lifelesshey, how does rickspencers script detect 'canonical staff' ?00:41
RAOFBah.  Why is squid-deb-proxy failing to start :?01:11
lifelessRAOF: anything in cache.log ?01:28
RAOFlifeless: No errors; just a list of Making directories01:28
lifelessthat should only have happened the once (for squid -z)01:29
RAOFAnd, indeed, that's timestamped 2012/02/02, which was the last time I tried to fix this by blowing away all the directories :)01:29
RAOFNothing seems amiss in either access.log.1 nor store.log.1, either.  Although there doesn't appear to be an access.log nor a store.log, which suggests that squid is dying before writing them?01:31
RAOFHm.  squid3 itself is perfectly happy to start; there's clearly something wrong with the squid-deb-proxy config.01:43
lifelessRAOF: squid3 -k parse -f <whatever>01:44
RAOFAha!  mirror-dstdomain.acl is bungled, apparently.01:45
RAOFIt doesn't like ddebs.ubuntu.com to be duplicated in the allowed domains.01:48
RAOFIt would be slightly nicer if this error made it *anywhere*; as far as I can tell the only thing that hits the logs is “[33760.219160] init: squid-deb-proxy main process (20601) terminated with status 1”01:49
lifelesscan't write to cache.log until you know where cache.log will be01:52
lifelessso it goes to std* until that point in the config01:52
lifeless(process, not parsing - its a two stage thingy)01:52
lifelessanyhow, that's probably a change for strictness01:53
broderis upstart's job logging still turned off? if squid is spewing to std*, there might be something there01:58
broder(maybe check /var/log/upstart?)02:01
broderoh well02:01
pittiGood morning05:11
BigWhaleGood Morning.05:55
RAOFBoo, hiss.  Why isn't dbus-test-runner in Debian :(05:56
RAOFBigWhale: Good morning!05:58
BigWhaleHey RAOF. :)05:58
micahgRAOF: you can fix that ;)05:59
RAOFNot before I get at least DM.06:00
micahgraoF: I think we have plenty of sponsors around06:01
RAOFWe do, I just find it quite demotivating to have to get sponsorship.06:02
RAOFI don't think Debian sponsorship works as well as our processes.06:02
broderRAOF: it works pretty well when you just pick one person and harass them :)06:03
micahgraoF: join pkg-utopia?06:03
micahgteam based sponsorship seems to work pretty well06:03
RAOFYeah, it does.  What's pkg-utopia about?06:04
RAOFSpeaking of sponsorship - pitti, could I get colord (0.1.16-2) up again?  This time with significantly more installability!06:09
pittiRAOF: sure, doing06:10
RAOF(Note to self: not only check that purging and reinstalling works, but also that installing in a clean chroot works!)06:10
robert_ancellpitti, perhaps you can help me with this one - do you know why using a GLib main loops stops signal handlers working? http://paste.ubuntu.com/827320/06:20
robert_ancellthey revert to the KeyboardInterrupt exception06:20
pittihmm, not off the back of my head06:21
broderrobert_ancell: the behavior you describe is exactly what pygobject's wrapper around GMainLoop sets up06:23
robert_ancellpitti, alternatively do you know if you can access g_unix_signal_add from python via gia?06:23
broderrobert_ancell: it opens a pipe and passes it to PySignal_SetWakeupFd, which has a byte written to it any time the process receives a signal06:24
broder(and a GSource watches the other end of the fd, and fairly blindly raises a KeyboardInterrupt when something gets written to it)06:25
broderit's all in gi/_glib/pygmainloop.c in the pygobject source06:25
brodermy guess would be that PySignal_SetWakeupFd is a heavier hammer than is actually necessary, but i don't know06:25
robert_ancellbroder, thanks.  I have something similar for lightdm, but this is a small program and it seems overkill to do that :(06:27
broderrobert_ancell: if, uh, you use Gtk.Main() instead of GLib.MainLoop().run(), you can avoid this pygobject "feature" :)06:28
broder(unless that's been fixed - there's an open bug asking for Gtk.Main() to throw KeyboardInterrupt like GLib.MainLoop does)06:29
robert_ancellthat seems like the wrong way around - you probably don't care too much about signals when using GTK+, but you do when you're running from a plain main loop06:29
pittirobert_ancell: sorry, no; g_unix_signal_add doesn't seem to be in any GIR06:29
smspillazhey hey didrocks06:31
didrockshey smspillaz :)06:31
broderrobert_ancell: right. my point is that, at present, GLib.MainLoop asserts that it wants to be in charge of all signals while Gtk.Main does not, so if you want to claim some signals, Gtk.Main allows that while GLib.MainLoop does not06:31
robert_ancellbroder, so it's impossible to handle signals when using GLib.MainLoop?06:32
didrocksgood morning robert_ancell :)06:32
robert_ancelldidrocks, hey06:32
didrocksrobert_ancell: btw, the ubuntu-desktop merger machine is already ready for 3 weeks now, we just need to talk about how it works :)06:33
robert_ancelldidrocks, cool, have to tackle that one next week06:34
didrocksrobert_ancell: exccellent :)06:34
broderrobert_ancell: right. PySignal_SetWakeupFd bypasses the signal module's mechanism for recording that a signal was received so that it can call the handler once it has an opportunity to reacquire the GIL and so forth06:35
broderso GLib.MainLoop's code gets called any time that the signal module would otherwise call a signal handler06:35
robert_ancelli see06:36
robert_ancelloh well, have to tackle that next week.  Thanks broder, pitti06:39
=== bschaefer_ is now known as bschaefer
didrocksoh, interesting: gzip-file-is-not-multi-arch-same-safe lintian warning08:24
sil2100didrocks: it seems yesterday my unityshell doesn't start at all - is there some easy way I could enable some debugging in both unityshell and compiz to find out what's going on when the plugin is loaded?08:31
didrockssil2100: hey08:31
didrockssil2100: when all plugins are loaded08:31
didrocksdo you see "unityshell" as being one of them?08:31
sil2100didrocks: should compiz print that out during start?08:32
didrockssil2100: compiz prints the plugins that are loaded (in the previous versions at least) at start, yeah08:32
didrocksmaybe you need --debug now08:32
didrocks(to first ensure it tries to load it :))08:32
sil2100hm, it doesn't seem to list the plugins that its loading08:36
didrockspitti: hey! I'm getting a gzip-file-is-not-multi-arch-same-safe lintian warning. (yeah, see how gzip on multiarch is a hot topic!). I thought the option to use -n for gzip was made on the platform, isn't it? should I override it manually in my package?08:36
sil2100With --debug I only get a list of failed stat() operations on .so's from .compiz-1/plugins08:36
didrockssil2100: you shouldn't have any in ~/.compiz-108:36
didrockssil2100: remove that dir08:37
didrocksseems you have a local build08:37
sil2100didrocks: I think he's also looking for them in the global directory too08:39
sil2100Since I remember once getting a stat() error of a not found library in /usr/lib/compiz/ or something, when I forgot installing one package08:39
didrockssil2100: yeah, but you maybe have a locally installed unityshell plugin which override the system one08:39
sil2100didrocks: I removed .compiz-1, but it didn't seem to have a plugins directory at all08:40
didrockssil2100: I think you should see with smspillaz then if the debug messages are not around anymore08:41
sil2100Ok, will do!08:41
didrocksthanks :)08:41
dpmgood morning desktop folk!. After the last dist-upgrade it seems the indicator menus don't follow the theme. They appear in the standard grey gtk3 theme instead of them being black as usual. I've already logged out and back into the session, but that did not seem to solve it. Any ideas?08:41
pittisounds like a regression from the new GTK08:44
pittiseb128 is not online yet, I wonder whether he gets that, too08:44
pittilooks fine here, but I'm using Radiance08:44
didrockspitti: snif, ignoring me? :)08:45
* pitti hugs didrocks, why?08:45
* didrocks hugs pitti back :)08:45
pittisorry, didn't see your ping08:45
didrocksno worry, just jocking :)08:45
pittididrocks: is that for buildinfo sor somethign?08:45
didrockspitti: yeah, local build of bamf08:46
pittididrocks: that's one known case which we don't need to worry about, pkgbinarymangler removes them08:46
pitti.buildinfo files are rather useless anyway08:46
pittiand yes, they would cause multi-arch problems08:46
didrockspitti: great! ok, I won't add the override then08:46
didrocksyeah, I heard about some gzip fun you had :)08:46
pittithat's unrelated, t hough08:47
didrocksit's only timestamp in that case, not content, isn't it?08:47
didrocksdpm: pitti: I saw that kenvandine uploaded light-themes08:48
didrocksdpm: pitti: the according fix for it is coming with the new unity release08:48
pittididrocks: not sure, but buildinfo files sound like they could potentially differ between arches, too08:48
pittididrocks: ok, which is planned today, right?08:48
didrockswhich is even happening… now! :)08:49
* dpm hugs didrocks ;)08:49
dpmthanks guys08:49
* didrocks hugs dpm08:49
didrocksdpm: tell me if it doesn't fix it for you08:49
didrocksdpm: just to confirm, you didn't installed the unity-team ppa for 5.2, isn't it?08:49
dpmdidrocks, no, I'm on standard packages from the archive08:50
didrocksyeah, so makes totally sense :)08:50
didrocksdpm: tell me once you get the new unity that it actually fixes it for you08:50
dpmsure I'll let you know if it's fixed once I update unity08:50
didrocksgreat :)08:51
pittihey seb12808:58
seb128hey pitti, wie gehts?08:59
pittiseb128: gut, danke!08:59
pittiseb128: und Dir?08:59
seb128pitti, zehr gut, danke!08:59
seb128pitti, nice work on the language selector stack, impressive changelogs09:00
pittiheh, thanks09:00
pittiseb128: I blogged about the new PK/aptdaemon API09:00
seb128though the number of changes we have to accountsservice start to scare me a bit :p09:00
pittiyes, this stuff needs desperate cleaning09:00
pittiour shell scripts there are hackish09:00
didrockssalut seb128, ça va ?09:00
pittibut I'm more concerned about the lack of any feedback on the upstream bugs, for the API changes09:01
seb128lut didrocks, nickel, et toi?09:01
didrocksseb128: ça va :)09:01
seb128pitti, I guess somebody should direct ping them on IRC09:01
pittiyes, Gunnar was meaning to09:01
chrisccoulson_good morning everyone09:05
chrisccoulson_gah @ bug 92590709:05
pittididrocks: so that should fix the two light-themes bugs that jibel just assigned to us?09:05
ubot2`Launchpad bug 925907 in light-themes "12.04 thunderbird colour theme is unreadable" [High,Triaged] https://launchpad.net/bugs/92590709:05
pittihey chrisccoulson_09:06
seb128pitti, I reassigned them to Cimi09:06
pittireally, can we please avoid this kind of premature upload?09:06
didrocksseb128: isn't Cimi's bug without the new unity?09:06
seb128pitti, talk to ken, he's the one who did the snapshot and upload09:06
chrisccoulson_hey pitti, how are you?09:06
seb128didrocks, I doubt it, at least not this tb bug09:06
pittichrisccoulson_: quite fine, thanks!09:07
didrocksseb128: no, I'm speaking about the indicator one09:07
seb128didrocks, the fix is in unity 5.209:07
pittiseb128: sounds like a good case for a revert -- three major bugs in one day; want me to uplaod a revert?09:07
didrocksyeah, that's what I meant :)09:07
seb128pitti, no please09:07
seb128pitti, we knew that some issues would be there until unity 5.2 would be uploaded this morning09:07
seb128pitti, revert will create other bugs with the new gtk09:08
seb128like the icons having a square solid background and some menus issue as well09:08
seb128pitti, do you confirm the slider bug with unity 5.2?09:09
pittihm, not in my current session; let me start a guest session, I didn't restart since this morning's dist-upgrade09:10
seb128pitti, ok, so what didrocks said, part of the issue will be fixed with unity 5.209:10
pittiseb128: yes, confirmed09:11
seb128the indicators issue09:11
seb128we need the new theme for the new gtk so revert is not a good idea09:11
pittimvo: guten Morgen09:11
seb128the tb issue is lack of testing and will need to be fixed today though09:11
pittimvo: would you mind doing a s-c-aptdaemon-plugins upload? I can't push to bzr, so I didn't yet09:11
pittiugh @ http://people.canonical.com/~ubuntu-archive/testing/precise_probs.html09:12
pittipoor armel buildds :/09:12
mvopitti: sure, doing that now09:13
pittimvo: danke09:13
=== chrisccoulson_ is now known as chrisccoulson
chrisccoulsonseb128, pitti, so i guess from the scrollback that thunderbird isn't the only thing broken b the light-themes update?09:17
chrisccoulsoni haven't updated it yet09:17
chrisccoulsonbut the screenshot looks pretty unusable ;)09:17
seb128chrisccoulson, sort of, the unity issue is known and fixed in 5.2 which is about to be uploaded09:17
seb128to be fair one part of the issue is that gtk3 keep changing theming details this cycle which break the theme, we needed an update for the new gtk09:18
chrisccoulsonseb128, right. but thunderbird is using gtk2 :)09:18
seb128right, should be easy to figure what changed there, I guess most of theming work is for gtk309:19
seb128chrisccoulson, tb is fine for me but I use the light theme, let me try with the default one09:22
seb128yeah, I can confirm the issue09:24
seb128well it's not a stopper, you can read stuff when you mouseover, still should be fixed today, I will ping Cimi when he gets online09:24
chrisccoulsonand it didn't break the tabs in firefox too?09:25
chrisccoulsonactually, 'i'm upgrading now09:25
chrisccoulsonlets have a bit of breakage for a chance :)09:25
chrisccoulsonhmmm, you can tell i really need coffee this morning09:26
seb128could be that it did, who use tabs? ;-)09:26
seb128joke aside I never opened tabs in tb out of mistakes ;-)09:26
chrisccoulsoni can't believe jo has taken the car to go out, leaving me in a house with no coffee09:31
pittichrisccoulson: are you guys freezing to death ATM, too? -14 celsius here for several days now..09:43
chrisccoulsonpitti - yeah, it's certainly colder this week than it has been at any other point during this winter, and people are talking about the cold a lot, but it feels quite normal for this time of year here :)09:44
chrisccoulson-14 is pretty cold though ;)09:45
chrisccoulsonoh, i think it was -9 here last night, which is a bit colder than average09:45
chrisccoulsoni quite like the cold. it's refreshing :)09:46
chrisccoulsonw00t, just upgraded to thunderbird 13, unlucky for some ;)09:48
chrisccoulsonlol @ mark's reply on bug 820034 ;)09:50
ubot2`Launchpad bug 820034 in ubuntu-font-family-sources "Expansion: Miscellaneous Symbols and Pictographs U+1F4A9" [Wishlist,Confirmed] https://launchpad.net/bugs/82003409:50
chrisccoulsonseb128, b'ah, the light-themes update has broken the awesomebar styling :(09:53
chrisccoulsonwhich i had actually fixed in firefox a couple of weeks ago to make it not suck with ambiance ;)09:54
seb128chrisccoulson, well you are lucky, Cimi is online now09:56
seb128chrisccoulson, so just ping him to complain ;-)09:56
seb128he just reassigned the slider bug to ido09:56
seb128chrisccoulson, he also reassigned the tb bug to tb and invalidating the light-theme part of the bug :p09:57
jincreatorpitti: Hi. Did you see my comment in https://bugs.launchpad.net/bugs/916847 ?09:59
ubot2`Launchpad bug 916847 in ttf-unfonts "Please merge fonts-unfonts-core, fonts-unfonts-extra from (universe) from Debian unstable (main) " [Undecided,In progress]09:59
chrisccoulsonseb128, oh, it's funny how thunderbird is to blame when i didn't change anything in it yesterday ;)10:02
seb128chrisccoulson, yeah, I found as well ;-)10:03
pittijincreator: yes, we need to make them as small as possible; if the package baloons from 7.5 MB to 19, we just can't fit it on the CD any more and have to drop it from the default install10:19
didrocksjust trying logout/login, bbiab10:20
pittijincreator: if you install with network, they will be installed for Korean users, but they wouldn't be on the live system any more, or for non-network installs, so not really ideal10:20
jincreatorpitti: Well, but ttf-unfonts-core(fonts-unfonts-core) doesn't ship with Cd and default font is fonts-nanum.10:25
jincreatorpitti: Because fonts-nanum is at live system, unfonts and its related packages can install with network, I think.10:27
seb128didrocks, gnome-classic as a session is an Ubuntuism right?10:31
didrocksseb128: indeed10:32
didrocksit's compiz + gnome-panel10:32
seb128didrocks, thanks10:32
didrocksyw :)10:32
ricotzseb128, hello, you forgot to actually add libelf-dev to glib10:44
seb128ricotz, arg10:44
seb128ricotz, what difference does it make? is that only for the gresource utility?10:45
ricotzyes, it wont support reading elfs10:45
ricotzlike mentioned in the changelog ;)10:45
seb128ok, that's fine10:45
seb128it's only a play tool, that can wait the next upload10:45
seb128that doesn't make any difference for applications or users10:46
chrisccoulsonisn't the plural of "elf" "elves"? ;)10:46
seb128sorry about that, I'm fixing it in the vcs now, I will just not do another upload only for it10:46
ricotzseb128, as long no app depends on it to resolve some information on this way10:46
seb128ricotz, I go sidetracked by the gzip stuff and finished by dropping the old changelogs rather than using your workaround10:46
seb128ricotz, well, we will probably have a glib upload next week, I think precise users are fine for the w.e ;-)10:47
ricotzyeah, dropping them was better10:47
ricotzright ;)10:47
didrocksunity 5.2 is released *phew*10:47
seb128didrocks, \o/10:48
chrisccoulsonnice :)10:48
pittijincreator: hm, indeed; they are on xubuntu and edubuntu, but that shouldn't hurt them much10:48
seb128ricotz, btw if you were doing merge requests, I could just bzr merge and we would avoid such errors :p10:48
ricotzhehe (or forget about them)10:49
seb128I'm lazy enough that I would not forget about the useful ones ;-)10:49
seb128I don't like to redo work manually :p10:49
ricotzpatch -R -p0 < diff seemed easy enough10:50
seb128if you have a diff yes10:50
seb128but you don't provide that10:50
ricotzi gave you the link10:51
ricotztwice ;)10:51
seb128right, which had stuff like changing -c4 -c1 as well10:51
seb128or dropping some of our patches10:51
seb128or changing build-depends and break revisions10:51
seb128i.e I couldn't apply it, I had to cherry pick the actual fixes10:51
ricotzthis is the one i gave you10:52
ricotzlooks pretty clean to me10:52
seb128ah that one, yeah, there was some weird reshuffling in the .install, the rules hack was not good and it was reversed10:52
seb128but otherwise it was ok ;-)10:53
seb128well anyway fixing the vcs, thanks10:53
tkamppeterpitti, hi10:54
pittihello tkamppeter10:54
tkamppeterpitti, it is about bug 911436 and bug 925227. A problem in p11-kit causes a lot of crashes, mainly in CUPS and the bugs are accumulating duplicates.10:57
ubot2`Launchpad bug 911436 in p11-kit "https crashed with SIGSEGV in lookup_or_create_bucket()" [Critical,Triaged] https://launchpad.net/bugs/91143610:57
ubot2`Launchpad bug 925227 in cups "package cups 1.5.0-16 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,Invalid] https://launchpad.net/bugs/92522710:57
pittitkamppeter: yes, I noticed10:57
pittitkamppeter: I just replied to the second bug10:57
tkamppeterpitti, is someone maintaining the p11 stuff?10:58
pittiit's assigned to foundations, but no progress on it yet10:58
tkamppeterpitti, I have rasied the first bug to "Critical".10:59
pittichrisccoulson: ah, so bug 805136 got marked as fixed upstream, but it looks it was mainly tested on windows; do you know whether that fix applies to linux as well?10:59
ubot2`Launchpad bug 805136 in thunderbird "accounts window doesn't fit on screen and no scrollbar to show hidden fields" [Medium,Fix released] https://launchpad.net/bugs/80513610:59
chrisccoulsonpitti - yeah, it's got scrollbars and is resizeable on the current nightly build here11:01
pittisweet, thanks11:01
chrisccoulsonoh, it landed in september. i wonder which release that's in now11:02
chrisccoulsonit should be in thunderbird 9 then11:02
jincreatorpitti: I didn't know that these distros are not maintained by same ubuntu-seeds. If than, should I report bug to request change default fonts?11:08
pittijincreator: for xubuntu, yes; edubuntu pulls in pretty much all lang support, they have a DVD and can affort it11:09
pittichrisccoulson: hm, we do have that11:10
* pitti hugs didrocks11:12
* didrocks hugs pitti back :)11:13
didrocks(sorry, will be hard to reach my number of closed bugs now!) :)11:13
jincreatorpitti: OK. Then I'll report bug and if Xubuntu developers change it, I'll submit comment at my merge request bug.11:13
pittididrocks: http://reports.qa.ubuntu.com/reports/bug-fixing/precise-fixes-report.html11:16
pittiand I had almost got you!11:16
pittijust one more to go11:16
pittijincreator: thanks; but I think that merge is fine now anyway11:16
seb128pitti, that was before he added an another hundred to his counter? ;-)11:16
pittiseb128: before11:16
jincreatorpitti: Ah, great.11:16
didrockspitti: toooooo late! :)11:16
pittijincreator: so, if anyone wants to sponsor this, I left an "OK" comment11:17
pittijincreator: otherwise I'll look into it on Monday (need to run in a few mins)11:17
pittiseb128, didrocks: I'll spend the afternoon in a train, working offline; do you still need me to do anything in the next few mins?11:17
* didrocks thinks that mvo's results are quite disappointing, he doesn't whip the crack enough to get his bug fix counter at the top :)11:17
jincreatorpitti: I see.11:18
didrockspitti: all sounds safe and good! :) Enjoy your travel11:18
seb128pitti, I think we are fine, is there anything we should watch for?11:18
seb128pitti, have a good w.e!11:18
pittiseb128: noting in particular from my side11:18
seb128ok, great11:18
pittiwe don't have a release meeting today11:18
seb128see you on monday then ;-)11:18
pittididrocks: thanks11:18
pittiseb128: merci11:18
pittienjoy the weekend, y'all :011:18
didrockshave a good week-end pitti :)11:19
pittifinally have some time to work on that "report disk wakeup culprits" program11:19
didrocksoh great :)11:19
pittiseb128: I looked at bug 890592, but not very successfully so far; need to debug that harder, kernel is not giving me cookies :(11:20
ubot2`Launchpad bug 890592 in udev "Ejecting a CD or DVD manually does not unmount it" [High,In progress] https://launchpad.net/bugs/89059211:20
seb128pitti, I noticed your comment, thanks for looking at it11:20
seb128pitti, good that you get it as well at least, it will make debugging easier ;-)11:20
seb128it's quite disturbing to be still able to browse a CD in nautilus, i.e enter dirs and everything when the CD has been ejected and is on your desk11:21
pittiI can easily fix this part11:22
pittiand already did so locally11:22
pittibut now the eject button doesn't work at all any more11:23
pittii. e. you need to use the nautilus one11:23
pittithat's the kernel part11:23
didrockssil2100: keep me up to date on the compiz progress, will be nice to push to a ppa for user testing11:24
seb128pitti, ok11:24
didrockssil2100: also, I think that now unity 5.2 is ready, your team needs to bump unity in the hud ppa11:24
didrocks(this is something to see with gord to take his latest code)11:25
gordah yes11:25
gorddidrocks, sil2100 - will wait for later today, hud branch dep's on another branch that conflicts with trunk. need to wait for the americans to awaken11:29
didrocksgord: let's wait on the us then! :)11:29
* pitti waves good bye, enjoy the weekend everyone!11:29
didrockssee you pitti :)11:30
seb128pitti, have fun, see you11:30
sil2100didrocks: working on it, it seems I was missing some patches to make unity and upstream compiz work11:31
sil2100didrocks: it's rebuilding now11:31
* sil2100 sighs12:01
sil2100Could anyone take a look ;)?12:08
sil2100stdout and stderr from my 'buggy' unity compiz session12:09
sil2100The panel, launcher and dash are there, but I can't see any graphics12:09
sil2100Instead, I get black rectangles everywhere, blinking from time to time12:10
sil2100During moving windows it's the same - and it only happens with the unityshell plugin enabled12:10
=== MacSlow is now known as MacSlow|lunch
=== s9iper1_ is now known as s9iper1
sil2100didrocks, smspillaz ;)? ^12:49
didrockssil2100: i think you should see that directly with upstream, so sam :)12:50
=== dduffey_afk is now known as dduffey
=== greyback is now known as greyback|lunch
Berethi all13:05
Beretare the depends for unity currently broken?13:06
didrocksBeret: bamf FTBFS13:06
didrocksBeret: because of the new glib, just fixed it13:06
seb128didrocks, not because of glib, don't blame glib for use of outdated apis in dx code ;-)13:07
didrocksyeah, it's all dbo's fault! :)13:07
=== MacSlow|lunch is now known as MacSlow
ockhamhi; this merge request is already approved, but doesn't seem to have been merged yet:13:10
ockhamit would be great to have it in precise...13:11
ockhamanyone know why that hasn't been done yet?13:11
seb128ockham, it will be in precise don't worry, they just go busy and didn't roll tarballs for a while13:11
ockhamseb128: great to hear!13:11
seb128that's starting again, we got a new libdbusmenu and indicator stack being uploaded13:11
cjwatsonhey, I'm trying to fix bug 925427, and am stuck on a stupid problem: I can't get GTK to hide the contents of a widget without causing its container to want to resize13:14
ubot2`Launchpad bug 925427 in ubiquity ""Ready when you are" message is misleading and incorrect" [Medium,Confirmed] https://launchpad.net/bugs/92542713:14
cjwatsonis there some way I can get a widget to just not display anything, but keep its existing size?13:14
cjwatsonthis is frustrating since I'm sure it should be easy, I just can't find it13:15
cjwatsonor maybe I want to freeze the container's size allocation or something13:17
seb128cjwatson, hum, not sure how to do that, you might have a better chance to get a reply on #gtk+ or #gnome-hackers on irc.gnome.org13:20
Beretdidrocks, sweet thanks13:20
seb128try the gtk channel you usually have a good chance to get a reply there13:20
cjwatsonfair enough, thanks13:20
cjwatsonI have one last idea I'll try first :)13:20
cjwatson(temporarily set the container's size request to its current allocation)13:21
ockhamcjwatson: if you succeed, can you post a link to a vcs commit or whatever to let others see how you did it?13:23
cjwatsonbingo!  cardboard programmer strikes again13:34
cjwatsonseb128,ockham: there's some more work to do before I commit this (notably, I now need to do the same for KDE), but it's essentially http://paste.ubuntu.com/827602/13:35
seb128cjwatson, ok, seems to make sense, thanks ;-)13:36
mdeslaurheh, "cardboard programmer"...first time I've heard that :)13:37
cjwatsonjodh used the term "confessional debugging" to me yesterday which is a synonym but I hadn't heard that one either ...13:38
mdeslaurhehe, good one too :)13:38
seb128chrisccoulson, can I get your opinion on something? ;-)13:39
seb128chrisccoulson, http://git.gnome.org/browse/gnome-desktop/commit/?id=a81bd53c44a0785373ba9fba543556e3c4e12431 do you see that as an api change?13:39
sil2100didrocks: I'll put those packages I compiled up somewhere13:40
didrockssil2100: are they working?13:40
seb128chrisccoulson, do you think it will make g-s-d bug if we take that without updating gsd?13:40
sil2100didrocks: well... ;) Not on my desktop, that's why I would like to see if it's the same on other platforms13:40
sil2100Since it seems to be a driver issue13:40
chrisccoulsonseb128, it should be fine if we drop our patch from g-s-d13:41
didrockssil2100: you can maybe see with your fellows on your team then, to see if it works? :)13:41
chrisccoulsonseb128, yeah, we should drop this one if we take the new gnome-desktop: http://bazaar.launchpad.net/~ubuntu-desktop/gnome-settings-daemon/ubuntu/view/head:/debian/patches/46_share_rr_screen.patch13:42
seb128chrisccoulson, I'm pondering backporting the work federico did to improve dock stations support in the xrandr code, but first we would need to update gnome-desktop, so trying to figure in what order we should do things13:42
seb128chrisccoulson, http://git.gnome.org/browse/gnome-settings-daemon/commit/?id=6e17bc786f55f283f2c721249197e7740174fd43 has "This requires gnome-desktop 3.3.4 or later, as the semantics of gnome_rr_screen_new() are slightly different there - now it returns a singleton instead of a new object every time."13:43
seb128I was wondering if we would have issues without that13:43
sil2100didrocks: will do so ;)13:43
seb128chrisccoulson, right, that patch is to drop for sure, I'm trying to figure what will happen to g-s-d without the patch and the new gnome-desktop13:44
seb128I guess I should just try ;-)13:44
chrisccoulsonseb128, oh, the relevant part of http://git.gnome.org/browse/gnome-settings-daemon/commit/?id=6e17bc786f55f283f2c721249197e7740174fd43 is the bit which drops the change that is equivalent to our patch13:44
chrisccoulsonit seems that rodrigo committed our change upstream13:45
chrisccoulsonand this commit reverts it for the new gnome-desktop13:45
chrisccoulsonso, i think we should be fine13:45
didrockssil2100: excellent, let's cross fingers then :)13:45
seb128chrisccoulson, oh, I see, http://git.gnome.org/browse/gnome-settings-daemon/commit/plugins/xrandr/gsd-xrandr-manager.c?id=4cb1515ebaa635e7686d153eba977a206b25651413:45
seb128chrisccoulson, thanks13:46
=== greyback|lunch is now known as greyback
Beretdidrocks, still get - http://paste.ubuntu.com/827617/13:55
didrocksBeret: you will need to dist-upgrade13:55
Beretdidrocks, that was from a dist-upgrade13:55
didrocksBeret: but again, new unity is not built yet13:55
Beretoh ok13:55
didrocksBeret: get apt-get update and apt-cache policy at some point will tell you that unity 5.2 is available13:57
Beretdidrocks, oh ok, thanks, I was under the impression it was built already14:05
Beret's all good14:05
GunnarHjCurrent unity dependency issue:14:10
GunnarHj~$ sudo apt-get install ubuntu-desktop14:10
GunnarHjReading package lists... Done14:10
GunnarHjBuilding dependency tree14:10
GunnarHjReading state information... Done14:10
GunnarHjSome packages could not be installed. This may mean that you have14:10
GunnarHjrequested an impossible situation or if you are using the unstable14:10
GunnarHjdistribution that some required packages have not yet been created14:10
GunnarHjor been moved out of Incoming.14:10
GunnarHjThe following information may help to resolve the situation:14:10
GunnarHjThe following packages have unmet dependencies.14:10
GunnarHj ubuntu-desktop : Depends: unity but it is not going to be installed14:10
GunnarHjE: Unable to correct problems, you have held broken packages.14:10
didrocksdesrt: hey, are you around?14:16
didrocksGunnarHj: not an issue, you get nux installed, but unity is not yet published14:16
didrocksand there is an ABI break14:16
seb128GunnarHj, it's a transition, you need to wait for unity to build14:16
GunnarHjdidrocks, seb128: Ok, thanks for letting me know. Just wanted to mention it in case ... :)14:17
didrocksno worry, should be published soon :)14:17
* kenvandine installs thunderbird14:20
desrtdidrocks: yup14:35
desrtgood morning, everyone14:36
didrocksdesrt: good morning!14:36
didrocksdesrt: remember this discussion about the "default" value?14:36
desrtyesish :)14:36
desrtthe one about indicating on a slider, etc?14:36
desrthow can i help you?14:37
desrtseb128: poke14:37
didrocksso basically, I tried that: http://pastebin.ubuntu.com/827651/14:37
seb128desrt, hi14:37
desrtseb128: remember how i asked you to package the beta automake?14:37
seb128desrt, yes?14:38
didrocksunfortunatly, it seems that default_gtk_theme doesn't really contain the default14:38
seb128desrt, I hate you for that :p14:38
didrocksbut the current selection14:38
desrtseb128: why?14:38
seb128desrt, I've just waster 1 hour try to debug why the current debian version I upload to precise fail in the testsuit14:38
desrtseb128: doko hit you over the head for the NMU?14:38
seb128it hits Makefile.am: required file `./NEWS' not found14:38
seb128and exit 114:38
desrtshouldn't uloading a debian src package direct to precise cause no trouble at all?14:39
seb128desrt, it should, dunno why the testsuit work for them and not for us14:39
desrtsorry :)14:39
seb128desrt, https://launchpad.net/~ubuntu-desktop/+archive/ppa/+build/3183452 is your new version as well14:39
seb128let's see if that one builds14:39
desrtdidrocks: so one note: g_settings_revert() here is probably not doing what you think14:39
seb128desrt, debian has 1.11.2 not 1.11.314:40
desrtseb128: ya.  1.11.3 only came out 2 days ago14:40
didrocksdesrt: oh?14:40
desrt(ie: no point in trying to package the beta anymore)14:40
desrtdidrocks: once a gsettings object is delayed, it is delayed forever14:40
desrtdidrocks: lemme try this code locally14:40
desrtit 'looks fine'14:40
didrocksinteresting :)14:40
desrtother than the revert thing14:40
seb128desrt, hum, so you don't need 1.11.3?14:41
seb128desrt, I though you wanted the new version14:41
desrtseb128: no.  1.11.3 is the one i want.14:42
seb128desrt, well anyway it's uploaded14:42
desrt1.11.2b was a beta release for 1.11.314:42
desrtseb128: nice :)14:42
seb128desrt, it's the one in https://launchpad.net/~ubuntu-desktop/+archive/ppa/+build/318345214:42
desrtseb128: thanks very much.  i'll test that as soon as it is built.14:43
seb128desrt, it will not build, seems it has the same issue :-(14:43
* desrt just got a similar email this morning from someone else: http://koji.fedoraproject.org/koji/taskinfo?taskID=375879614:43
seb128Makefile.am: required file `./ChangeLog' not found14:43
seb128Makefile.am: installing `./COPYING' using GNU General Public License v3 file14:43
seb128Makefile.am:     Consider adding the COPYING file to the version control system14:43
seb128Makefile.am:     for your code, to avoid questions about which license your project uses.14:43
seb128Makefile.am:1: installing `./texinfo.tex'14:43
seb128+ exit_status=114:43
desrt'touch'? :)14:43
seb128desrt, well it's the test suit, I need to understand how it creates the test subdirs and why it's working in Debian14:44
desrtseb128: if it's too much hassle, i don't blame you if you want to give up14:45
desrtthis is really not your responsibility at all14:46
seb128desrt, well I've synced the current version to precise and we will need to sort it anyway, the previous version was failed to build as well it seems, so probably something else changed14:46
desrti've been in touch with the debian maintainer.  he's pretty responsive and promised me some speedy packaging when it came out14:46
desrtso i'll ping him back14:46
seb128desrt, don't worry one way or another I need to fill my friday afternoon ;-)14:47
desrtnice canadian guy :)14:47
desrtafter that, once debian has it updated, i feel i have better grounds to harass doko14:47
seb128desrt, hum, if you email him can you drop a "btw the current version failed to build in Ubuntu: https://launchpadlibrarian.net/91800731/buildlog_ubuntu-precise-i386.automake1.11_1%3A1.11.2-1ubuntu1_FAILEDTOBUILD.txt.gz, do you have a clue about what could be wrong"?14:48
desrtseb128: i don't want to tell him i'm involved with ubuntu.  he may stop responding to my mails :p14:48
seb128desrt, lol14:48
desrt(actually, i already sent the mail so it's just too late) :p14:48
desrtdidrocks: huh14:50
desrtdidrocks: i suspect you have found a bug14:50
desrti get to avoid working on the hud for another day!14:52
* desrt checks to make sure olli is not around14:52
didrocksdesrt: ahah, "emergency fix for desktop?" :)14:53
desrtyes.  highly critical issue, clearly.14:53
desrtno.. but seriously... i can't imagine this takes more than half an hour to track down14:53
desrthopefully it's something easy to fix14:53
didrocksgreat :-)14:58
desrtya.  it's a dconf bug with an obvious (but slightly tricky) fix15:01
desrti'm surprised you're the first person to notice this15:01
didrocksyeah, as you were telling that it's a quite well known trick :)15:03
dobeydidrocks: http://pastebin.ubuntu.com/827689/15:04
dobeydidrocks: what's up with upgrading nux removing unity? :)15:04
didrocksdobey: nux has an ABI break, apt-cache policy unity should show you you don't have 5.215:05
didrocksit's published now, so should be soon available15:05
desrtdidrocks: http://git.gnome.org/browse/dconf/commit/?id=33f32ba175c6192d500ed7081d3cf973e724379a15:06
dobeyoh looks like it's available now15:06
didrocksah :)15:06
desrtdamn.  that only took 15 minutes.15:06
didrocksdesrt: what, even not half an hour? :)15:06
dobeyit wasn't 10 minutes ago when i last did apt-get update15:06
didrocksdesrt: ok, I see that you need to know the layering level of dconf for it :)15:07
desrtya.. as i said, tricky15:07
desrtbut obviously trivial15:07
desrtanyway.. next release will be on feb 615:07
desrtso you can wait for that or you can vendorpatch until then15:08
* desrt thinks that package build priority should be related to historical time-to-build15:12
desrt'oh?  you did a dconf upload?  well, that won't take long at all.... here... go first."15:12
desrt"oh... your name is Sweetshark?  i think i have a free slot next week..."15:13
cjwatsondesrt: well volunteered to work on the LP build scheduler15:15
didrocksdesrt: I think this can wait anyway :)15:18
seb128desrt, ok, it's your day15:38
desrtseb128: sup?15:40
didrockssecond bug? :)15:40
desrtdidrocks: by that measure, every day is my day :)15:40
seb128desrt, https://launchpad.net/ubuntu/+source/intltool/0.50.0-0ubuntu315:40
didrocksdesrt: heh :)15:40
desrtseb128: damn.  i just filed a new bug against intltool :)15:40
desrtseb128: thanks :)15:41
seb128desrt, and I hate automake, the tests behave different when ran manually, the test which failed on the buildd works for me, I retried and it's ok now on the buildds15:41
desrtseb128: must be the old kernel version or something ;)15:41
seb128desrt, seems it was a random error, looks like you will get your deb in a bit15:41
seb128desrt, ;-)15:41
seb128desrt, automake takes longer to build that glib and gtk together15:41
desrtseb128: it's all tests...15:42
seb128well it takes like 15 seconds to build but it tests for 45 minutes15:42
seb128desrt, yeah, their tests are taking ages15:42
* desrt needs to keep up the pressure on danilo15:42
seb128I've a new respect for automake, we might not like it but I didn't know it was that much test covered :p15:42
desrtseb128: you merely have to justify that by telling yourself that they must be really really really bad at writing testcases :p15:43
seb128desrt, I think your bug is a dup of https://bugs.launchpad.net/intltool/+bug/90334015:44
seb128desrt, https://code.launchpad.net/~hiberis/intltool/bug-903340/+merge/8908815:44
ubot2`Launchpad bug 903340 in gentoo "intltool.m4 should use --no-translations only if >=intltool-0.50.0 is detected" [Medium,Won't fix]15:44
desrtah yes.  probably.15:44
desrtuh.  wontfix?15:44
seb128desrt, the gentoo line15:45
seb128desrt, launchpad can be confusing like that :p15:45
dobeyoh hmm15:45
pgranerdidrocks, are you seeing compiz and unity-panel service segfaults on the new packages today?15:45
seb128pgraner, no, stacktrace? can you pastebin your .xsession-errors?15:45
desrtseb128: i believe my bug stands15:46
didrockspgraner: I got not issue in the user testing as well15:46
desrtthe bug danilo marked as wontfix is only 1 of my 3 suggested solutions15:46
didrocksand nothing relevant on lanchpad15:46
seb128desrt, it seem a similar issue, though the solutions suggested are different15:46
seb128didrocks, I bet you a beer that it's partial upgrades somewhat :p15:46
didrocksyeah :)15:46
didrockslike no unity-services15:47
* desrt will fight with him about it :)15:47
didrockspgraner: apt-cache policy unity-services15:47
pgranerseb128, didrocks: https://bugs.launchpad.net/ubuntu/+source/unity/+bug/92613715:49
ubot2`pgraner: Error: <Bugtracker.plugin.Launchpad instance at 0xa5ae8cc> bug 926137 not found15:49
didrockspgraner: waow, the stacktace looks weird15:50
didrockslet's wait for the retracer15:50
didrockspgraner: do you have that constantly?15:50
seb128didrocks, it retraced and failed15:50
didrocksoh right, second comment15:51
pgranerpgraner@beavis:~$ apt-cache policy unity-services15:51
pgraner  Installed: 5.2.0-0ubuntu115:51
pgraner  Candidate: 5.2.0-0ubuntu115:51
pgraner  Version table:15:51
pgraner *** 5.2.0-0ubuntu1 015:51
pgraner        500 http://us.archive.ubuntu.com/ubuntu/ precise/main amd64 Packages15:51
pgraner        100 /var/lib/dpkg/status15:51
didrocksok, sounds fine…15:51
seb128didrocks, side effect of having ddebs out of soyuz, it takes some hours to get the new ddebds15:51
seb128didrocks, pgraner: sees to be 91622815:52
seb128not new in 5.2 and quite some duplicates15:52
seb128somebody should hit on dx for having ignored it for 5.215:53
didrocksI'll add it as critical on unity distro priority15:53
didrocksas it seems they don't look at milestones15:53
pgranerseb128, yea by box is borked had to move to laptop15:53
didrockspgraner: you didn't get it at all with 5.0 before?15:54
seb128pgraner, does it happen all the time?15:54
pgranerdidrocks, everytime, I had filed a compiz bug with stacktrace  but its not in lp now15:54
didrockspgraner: ok, I've added it as critical on the unity distro priority list15:54
didrockspgraner: I'll make sure Tim is aware of it15:55
didrocks(and eventually distro-patch is)15:55
pgranerdidrocks, refiling the compiz one now15:56
didrockspgraner: yes please (I have good feeling it's due to nux or unity as compiz didn't change)15:56
seb128pgraner, you can try your luck on #ubuntu-unity to see if anyone is wanting to look at the issue15:56
seb128pgraner, it's not a new bug, it's opened for some time and was milestoned but they ignored it so far, maybe you can provide them useful infos since you get the issue15:56
pgranerdidrocks, its nux ... bug #92614915:57
seb128pgraner, 916228 is the bug you hit15:57
pgranerseb128, sure just tell me, hell I'll give em' remote access to the box15:58
didrockspgraner: is 926149 a typo? I don't find it or lost access :)15:58
seb128pgraner, well, it's really a dx issue, try #ubuntu-unity15:58
seb128didrocks, it's probably not retraced yet15:58
didrocksright, the team shouldn't be subscribed to it yet15:59
seb128didrocks, the new apport bugs are private for us until retracing15:59
pgranerdidrocks, yep thats the correct number15:59
didrockspgraner: for your first bug, I flagged it for 5.2, and set it as high (then Tim set it as critical)15:59
didrocksshame that it was ignored15:59
seb128pgraner, can you go on #ubuntu-unity and ask them if you can help providing infos for 91622815:59
didrocksseb128: understood!15:59
seb128or DBO maybe can help you15:59
pgranerseb128, sure thing15:59
seb128but better on #ubuntu-unity anyway16:00
seb128we only do the packaging, that issue is in the unity code16:00
seb128tjaalton, hey16:07
seb128tjaalton, did you forget to push your totem upload to the vcs? it's still UNRELEASED there, you should copy the upload changelog, debcommit -r and push ;-)16:07
tjaaltonseb128: what about now?16:10
seb128tjaalton, thanks ;-)16:11
tjaaltonprobably just forgot the second push, did it now and it seemed to do "something" :)16:11
seb128tjaalton, it's all good now16:11
=== fenris is now known as Guest21494
=== Guest21494 is now known as ejat
desrtseb128: so... how do you feel about orlando in september?16:38
seb128desrt, ?16:38
desrti'm actually seriously considering going16:39
desrti think we should send a large contingent :)16:39
desrt"uh.  hi.  this is an _ubuntu_ hotel."16:39
jbichaLinux for Humans not Lizards!16:39
seb128desrt, it's a short fly for you, I don't consider crossing the ocean just to troll vuntz :p16:39
* desrt would go great distances to troll vuntz16:40
desrtdidrocks: you in? :)16:40
seb128desrt, I'm sure vuntz likes this hotel and remember fondly his UDS there :p16:41
didrocksdesrt: yeah, I'm here :)16:42
desrtseb128: the first part is probably true. ;)16:42
didrocksoh well, I'm in for sure!16:42
desrtdidrocks: seriously?16:42
desrtbecause i'd seriously do this16:42
didrocksahah ;)16:42
didrockstrolling vuntz at any cost!16:43
desrtit's not just about vuntz, you know16:43
desrtlizards are people too16:44
* didrocks still remind every that he's sharing his bed with a Novell's lizard because Julie likes it :)16:46
* desrt has such a lizard, but avoids sleeping with it16:47
sil2100didrocks, smspillaz: just a quick question16:47
sil2100For the packages I was to prepare, what version of libnux should I use?16:47
didrockssil2100: hum, you mean, unity in general,16:48
didrockssil2100: you should just take the version in the distro16:48
didrocksof unity16:48
didrocksreapply my changes16:48
didrocksand build from that16:48
didrocksdesrt: I'm sure it was a vuntz evil plan when he gave me one at guadec!16:49
sil2100Since I'm trying to make someone from my team install my packages, but it seems he has a too 'recent' libnux2.0-0, which makes installing my unity packages impossible16:49
didrockssil2100: is he on precise?16:49
sil2100didrocks: yes, and he seems to have a newer libnux2.0-0 - in version 2.2-0ubuntu116:50
didrockssil2100: yeah, that's why you should build with my latest unity branch16:50
sil2100While I have (and the packages I built require) libnux2.0-0 in version 2.2-0ubuntu216:50
didrockssil2100: lp:~ubuntu-desktop/unity/unbuntu16:50
sil2100But wait, I think I did16:51
didrocksthen takes the branch with my changes I pointed to you (the sed change in debian/rules)16:51
didrocksand apply it on top of this branch16:51
sil2100hmmm, ok16:52
* sil2100 got a bit lost in all the packages, patches and versions16:52
sil2100One moment then ;) Thanks16:53
desrtdidrocks: https://bugzilla.gnome.org/show_bug.cgi?id=66932416:57
ubot2`Gnome bug 669324 in gsettings "API for getting 'reset' GSettings value" [Normal,New]16:57
desrtdidrocks: basically a bug talking about the reasons for not wanting that API.  i welcome your thoughts.16:58
didrocksdesrt: oh sure, sorry, kind of pingy machine right now, but I'm opening the tab :)16:58
desrtno worries.16:59
popeycjwatson: is it expected behaviour during 12.04 install to 'wipe swap space for security' even on a brand new laptop which only ever had windows 7 on it?17:23
seb128does somebody understand what's going on there?17:35
seb128desrt, ^ maybe you? ;-)17:35
seb128it's basically filesystem stuff but I don't get it, it's a file it would let me rm17:35
desrtseb128: i assume you are 'user' here17:36
cjwatsonpopey: yes, if you're using encrypted swap17:36
seb128I start thinking it's a libc, coreutils bug or something17:36
cjwatsonor rather an encrypted home directory or whatever it is17:37
desrtseb128: there's no good reason for why that should be happening17:37
seb128desrt, I can't recreate the issue myself by creating a dir and file with the same permission17:37
desrtseb128: it looks like a weird filesystem issue17:37
seb128desrt, that's the automake testsuit issue17:37
desrtor maybe apparmor is doing something17:37
seb128yes, wtf17:38
desrtall you need for 'rm' is write access to the directory17:38
seb128I don't get it17:38
apwdesrt, it worked for me17:38
seb128and I can't recreate it manually17:38
desrtseb128: there's something weird going on on the builders17:38
desrt(as usual)17:38
seb128somewhat the automake script is creating it in a weird way17:38
seb128desrt, no, I get it on my disk17:38
seb128it's my box17:38
seb128I can cp the dir around and it's still buggy17:38
desrtbut when you create it yourself, not17:39
seb128but if I tar and untar it the untarred dir works17:39
desrtmaybe ACL?17:39
desrtaccess control lists17:39
seb128right, but how do I check that?17:39
desrtlsattr might be a place to start?17:40
seb128thanks I was trying to remember that command :p17:40
* desrt knows so very little about this stuff17:40
seb128$ lsattr bug17:40
seb128-------------e- bug/file217:40
desrtya.  that's normal.17:40
desrtcheck the dir?17:40
seb128bug is the dir17:40
desrtya.  what's the dir say?17:40
desrtlsattr -d bug17:40
seb128$ lsattr -d bug17:41
seb128-------------e- bug17:41
desrtseb128: and what happens if you use 'tar' to do the copy?17:41
desrtand can you send me the result?17:41
seb128desrt, <seb128> but if I tar and untar it the untarred dir works17:41
seb128desrt, I tried, that "fixes" it17:41
apwseb128, where did this come from this directory ?17:41
desrtseb128: can you paste the result of 'strace rm -rf bug' ?17:41
apwseb128, and have you checked dmesg17:41
* desrt wants to see the unlink() call failing17:42
seb128apw, the dir comes from the automake1.11 test suit17:42
seb128a make distdir test case17:42
seb128the same testsuit works on debian17:42
apwan strace sounds good here, pls pastebin it17:42
seb128apw, nothing weird in dmesg17:42
desrtseb128: sounds like apparmor, then?17:43
seb128apw, well "dmesg" is useless, it's spammed by tons of17:43
seb128[31734.002164] Either the lower file is not in a valid eCryptfs format, or the key could not be retrieved. Plaintext passthrough mode is not enabled; returning -EIO17:43
sil2100(don't mind me)17:43
apwseb128, is bug in the ecryptfs ?17:43
desrtthis is on a cryptfs homedir?17:43
seb128desrt, that's just spam, nothing to do with this bug, buildds don't use ecryptfs17:43
jdstrandgrep DEN /var/log/kern.log17:43
jdstrandthat doesn't sound like apparmor17:43
seb128apw, no, the build fails on buildds17:43
popeycjwatson: oh, encrypted home implies encrypted swap? makes sense ☺17:44
jdstrandwe don't profile the shell or 'rm'17:44
apwbut right now, is this bug directory in there now17:44
seb128that's the strace17:44
apwseb128,what filesystem type is 'bug' in17:45
seb128apw, ext417:45
cjwatsonpopey: yep17:45
apwseb128, no overlays or anything17:45
jdstrandsounds like ecryptfs too17:45
seb128right, my user dir is ecryptfs, but the same test fails on the builders17:46
seb128let me try to get out of the ecryptfs for debugging17:46
apwseb128, so this isn't ext4 at all then ?17:46
apwseb128, so we may have a different local reason for failure here17:46
desrtseb128: can you try on your local machine inside of a native ext4 fs?17:46
apwcan you confirm whether you get an new ecryptfs report for each rm ?17:46
desrtor on a tmpfs, even?17:46
apwas the error there means you have a broken encrypted file on disk, often a 0 length file in the underlying fs17:47
desrtseb128: also... buy a thinkpad :)17:48
seb128sorry guys, trying to get the issue out of ecryptfs17:48
seb128apw, no, I don't get new ecryptfs error every time I try to rm it17:49
seb128that's in my /tmp17:52
seb128i.e out of my user ecryptfs dir17:53
desrtseb128: okay.  next show us a strace of you cping the file17:53
jdstrandseb128: your lsattr output is interesting17:53
desrtstrace cp -a bug /tmp17:53
jdstrandthe 'e' is not normal. I don't think tar will preserve extended attributes17:53
seb128desrt, http://paste.ubuntu.com/827847/17:53
desrtthis is getting awesome17:53
jdstrandman chattr tells me that 'e' is 'extent format'17:54
desrtfsetxattr(4, "system.posix_acl_access", "\x02\x00\x00\x00\x01\x00\x04\x00\xff\xff\xff\xff\x04\x00\x04\x00\xff\xff\xff\xff \x00\x04\x00\xff\xff\xff\xff", 28, 0) = 017:54
seb128jdstrand, hum17:54
desrtseb128: apt-get install attr17:54
=== fenris is now known as Guest21540
desrtseb128: and run getfattr -d on some things17:55
=== Guest21540 is now known as ejat
desrtseb128: this is almost certainly some posix acl thing after all17:56
jdstrandseb128: actually, that might be a flase alarm17:57
seb128desrt, getfattr -d outputs nothing17:57
jdstrandseb128: looks like 'e' is used when it is on ecryptfs17:57
desrtseb128: even on the file in /tmp ?17:57
desrtafter copying...17:57
seb128$ getfattr -d /tmp/unity_support_test.017:58
desrtthis is very annoying!!17:58
desrtseb128: i guess root can delete it anyway, right?17:58
seb128desrt, yes17:58
desrtseb128: it strikes me as a bit odd...17:59
desrtbecause clearly fsetxattr is being called here17:59
seb128desrt, I will try to get a sharable testcase17:59
seb128well you can apt-get source automake1.11 and build it if you want17:59
desrtseb128: try with -m -17:59
desrtto fgetattr17:59
desrtor getfattr (dumb command name)17:59
desrtseb128: looks like the default behaviour is to filter out 'system' items from the output of -d18:00
desrtseb128: looks like there is also a getfacl command18:01
seb128still no output18:01
seb128do you get output from it?18:01
desrtseb128: i get nothing from it, but my filesystem isn't filled with evil files :)18:02
seb128$ getfacl file318:02
seb128# file: file318:02
seb128# owner: seb12818:02
seb128# group: seb12818:02
seb128$ rm file318:02
seb128rm: remove write-protected regular empty file `file3'? y18:02
seb128rm: cannot remove `file3': Permission denied18:02
desrtbehold.  totally normal.18:02
desrtseb128: this is amazing...18:04
desrtseb128: what if you copy without -a ?18:05
* desrt would be very surprised if the problem still happened in that case18:05
seb128desrt, wait, I'm very confused now, I will redo a proper build in /tmp to be sure to have nothing weird from my ecryptfs18:06
desrtseb128: it shouldn't matter18:06
desrtseb128: if there is no acl being copied to /tmp then i can't imagine what the issue would be18:06
desrtseb128: but yes.  please try that :)18:06
stgraberseb128: you tried lsattr on it already I guess? (should be -------------e- if nothing special)18:07
seb128stgraber, yes18:07
* didrocks waves good evening and good week-end!18:09
seb128'night didrocks18:09
desrtdidrocks: good night18:09
didrocksmerci seb128, toi aussi! You too desrt :)18:09
seb128desrt, seems it doesn't bug the same way out of my user dir18:10
desrtseb128: add to it the fact that debian seems fine...18:10
desrtand fedora too18:10
desrti wonder if the builders have some weird ecrypt-like issue18:11
desrtmaybe linux containers or something, causing trouble?18:11
seb128desrt, well I would assume the debian buildders have similar containers and automake built on our builders before18:11
desrtseb128: you said 2.11.2b had the same problem, right?18:12
seb128desrt, that's 1.11.2 we are talking about18:12
desrtwell, afaik nobody actually packaged 1.11.2?18:12
seb128desrt, it's the version debian has and that I synced today and that failed to build18:12
desrtright.  okay.18:13
seb128desrt, .3 was for the ppa but it hits similar issues18:13
desrt.2 must be recently packaged18:13
seb128desrt, 2012-01-1518:13
desrti'm guessing that was in response to my ping to the maintainer asking about packaging .3 when it comes out18:14
desrt"oh.. i didn't even do .2... let's do that now"18:14
desrtseb128: did you consider trying a build of .1 on your machine?18:14
desrtor on the builders?18:14
seb128desrt, yes, .1 failed the same way on my machine18:14
desrtbecause... like... it could be automake that changed, between .1 and .218:14
desrtor it could be something else18:14
seb128but I start thinking my issue is ecryptfs related18:15
desrtso that's very interesting18:15
desrtwell... maybe not18:15
desrt.1 used to work on the builders, and presumably used to work on ecryptfs (let's assume)18:15
desrtnow it doesn't18:15
desrtthe builders and you don't share a kernel, but they do share a userspace18:15
desrti wonder if something in precise like glibc or libacl or something like that have some weird issue that is causing troubles18:16
desrtcan you try submitting .1 to the builders?18:17
desrtunder precise18:17
seb128desrt, yes, will do that18:17
desrtlooks like doko may end up with some work to do after all ;)18:18
seb128I also started a build on a porter box to make sure it's not ecryptfs18:18
desrtseb128: does the plain upstream tarball cause the issue?18:18
seb128desrt, if you don't need to much cpu for your work you can just apt-get source automake1.11 cd here and debuild it18:18
desrtokay.  i'll do that.18:18
desrti have an encrypted filesystem... but not a bad one :)18:19
seb128desrt, it's distdir.test fail fails there when doing the make check-TESTS18:19
* desrt is on acloca1718:20
seb128desrt, yeah, it can take a while18:20
mdeslaurseb128: do you need to do something special to get the user wallpapers in lightdm?18:21
seb128desrt, you can try to stop it and just run ./distdir.test18:21
seb128in tests18:21
seb128mdeslaur, define user, it doesn't work with wallpapers stored in your user dir atm and you need to change the background in the control center18:22
seb128mdeslaur, lightdm only has the path to the image, if it's in an user dir it can't read that doesn't work18:22
mdeslaurseb128: ah, ok, cool18:24
seb128desrt, that specific issue is ecryptfs for sure then18:24
seb128I don't need a full build18:25
seb128- apt-get source automake1.1118:25
seb128cd automake1.11-1.11.118:25
seb128./configure && make18:25
seb128cd tests18:25
seb128 18:25
seb128it breaks in my user dir18:25
seb128works in tmp18:25
seb128let's see what the build on the porter box do18:25
desrti just don't understand how you can copy the file from your userdir to tmp and still have the problem18:26
seb128I guess the buildds issue is different from mine18:26
desrtyet supposedly there are no extended attributes or anything18:26
desrtseb128: did you look at the log?18:26
desrtlike... does it fail the same test in the same way?18:26
desrtor just.... some issue?18:26
seb128$ rm -rf /tmp/some/file18:27
seb128rm: cannot remove `/tmp/some/file': Permission denied18:27
seb128yet ...18:27
seb128desrt, the log has no details, it has a line PASS: or FAIL: <testname> for each test18:27
seb128so I just know it failed18:27
desrtsame test, tho18:27
seb128in 800 tests18:28
desrta mystery!18:28
seb128one fail and it's the same one18:28
desrtPASS: distdir.test18:28
desrtseb128: get a thinkpad :p18:29
seb128desrt, lol18:29
=== gord is now known as Guest53691
seb128desrt, ok, thanks for the help18:34
desrtseb128: i'm the one that gave you the problem in the first place18:35
desrtreally, doko should be thanking me ;)18:35
seb128desrt, the retry failed on another test so I think the buildd issue is a different one18:35
seb128desrt, and my local build issue is an ecryptfs issue for sure18:35
seb128desrt, I will upload a package with verbose output for the buildd issue18:36
desrtseb128: cool18:36
desrtmaybe it was just ETOOMANYTESTS18:36
seb128I should give it to chrisccoulson18:36
seb128he's used to those :p18:36
seb128he got lot of fun firefox testsuit issues18:37
seb128chrisccoulson, https://launchpadlibrarian.net/91811783/buildlog_ubuntu-precise-i386.automake1.11_1%3A1.11.2-1ubuntu1_FAILEDTOBUILD.txt.gz18:37
seb128chrisccoulson, want some testsuit fixing fun :p18:37
chrisccoulson"la la la la la"18:37
* chrisccoulson puts fingers in ears18:37
seb128chrisccoulson, I just wasted an hour to notice that my local issue is due to ecryptfs18:37
seb128chrisccoulson, it won't let me delete a file18:38
chrisccoulsonthat sucks ;)18:38
seb128empty file in a read only dir, which is fine out of ecryptfs18:38
desrtseb128: btw: debian maintainer says he will get around to having .3 in debian unstable over the weekend18:40
seb128desrt, excellent, I can probably sync that on monday then ;-)18:42
desrtseb128: assuming you can figure out the builder issues18:43
desrtseb128: got a link to the .1 upload?18:43
* desrt is curious how that will end18:43
seb128desrt, not yet, I will let you know the details when I'm done with it18:46
kenvandineseb128, i uploaded g-c-c to the ubuntu-desktop ppa19:03
kenvandinelet the carnage begin :)19:03
desrtoh.  party time.19:08
desrtsmspillaz: paying attention? :)19:08
* kenvandine heads to lunch, bbiab19:08
desrtthat's 3.219:08
desrtkenvandine: where's the carnage?19:08
jbichadesrt: I know, I was disappointed too ;)19:10
kenvandinenew sound settings panel for control-center19:10
kenvandinein the ubuntu-desktop ppa19:10
desrtthat's not carnage19:10
desrtthat's merely sad19:10
* desrt disappears for now19:11
dobeywhy is unity-greeter reading my user settings?19:20
jbichadobey: which settings? reading your wallpaper is a feature now19:24
dobeythe wallpaper19:24
dobeyi'll file a bug19:24
jbichabut that's not a bug19:25
dobeyit's a security/privacy issue, and it's a misfeature.19:25
jbichadobey: you probably should read bug 84408119:29
ubot2`Launchpad bug 844081 in accountsservice "Unity Greeter - Background of the Unity Greeter should reflect the background chosen by the user that is currently selected" [Wishlist,Confirmed] https://launchpad.net/bugs/84408119:29
dobeyi should probably file a bug, because any bug with 44 comments and which is considered fixed released (which is technically true, because the bug isn't a bug, but a work statement, and the work was completed).19:33
jbichadobey: comment 19 might be what you want but it looks like sabdfl doesn't like the idea19:37
dobeynot really19:39
mdeslaurdobey: why is that a security/privacy issue? if you don't want local users to read your wallpaper, set appropriate permissions on the file...19:41
jbichayou could then use a .gschema.override to make it system-wide by default? or how about #619:41
dobeyone thing i do want, is rationale from design for having such a "feature" in the first place. "it would be nice" or "we should do it" are not rational reasons to implement it.19:41
jbichahow about "it looks cool" and makes the computer "more personalized"19:42
dobeymdeslaur: ~/.config/dconf is only readable by me.19:43
dobeyjbicha: no19:43
jbichaI am a bit skeptical of it being a privacy issue, it's not like it's showing stuff from hidden directories in ~/Pictures19:43
dobeyjbicha: well, not unless it's set as your background anyway19:43
jbichathen it's not very private, is it? ;)19:43
dobeysure, unless you are choosing to set it as your greeter background as well19:44
dobeywhich i haven't done19:44
dobeyalso, it's not clear how it technically works (i haven't looked at the code yet). but it's either doing evil and reading my stuff somehow, despite the fact it shouldn't be allowed to, or it's copying my data somewhere it shouldn't be.19:45
mdeslaurdobey: when you change backgrounds, the filename gets sent to AccountsService which stores it in /var/lib/AccountsService19:45
dobeyyeah it shouldn't do that19:46
mdeslaurI'm sorry, I don't quite understand...your background is also displayed when your screen is locked, and is usually in a file located in your home directory which is readable by other users19:47
dobeymdeslaur: and i think displaying the background on the screen unlock is also a problem19:48
dobeybut at least the screen lock isn't copying my *user* settings to a systemwide location for peruse later by some other tool that's not the user19:49
dobeyif it were opt *in* that would be different19:51
seb128desrt, https://bugs.launchpad.net/ecryptfs/+bug/926292 for the record19:51
ubot2`Launchpad bug 926292 in linux "automake distdir.test fails because of an EPERM error" [Medium,Confirmed]19:51
seb128dobey, that's copying nothing, that's putting a filename in a database, if you userdir or the directory of the image is not readable by the greeter user no background is being displayed19:54
seb128dobey, if the file you use is wordreadable it's that you don't care so much about it being private19:54
dobeyseb128: my DCONF directory is not readable by the greeter. it is copying a value out of one database which is not readable, into another which is19:55
seb128dobey, it's copying a filename and no it's not copying it, the control center is setting it in another database19:55
seb128dobey, no copy is done19:56
seb128you just a tool which is writting setting in accountsservice19:56
seb128if you don't like it don't use the tool19:56
dobeyseb128: fine, it's piped to N places, which don't have the same read permissions, and violates the default permissions19:56
dobeydon't use what tool?19:56
seb128dobey, indeed, don't use it if you don't like it19:56
dobeydon't use what?19:56
seb128dobey, it's also writing your locale and keymap in that same public database19:57
seb128dobey, gnome-control-center19:57
dobeyi should go buy macos then?19:57
mdeslaurand user icon19:57
seb128and username19:57
seb128dobey, if you think macos is giving better control and a better view of what the os is doing from your datas sure19:57
seb128why not? nobody force you to use open tools, you are fine to use closed source and trust apple19:58
seb128who knows they might collect your background image over the internet without telling you ;-)19:58
seb128dobey, it's only the path to a file, it's not even datas or content, what's the big deal?19:59
dobeyfor all i know, accountsservice is doing that19:59
seb128dobey, well you know little, at least you can read the source in this case ;-)19:59
dobeyi'd have to read the code and watch the tcp traffic to determine it's not19:59
dobeythe big deal is it's broken20:00
seb128dobey, so you assume people mean bad by default there?20:00
seb128how so?20:00
dobeyno, i don't assume people mean bad. i don't assume anything.20:00
dobeywell, what if my background is a 32x32 image which is tiled? what if it's a small image that's centered? what about the background colors and gradient? opacity? and it makes non-plain backgrounds look like crap with the greeter's overlay having all the white dots20:02
dobeyand it's the first setting. if it's ok to have a copy of that setting somewhere, when is it not ok to write a copy of a setting to that world readable database?20:03
dobeyi never explicitly chose to have my configuration written there. i chose to have my configuration written in XDG_CONFIG_HOME which is set to ~/.config/; and which is not world-readable20:03
seb128dobey, you don't have your configuration there, you have a filename20:04
dobeyseb128: how did it get there?20:04
seb128dobey, do you have issues with having your locale and keyboard layout there?20:04
dobeyi changed some configuration and it was placed in 2 places. one which is not world readable, and one which is.20:05
dobeyseb128: yes20:05
seb128dobey, the same way that your icon, keymap, locale, etc get there for yers20:05
dobeynow that i know they are being written there, i have a problem with them being written there20:05
dobeyaccountsservice is not "years" old20:05
seb128the login screen has access to your keymap and locale for years20:05
seb128right, before there was a copy of .dmrc in /var20:06
seb128which is an user config with your session, locale, keymap etc20:06
dobeyyou mean when i choose the session/keymap/locale *in* the dm?20:06
seb128how do you think gdm was getting your locale,keyboard infos?20:06
seb128the user session stuff update the same config20:06
dobeyif i choose something in the *DM* it's not a user config20:07
seb128the dm doesn't always have access to your user dir to write those20:07
dobeyand really it should store it via some pam thing in that case20:07
seb128same problem20:08
mdeslauryour user image icon has been copied system wide for years20:08
seb128if you user dir is encrypted how can the login screen get your keymap right?20:08
mdeslaurdobey: please file a bug, a bug discussion would be better20:08
dobeythe user face avatar thing i don't care about. it's something that was always explicitly specified as being the thing that appears in the login screen.20:08
dobeyseb128: keymap really should be stored via pam20:09
dobeyanyway, yes i will file a bug20:09
seb128dobey, there is already a bug20:11
seb128dobey, it will be dupped20:11
seb128dobey, it's wontfix in any case20:11
mdeslaurso, by default, when you select a  background in gnome-control-center, it copies it with a hashed name into your .config directory. That directory is only readable by you by default, and the original filename is mangled.20:11
seb128that seems like a "make things not working"20:12
seb128I will object to that20:12
dobeymake things not working?20:12
mdeslaurseb128: object to what? that's the current behaviour20:12
seb128what you suggest is broken yes20:12
dobeyi haven't suggested anything exactly20:12
seb128mdeslaur, no, the current behaviour doesn't require access to your userdir20:12
dobeyseb128: and what mdeslaur just said, is how dconf works20:12
mdeslaurseb128: when you select a custom image in gnome-control-center, currently, that's what happens20:13
seb128I've picked an image from the default system set for my ecryptfs user and it's reflected on the login screen20:13
mdeslaurseb128: if you select from the default system images, it gets displayed on the login screen20:13
seb128which seems the correct behaviour to me20:13
mdeslaurthis isn't exactly a privacy leak, it will only show what your preference of filtered system images are, not some random pr0n from your home directory20:14
seb128right ;-)20:14
seb128I'm fine with that20:14
mdeslauryes, me too20:14
seb128I just don't get what dobey wants to open a bug about20:14
seb128I think the current behaviour is fine20:14
seb128if you select a system image it's not private20:14
seb128if you select a personal photo it is20:14
seb128that seems sane enough20:15
mdeslaurseems sane to me20:15
dobeynot entirely true20:17
seb128dobey, what is not true?20:18
seb128dobey, well read the discussion on the bug pointed before before opening a new one in any case, also note that nobody force you to use the default config or greeter20:19
dobeyso it looks like gnome-control-center copies the background into a hashed name in the XDG_CACHE_HOME folder; and this filename gets written to accountsservice, when it's not a system background20:21
seb128that directory is only readable by your user20:22
dobeyyes, but it shouldn't write the path to the system in that case either20:23
dobeyand even if it didn't do that, it doesn't resolve all of the other problems with the "feature"20:23
RainCThey seb128, still working?20:24
seb128not sure what's the intend there but I think if the original file is world readable the greeter should reflect it20:24
seb128RainCT, hey, sort of, rather IRC arguing than "working" ;-)20:24
seb128dobey, what problems?20:24
RainCTseb128: Heh. I've got a Nautilus patch for you :). But I guess having it reviewed by someone first wouldn't hurt, so I'll bug you on Monday about it :P20:25
seb128RainCT, ok ;-)20:25
dobeythe dot overlay, tile/zoom/color options missing (which would increase privacy concern if they were also copied); and *WHY* having this "feature" is even a good idea20:25
ryehello, bug #926152 - adding ppa is broken now20:38
mdeslaurrye: argh, thanks20:40
ryemdeslaur, it was unassigned and it looks i can't add my own ppa :)20:43
mdeslaurrye: I broke it :)20:44
mdeslaurrye: fix uploaded, thanks21:09
ryemdeslaur, wow, thank _you_!21:12
=== dduffey is now known as dduffey_afk
dobeykenvandine: are there any special requirements under the new upstream-is-canonical upload stuff, for when a package adds new binary packages?21:43
kenvandinenothing new21:43
dobeyso i should just upload it and poke someone to approve it?21:43
kenvandineit'll still go to binNEW21:43
dobeyalright, cool21:44

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!