.ubuntu.com/ubuntu-precise/canonical-desktop-team.html and will just do a mass dropping [07:53] seems RAOF own most of the remining ones [07:53] 5 todo and 1 in progress [07:54] there are about 30 left [07:54] and Luke has 6 [07:55] pitti, well most of others are 1 by people and mostly not-so-precise-related items [07:55] like rickspencer3 has 2 [07:55] rickspencer3, update your work items! ;-) [07:56] pitti, I just handled my 2 remaining ones [07:58] thanks [07:59] [mvo] backport apt-clone to lucid (lp:~mvo/apt-clone/lucid-backport): INPROGRESS [07:59] mvo: ^ this looks DONE? or is there more to it? [08:01] pitti: uh, last time I looked at this iirc there was a missing dh_python2 in lucid [08:02] missing as in has not been backported yet? [08:02] iirc yes [08:02] this is true :) [08:03] Bug #788524 [08:03] Launchpad bug 788524 in python-defaults "backport dh_python2 to lucid (and maverick if appropriate)" [High,In progress] https://launchpad.net/bugs/788524 [08:03] ok [08:07] mvo: ah, ok [08:07] mvo, hey, how are you? [08:07] hey seb128 [08:07] good, busy, good, release time! [08:07] pitti: I guess I could simply convert, let me look at the code again [08:08] mvo, ;-) for me release time is quiet time ;-) [08:08] they wouldn't let me get fixes in now, so I can sleep :p [08:08] mvo: I'm fine with "postponed", too, I'm just running through the remaining ones to check that we don't drop something urgent/important [08:08] seb128: haha [08:08] -proposed is wide open! [08:08] pitti: ok, if you have it in front of you, just set it to postponed, otherwise I will do that [08:09] (but need to hunt the blueprint first ;) [08:09] mvo: doing [08:09] mvo: https://blueprints.launchpad.net/ubuntu/+spec/foundations-p-lts-upgrades FYI [08:10] ta === chaoticuk_ is now known as chaoticuk [08:49] * pitti hugs apt [09:03] good morning everyone [09:12] How can this make it into Ubuntu? https://github.com/engla/keybinder/tree/keybinder-3.0 [09:12] keybinder that is now included is Gtk2 only [09:27] BigWhale: You know, the ideal way to get that into Ubuntu is to have GTK suck it up. [09:27] BigWhale: It's *still* stupid that GTK doesn't have an API for that :) [09:27] I know. [09:27] :/ [09:28] Just as long as we're on the same page :) [09:28] Incidentally, I didn't know that code had got itself libraryised. [09:28] I should probably rip out the copy in Do an depend on that instead. [09:29] Oh. GPL? Really? [09:31] I never bothered to research a bit more into this. If I did, I'd probably repackage keybinder for Precise ... Crap. [09:34] Oooh, there's exactly one rdepend of the existing libkeybinder0. [09:50] RAOF, what is it? [09:51] BigWhale: The xfce volume daemon. [09:51] (As in: sound volume) [09:51] oh [09:51] chrisccoulson, hey, how are you? [09:52] seb128, yeah, good thanks. although, my daughter is not too well today [09:52] how are you? [09:53] chrisccoulson, oh? she got a cold or something? [09:53] chrisccoulson, I'm good thanks [09:53] seb128, yeah, she's just got a cold [09:53] https://twitter.com/#!/chrisccoulson/status/193067610122887168/photo/1/large ;) [09:54] chrisccoulson, I should subscribe to twitter and follow you ;-) [09:55] Good to see that the prime directive of the internet is still in effect :) [09:55] heh :) [09:55] hey RAOF, how are you? [09:55] Hey seb128. Good! [09:55] seb128, you like cats don't you? [09:55] https://twitter.com/#!/chrisccoulson/status/193268814723547136 [09:55] ;) [09:55] chrisccoulson, I love cats! [09:55] lol [09:56] nice picture ;-) [09:56] it seems one of ours decided to sleep in a an old flower pot ;) [09:56] heh :) [09:56] oh [09:56] i've just noticed from that photo that he's lost his collar! [09:57] gah [09:57] he keeps doing that [09:58] why do you put him a collar? [09:59] so people don't think he's a stray. and the bell also frightens the birds away, to stop him from sneaking up on them ;) [09:59] we end up with dead brids all over the garden when he hasn't got a collar on ;) [09:59] poor cat, can't even get a bird for lunch? [09:59] heh [09:59] we do feed him [09:59] ;-) [10:50] is didrocks on vacation today ? [10:51] * ogra_ has a compiz change he would need to upload, but there are un-uploaded changes in the branch [11:00] ogra_: yes [11:00] k :( [11:03] pitti, he swapper his easter monday for today [11:03] swapped [11:03] ups [11:03] ogra_, ^ [11:03] ITYM ogra [11:04] pitti, indeed ;-) [11:04] well, thanks, i'll talk to the release team if monday is still ok for a 0 day SRU [11:05] i just dont want to upload stuff he didnt nod off [11:05] ogra_, yeah, you will be up for a rant again about wth you guys are doing coming so late in the cycle with non working stuff [11:06] seb128, it would have worked without that werid branch stuff :P [11:06] lol [11:06] it would have worked if had landed 3 months earlier [11:06] well, i was promised that patch for opening of precise ... [11:07] somehow there was some kind of a slight slowdown :) [11:08] (though i also was promised the same patch for the opening of oneiric ... dunno what made me hope for it to be ready when 12.04 opens :P ) === MacSlow is now known as MacSlow|lunch [11:15] anybody here able to give a hand with an issue i'm seeing in a scope? === greyback is now known as greyback|lunch [12:16] chrisccoulson: hey, just wondering if you updated the patch in https://bugzilla.gnome.org/show_bug.cgi?id=608402 to fix my comments? [12:16] Gnome bug 608402 in general "Add support for delaying autostart applications" [Enhancement,New] [12:21] pitti, I got a fix for bug 980673 in cups-filters. [12:21] Launchpad bug 980673 in cups-filters "pdftopdf crashed with SIGSEGV" [Critical,Triaged] https://launchpad.net/bugs/980673 [12:23] pitti, probably now it should go as SRU. I will put cups-filters 1.0.17 onto the Debian BZR, can the SRU also be 1.0.17 or should it have the fix in debian/patches (1.0.17 differs from 1.0.16 only by this patch)? [12:25] tkamppeter: no, it can also be 1.0.17 [12:25] tkamppeter: right, just upload to -proposed [12:35] pitti, OK, it is on its way now. === MacSlow|lunch is now known as MacSlow === greyback|lunch is now known as greyback [13:03] pitti, fix for bug 980673 is uploaded to -proposed for an SRU now. It is also committed to the Debian BZR. [13:03] Launchpad bug 980673 in cups-filters "pdftopdf crashed with SIGSEGV" [Critical,Fix committed] https://launchpad.net/bugs/980673 [13:23] happy friday [13:28] desrt, hey, happy friday! how are you? [13:28] us usual, i'll be able to tell after coffee :) [13:28] ;-) [13:45] hello everyone! would you please remind me which is the proper channel to report unity issues? [13:45] nessita, #ubuntu-unity [13:46] kenvandine: thanks! [13:46] anytime! [14:00] dpm: ping [14:03] dpm: n/m [14:10] mterry: apparently you've been naughty? [14:23] desrt, is that the autostart delay stuff? [14:24] yes [14:24] desrt, it was a "let's assume GNOME doesn't have reason to reject that simple patch and not go through a temporary name and transition" [14:24] if you want the reason to use X-GNOME- [14:24] but the patch has been sitting in bugzilla and never went it [14:25] desrt, hello. I have?! [14:25] mterry, in fact chrisccoulson had [14:25] desrt, https://bugzilla.gnome.org/show_bug.cgi?id=608402 for the record [14:25] Gnome bug 608402 in general "Add support for delaying autostart applications" [Enhancement,New] [14:25] phew :) [14:25] or vuntz had [14:25] desrt, "Vincent Untz [gnome-session developer] 2010-01-29 01:03:03 UTC [14:25] Please prefix the key name with X-GNOME" [14:25] * mterry doesn't want to ruin his Christmas haul [14:26] seb128: on the assumption that the patch would go upstream that makes sense [14:26] desrt, I think it's not fair to blame us to do what upstream asked on the bug review [14:26] seb128: nobody is blaming [14:27] desrt, " what I most strongly dislike is the use of X-GNOME for something that is not, in fact, a gnome thing" [14:27] desrt, but well, let's not argue on details ;-) [14:27] well, that part is true :) [14:27] it's clear what happened, though [14:27] ... [14:27] you took the patch on assumption that it would one day make it upstream [14:27] I think they is no bad willing there [14:27] it never did [14:27] i agree [14:27] there [14:27] for reference, it would help to have the latest version of the patch in the bug :-) [14:28] vuntz, I can do that ;-) [14:28] since the X-GNOME thing was requested in the last review, and there's no patch uploaded after that [14:28] bonus point if everything mentioned in the review is fixed [14:29] then the solution is that gnome takes the patch and everyone is happy [14:29] right [14:29] well mclasen didn't seem so happy about GNOME taking that patch [14:29] seb128: i'm not thrilled about the approach myself [14:30] better suggestions that work today or would work with reasonable efforts are welcome ;-) [14:30] * vuntz is not thrilled but doesn't care enough to argue against or for :-) [14:30] seb128: 'take the patch' ;) [14:30] "let's do nothing until we fix that will in 3 years" not so good ;-) [14:30] will->well [14:39] dobey, sorry, I was on the phone. Is there anything I can still help with? [14:39] if it's the bug with untranslated Yes/No/Cancel, I'm talking about it with mandel [14:40] dpm: it was that bug, yes [14:40] ok [14:42] vuntz, I've attached the current version, chrisccoulson should still update it to address your review comments though ;-) [14:42] hi! [14:42] oh, that one again ;) [14:42] perhaps i should do non-firefox desktop work for a cycle :) [14:44] chrisccoulson, when we switch to midori you can :) [14:44] heh [14:50] kenvandine: what, are we going to just ship elementary now? [14:50] :) [15:08] seb128: (looking over bug mails) the nfs/ecryptfs homedir thing is what causes the crashes in refresh() [15:08] seb128: the segv one is the one that i think is happening on logout [15:08] desrt, did I comment asking on the wrong one? ;-) [15:09] seb128: i already asked it on the other one [15:09] and the reporter confirmed that they have ecryptfs [15:09] desrt, oh ok, I got confused, thanks ;-) [15:09] desrt, how does ecryptfs lead to that issue? [15:09] seb128: mmap on ecryptfs seems unreliable [15:09] hum [15:10] never got an issue here [15:10] for the same reason that it's unreliable on nfs, i guess [15:10] ya... not will-always-fail [15:10] just not totally reliable [15:10] enough that we get a few bugs about it... [15:11] ok, I'm a bit surprised still, but if you are confident it's that ;-) [15:11] anyway... it's sort of an interesting point. we may want to treat fuse-based filesystems in the same way we will treat nfs file systems (ie: don't trust that they will work properly) [15:11] well [15:11] i guessed, and had my guess confirmed [15:11] so i'm not sure, but i'm fairly happy with that explanation [15:12] the fix to this problem is very very simple: we really need XDG_RUNTIME_DIR [15:12] that's not a "fix" for ecryptfs not being reliable [15:12] I'm not sure I want to store my datas on a non reliable fs [15:12] i.e that raise questions out of that bug ;-) [15:13] it's a bit of a disturbing situation [15:13] it's not the data that's the problem [15:13] it's the shm file (ie: a one-byte signaling mechanism we use for synchronising the service and the apps) [15:13] well, how many programs will have the same issue [15:14] anything that tries to create shared memory handles using the home directory as a backing store? [15:14] this is exactly the reason that XDG_RUNTIME_DIR was invented and why we are trying to use it for all things like this [15:15] ok [15:15] we need a reliable local native fs (like ext3, ext4, btrfs, tmpfs, etc) [15:15] not 'fake' filesystems like ecryptfs/fuse or nfs [15:15] I'm still unsure why shm file on ecryptfs are unreliable [15:15] and why we think that's ok ;-) [15:15] because it's impossible for it to be 100% perfect [15:16] the hardware doesn't notify the kernel on every single write [15:16] so the data can't be pushed down into userspace on every single write either [15:16] also: it has the classical use-after-delete thing going on [15:16] and the underlying file in the unecrypted filesystem disappears [15:16] if the data gets pushed out of the cache in the meantime then when it comes time to refetch the backing store, it's already gone [15:17] hum, alright [15:17] but yeah, agreed, we really need the xdg_runtime_dir next cycle [15:18] it seems a few things are starting to use it [15:19] udisks2 do user mounts there right? [15:19] no, I don't think it does [15:19] NEWS: Avoid using $XDG_RUNTIME_DIR/media for now [15:19] NEWS: Use /run/media/$USER for mounting [15:20] NEWS: Prefer mounting in /run/user/$USER/media instead of /media [15:20] pitti, ok, ok [15:21] "This is because of security concerns - it is way too dangerous to let a system-daemon such as udisks manage directories in a user-controlled location such as $XDG_RUNTIME_DIR. So now udisks2 is using /run/media/$USER instead, see http://cgit.freedesktop.org/udisks/commit/?id=aa02e5fc53efdeaf66047d2ad437ed543178965b" [15:21] that's from the corresponding gvfs commit [16:02] pitti, the pdftopdf bug is verified ... [16:02] tkamppeter: with the packages from -proposed? [16:03] pitti, I think so. [16:04] tkamppeter: cool, I'll copy it into precise release then [16:36] pitti, can you have a look at bug 982675? [16:36] Launchpad bug 982675 in cups-filters "pdftops crashed with SIGABRT in __libc_message()" [Medium,New] https://launchpad.net/bugs/982675 [16:39] pitti, there happens a segfault and I do not understand why. There is a call 'numvalues = sscanf(resolution, "%dx%d", &xres, &yres)' with xres and yres being integer variables and resolution being a static char[128] array containing "4800CQ". The sscanf should set xres to 4800, numvalues to 1 and leave yres untouched. [16:42] tkamppeter: it's not a segfault, it's an assertion failure [16:43] looks like double-free [16:43] sscanf (string=0x7fff072e67a0 "4800CQ", format=0x7f509f1c168b "%dx%d", args=0x7fff072e2168) at iovsscanf.c:45 [16:43] now, that indeed does not match [16:44] pitti, but also strange, why should this simple sscanf cause a double free? [16:44] well, something went wrong in memory management, not necessraily double free [16:44] tkamppeter: but yes, sscanf should just stop scanning at the 'x' [16:45] The "4800" matches the first %d, so it should go into xres then the "CQ" does not match "x" which should make sscanf stop, wuith the counter on 1 and yres untouched. [16:45] yes, indeed [16:45] pitti, so it is not a bug of cups-filters. To which package I should assign that? [16:46] tkamppeter: well, it most likely is [16:46] there might be some memory corruption happening before this [16:46] which just blows up in sscanf [16:47] pitti, but the traceback shows that the input parameters are still OK. [16:47] right, but memory state might not be [16:47] tkamppeter: you could write a small C program with just this sscanf and see whether it works [16:47] tkamppeter: or disable parts of the code until you don't get the crash any more, and then see what causes it [16:54] good night everyone, have a nice weekend! [16:58] pitti, 'night [17:25] 5539 seb128 20 0 1308m 551m 36m S 2 15.9 65:50.91 firefox [17:25] could firefox stop being a dog? ;-) [18:11] mterry: any idea why I only get a while circle instead of the Kubuntu logo for the KDE session in unity-greeter? [18:12] *white [18:12] jbicha, there is a kde_badge.png installed... [18:13] yes, how long has that badge been there, because I've always seen the white circle instead [18:13] jbicha, for quite a while. hold on, let me see what the kde session is called, maybe there's a mismatch [18:14] jbicha, what do you have in your /usr/share/xsessions/ ? [18:14] kde-plasma.desktop [18:15] jbicha, ah... that's why. unity-greeter is expecting just kde [18:15] jbicha, did that change or is unity-greeter just being dumb? [18:15] we should probably ask in #kubuntu-devel or something to make sure, I don't use the kubuntu-desktop metapackage but pick and choose parts of KDE === fenris is now known as Guest89753 === Guest89753 is now known as ejat [18:19] jbicha, seems kde.desktop doesn't exist in oneiric or precise. I'll patch to add kde-plasma [18:20] mterry: thanks [18:20] I just renamed kde-plasma.desktop to kde.desktop to test it, & the logo looks nice [18:42] seb128, are we done with uploads? [18:46] pitti, ^ I'm not clear on when the no-respin cutoff is [18:47] mterry: is Kubuntu the only official Ubuntu flavor that's not using lightdm for 12.04? [18:47] seb128: i'm done with hud-service scoring bugs [18:47] jbicha, I don't know [18:48] it's impossible to make everyone happy [18:48] desrt, if it makes you feel better, i am happy :) [18:48] and users expecting to type 2 letters "sh" to match "show only glyphs..." without wanting to type the extra 'o' is pretty.... low on my pity list :p [18:49] kenvandine: it does :D [18:49] glad to help :) [18:49] kenvandine: as the distro packager of this stuff maybe you can comment on the bug that you feel that the current algorithms are producing good results for 99% of cases and that it's not worth risking the change at this point [18:50] desrt, actually i haven't really looked very closely at the relevance of the results, i've just been trying hard to remember to attempt to use the hud :) [18:51] hard to teach an old dog new tricks :) [18:51] i do feel it is too late to risk change... [18:51] kenvandine: your "i am happy" comment no longer makes me feel better :p [18:52] you're only happy because you never use it :) [18:52] hehe [18:52] i use it at least a couple times per day :) [18:53] just not really looking closely at the scoring [18:53] it just seems to work :) [18:58] desrt, heh, that description is not an accurate one ;-) [18:58] desrt, I would have closed it otherwise :p [18:59] desrt, it does match "indicator > shutdown" before "view > show", when "sh" is an exact mathing of 2 letters for each ;-) [19:00] huh? [19:00] quoting you: "sh" lists "Device > Shutdown" first and then "View > Show only [19:00] desrt, oh, just read your comment, the different is the trailing string length [19:00] glyphs..." from gucharmap [19:00] ya [19:00] right [19:00] *Sh*utdown [19:00] we had to do that for "savE" to come before "save as" and "save all documents" [19:00] *Show* [19:01] ups [19:01] *Sh*ow [19:01] "only glyphs..." [19:01] like they both have a perfect start match [19:01] there is a penalty for the extra words [19:01] that's what makes "save" match "save" better than "save as" [19:02] * desrt loves the ambiguity in that sentence [19:02] desrt, ok, fair enough, I was unsure if the "extra char" penalty was on a word or on the sentend [19:02]