/srv/irclogs.ubuntu.com/2014/08/22/#ubuntu-devel.txt

pittiGood morning03:43
pittiinfinity, xnox: if you install the hook as "source_plymouth.py" it will apply to any binary package of the "plymouth" source03:44
pittiah right, xnox said that03:44
pittibdmurray: saw your mail, will look and reply03:45
infinitypitti: Right, my complaint was exactly that it was keying off filename. ;)03:49
infinitypitti: (which is what leads to the bug where you can't have it in a library package)03:50
infinityAnd if the library is the lowest common denominator of your source package stack, that seems the sensible place to have the source-wide hook.03:51
pittiright, you can't put a source_ hook into a lib pacakge, but certainly a hook for that lib package03:51
infinityAnyhow, I think xnox is just going to move it from libplymouth to plymouth and call it good enough.03:51
pitticjwatson: the perl transition now caused a number of failures due to a new warning "push on reference is experimental" on stderr; is that expected and should the tests be adjusted to ignore this, or should this be fixed (or suppressed)?04:09
cjwatsonpitti: tests should be fixed not to do that; if they are really sure they want to use that feature then there's a switch available in perl to suppress the warning.  http://search.cpan.org/dist/perl-5.20.0/pod/perldelta.pod#New_Warnings04:28
slangasekxnox: it's in the library package because that's the base package that all other packages depend on, so ensures that it's always present.  A simple Replaces: would have sufficed to handle the file conflict (without even a Breaks:)  But if you want to move it from the libplymouth package to the plymouth package, that seems ok04:28
pitticjwatson: oh wow, didn't expect you at that hour :)04:28
pitticjwatson: ah wait, debconf04:28
cjwatsontravelling today04:28
slangasek:)04:30
infinityslangasek: Using replaces that way is generally considered wrong, though.04:54
infinityslangasek: Since it can lead to missing files (though, I guess in a library ABI bump, the odds of going backwards are slim).04:55
slangasekinfinity: as the only replaced file here was an apport hook, going backwards doesn't break the package's main functionality anyway04:59
infinityslangasek: True.04:59
=== doko_ is now known as doko
dholbachgood morning06:39
dokojamespag`, please could you have a look at the mysql-5.5 autopkg test failure? blocks perl (and mysql-5.6 autopkg tests always failed)06:40
pittiwgrant: nevermind mangling the utopic langpack cronjob; I did a workaround yesterday for updating the touch langpacks straight out of the upstream trunks06:50
LocutusOfBorg1Hi doko do you think debian 727330 can be closed? The package build also on ppc64el and arm6406:52
ubottuDebian bug 727330 in src:binutils-msp430 "binutils-msp430: run dh-autoreconf to update config.{sub,guess} and {libtool,aclocal}.m4" [Normal,Open] http://bugs.debian.org/72733006:52
LocutusOfBorg1bdrung, you there?06:54
LocutusOfBorg1I would like to merge vlc master-daily (ppa) with the debian git, to fix the FTBFS06:56
jfi_Hello, what is the correct way to checkout the utopic branch of a package? bzr branch lp:ubuntu/utopic/<pkg> retrieves the trusty branch.07:37
geserjfi_: which package? have you checked if it's listed on http://package-import.ubuntu.com/status/ having issues?07:39
jfi_geser, the pkg that I am trying to checkout (to fix a major bug) is: psensor (https://code.launchpad.net/~ubuntu-branches/ubuntu/utopic/psensor/utopic)07:40
jfi_geser, the pkg (psensor) is in this list07:40
geserhttp://package-import.ubuntu.com/status/psensor.html#2014-04-08%2014:18:53.94018207:41
pittidoko: libtbfs-perl apparently needs to be rebuilt for new perl? it's uninstallable07:43
pittidoko: lintian4python, adequate, and liblingua-en-numbers-ordinate-perl succeed again with fixed adequate07:44
jfi_geser, hum, it means that the bzr repository is corrupted for this package and I need to report a bug to get some attention about it?07:44
pittidoko: hmm, seems https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=735623 is in the way for libtfbs-perl, as its build-dep pdl fails on powerpc; we might just remove it from powerpc then?07:45
ubottuDebian bug 735623 in pdl "pdl: FTBFS on mips, powerpc, s390x, sparc, (ia64): testsuite failures" [Serious,Open]07:45
geserjfi_: yes, it's at least out-of-date (which explains why you got the trusty version from the utopic branch). I'm not sure what's the correct process to get this fixed07:46
Saviqmardy, hey, I got "Ubuntu One" greyed out in accounts after account removal07:46
Saviqmardy, guys in London saw that last week, too, anything I can get you to debug? how can I recover?07:46
mardySaviq: weird, it gets greyed out once there is an U1 account07:47
mardySaviq: can you do "account-console list" and see if there's still an U1 account?07:47
Saviqmardy, yeah, ssweeny saw that the account didn't get removed properly07:47
Saviqmardy, hmm no, empty07:47
Saviqmardy, ah now it's back, but doesn't open...07:48
mardySaviq: how did you make it come back? did you exit from system settings?07:48
Saviqmardy, yes07:48
dokopitti, now synced pdl, although this won't fix things. I'll check to demote/remove pdl/libtfbs-perl07:48
Saviqmardy, it's meant to work via trusted sessions now is it?07:49
mardySaviq: yes07:49
Saviqmardy, so ok, the fact that doesn't open seems to be our fault (testing silo)07:49
mardySaviq: so, you click, but nothing happens?07:49
Saviqmardy, yeah, looks like a problem in our silo07:49
pittidoko: only rdepends of libtfbs-perl is med-bio-dev (only a recommends), so removing that on powerpc seems fine07:49
Saviqmardy, and I can't reproduce the greyed-out with another phone...07:50
jfi_geser, Thanks for pointing me to the pkg import page. I am going to submit my patch as a debdiff and try to find the best way to report the bzr/lp issue.07:50
Saviqmardy, ok, will report back if I see it again, for now, as you were ;)07:50
dokopitti, otoh, maybe just lets ignore the test results on powerpc?07:52
dokoResult: FAIL07:52
dokoFailed 3/124 test programs. 0/1775 subtests failed.07:52
pittidoko: hm, I'd avoid sourceful changes and just remove it from ppc -- who cares..07:52
pittibut either way07:53
pittibut like that, as soon as it gets fixed in Debian it'll automatically come back07:53
doko$ apt-cache rdepends pdl07:53
dokopdl07:53
dokoReverse Depends:07:53
doko  pdl:i38607:53
doko  science-viewing07:53
doko  science-numericalcomputation07:53
doko  science-astronomy07:53
doko  libtfbs-perl07:53
doko  libpdl-stats-perl07:53
doko  libpdl-netcdf-perl07:53
doko  libpdl-linearalgebra-perl07:53
pittidoko: oh, you mean pdl07:53
doko  libpdl-io-matlab-perl07:53
doko  libpadre-plugin-pdl-perl07:53
pittiI was thinking of libtfbs; pdl is already not available on powerpc, so no need to change anything there07:54
dokopitti, ok, done, and libtfbs-perl uploaded08:02
=== Mikee_C_afk is now known as Mikee_C
highvolt1gedayangkun:08:05
pittidoko: cool, thanks! still looking into orafce, I have a reproducer and filed an upstream bug now (https://github.com/orafce/orafce/issues/15)08:30
tkamppeterdoko, is printer-driver-brlaser seeded in Utopic now (bug 1355136)?09:18
ubottubug 1355136 in xubuntu-meta (Ubuntu) "Add recommends printer-driver-brlaser package in -desktop installations" [High,Confirmed] https://launchpad.net/bugs/135513609:18
pittitkamppeter: I thought we had some magic to auto-install printer driver packages?09:19
=== Trevinho|holiday is now known as Trevinho
tkamppeterpitti, for  Ubuntu packages of printer drivers?09:23
tkamppeterpitti, I see a lot of Recommends: for printer drivers in ubuntu-desktop.09:23
pittitkamppeter: ubuntu or openprinting.org, etc.09:23
pittiyes, I know; but that has  always felt like a workaround09:23
pittiwe install video codecs and other stuff on demand, for printer drivers this would be even more appropriate09:23
tkamppeterpitti, there is a mechanism to auto-install drivers via OpenPrinting, but these are LSB packages or PPDs directly from manufacturers.09:24
pittitkamppeter: ah, I thought we had something similar for ubuntu packages09:24
pittiprobably mixed it up with ubuntu-drivers-common for other hw09:25
=== henrix_ is now known as henrix
tkamppeterpitti, can you add printer-driver-brlaser to ubuntu-meta when it is not yet done so (bug 1359137, bug 1355136). Thanks.09:29
ubottubug 1359137 in brlaser (Ubuntu) "[MIR] brlaser" [High,Fix released] https://launchpad.net/bugs/135913709:29
ubottubug 1355136 in xubuntu-meta (Ubuntu) "Add recommends printer-driver-brlaser package in -desktop installations" [High,Confirmed] https://launchpad.net/bugs/135513609:29
pittitkamppeter: seeded to platform.utopic desktop-common, so all the metapackages will pick it up on next rebuild09:32
tkamppeterpitti, thank you very much.09:34
Saviqjodh, hey, could you please help mzanetti with an upstart job based on the file bridge, he's not getting delete or change events, just create09:51
Saviqmzanetti, can you give more details ↑?09:51
mzanettijodh: have this job for a test: http://paste.ubuntu.com/8112758/09:52
mzanettijodh: doing a "start unity8-filewatcher" works fine... but doesn't get executed when doing touch/rm ~/.local/share/applications/foo.desktop09:52
zygamvo_: hey, do you have a moment to look at interesting apt related case?10:27
zygamvo_: https://code.launchpad.net/~roadmr/checkbox/ppppcc-glmark2-es2/+merge/23178810:27
zygamvo_: look at the dependencies there10:27
zygamvo_: is that the right way to handle this?10:28
jodhmzanetti: hmm - looks like there might be a problem with globs coupled with tilde expansion. Please can you raise an upstart bug. Meantime, you could use something like http://paste.ubuntu.com/8113277/10:44
mzanettijodh: hey, thanks! yes, I'll try your suggestion and file a bug10:45
mzanettiindeed, hardcoding /home/phablet/ makes it work10:48
mzanettiI wonder if/how url-dispatcher-update.conf works then10:49
mvo_zyga: I need a moment to look into the detail (a bit busy right now due rtm/debconf)11:10
Saviqmzanetti, we've had it *not* work from time to time, too11:19
Saviqmzanetti, so that could explain it11:19
=== tester56 is now known as kdeuser56
kdeuser56does partman allow reusing partitions somehow?11:20
mzanettiSaviq: ok... this should improve the situation: https://code.launchpad.net/~mzanetti/unity8/update-launcher-on-appdate/+merge/23186611:23
Saviqmzanetti, I think James suggested FILE=~/.local/share/applications MATCH=*.desktop11:24
mzanettiSaviq: not directly, but I agree worth a try11:24
Saviqmzanetti, although manual does not list that as a possibility...11:25
mzanettiSaviq: well, let me try. would be better than hardcoding to phablet11:26
Saviqmzanetti, you could also try .local/share/applications/*.desktop11:26
Saviqmzanetti, ~/ is implicit11:26
Saviqmzanetti, ah and it needs to have a trailing slash if it's a dir11:27
=== MacSlow is now known as MacSlow|lunch
zygamvo_: sure, no rush11:30
mzanettiSaviq: nope... both not working11:34
Saviqmzanetti, :|11:34
=== MacSlow|lunch is now known as MacSlow
=== ogasawara_ is now known as ogasawara
Mirvmlankhorst: hey there! the SDK team still has a regression with the new LLVM 3.5, so unfortunately the x86 tests are now again failing with LLVM error https://launchpadlibrarian.net/182951835/buildlog_ubuntu-utopic-i386.ubuntu-ui-toolkit_1.1.1206%2B14.10.20140822-0ubuntu1_FAILEDTOBUILD.txt.gz12:31
dokoRiddell, ScottK: why aren't you working on the smokeqt ftbfs, but trying to work around it?12:46
dokopitti, so just mysql-5.5 is failing ...12:46
dokodobey, jamespag`: mysql-5.5 ping12:47
tvossmlankhorst, mind jumping over to #ubuntu-ci-eng12:55
tvoss?12:55
tvossmlankhorst, seems like we have another llvm related issue12:55
pittitvoss: "LLVM ERROR: Do not know how to split the result of this operator!12:56
pittitvoss: by any chance?12:56
Mirvtvoss: pitti I just filed bug #136024112:57
ubottubug 1360241 in mesa (Ubuntu) "[Regression] LLVM error in executing Ubuntu UI Toolkit tests on x86" [Undecided,Confirmed] https://launchpad.net/bugs/136024112:57
Mirvpitti: that exactly12:58
pittiack12:59
mlankhorstpitti: again?12:59
mlankhorstwhere this time12:59
seb128bah, why was the llvm version change reverted?12:59
mlankhorsti moved back to llvm 3.5 because my original testcase worked12:59
mlankhorstshould I revert again for now? :P13:01
mlankhorstdoko: ^13:01
seb128yes13:01
seb128it's friday afternoon, not likely it's going to be resolved properly today13:01
dokoI'm afk now13:01
pittiOOI, why don't we build with gcc like everything else?13:02
dokobuilt what?13:02
pittiI seriously doubt that some .1% performance difference is noticeable in things like telephony-service or indicators?13:02
dokono, it's not using clang, it's the jit13:03
pittidoko: the bug that tvoss pointed to (UI toolkit etc.)13:03
Mirvpitti: I think it's mesa using LLVM failing on OpenGL tests, not that we use llvm to compile software?13:04
pittior are these like 'test guinea pigs' for a future complete migration to llvm?13:04
pittiMirv: ah, ok; thanks13:04
mlankhorstit's llvmpipe that's regressed, but on i386 only.13:04
mlankhorstit's the swrast13:04
Mirvand on x86 only, but since it's in unit tests UITK fails to build13:04
mlankhorsteven more specific, isn't it i386 only?13:05
mlankhorstnever seen an issue on amd64 related to it13:05
pittiI've seen it on arm and arm64, too13:05
mlankhorstodd, not me13:05
dokoMirv, mlankhorst so what about ignoring this test when it is i386 only?13:05
mlankhorstafaik it's in arch specific code13:06
Mirvbzoltan mentioned he wouldn't like to start selectively going through the tests and disabling them on x86, but sure that's one option13:06
Mirv(but bzoltan isn't on this channel)13:06
* Mirv needs to afk13:06
dokoanyway, I'm afk now13:07
mlankhorstfwiw updating mesa won't help, the problem is in llvm13:07
dokowe need somebody ro report this ...13:07
mlankhorstdo we have someone working on llvm?13:07
dokoshould be the desktop team, they are using llvmpipe13:08
Riddelldoko: I'll be looking at it now13:08
mlankhorstfwiw I think I will push the revert for now, i still want 3.5 but they're annoying13:08
mlankhorstno need to stall everything while we find a more permanent solution13:08
dokosure13:09
mlankhorstok uploaded13:12
seb128doko, desktop team is using llvmpipe? where?13:18
seb128don't thing we are13:19
seb128but we are not looking at llvm for sure13:19
pitti. o O { hot potato }13:19
seb128indeed ;-)13:20
seb128if it's a desktop issue I just vote for staying on 3.4 which works13:20
seb128if somebody wants 3.5 they can deal with the update and the issues it creates...13:20
zbenjaminogra_: ping, didn't we have gdbserver on the image by default?13:29
ogra_zbenjamin, hmm i thought we seeded it once ... but i cant see it ...13:31
zbenjaminogra_: yeah it seems to be gone13:32
ogra_(and we are getting very short on space ... we need to start gatekeeping what goes in soon)13:32
ogra_(else i would just re-add it )13:32
ogra_(and start dropping debug tools for the final product)13:32
zbenjaminogra_: well for developer expirience it should go in, it would be a bit of a dealbreaker if a developer needs to void hos warranty13:32
pittiubuntu touch images aren't going to be a developer platform anyway13:33
zbenjaminogra_: we would force devs to make the image writeable13:33
ogra_we are at 493 of 500MB allowed for the rootfs13:33
zbenjaminpitti: why not?13:33
ogra_pitti, err13:33
pittiit's already way bigger than it should be, and still doesn't even have the basic dev tools13:33
ogra_pitti, it should have all debug tools you need for click package development13:34
zbenjaminpitti: i don't need gdb or gcc, we need gdbserver13:34
ogra_i.e. we have strace13:34
pittiwell, for development you need quite a bit more than that :) for getting crash reports those are enough, yes13:34
ogra_gdbserver and valgrind were added ... i know valgrind was dropped again because it pulled in libc6-dbg13:35
ogra_not sure why gdbserver is gone now13:35
zbenjaminogra_: valgrind is not supported by the SDK yet anyway13:35
ogra_zbenjamin, and never will be on the image ...13:35
zbenjaminpitti: compiling will happen on the host, as well as debug information can be on the host, what else would you need on the phone?13:35
zbenjaminogra_: fine for me ;)13:35
ogra_way to big :)13:36
pittizbenjamin: my thought exactly :)13:36
pittizbenjamin: well, this will become more interesting in a converged world when devs will actually develop *on* the phone; but still a bit far out13:36
=== desrt is now known as desrt|pdx
pitti(looking forward to that, though! :) )13:37
ogra_pitti, by that time the desktop and phone installs should already be identical ;)13:37
ogra_so we will have solved the issue already ;)13:37
pittino more carrying laptops around to family or friends, just plug in phone into their laptop or TV13:37
* ogra_ just wants a system-image install on his laptop ... 13:38
ogra_full release upgrade in 10 min and so on ;)13:38
ogra_and no more upgrade breakage possible13:38
pittiogra_: I doubt you'd stay happy for very long with the current images, though :)13:38
ogra_pitti, well thats the plan13:38
=== pete-woods is now known as pete-woods|lunch
pittiogra_: (sure, sure -- just Friday trolling)13:39
ogra_true for the current qualiity though :)13:39
pittiogra_: no vim, no mutt, no cookies!13:39
ogra_we ship vi-runtime13:39
ogra_;)13:39
ogra_or vim-runtime13:39
ogra_and who needs mutt ... use dekko !13:39
ogra_the best since sliced bread ;)13:39
pittiogra_: more seriously, do we actually have an email client? I looked for one the other week and didn't find one13:40
pitti(in the store)13:40
ogra_yeah, dekko13:40
ogra_a fork of trojita13:40
* pitti installs13:40
ogra_the current version has issues with creating a new account though (afaik, i use an old config )13:40
ogra_you can find DanChapman (teh dev) in #ubuntu-app-install if you have questions/issues13:41
ogra_err13:41
ogra_#ubuntu-app-devel13:41
ogra_sorry13:41
ogra_pitti, soo ... i asked you before but my question always drowned in other discussions we had ... i have that whishlist bug for RZM to add a popup for low diskspace ...13:42
ogra_*RTM13:42
ogra_is there a backend i could easily use ? (i knwo i asked you in malta. but forgot what the solution was)13:43
pittiogra_: oh the dekstop that's {gnome,ubuntu}-settings-daemon; we don't have that on touch, so I don't think so13:43
pittiogra_: but it's not terribly difficult to call df, I figure13:43
* pitti checks how g-s-d is doing that, whether there's some clever kernel notification or so13:44
pittiogra_: wow, how big is that? has been downloading for like 4 mins now13:45
pittiit comes with all my mails pre-included :)13:45
* pitti doesn't get that far, no "continue" button after entering password; anyway, something to try later on13:47
ogra_pitti, lol, yeah13:47
ogra_oh. you could create an account ?13:47
pittiogra_: ah, asked Dan in #u-a-d; thanks for pointing out!13:49
smoserhi, server team has 4 packages in New queue for utopic that we'd like to see let in. (python-lxc, openstack-granite, and percona-xtradb-cluster-5.6 and  percona-xtradb-cluster-galera-3.x )14:49
smoseranyone able o help us out ?14:49
pittimvo_, cjwatson: FYI, latest click upload mightily regresses its tests15:26
pitti(no email notification, as it's a train upload)15:26
mvo_pitti: oh? do you have a url? i check it out15:43
pittimvo_: https://jenkins.qa.ubuntu.com/job/utopic-adt-click/29/ARCH=amd64,label=adt/15:43
mvo_thanks pitti15:48
=== pete-woods|lunch is now known as pete-woods
Riddelldoko: I conclude the smokeqt issue is to do with gcc 4.9 as one of the errors goes away when you force using gcc 4.8 (can't just set enviornment variables, I suspect smokegen isn't smart enough) but there's another error I can't work out16:26
Riddelldoko: see e-mail16:26
=== Mikee_C is now known as Mikee_C_afk
=== roadmr is now known as roadmr_afk
* Elbrus is wondering about the gnustep-gui transition17:54
ElbrusI thought it had to be done before the import freeze, but now I see that gnustep-gui is imported into proposed17:54
Elbrusreading http://people.canonical.com/~ubuntu-archive/transitions/html/gnustep-gui.html says no action is taken to rebuild anything after that17:55
Elbrushow can the rebuilds of all the dependecies be scheduled?17:56
Elbruscjwatson: could you do this somewhat automated? (in Debian I would be requesting binNMU's but I was told that does not exist (yet) in Ubuntu17:57
mvo_pitti: I did a 0.4.31.2 now that should fix the failures (it did for me locally at least :)18:06
=== roadmr_afk is now known as roadmr
Unit193Linked to a couple dsc files on LP 992068 and LP 135877019:25
ubottuLaunchpad bug 992068 in torsocks (Ubuntu) "[needs-packaging] broken in 12.04 and 12.10: libtorsocks: many symbols not found" [Wishlist,Confirmed] https://launchpad.net/bugs/99206819:25
ubottuLaunchpad bug 1358770 in remmina (Ubuntu) "ubuntu utopic: missing dependency to libvncclient.so.0" [Undecided,New] https://launchpad.net/bugs/135877019:25
=== arrrghhhAWAY is now known as arrrghhh
=== robru is now known as robru|lunch
=== robru|lunch is now known as robru
NoskcajTo rebuild a package in trusty, do i just file an SRU as normal or is there a easier way?21:13
jtaylorits like a normal sru21:15
Noskcajok21:22
=== arrrghhh is now known as arrrghhhAWAY

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