=== timchen1` is now known as timchen119 [08:15] Goooood morning, desktoppers! [08:16] You write C or C++? You want to have opinionated discussions with your compiler? Or just be scared by what "undefined behaviour" really means? Read this: http://blog.llvm.org/2011/05/what-every-c-programmer-should-know_14.html [08:16] llvm.org bug 2011 in Loop Optimizer "Loop Index Splitting Pass asserts on SPASS" [Normal,Resolved: fixed] [08:16] desrt: as discussed yesterday ^^ [08:30] seb128: some new comments in https://bugs.launchpad.net/ubuntu/+source/whoopsie/+bug/1437633 [08:30] Ubuntu bug 1437633 in lxc-android-config (Ubuntu RTM) "Choosing not to report crashes and errors setting reverts" [High,Confirmed] [08:30] if you're interested [08:30] Laney, I am, thanks === qengho is now known as CardinalFang === CardinalFang is now known as qengho [08:45] Sweet5hark: ouch, that really *is* scary [08:50] ochosi: hrhr, yeah -- an unrelated bug might cause your guards to be happily optimized away ... [08:53] Sweet5hark: yeah, i'm trying to forget that again now so that i can continue to live happily (more or less) [08:54] also: you put your best coders on the security relevant code and they do all h. but then your less experienced coders use that code and since you [08:56] * desrt drops https://wiki.gnome.org/RyanLortie/BugsInC [08:56] also: you put your best coders on the security relevant code and they do all the checks and guards. And then your less experienced coders use that code, have undefed behaviour around it, and since its inlined, the guards/checks get optimized out .... [09:01] Laney, seb128 ... [09:01] phablet@ubuntu-phablet:~$ sudo whoopsie-preferences [09:01] Could not write configuration: Failed to create file '/etc/whoopsie.LRHGXX': Read-only file system [09:01] Lost the name: com.ubuntu.WhoopsiePreferences [09:02] that is with the fix ... that cant really work :/ [09:02] talking about bug 1437633 [09:02] bug 1437633 in lxc-android-config (Ubuntu RTM) "Choosing not to report crashes and errors setting reverts" [High,Confirmed] https://launchpad.net/bugs/1437633 [09:03] could we move the file to /etc/whoopsie/whoopsie ? [09:03] or some such [09:08] ogra_: read like 5 lines above your sentence :-) [09:08] Laney, ah, heh, i'm blind [09:08] are you talking about the rtm or the vivid version? [09:08] vivid I guess [09:08] rtm is dead [09:08] nice [09:08] NICE [09:09] (unless there is somethin super critical found) [09:09] so I didn't work on the vivid implementation & assume Seb is looking at it [09:09] wait for him I guess [09:09] but then only a fix for that one thing would land ... else we'll wait for vivid stabilization now [09:15] desrt: all those are resolved in C++11 at least, #justsayin' [09:28] larsu: can you take a glance at https://bugzilla.gnome.org/show_bug.cgi?id=744282 before i push? [09:28] Gnome bug 744282 in gio "gvfs-open for application/x-virt-viewer changed behaviour between 2.40 and 2.42" [Normal,New] [09:29] since nobody really properly reviewed this yet -- only tested it [09:29] desrt: new dotty record! [09:29] desrt: I mean: sure, no problem [09:30] larsu: sounds like i pinged at a good time ;) [09:32] larsu: di da dotty (sing this to the melody of schni schna schnappy) [09:32] shit. instant earworm [09:32] THANKS Sweet5hark [09:33] sure glad i didn't get this when i'm a kid [09:33] desrt: you're sentence didn't made any sense grammatic [09:33] i know [09:33] i knew it as i was typing it [09:33] was too lazy to fix it [09:34] I didn't get this as a kid either [09:34] it's not that old [09:35] "which is nice" <-- actual fragment from desrt's commit message [09:36] we totally need a commit message of the year competition [09:37] desrt: setting GPtrArray.pdata directly is smelly [09:44] there is no API to do that [09:44] but i could change it to run over the array and dup it after i free the ptrarray [09:44] but.... why? [09:44] smell-avoidance? [09:45] doesnt gcc have a deodorant plugin for smell avoidance? [10:38] ogra_, yeah, it's annoying, whoopsie edit the file in a safe way, e.g to the create a new file and rename way [10:38] ogra_, which of course doesn't work with our ro fs [10:39] right, that is fine as ling as we can do it in a subdir [10:39] *long [10:39] (which i actually thought it did ... i understood /etc/whoopsie to be a dir in your patch) [10:45] ogra_, so maybe my patch is wrong, it's a file [10:45] ogra_, I though you could list specific files in writable-paths [10:45] seb128, yes, you can ... [10:45] k, so it's a file [10:45] but for atomic writes you would need /etc writable [10:45] I guess we would need to change whoopsie then [10:45] to write to a subdir [10:46] yeah [10:47] it's a bit suboptimal that we need to change whoopsie just to workaround design limitation of our ro images [10:47] we need a proper overlayfs ;-) [10:48] bregma, could somebody from your team look at bug #1405349? I guess that's one for andyrock (I would ping him but he's not on IRC it seems?) [10:48] bug 1405349 in unity (Ubuntu) "switching to activated guest session causes screen blank" [Undecided,Confirmed] https://launchpad.net/bugs/1405349 [10:59] bregma: http://summit.ubuntu.com/uos-1505/meeting/22447/user-documentation-for-unity-8/ [11:07] happyaron, what's the status of https://bugs.launchpad.net/ubuntu/+source/ibus/+bug/1439202 ? [11:07] Ubuntu bug 1439202 in ibus (Ubuntu) "/usr/lib/ibus/ibus-ui-gtk3:11:XKeysymToKeycode:keybinding_manager_bind:panel_keybinding_manager_bind:panel_bind_switch_shortcut:panel_construct" [High,In progress] [11:08] seb128: cannot determine which side is wrong yet, but traces down to glib... [11:08] happyaron, you have desrt 1.5m away from you if needed ;-) [11:08] yep, :) [11:09] seb128: will deal with that a bit later, struggling with canonistack atm [11:09] k, thanks [11:10] it's stucking my instance, hope it won't get swallowed [11:18] seb128, why would you add that patch to systemd ? systemd isnt used on phones [11:19] ogra_, I said a similar change to whoopsie, to rename the file in /etc/writable rather than /etc [11:19] oops,, sorry [11:19] yeah === MacSlow is now known as MacSlow|lunch [11:29] didrocks: what is ubuntu-desktop-next? [11:30] Riddell: it's the experimental unity8 image we had in the past, I guess the first image was about 18 months ago [11:31] Riddell: are you going to review my changes? :) [11:35] didrocks: I wouldn't trust myself to be able to :) === MacSlow|lunch is now known as MacSlow [14:40] pitti: about bug 1443667, the cause of this is http://anonscm.debian.org/cgit/pkg-openoffice/libreoffice.git/commit/?h=debian-experimental-4.4&id=b0002aad6276ac1f21828e5bea1d4c319a345b58 which has been around since libreoffice 4.4.0 and _rene_ says "its only a dir, so dpkg shouldnt care" ... [14:40] bug 1443667 in libreoffice (Ubuntu) "Installer crashes installing libreoffice" [High,Triaged] https://launchpad.net/bugs/1443667 [14:42] ... never seen that happen for an update here either. Sure can add a B/R, but am unsure on how to verify the fix ... [14:43] * Sweet5hark should maybe try a manual update sequence to trigger this. [14:43] Sweet5hark: does that also apply to an utopic->vivid upgrade, or only intra-vivid? [14:46] pitti: I assume it might also affect utopic->vivid. change was in the first version of 4.4.0. [14:47] pitti: But in that case it should have possibly affected every update from libreoffice 4.3 to 4.4, so I wonder if this is some kind of heisenbug. [14:48] Sweet5hark: it depends on the unpack order really [14:49] pitti: yep, that makes it so much "fun" ... [14:53] pitti: heh. _rene_ now wants to put the symlink back to -common, while keeping the dir in -core. (change was done for https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=775150 btw) [14:53] Debian bug 775150 in libreoffice-core "libreoffice-core: removal of libreoffice-core makes directories disappear from libreoffice-common" [Important,Fixed] [14:53] Sweet5hark: hm; please don't keep replacing dirs with symlinks or vice versa, that confuses dpkg [14:55] pitti: well, if I wouldnt sync from debian, there wouldnt be any symlinks and wild moving around of dirs anyway. [14:56] * Sweet5hark isnt a fan of this symlinking and moving either. [14:57] * ogra_ is looking forward to a snap/click of openoffice ... that will never have such issues :) [15:00] ogra_: thats true. because there will never be any updates of openoffice anyway. security or otherwise. [15:01] Sweet5hark, so you plan to stop maintaining it ? [15:02] I stopped maintaining openoffice four years ago ... [15:04] lol [15:04] my brain stopped switching to the right name years ago too it seems [15:04] sorry :) [15:05] ogra_: np [15:05] * Sweet5hark now carries on with breaking libreoffice ... [15:07] pitti: sooo, I agreed with _rene_ that we just add a replaces vs. -common for -core. Is the breaks: really needed? Neither _rene_ nor me see a reason for that. [15:08] Sweet5hark: it might work without, it's just standard practice so that apt unpacks the newer version of the broken package first [15:08] but if you try to replace a symlink with a dir you need a preinst anyway, and hten definitifely a breaks [15:09] pitti: nono, we are not replacing the dir with a symlink [15:09] (not sure which way aroud you changed things, only following with half a brain cell) [15:13] pitti: /usr/lib/libreoffice/share/uno_packages moved from -common (in 4.3) to -core (in 4.4). _rene_ assumed the dir to be empty (and thus needing no replaces:), but its not as there is a symlink called "cache" in it. So just adding a "Replaces: libreoffice-common (<< $ver)" in libreoffice-core should make this work out. [15:15] heh, as it goes there is a Replaces: in there already anyway, just needs the version bumped ... [16:08] seb128: http://people.canonical.com/~bjoern/vivid/4.4.2/libreoffice_4.4.2-0ubuntu2_source.changes no l10n update needed, no sign-all needed [16:16] Sweet5hark, uploaded [16:19] seb128: thx [16:19] Sweet5hark, yw!