/srv/irclogs.ubuntu.com/2014/04/02/#ubuntu-devel.txt

=== cjwatson_ is now known as cjwatson
=== alesage is now known as alesage|afk
cyphermoxxnox: tomorrow morning I'll test removing it to see what breaks, and I tell you then whether it's ok to remove?00:38
cyphermoxmy tomorrow morning is in about 12 hours.00:38
=== seelaman` is now known as seelaman
FourDollarscyphermox: Could you help me to review the patch of https://bugs.launchpad.net/ubuntu/+source/bluez/+bug/1300623 ?01:35
ubottuLaunchpad bug 1300623 in bluez (Ubuntu) "bluetoothd crashs after resuming from Suspend to RAM. " [Undecided,New]01:35
mwhudsoninfinity: hey did you stick that libc patch into a ppa?01:48
mwhudsonshould have built or failed by now i think...01:49
infinitymwhudson: Nope.  I'm like a kitten with a laser pointer today.01:49
mwhudsonheh01:49
mwhudsoninfinity: anything i can do to make it more likely?01:50
infinitymwhudson: Uploading nowish.01:52
infinityIn theory.01:52
infinityWhen the source builds.01:52
* infinity smacks his laptop.01:52
mwhudsoninfinity: which ppa?01:53
infinitymwhudson: adconrad/ppa01:53
mwhudsoncool01:54
infinitymwhudson: And building.01:54
mwhudsoninfinity: awesome, thanks01:55
mwhudsontakes about 90 mins on arm64, right?01:56
infinitymwhudson: So, if this works and fixes your issue, we have a few things to do.01:56
infinitymwhudson: 1: nag Will harder about upstream review (but I won't block on this)01:56
infinitymwhudson: 2: Fish Will's testcase patch out of libc-alpha and give me that to add with this other patch.01:56
infinitymwhudson: 3: Figure out WTF our distro kernel chokes on tst-tls5. :/01:56
mwhudson1: i cc:ed you on a mail to will earlier i think?01:57
mwhudsoni meant to, anyway01:57
infinitymwhudson: Yeah, you did. :)01:57
mwhudsoncool01:57
mwhudson2: ack01:57
mwhudson3: yes :/01:57
mwhudsondo you know offhand what that test is testing?01:57
mwhudsonother than thread local storage01:57
infinityHaven't looked at it at all.01:57
mwhudsonkay01:57
infinityBut the failure doesn't look like something we should XFAIL either, without understanding why.01:58
mwhudsonindeed01:58
mwhudsontotally a WAG but... how do i tell if i have a 64k page kernel?01:58
infinitygetconf01:58
infinity(base)adconrad@cthulhu:~$ getconf PAGESIZE01:58
infinity409601:58
mwhudsonok same01:59
mwhudsonboring01:59
infinitymwhudson: And about 70 minutes to build on arm64.02:00
infinitymwhudson: At least, on my snazzy 2.4GHz buildds. :P02:00
mwhudsonheh02:00
mwhudsonwhat are the porter boxes?  2.0?02:01
infinitymwhudson: If they're still A2s, then yeah.02:01
mwhudsonhm, 'make check' just succeeded in ~/eglibc-builds/eglibc-2.19/build-tree/arm64-libc02:02
mwhudsonhow do you run the tests?02:02
infinitymwhudson: Well, make check would have had nothing to re-make...02:03
mwhudsonoh ok02:03
infinitymwhudson: I go hunt through the log for the exact invocation of the test that failed and just run that massive commandline over and over.02:03
infinityI'm sure there are better ways. :P02:03
infinitymwhudson: So, I have no idea what am1 is, cause the new cpuinfo is completely uninformative.02:05
infinitySpecial.02:05
mwhudsoninfinity: hey, it tells you that there are 8 cores02:05
mwhudsonbut yes02:05
infinitymwhudson: The lack of model name and speed seems weird.02:07
mwhudsoninfinity: ok02:09
mwhudsonso02:09
mwhudsonthe test program prints WRONG ALIGNMENT02:09
mwhudson(and fails)02:09
mwhudson_if_ you have ulimit -s unlimited02:09
mwhudson(which makes memory be allocated in the other direction)02:10
mwhudson(the fact that i know this bothers me i think)02:10
NilesCould it be a PAE problem?02:10
mwhudson64 bit user space on 64 bit processor?  i hope not02:10
NilesOh wait, that's dumb02:10
NilesNevermind02:10
NilesOnly case I know where memory gets allocated backwards02:11
mwhudsoni've only read the relevant bits of the kernel for arm64 :)02:11
mwhudsonso um yeah02:12
mwhudsondoes anyone happen to know where THREAD_SELF is defined/02:20
mwhudson?02:20
mwhudson(on aarch64)02:21
barryxnox: that would rock :)02:21
* mwhudson hugs the glibc source layout02:22
sarnoldI've never seen an ironic hug before, but there it is :)02:23
mwhudsonnow __builtin_thread_pointer02:24
mwhudsoni guess that's a gcc thing02:24
mwhudsonsarnold: :)02:24
infinitymwhudson: I'd I'm reading the log-in-process right, tst-tls5 didn't fail in my PPA.02:40
infinitys/process/progress/02:41
mwhudsoninfinity: i think if my patch was causing this failure i'd give up and start looking for lawn mowing jobs02:41
infinityLawn mowing is good, honest work.02:41
infinityYou might need to buy some shoes, though.02:42
infinitysarnold: You don't like the glibc source tree?02:42
infinitysarnold: Given the complex concepts it needs to encapsulate, I feel it does a pretty good job.02:43
infinityDoesn't take long to wrap your head around the cascading include paths, etc.02:43
mwhudsoni find the fact that the eg s390 platform specific code is in ./nptl/s390 but the aarch64 specific stuff is in ./ports/sysdeps/unix or somewhere constantly confusing02:44
infinitymwhudson: Oh, well, ports is gone in 2.2002:45
mwhudsoninfinity: \o/02:45
infinitymwhudson: Or will be, once all the ports are moved.  Most are.02:45
infinityIn fact, all but hppa look to have moved.02:45
infinityCarlos is a slacker.02:45
mwhudsonoh02:48
mwhudsonum02:51
mwhudsondoes libpthread do some crazy magic in its .init section?02:51
mwhudsonbecause i now have a test program that fails _if_ it's linked against libpthread02:52
mwhudsonand not otherwise02:52
mwhudsoninfinity: can you run a small test case on a buildd (or some other vintage kernal machine i guess)02:53
infinitymwhudson: Sure.02:55
mwhudsoninfinity: http://paste.ubuntu.com/7192634/02:55
infinitymwhudson: http://paste.ubuntu.com/7192638/02:57
mwhudsonho strange02:58
infinitymwhudson: Well, seems to be the same as the results you got?02:58
mwhudsonyes02:58
infinitySo, not a regression.02:58
infinityBut quite possibly a bug.02:58
mwhudsonmaybe i was just building eglibc in a shell that had ulimit -s unlimited?02:59
infinitymwhudson: Seems plausible.02:59
mwhudsonit would be a good explanation, but i don't recall changing that :/03:00
mwhudsonah, i remember why i didn't quilt import this patch: it isn't rooted at the base of the tree03:05
infinitymwhudson: Well, quilt import doesn't work so well in eglibc probably anyway.03:07
infinitymwhudson: But just stuffing the git-formatted patch in debian/patches/ubuntu/ works fine. :P03:08
mwhudsonok i'm now rebuilding in a shell that _definitely_ does not have ulimit -s set03:09
mwhudson(well rather, not set to unlimited)03:09
mwhudsonand yes, this looks like a bug03:09
mwhudsonglibc tests that pthread_self is always 16 byte aligned, and on aarch64 with unlimited stack it isn't03:10
mwhudsonso is the bug in the test or the ... whatever sets the tls pointer03:10
infinitymwhudson: Worth an upstream bug report and a nudge to the maintainers.03:13
mwhudsonyeah03:13
mwhudsonglibc bugzilla?03:14
infinityYeah.03:14
mwhudsoni'll ping linaro-toolchain about it too03:14
mwhudsoninfinity: https://sourceware.org/bugzilla/show_bug.cgi?id=16796 (i added you to cc so maybe you got mail already)03:19
ubottusourceware.org bug 16796 in nptl "[aarch64] pthread_self not aligned to 16 bytes when heap grows up" [Normal,New]03:19
cyphermoxFourDollars: gladly. patch looks fine, perhaps it should just also happen in saucy and trusty?03:39
FourDollarscyphermox: The codebase changed too much in saucy and trusty so this patch may only work in precise.03:43
cyphermoxok03:44
cyphermoxwell then I'd like to give it a shot in trusty tomorrow03:45
cyphermoxand if I can't make it work, then I'll just say so on the bug so there's no questions03:45
cyphermoxFourDollars: want to update the bug for the SRU process?03:46
FourDollarscyphermox: yes03:46
=== cr3_ is now known as cr3
cyphermoxFourDollars: thanks03:48
=== cr3 is now known as Guest29639
cyphermoxFourDollars: I'll sponsor the patch first thing tomorrow morning03:49
FourDollarscyphermox: Sorry. I just checked saucy and trusty. I think this patch can also apply on them.03:49
FourDollarscyphermox: Thanks a lot.03:49
cyphermoxFourDollars: alright, np :)03:57
FourDollars:)03:57
cyphermoxif it turns out it doesn't work for you, I'll look myself in the morning to see if I can make sense of it03:58
FourDollarsOK03:58
sarnoldinfinity: well, I can't really argue one way or another about glibc source tree except to say that when I try to figure out how something is implemented in the libc wrappers in linux, it always takes me two or three times longer to find it than I'd expect05:27
pittiGood morning06:17
ionthat06:17
infinitymwhudson: Well, that was a quick response to your bug.06:22
=== jackson is now known as Guest51547
infinitymardy: You around?06:39
infinitymardy: Your dpkg-maintscript-helper calls in account-plugins should have a version argument on them, so it's not attempted on every single upgrade forever.06:41
infinitymardy: I'm rejecting based on that.  Please fix.06:41
dholbachgood morning06:41
mardyinfinity: OK06:48
mardyinfinity: I should put the version where the script was last useful, right?06:49
infinitymardy: It should be the version just before the one being uploaded.  The easiest way to do that is just to tack a ~ on the end of the version you're uploading.06:50
infinitymardy: ie: if you're uploading 1.2.3-1 that removes the conffile, the version passed to rm_conffile should be 1.2.3-1~06:50
infinitymardy: The manpage explains why, and how.06:50
mardyinfinity: ok, thanks06:51
infinitymardy: (Given that you're doing daily builds that tend to have a large gap between uploads, though, you could probably just fudge it as "$date -1" or something if you wanted to)06:51
infinitymardy: But generally, the goal is to not screw over people who rebuild, do local test builds, derive in other distros, etc, etc.06:52
=== doko_ is now known as doko
dokoSweetshark, could you have a look at https://launchpad.net/ubuntu/+source/accessodf/0.1-4/+build/5866883 ?  looks like the only difference is the lo version07:09
=== psivaa is now known as psivaa-afk
mardydbarth: so, infinity suggested a change in the account-plugin-flickr debian scripts07:57
mardydbarth: quoting from this morning:07:57
mardy09:39 < infinity> mardy: Your dpkg-maintscript-helper calls in account-plugins should have a version argument on them, so it's not attempted on every single upgrade forever.07:57
mardy09:39 < infinity> mardy: I'm rejecting based on that.  Please fix.07:57
dbarthmardy: ah ok08:04
dbarthmardy: then can you file a bug with infinity's concerns, and i'll land asap08:04
mardydbarth: why a second one? account-plugins didn't land, did it?08:05
Noskcaj_Can someone check if modemmanager is syncable?08:05
Noskcaj_There's some fairly important upgradeing fixes08:05
dbarthmardy: it did08:05
dbarthmardy: was tested and published yesterday08:05
dbarthmardy: i merged them this morning, since we couldn't take back the packages anyway08:05
mardydbarth: oh, you are right08:06
mardydbarth: but the bug is still open08:06
mardydbarth: and the merge proposal is not in "merged" status yet08:07
mardydbarth: even though I see that the changes are now in trunk... weird...08:07
Sweetsharkdoko: weird. yeah, Ill take a look.08:15
Noskcaj_Logan_: http://paste.ubuntu.com/7193182/08:30
seb128Noskcaj_, hey08:32
=== psivaa-afk is now known as psivaa
Sweetsharkdoko: so, have it fixed locally, wonder why it ever build. investigating if we changed the API there upstream ...08:49
Sweetsharkoh, we did. Well actually those cowboys at AOO changed the API and we followed suit ...08:51
brendandhi, this is a slightly unusual situation, but if we had a feature in a package and then had to drop it for while and now want to bring it back, will we need a feature freeze exception?08:54
pittiinfinity: ah, lintian autopkgtest succeeds now; thanks for pointing out the dep parsing bug08:56
infinitypitti: Yeahp, I got the spam about it being fixed.  Thanks. :)08:57
didrocksxnox: are you going to reupload your g+ apps or should I do that?08:59
pittiinfinity: so that was either a bug in lintian or in python-debian; I suspec that the comment in lintian isn't entirely legal, but then again this is probaly underdefined (comments starting at the first column but don't break the continuation line)09:02
infinitypitti: The comment does start at the first column...09:03
infinityOh, I missed the "but" there.09:03
pittiinfinity: right, but it's in the middle of a continuation line09:03
pittiinfinity: so, I don't know whether that's legal or a bug in python-debian, but it just works around that now09:03
pittiit == adt-run09:03
infinitypitti: Yeah, it's arguable if it's allowed or not, but I'm not sure I've ever read anything that says it's not, and dpkg parses it correctly.09:03
infinitypitti: And I'd say that dpkg is the authority on how to parse debian/control.09:04
pitti*nod*, hence I filed a bug against p-debian09:04
infinitypitti: I'd have just torn the comments out in our delta, but then it would still be broken for ci.debian, plus it can't be the only package that does that.09:05
infinitypitti: So, yay, thanks for the bug and the workaround. :)09:05
xnoxdidrocks: i can republish it, if the container/ua is fixed now.09:07
xnoxcyphermox: that would be awesome! you're a start =) thanks.09:08
didrocksxnox: yeah, you need to target the 14.04 dev1 framework though09:08
didrocksxnox: to get oxide09:08
xnoxdidrocks: right, yeah i see ping from davmor2 about that.09:09
davmor2xnox: see everyone loves the google+09:22
ogra_wow, G+ in the new browser is just lovely to use09:25
ogra_working javascript makes such a big difference09:25
davmor2ogra_: so easily pleased now get off G+ and get some work done ;)09:26
ogra_haha09:26
davmor2ogra_: and none of this I'm just "testing oxide" milarky either that's my excuse :P09:27
ogra_heh09:28
xnoxis there a way to get oxide on the desktop?09:35
* xnox dist-upgrades first09:35
ogra_xnox, i think that should get it for you (at least for the webapps)09:35
happyaronseb128: why unity-control-center depends on ibus?09:36
seb128happyaron, because it writes to its gsettings schemas09:37
seb128I though of you when that was added09:37
seb128I guess we could handle the missing schemas and lower to a recommends or something09:37
happyaronif recommends is doable then that'd be great.09:37
seb128happyaron, https://code.launchpad.net/~attente/unity-control-center/1288717/+merge/20998209:38
seb128happyaron, btw https://bugs.launchpad.net/ubuntu/+source/unity-control-center/+bug/1288717/comments/309:39
ubottuLaunchpad bug 1288717 in unity-control-center (Ubuntu) "/usr/bin/unity-control-center:6:g_assertion_message:g_assertion_message_expr:ibus_config_get_value:orientation_combo_changed:_g_closure_invoke_va" [High,Fix released]09:39
seb128happyaron, but you never bothered replying to my comment09:39
seb128happyaron, if you had you would have seen the change/new depends...09:39
seb128happyaron, anyway, I don't think making that a recommends is on top of you priority list, but patches are welcome if somebody wants to work on that09:40
happyaronseb128: well I'm being so busy recently dealing with all kinds of stuff...09:41
seb128happyaron, we all are...09:42
happyaron:)09:42
seb128happyaron, anyway you have the why you asked for now ;-)09:43
=== tkamppeter_ is now known as tkamppeter
tkamppetermdeslaur, I have issued cups-filter 1.0.51 upstream now and also packaged and uploaded it for Trusty, so the vulnerability is fixed now. Thanks for the report.09:49
happyaronseb128: well the good thing is we got this fix: https://bugs.launchpad.net/ubuntu/+source/im-config/+bug/1297831/comments/1809:50
ubottuLaunchpad bug 1297831 in Ubuntu Kylin "Input method set to ibus in Ubuntu Kylin, while fcitx is the desired default" [Critical,Fix committed]09:50
happyaronseb128: so you are right, it's not on the top of my list.09:51
seb128happyaron, ok, good09:53
brendandseb128, this is a slightly unusual situation, but if we had a feature in a package and then had to drop it for while and now want to bring it back, will we need a feature freeze exception?09:53
seb128brendand, I would assume so, but check with #ubuntu-release09:54
Sweetsharkdoko, seb128: fix for the build breaker at http://people.canonical.com/~bjoern/trusty/accessodf_0.1-1.3ubuntu2_amd64.changes -- the diff is next to it ...10:04
seb128Sweetshark, thanks, is that something we can get in Debian and sync?10:06
Sweetsharkseb128: its a clearcut buildbreaker -- I guess debian would want that. Although the fix doesnt recreate the messagebox type (infobox, questionbox etc.), but IMHO thats something for upstream (not debian) to take care of.10:09
pittidoko: s/langpack-en-us10:10
pittidoko: ... /language-pack-en-us/ :)10:10
Sweetsharkseb128: Ill hint rene at it (although he is not the maintainer there)10:11
pittidoko: sorry, I mean language-pack-en; we don't have by-country packs except for -zh10:11
pittidoko: FYI, infinity wants to move langpack-locales back into eglibc and also build locales-all (hopefully next cycle), so that's going away at some point10:12
seb128Sweetshark, do you still want sponsoring to trusty? we can sync over from Debian once they fix it10:12
Sweetsharkseb128: lets wait for renes reaction first ..10:14
seb128Sweetshark, k10:15
apacheloggerpitti: https://launchpad.net/ubuntu/+source/gpgme1.0/1.4.3-0.1ubuntu2 seems to have broken kdepim s/mime support as per bug 129370410:21
ubottubug 1293704 in kdepim (Ubuntu) "Kleopatra don't support s/mime" [High,Confirmed] https://launchpad.net/bugs/129370410:21
=== vrruiz_ is now known as rvr
shadeslayerpitti: https://bugs.launchpad.net/ubuntu/+source/ubuntu-drivers-common/+bug/130130710:44
ubottuLaunchpad bug 1301307 in ubuntu-drivers-common (Ubuntu) "Empty vendor string for driver" [Undecided,New]10:44
shadeslayertseliot: ^^10:45
xnoxogra_: are there any webapps that work as contained clicks with oxide?10:45
xnoxcause updated g+ fails like shown in bug #130130510:45
ubottubug 1301305 in webbrowser-app (Ubuntu) "oxide fails to launch contained webapp" [Undecided,New] https://launchpad.net/bugs/130130510:45
ogra_xnox, yes, the default shipped ones (gmail, facebook etc) ... and the browser itself is using it now10:46
ogra_only webapps that havent been re-uploaded with the new framework added use the old way still10:46
tseliotshadeslayer: I don't think there's a single vendor for that10:46
shadeslayertseliot: right, but then the UI just looks terribly broken10:47
shadeslayeror atleast the KDE one does10:47
tseliotshadeslayer: any pictures?10:47
shadeslayermoment10:47
ogra_xnox, using the gmail app and then clicking the G+ menu option in there works too ... just has the gmail top bar then :)10:47
xnoxogra_: maybe i need "webview" policy group. Let's see.10:48
shadeslayertseliot: http://im9.eu/picture/a1986210:49
shadeslayertseliot: something else I noticed10:50
tseliotshadeslayer: you might want to have a fallback such as "Unspecified vendor" for when that happens10:51
shadeslayertseliot: the pci values for Airport extreme is a subset of the one from linux-firmware-nonfree10:51
shadeslayer/sys/devices/pci0000:00/0000:00:15.0/0000:02:00.0/ssb0:0 vs /sys/devices/pci0000:00/0000:00:15.0/0000:02:00.010:52
shadeslayerso maybe it would make sense to group them together?10:52
tseliotshadeslayer: does the Airport extreme require linux-firmware-nonfree?10:53
shadeslayertseliot: yep10:53
shadeslayerthat or the other b43 driver10:53
shadeslayerboth work10:53
tseliotshadeslayer: and what do you mean by "grouping them together"?10:54
shadeslayertseliot: under the same device :)10:54
shadeslayerso that ubuntu-drivers devices shows Broadcom Airport Extreme with 2 drivers : b43 and linux-firmware-nonfree10:55
tseliotoh, I see what you mean now10:55
tseliotI guess that would make sense10:56
tseliotpitti: ^10:56
ogra_oh !11:08
ogra_oh ! oh !11:08
ogra_video streaming inside the phone bowser works now !!!11:08
xnoxnew g+ is in the store review queue.11:17
mdeslaurtkamppeter: awesome, thanks!11:18
Riddell6/win 1511:20
Riddelltsk11:20
=== MacSlow is now known as MacSlow|lunch
addikshi, does anybody know which exact component is responsible for giving input-events to the applications?11:42
dobeydpm_: ping11:46
dpm_hi dobey11:54
dobeydpm_: hi. do you know what the recommended way to build translations with under cmake would be?11:54
mardyis there a directory which can be used to store files which are cleared after a user's session ends?11:54
dobeyintltool doesn't integrate with cmake11:55
mardyI've thought about XDG_RUNTIME_DIR, but that's not cleared on logout11:55
dobeymardy: delete the file when your process ends?11:56
mardymdeslaur: I see you are here, maybe you know? ^11:56
* ogra_ hugs xnox 11:56
mardydobey: I want the file to persist across my process invocations, while in the same user session11:56
mdeslaurmardy: I don't believe there are any11:56
dobeyyeah, there aren't any11:57
mardydobey, mdeslaur: that's also good to know :-)11:57
dpm_dobey, yeah, unfortunately there is no standard way of building translations with cmake, and intltool does not support cmake. We ended up writing a custom rule to build them for core apps, e.g.: http://bazaar.launchpad.net/~ubuntu-weather-dev/ubuntu-weather-app/trunk/view/head:/po/CMakeLists.txt11:57
mdeslaurmardy: you can probably store something in XDG_RUNTIME_DIR, and then check the timestamp on it vs. when the user session started11:57
ogra_xnox, external links dont open in the G+ app ... (not sure thats the app or the container)11:57
mardymdeslaur: mmm... interesting! How can I know when the session started?11:57
dobeydpm_: ok, thanks11:57
xnoxogra_: i had to enable urlpatterns. At the moment login screens and google plus itself should stay in google plus.11:58
mdeslaurmardy: I...uh....well....perhaps you can ask logind? /me isn't sure11:58
xnoxmdeslaur: i thought we do cleate a fresh XDG_RUNTIME_DIR on each login... that's the whole point of it.11:59
dpm_dobey, in addition to that, we also use a bit of a hack to extract translations from the desktop.in file: http://bazaar.launchpad.net/~ubuntu-weather-dev/ubuntu-weather-app/trunk/view/head:/CMakeLists.txt#L5811:59
ogra_xnox, well, it doesnt open any external pages ... (though i assume thats a drawback of the new browser)11:59
xnoxogra_: oh, well... not sure =)12:00
ogra_xnox, oh, and i can play embedded videos when i open G+ in the browser, i guess your app musses the video apparmor profile12:01
ogra_*misses12:01
ogra_(i cant play it in the app)12:01
xnoxogra_: i have video profile.12:01
ogra_weird12:02
xnox    "policy_groups": [12:02
xnox        "networking",12:02
xnox        "audio",12:02
xnox        "video",12:02
xnox        "location",12:02
xnox        "webview"12:02
xnox    ],12:02
dobeydpm_: ok.12:02
mdeslaurxnox: doesn't look like it...stuff stays there12:03
ogra_xnox, well, open G+ in the browser and tap a youtube video, it plays fine ... doesnt in the app12:03
* ogra_ was just trying that with his recent post12:03
ogra_in the app i see the loading animation ... and then it stops12:04
ogra_dbarth, ^^^12:04
mdeslaurxnox: nope, if the directory exists, it's kept as-is12:13
dbarthhmm, just reading the backlog12:21
dbarthxnox: do you have apparmor rejects when trying to play that video?12:22
dbarthxnox: it may also be that the app tries to play video fullscreen and fails12:23
dbarththe youtube player i mean12:23
ogra_dbarth, http://paste.ubuntu.com/7194027/12:23
dbarthpreviously it was not even trying, cause we were not detected as a "good enough" webview12:23
ogra_smells like you need some apparmor love from jdstrand12:23
dbarthogra_: weird12:24
ogra_(this is what i get when trying to play an embedded yourtube video in G+ ... the same thing works when using G+ in the browser)12:24
dbarthdo you have policy_version 1.*1* ?12:24
ogra_hmm, no idea12:24
dbarthi see you have the webview group already12:24
ogra_i'm on the latest image12:25
dbarthyou need both, ie add the webview p.g. and bump the policy_version to 1.112:25
ogra_where do i check that ?12:25
dbarthand declare framework 14.04-dev112:25
dbarthmanifest.json and <app>.json12:25
ogra_yes, i think thats what xnox did with his last upload12:25
dbarththen it's a bug12:26
ogra_http://paste.ubuntu.com/7194044/12:26
dbarthcan you file and affect webbrowser-app, use [webapp-container] as a prefix12:26
ogra_looks all fine12:26
dbarthand i'll affect oxide and other if neede12:26
dbarthd12:26
ogra_(this was /opt/click.ubuntu.com/net.launchpad.click-webapps.gooogleplus/6/app.json)12:27
jdstranddbarth: where have these rejects been happening?12:29
jdstranddbarth: the silo or the archive?12:29
ogra_dbarth, bug 130135112:30
ubottubug 1301351 in webbrowser-app (Ubuntu) "[webapp-container] embedded videos do not play in G+ app while they play fine in the browser" [Undecided,New] https://launchpad.net/bugs/130135112:30
ogra_jdstrand, latest image with the latest G+ app12:30
ogra_(both updated today, using oxide now)12:31
ogra_jdstrand, see the bug above12:31
jdstrandwell, it should be against apparmor-easyprof-ubuntu12:32
* jdstrand changes12:32
ogra_jdstrand, well, not sure, it might be two bugs ... (external links ... and video playback)12:32
jdstranddbarth: fyi, I was trying to work out the final profile changes when I hit that crasher bug, so was blocked12:32
jdstrandbut I'll get it worked out12:33
=== _salem is now known as salem_
ogra_dbarth, m.youtube.com doesnt play back videos at all ... tapping on play just reloads12:35
ogra_(funny because the same video embedded in a G+ post plays just fine)12:35
xnoxjdstrand: would i need to reupload the app? i believe i have all the correct intentions declared in the policy groups.12:42
=== MacSlow|lunch is now known as MacSlow
jdstrandxnox: for that bug ^? no12:48
jdstrandxnox: I'll fix it and the policy will be regenerated on the first boot of the new image. if you are impatient, you can install apparmor-easyprof-ubuntu directly12:50
xnoxjdstrand: ah cool =) didn't know how/when policies are regenerated =) if all gets fixed with an image update, that's cool! =)12:50
xnoxlike, really _cool_ =)))))12:50
smoserxnox, are you still looking at bug 1160079 ?12:51
ubottubug 1160079 in plymouth (Ubuntu) "plymouth aborts in cloud images" [Medium,Confirmed] https://launchpad.net/bugs/116007912:51
xnoxsmoser: yes, i can upload a fix with just the cloud bits. Cause my other changes on top need some further work (e.g. on my machine plymouth crashes on shutdown and i haven't debugged it yet)12:54
xnoxsmoser: i'll upload the cloud fix later today.12:54
smoserxnox, thanks, that would be awesome.12:55
hallynseb128: would you mind looking at the 1.1.1-0ubuntu8.9 libvirt awaiting acceptance for saucy-proposed?13:56
seb128hallyn, I'm not part of the SRU team13:56
hallynd'oh, i was looking at the wrong comment, sorry13:56
hallynbdmurray: ^13:56
seb128no worry ;-)13:57
jdstrandogra_: do you have a couple minutes to help me with the apparmor denials?14:05
ogra_jdstrand, i can try14:05
jdstrandogra_: ok, in /var/lib/apparmor/profiles/net.launchpad.click-webapps.googleplus_googleplus_614:06
jdstrandogra_: you'll see these rules:14:06
jdstrand/usr/lib/@{multiarch}/oxide-qt/oxide-renderer Cx -> oxide_helper,14:06
jdstrand/usr/lib/@{multiarch}/oxide-qt/chrome-sandbox cx -> oxide_helper,14:06
jdstrandogra_: can you change them to:14:06
jdstrand/usr/lib/@{multiarch}/oxide-qt/oxide-renderer Cxmr -> oxide_helper,14:07
jdstrand/usr/lib/@{multiarch}/oxide-qt/chrome-sandbox cxmr -> oxide_helper,14:07
jdstrandogra_: when done, please do: sudo apparmor_parser -r /var/lib/apparmor/profiles/net.launchpad.click-webapps.googleplus_googleplus_614:07
jdstrandogra_: then relaunch the web app14:07
jdstrandogra_: oh, I forgot one. way down you'll see this rule:14:09
jdstrand/usr/lib/@{multiarch}/oxide-qt/oxide-renderer rmix,14:09
jdstrandunder it, can you add:14:09
jdstrand@{PROC}/sys/kernel/yama/ptrace_scope r,14:09
udevbotError: "{PROC}/sys/kernel/yama/ptrace_scope" is not a valid command.14:09
jdstrandogra_: then do 'sudo apparmor_parser -r /var/lib/apparmor/profiles/net.launchpad.click-webapps.googleplus_googleplus_6' and relaunch the app14:10
ogra_jdstrand, no change ...14:13
ogra_let me reboot the phone14:13
jdstrandogra_: are there denials?14:13
ogra_havent seen any when tailing syslog14:13
jdstrandok, then it is like you said-- two bugs14:13
ogra_but the video doesnt play14:13
ogra_well, the other would be "not opening external links"14:14
jdstrandheh, well, I mostly mean I can fix my bug14:14
ogra_jdstrand, i dont really get why it plays flawless when opening G+ in the browser ... might not be apparmor related though14:16
jdstrandit is probably the webapp-container14:17
jdstrandI'll direct you to alex-abreau and dbarth14:18
jdstrandlet's go to #webapps14:18
jdstrand(where they both are)14:18
ogra_well, dbarth asked me to file the bug, he is aware14:18
* dholbach hugs mvo14:19
ogra_no need to be pushy, it never worked before ... having it working in the browser now is a massive improvement already :)14:19
LaneyHow can I run a/the click user hook/s manually?14:20
LaneyI got an apport report from one of them14:20
Laneyclick hook run-user, got it14:23
barrymvo: ping14:25
mvobarry: hello! pong14:25
barrymvo: hi!  thanks for merging and uploading gdebi.  it'll be nice to kill another python2 dependency. :)  however i think there's a problem in sid:14:25
barryapt-get install gdebi-core14:26
barryThe following packages have unmet dependencies:14:26
barry gdebi-core : Depends: python3:any (>= 3.4~)14:26
barry 14:26
barrythat's in a fresh sid chroot, but on a debian system w/gdebi-core already:14:26
barry gdebi : Depends: python3:any (>= 3.4~)14:26
barry         Depends: gdebi-core (= 0.9.5) but 0.9.4 is to be installed14:26
barry 14:26
mvobarry: oh, let me have a look14:27
barrythanks14:27
barrymvo: ah, i see xnox committed (probably) a fix in bzr14:28
xnoxbarry: and uploaded...14:28
mvo:) thanks xnox !14:28
xnoxbarry: it seems that X-Python3-Verions: >= 3.4 was actually honored and impossible dependencies got generated =)14:29
barryxnox: thanks!  will you sync once it lands in debian?14:29
xnoxbarry: it's not needed on ubuntu.14:29
xnoxbarry: as our python3 is 3.414:29
barryxnox: right, it didn't break us, but it still might be good to keep in sync14:29
barryxnox, mvo anyway, thanks!14:30
Riddellgosh, an mvo14:31
mvohey Riddell14:40
=== salem_ is now known as _salem
pittijuliank: FYI, our jenkins now exports *-packages, like in https://jenkins.qa.ubuntu.com/job/trusty-adt-unattended-upgrades/45/ARCH=i386,label=adt/14:48
bdmurrayhallyn: should that supercede the existing libvirt (which appears to have ftbfs btw)15:19
hallynbdmurray: yes15:19
bdmurrayhallyn: got it15:19
hallyni goofed - the pach i was backporting in 8.8 had a hunk applied int he wrong effing fn15:19
hallynthanks15:19
jibelbarry, mvo latest pyflakes seems to have broken autopkgtest of unattended-upgrade https://jenkins.qa.ubuntu.com/view/Trusty/view/AutoPkgTest/job/trusty-adt-unattended-upgrades/ARCH=amd64,label=adt/46/artifact/results/dsc0t-run-tests-stdout15:37
barryjibel: thanks, that ought to be shallow.  let me see if i can craft a fix15:38
juliankpitti: That's great, but why do you tell me about it? Did I ask about that recently?15:42
juliankSeriously, I don't know. I forget many things I ask.15:43
mvojibel: thanks15:44
mvobarry: let me know if you want me to have a look instead15:44
barrymvo: if you have some time to get a fix uploaded, that would be great.  this ought to do it: http://paste.ubuntu.com/7194816/15:45
barrymvo: or something like it ;)15:45
mvobarry: aha, the new "shadowing" warning :)15:45
mvobarry: sure thing, happy to do that15:45
barrymvo: yep ;)  thanks!15:46
pittijuliank: hm, I thought you did; sorry if I mixed that up with someone else15:46
pittijuliank: sorry, I was15:46
pittijtaylor: FYI, our jenkins now exports *-packages, like in https://jenkins.qa.ubuntu.com/job/trusty-adt-unattended-upgrades/45/ARCH=i386,label=adt/15:47
juliankpitti: No problem.15:47
jibelbdmurray, FYI, I reproduced bug 1286161 with the clone from the OP15:55
ubottubug 1286161 in ubuntu-release-upgrader (Ubuntu Trusty) "13.10 -> 14.04 upgrade failed: initramfs failed to ugprade, udev is not configured yet" [High,Confirmed] https://launchpad.net/bugs/128616115:55
mvobarry: uploaded15:58
barrymvo: thanks!15:59
bdmurrayjibel: oh, anything interesting?16:04
jibelbdmurray, apart that it is reproducible with packages from the archive without any PPA, nothing else yet16:09
bdmurrayjibel: oh, you might interested in bug 1290584 regarding apt-clone16:17
ubottubug 1290584 in apt-clone (Ubuntu) "apt-clone restore should have a use my mirror option" [Undecided,In progress] https://launchpad.net/bugs/129058416:17
=== _salem is now known as salem_
pittixnox: all these "ValueError: invalid literal for int() with base 10:" in ubiquity, is that real, or just a transient thing?16:46
pittixnox: i. e. does it make sense to retry a few times, or is that really broken?16:46
pittixnox: sounds like it's trying to interpret a nonexisting debconf key as a number?16:46
pittibut why would removing the u1 bits affect partman..16:47
xnoxpitti: haven't looked. That error messages comes up for any reason when debconf backend goes away.16:49
xnox(and ubiquity still tries to interract with it)16:49
pittihallyn: is 2.0 still on the table for trusty? it gets kind of tight now16:56
pittihallyn: FWIW, I've been using it since your annoucement (and that crash fix), and have basically forgotten about it :) (IOW, all happy)16:57
hallynpitti: yeah, my plan is if upstream doesn't tag v2.0 this weekend i will take a git snapshot for .orig.tar.gz and push to trusty archive16:58
pittihallyn: ah, you were waiting for the final, not on testing any more?16:59
hallynpitti: what do you mean?16:59
pittihallyn: yes, then uploading a pre-release snapshot sounds fine16:59
hallynpitti: im' still uploading new versions to that ppa every other day or so16:59
pittihallyn: the diff from the snapshot to final is presumably much smaller than 1.7 to 2.0 final?16:59
hallynoh, yeah, i definately do not want to do a diff on top of 1.717:00
pittioh, need to run out, taxi going back to the hotel; good night everyone17:00
hallynpitti: good night17:00
=== bfiller is now known as bfiller_afk
=== Logan__ is now known as Logan_
=== bfiller_afk is now known as bfiller
=== roadmr is now known as roadmr_afk
=== mhall119_ is now known as mhall119
dokomlankhorst, tjaalton: xorg ping18:21
tjaaltondoko: xorg pong18:27
dokotjaalton, two things:18:28
dokohttps://launchpad.net/ubuntu/+archive/test-rebuild-20140307/+build/5735335  the xorg-gtest ftbfs18:28
dokoand the vnc4 ftbfs for arm64 and ppc64el (embedded xorg), https://launchpad.net/ubuntu/+source/vnc4/4.1.1+xorg4.3.0-37ubuntu5  do you have an idea how to fix that?18:30
tjaaltonugh18:30
tjaaltoni bet xorg-gtest builds fine on sbuild..18:33
tjaaltonlp is picky18:33
tjaaltonno idea about vnc4.. i'd just drop it but someone might whine18:34
bdmurraystgraber: could you have a look at bug 1300654?18:35
ubottubug 1300654 in isc-dhcp (Ubuntu) "Prompt to update unmodified configuration file during upgrade from Precise to Trusty" [High,New] https://launchpad.net/bugs/130065418:35
stgraberbdmurray: I'm aware of it18:36
=== roadmr_afk is now known as roadmr
bdmurraystgraber: great18:46
=== buxy_bak is now known as buxy
=== jackson is now known as Guest78494
=== Guest78494 is now known as Noskcaj_
mwhudsoninfinity: so the glibc build failed *again* for me, and i am very sure that ulimit -s wasn't unlimited19:50
* mwhudson tries to remember the other reasons why the heap might grow up19:51
infinitymwhudson: Try the patch in the bug you filed?19:51
mwhudsonwell yes, that will certainly help19:51
infinitymwhudson: (assuming it was tst-tls5 that failed again)19:51
mwhudsonor is certainly worth trying at least19:51
mwhudsonyeah, same failure19:51
mwhudsoni wonder if the newer kernel grows the heap up in some other circumstance than 3.819:52
infinitymwhudson: I'm really not interested in knowing all the reasons why that test might fail, if the 2-line patch fixes the bug. :P19:52
mwhudsonyeah, focus focus19:52
mwhudsonbuilding19:54
infinityhallyn: *poke*20:15
infinityhallyn: In that cgmanager upload, you moved pretty much EVERYTHING to /lib ... Really, only the actual library should move, the .so, .pc, and .a should stay in /usr/lib20:17
infinityhallyn: Also, you dropped the static build (intentional?), and your changelog claims you added -static when I think you meant -shared.20:18
infinitystgraber: ^ If you want to address any/all of that for hallyn...20:18
=== alesage|afk is now known as alesage
infinitymvo: Say, why does unattended-upgrades have .pyc files in the source?20:22
juliankinfinity: I believe the pyc files in unattended-upgrades were accidental See https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=741467 for the fix.20:46
ubottuDebian bug 741467 in src:unattended-upgrades "unattended-upgrades: source package contains generated files (*.pyc)" [Minor,Fixed]20:46
juliankWhile I'm not mvo, this might still help.20:47
infinityjuliank: Ahh, kay, good to know that it's been addressed.  I just noticed the binary diffs in the queue and was non-plussed, but it's not like this is a new regression in the packaging or anything. :P20:48
juliankMaybe mvo wants to merge 0.82.4, assuming it's not too different from 0.82.1ubuntu1. The "fix Bug in rewind_cache() that can cause unwanted removals of packages" in 0.82.2 makes 0.82.1 seem a bit scary20:50
jtaylorwhat are the chances of still patching libtool to not strip gccs -fsanitize stuff in trusty?20:51
jtaylorhm I guess its not super important as you can get the links indirectly in your main application20:52
infinityjtaylor: I'm not sure that changing the behaviour of libtool two weeks before release would be a bright idea.  Since nothing will be rebuilt with it, but all security updates will be.20:53
jtayloryes20:54
jtaylorjust tested it, indirect links seem to work alright anyway20:54
hallyninfinity: feh, so i should just be moving the actual file by hand in debian/rules?21:02
hallyninfinity: i dropped the static build bc libcgmanager.so didn't want to be built -shared unless I did so;  if i find a way to have both happen i'll re-add the static lib21:03
infinityhallyn: Sure, don't care much about the static lib, was just curious.21:04
infinityhallyn: And yeah, check the bit in expat's debian/rules which moves libexpat.so.* to /lib and then re-targets the symlink in /usr.21:04
infinityhallyn: Should be able to do that fairly cleanly in a dh_auto_install override.21:05
infinity(Credit to slangasek for finding the expat rules snippet)21:06
hallynok, thanks.  I'll change that tongiht and re-push.21:06
dokorobert_ancell, pyexiv2 ftbfs on ppc64el22:26
xnoxstgraber: not sure if you are arround, could i get a silo for: https://code.launchpad.net/~xnox/indicator-applet/drop-u1/+merge/213873 and https://code.launchpad.net/~xnox/webapps-applications/remove-u1/+merge/213853 ?23:21
xnoxor does anybody know who is US/west based who can allocate silos... cyphermox ?23:22
stgraberxnox: I'm not on the landing team myself, so I can add you to the spreadsheet but since those aren't foundations team packages, you may as well go straight to the ci channel23:27
xnoxstgraber: hm, let me check who those two should go via. Maybe i can ping about them tomorrow.23:29
stgrabercyphermox is my usual suspect but it's past our eastern time EOD so he may be gone for dinner or something23:29
xnoxi've been chatting with dbarth about the webapps one, not sure who is responsible for legacy applets.23:32
xnoxstgraber: excellent. "indicator-applet" has no owner =) but it has been reviewed by ~unitish people. Could you silo up that one?23:33
stgraberxnox: since they're related, they probably should go in the same silo, let me add you to the spreadsheet at least23:34
xnoxstgraber: all of them are stand alone. webapps is about dropping ubuntuone music webapp, indicator-applet is about dropping recommends to indicator-sync which will stop working.23:34
xnoxstgraber: the two unity7 branches are already inflight in silo5 - which do same as indicator-applet branch + twiddle default settings to drop u1 control panel from default launchers.23:35
xnoxunity7 is bundled with other bug-fixes however, so i wouldn't want to piggy back on top of that.23:36
xnoxbregma: can you release/land indicator-applet?23:36
stgraberxnox: ok, I added you to the spreadsheet, let's see if there's someone in #ubuntu-ci-eng to give you a silo23:36
xnoxbregma: this is assuming indicator-applet are ~= unitiish =)))23:37
xnoxstgraber: could you ask? i'm not in that channel at all.23:37
stgraberxnox: yeah, just joined and asked (since we pretty much never land anything through the citrain I don't idle in there)23:37
stgraberxnox: do you have any testsuite I should link in there?23:38
xnoxstgraber: it's desktop only, btw.23:38
xnoxstgraber: i have manual test cases to test that xubuntu works, and that shortcuts are not in the default unity launcher. I've tested both locally.23:38
xnox(xubuntu is the prime user of indicator-applet)23:38
stgraberxnox: ok, I think I've put as much details as I can in the spreadsheet, hopefully someone will flip the switch and give you a silo soon23:39
xnoxstgraber: thanks a lot! i'll watch that spreadsheet.23:41
robert_ancelldoko, what is the easiest way to test on ppc64el?23:44
dokorobert_ancell, ssh porter-ppc64el.canonical.com23:46
robert_ancelldoko, any chance of getting git installed on there?23:48
cjwatsonrobert_ancell: schroot -c trusty-ppc64el23:51
dokorobert_ancell, afk now23:51
cjwatsonrobert_ancell: git's in the schroot23:51
cjwatsonrobert_ancell: (and you can "sudo apt-get install" things)23:51
robert_ancellah, thanks23:52
stgraberxnox: we've got a silo23:57
xnox\o/23:58
xnoxcjwatson: i'm pretty sure i've installed a whole bunch of things in the chroot, to the point of getting build-conflicts now. sudo apt-get remove does not work.23:58
xnoxcjwatson: was i using our porter boxes somehow wrong? (e.g. a session/snapshot is not created for me that gets auto-cleaned up)23:59

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