/srv/irclogs.ubuntu.com/2016/09/30/#ubuntu-devel.txt

jbichanacc: this late in ubuntu's release cycle I suggest to go ahead and upload to Ubuntu now00:03
nacccoreycb: would your team be able to look at the ftbfs for python-eventlet and python-taskflow? I think both are openstack umbrella and seem to be test failures00:04
naccjbicha: yep, that makes sense, that's what i was trying to figure out, on the balance, thanks!00:04
jgrimmnacc, ah, good to know. thanks00:42
tsimonq2thanks sarnold ;)01:01
* tsimonq2 is testing that it installs cleanly atm01:01
sarnoldtsimonq2: btw, your patch had dos-style crlf line endings.. I'm curious how you managed to do that :)01:05
* tsimonq2 looks at .vimrc01:06
sarnoldpatch claimed to strip them all; I haven't verified that by hand yet..01:06
tsimonq2for some reason this is enabled... https://github.com/vim-scripts/PreserveNoEOL01:07
tsimonq2could that do it maybe?01:07
tsimonq2otherwise I'm not seeing what could be doing that01:07
sarnoldtsimonq2: hrm. could be..01:14
* tsimonq2 remove it01:14
tsimonq2*removes01:14
=== juliank_ is now known as juliank
tsimonq2sarnold: thank you03:56
sarnoldtsimonq2: and also thank you :)03:56
tsimonq2:)03:57
tsimonq2sarnold: have a good evening03:57
=== hikiko_ is now known as hikiko
ventricalcq , cq goodmorning05:08
Kinder-PingviHi guys! What is the difference between ubuntu 16.10 beta 2 and daily builds?06:48
Kinder-PingviDoes daily changes are not included to beta 2?06:48
ventricala beta 2 is like a relase candidate06:55
Kinder-Pingviventrical, what about new updates from daily? Does it will be included into release candidate?07:35
ventrical4:30 am here in Canada .. I gotta get some rest :)08:19
smbIs there any preference whether a Vcs-Git line should become XS-Original-Vcs-Git or XS-Debian-Vcs-Git when I add a Vcs-Git line that points to Ubuntu? The former seems more in line with what is done with maintainer. Lintian does not seem to acknowledge any of those...08:27
Unit193pitti: You ever tried usrmerge on Ubuntu?09:47
Unit193FWIW, Debian #839046 / #839162.09:53
ubottuDebian bug 839046 in debootstrap "debootstrap: enable --merged-usr by default" [Normal,Open] http://bugs.debian.org/83904609:53
cjwatsonKinder-Pingvi: betas / release candidates are snapshots.  new updates aren't included in those snapshots, though if you install from the snapshot and apply all network-available updates you'll generally get a similar result.10:52
cjwatsonsmb: We settled on XS-Debian-Vcs-Git many years ago.10:53
cjwatsonsmb: (and yes, different from Maintainer for some reason I forget)10:53
smbcjwatson, Ok, thanks.10:55
=== hikiko is now known as hikiko|ln
Kinder-Pingvicjwatson, thanks for answer! So, if I do 'apt full-upgrade' on ubuntu 16.10 I will receive that updated included into daily builds, right?11:12
cjwatsonKinder-Pingvi: yep, should do11:12
Kinder-Pingvicjwatson, thank you!11:13
coreycbnacc, yep will look at those11:24
pittiUnit193: not tried directly on ubuntu, no; but looking forward to it anyway :)11:35
jbichacjwatson: are you aware of any tools that use Xs-Debian-Vcs ?11:36
pittiUnit193: this makes the whole thing simpler, more robust, and finally puts a lid on these "but I want to have separate /usr without initramfs, and on NFS" kind of quarrels :)11:36
cjwatsonjbicha: not off the top of my head11:37
Unit193pitti: Heh, as long as nothing breaks I don't really mind it either. :D  (Tempted to try it in a VM.)11:37
jbichafor most of the ubuntu-desktop packages, we just dropped the Vcs-Svn (pointing to pkg-gnome) and added Vcs-Bzr11:38
pittiUnit193: I think there's 4 packages left that break11:38
pittiUnit193: https://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=usrmerge;users=md@linux.it11:39
pittiUnit193: so that's a case of throwing them out of testing / demoting to -proposed IMHO; the rest is fine11:39
coreycbpitti, if you have a chance can you take a look at the xenial python-pylxd upload again please?11:41
pitticoreycb: I'm at a conference, just quickly checking IRC; so not today, sorry11:42
coreycbpitti, ok11:42
coreycbtjaalton, would you be able to review the python-pylxd sru upload to xenial?11:43
tjaaltoncoreycb: sure11:45
coreycbtjaalton, thanks11:45
=== hikiko|ln is now known as hikiko
=== marcusto_ is now known as marcustomlinson_
caribouwhat is the best way to be sure that a main package will build w/o Universe dependancies ?12:43
caribouI'm using sbuild-launchpad-chroot with only updates+main so it will bomb on me if Universe dependancies exist12:44
cariboubut that covers both build-only dependancies and binary ones, right ?12:44
tumbleweedthat's a pretty good way12:45
caribouyes, but I need to differentiate b/w build-only & binary, that's what I'm missing12:47
tumbleweedI don't understand that question12:48
cariboutumbleweed: that's because it's a dump one ;-)12:48
tumbleweedbuild-only dependencise are presumably Build-Depends, which answers the "will build w/o universe" question12:49
cariboutumbleweed: all I need to do is check the control file12:49
jbichacaribou: are you concerned about runtime dependencies or build dependencies for universe?12:50
cariboujbicha: runtime deps; I uploaded clamav w/o noticing that I had added a runtime dep12:50
cariboujbicha: just want to trace back what I missed so I don't do it again & get chased by doko to MIR the runtime dep :)12:51
tumbleweedcaribou: you could run the autopkgtests to tell you that12:51
tumbleweedor just manually eyeball the diff for changes in Depends12:51
cariboutumbleweed: hmm, good idea12:51
jbichatrying to build with only main will cause problems since some things like documentation build dependencies are in universe now since they don't add universe binary depenencies12:52
tumbleweedreading diffs is always a goo didea12:52
Laneybinary debdiffs are quite nice12:53
Laneyany kind of reading is not foolproof though12:53
* Laney is a big enough fool12:53
tumbleweedsometimes they are too big to read, thoguh. Then filterdiff -i '*/debian/*' is handy12:53
xnoxcaribou, why do you care? universe is always enabled, and one can build-depend on universe packages.12:53
Laneyxnox: you need to know if runtime universe deps are generated12:53
xnoxthe restriction is no Depends, Recommends, Built-Using in the resultant packages12:53
caribouxnox: Laney is right, this  is what happened to me with clamav12:54
xnoxcaribou, use $ check-mir ?12:54
xnoxhowever, i wonder if check-mir is still correct12:56
Laneyit only looks at debian/control IIRC12:56
xnoxit would be nice to look at debian/$PKG/DEBIAN/control12:56
caribouxnox: Laney: yeah, looks like it only checks debian/control12:58
caribouok,  I think I know now why I missed it the first time :13:01
cariboulibclamav7 has Depends: ${misc:Depends}, ${shlibs:Depends} only13:02
caribouthat got resolved as a runtime dependancy to libtfm-dev13:02
caribouThis is not a problem on Debian but it becomes one for us13:03
caribou(libtfm-dev is the one in Universe)13:03
=== _salem is now known as salem_
jgrimmhallyn, are you still looking after https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1571209 ?13:42
ubottuLaunchpad bug 1571209 in libvirt (Ubuntu Xenial) "Sockfile check retries too short for a busy system boot" [High,New]13:42
hallynjgrimm: no?   i delete all libvirt related emails automatically13:50
hallynjgrimm: as this is sysvinit and upstart only i doubt you care too much?13:52
hallynand in fact the current solution isn't bad, just not perfect.13:53
hallyn(on mine it appears to be doing the 'sleep 2', which is fine.  o fcourse i also have mine overrided to not start to save battery :)13:54
jgrimmhallyn, no worries. I was just doing some housekeeping this morning and noticed that one just lingering out there14:02
hallynjgrimm: but that one is 'fix released' right?14:02
hallynit just has a 'maybe we should do it better one day' comment at theend,14:02
hallynbut really the 'do it better' should be to use socket activation in upstart14:02
jgrimmand your last comment was about planning to sru to trusty14:03
hallynoh14:04
jgrimmhallyn, right, reporter was on trusty and hoping to get it SRU'd back.  that's what i was noting14:04
hallynshould be a trivial SRU fwiw14:07
* hallyn bbl14:07
jgrimmhallyn, no worries.. that's what i was checking on, whether you'd lost interest or still interesting in seeing it through14:07
hallynjgrimm: it'll be very simple, and would be a good one for a new contributor.  if you find one feel free to have them ping me to sponsor15:38
jgrimmhallyn, ack. thanks!15:39
jgrimmhallyn, https://bugs.launchpad.net/ubuntu/+source/lxc/+bug/1285850   is there a SRU required to Xenial or just open task in case there is still an issue?15:45
ubottuLaunchpad bug 1285850 in lxc (Ubuntu Xenial) "interuppting lxc-clone can destroy source container" [High,Confirmed]15:45
hallynjgrimm: i'm 99% sure that's fixed in xenial as stgraber has sru'd newer pkgs15:46
naccto trigger a rebuild of python-docutils (which should not ftbfs now that there is a new xml-core), do I just upload a new version -- or will the next test rebuild just automatically pick up the fix?15:46
jgrimmhallyn, that's what it sort of sounded like. maybe we mark xenial/yakkety as Fix Released (with comment of please open a new bug if <blah, blah, blah>15:47
hallynyup15:48
jgrimmtx15:48
hallynthx - ttyl15:48
=== victorp_ is now known as victorp
=== dpm is now known as dpm-afk
naccjgrimm: i think you saw my update on heimdal last night -- how did you want to proceed?16:35
jgrimmnacc, which part? that debian is investigating dropping? or the current upload's failures?16:38
naccjgrimm: yeah, that debian is dropping it -- and that the failures aren't resolved yet (and i'm not knowledgeable enough to know why ... yet). The debian patch leads to a different failure and i've updated hte upstream bug, but there hasn't been any response yet (afaict)16:39
jgrimmoh let me go look at the different failure16:40
nacchttps://github.com/heimdal/heimdal/issues/175]16:41
naccbah, without the ]16:41
naccjgrimm: so the debian patch does fix something, but the tests still fail for me16:41
naccright at the very end of the test16:41
naccjgrimm: ha! ok, so i did a quick re-run manually and it passed16:43
naccjgrimm: then ran `make check` manually and it passed as well16:43
naccjgrimm: let me see if i can reproduce that, i wonder if the tests are running in parallel again16:44
jgrimmnacc, ah .. k16:44
=== alexisb is now known as alexisb-afk
naccso i'm trying to help migrate heimdal along from proposed. I've got a build fix for it that fixes the current issue, but it still fails to sbuild due to a test failure. However, if i drop to the schroot after failure and run the test manually (eitehr `./debian/rules dh_override_auto_test` or `make check`), it passes. Any suggestions on how to debug that?17:43
naccjgrimm: i can upload the fixed version, but i'm not sure what do about the above --^17:45
jgrimmnacc, ack17:50
dobeyhmm, how can i get sbuild to run a script from the host system, inside the builddir18:14
naccdobey: i think setup-hook does that, but it says its deprecated :)18:15
dobeyand it just adds that to --chroot-setup-commands; which afaik are not executed within the builddir, and not sure it copies scripts into the chroot first18:17
dobeyand pwd for --pre-build-commands is the wrong place too18:18
naccdobey: yeah, reading https://wiki.debian.org/sbuild external commands18:18
naccthe implication is that it can run in the chroot18:18
naccbut not obvious which one18:19
naccand not sure it's copied automatically as opposed to you need to copy it in first?18:19
naccman sbuild implies at elast pre-build-commands and chroot-setup-commands are run in the chroot18:19
naccalthough18:20
nacc'EXTERNAL COMMANDS' contradicts that18:20
dobeypre-build-commands is run outside the chroot18:20
naccchroot-setup/cleanup- are run in the chroot18:20
naccdobey: man sbuild implies the wdir for the chroot-setup-command should be the BUILDDIR18:22
naccdobey: but taht doesn't really help with your script question, i wonder if you coud in pre-build-commands copy the script in and then in the chroot-setup-command execute the script?18:24
dobeywell pwd is nowhere near the chroot with pre-build-commands, so it doesn't help at all afaict18:25
naccdobey: you can use %b18:25
naccmabye?18:26
dobey?18:26
naccto get to the build-dir in the command definition18:26
naccsbuild has some special escapes18:26
dobeyhmm18:26
nacc  %b, %SBUILD_BUILD_DIR18:26
naccand %p, %SBUILD_PKGBUILD_DIR18:26
naccwhich can be used for these external commands and i guess sbuild substitutes them for you18:27
* nacc is thinking something like `sbuild --pre-build-commands='cp /path/to/script/ %b/...'18:28
naccdobey: %e might also work18:28
naccdobey: and might be the fastest way, although i don't see any example :)18:28
dobeywhat's difference beetwee BUILD and PKGBUILD though?18:29
dobeylatter is destinatio nfor binaries?18:29
nacce.g., in my sbuild of heimdal18:29
naccbuild/heimdal-yPecJ8 is BUILDDIR18:30
naccbuild/heimdal-yPecJ8/heimdal-1.7~git20160703+dfsg18:30
naccis PKGBUILDDIR18:30
dobeyhmm ok18:30
naccbased upon the log output18:30
nacc%e's description implies it's the best way to copy data in/out from the host to the chroot18:31
naccah i see18:31
naccwhat that does is something like schroot -c {id} --run-session -q -u root -p -- ....18:32
naccso it runs a script defined in the host system but in the chroot18:32
dobeyi don't see %e in the man page18:32
naccoh im on 16.10 and it's new :)18:33
naccdobey: sorry about that18:33
dobeyno worries18:36
naccdobey: so it seems like ther are some options for 16.04's sbuild ... 16.10's makes it a bit easier :)18:37
dobeyoh nice. %p doesn't exist yet in --pre-build-commands18:38
naccbah18:39
dobeybut that might be ok18:39
naccdobey: any cahnce you want to run 16.10 :)18:39
dobeyno18:39
dobeyi think this is actually running on a 14.04 host18:39
naccah18:41
=== alexisb-afk is now known as alexisb
doko_nacc: were there still some outstanding merges / ftbfs from your side?19:00
naccdoko_: i think the ones i've figured out are all uploaded19:08
naccdoko_: the openstack ones are all in -proposed per coreycb19:08
doko_ok, will be mostly offline until Oct 0419:08
naccdoko_: ok, i'll try and get the others figured out soon19:08
naccdoko_: http://qa.ubuntuwire.com/ftbfs/ only is for packages in -proposed, or is it the archive but using proposed?19:11
dobey19:14:50 W: The %SBUILD_CHROOT_DIR and %r percentage escapes are deprecated and should not be used anymore. Please use %SBUILD_CHROOT_EXEC or %e instead.19:16
doko_nacc: everything is built using -proposed. however the test rebuilds find stuff which gets broken by some changes in the build dependencies19:16
dobeyah, hrmm19:16
naccdobey: heh19:16
naccdoko_: got it19:17
naccdobey: seems weird if the manpage didn't mention %e ...19:17
dobeynacc: i can't run man on the machine where the sbuild actually runs19:18
dobeynacc: it has sbuild 0.71 though19:18
naccdobey: oh ok that's whats in 16.1019:19
naccdobey: and afaict only in 16.1019:20
naccyeah, they got rid of SBUILD_CHROOT_DIR when they added SBUILD_CHROOT_EXEC, i think19:20
dobeynacc: and whatever PPA it was installed from. this server is definitely not running 16.10 :)19:20
naccdobey: ah true, that could be it19:21
dobeywell, it's deprecated, but not gone yet19:21
naccdobey: err, right, sorry, meant deprecated (one of the options at the time was to just outright remove it)19:21
dobey'cat blub.txt | %SBUILD_CHROOT_EXEC sh -c "cat > blub.txt"'19:21
dobeyewww19:21
naccit doesn't seem lik eyou should need to do that though19:22
naccmabye you do, but that's not how i read the %e bits19:22
dobeythat's the example in the man page19:22
naccah19:24
dobeyand yeah, the description of %e suggests one must funnel data through STDIN/STDOUT of one process in the host and the proccess being run in the chroot19:24
dobeywhich is kind of nasty19:25
naccit seems like if you just need to run a script in the chroot, you could do 'cat script | %e sh -c -s' ?19:26
naccif you just needed to run a script19:26
dobeywell i need to run a script at start up. but then i need to find files and copy them back out after the build19:26
dobeyif i could just dumpt them in %SBUILD_BUILD_DIR and sbuild magically copies them back out, that'd work fine, but i'm not sure that's what it does19:27
naccyeah, not obvious to me either ... i guess you could reverse the above command to extract from the chroot, if you konw what they are in post-build-commands19:28
dobeyhmm, i'll try --finished-build-commands to see if dumping files in build dir just works19:29
dobeynope. sbuild doesn't just copy everything out. only the binaries/changes from debuild it seems :-/19:34
naccdobey: hrm :?19:53
nacc*:/19:53
tjaaltonanyone bumped into an issue with autopkgtest on lxd failing with "tar: ./control: Cannot change ownership to uid 12345, gid 12345, invalid argument"?21:06
dobeybah, dpkg-source doesn't get run until after --starting-build-commands :(21:08
tjaaltonah, needed to use sudo with autopkgtest21:17
naccdobey: you mean the xtract step?21:51
naccjgrimm: did you have a srcpkg handy that needed the vcs field change?22:20
* jgrimm looks. been a while since I filed that bug22:20
jgrimmnacc, it was nspr22:22
naccjgrimm: thanks22:22
jgrimmnp22:22
naccjgrimm: reproduced the python-cffi ftbfs on diamond, will debug there22:27
nacci'm guessing it's an arch assumption22:28
naccbut not sure yet what22:28
naccballoons: fyi, juju-core ftbfs on yakkety-proposed (http://qa.ubuntuwire.com/ftbfs/#ubuntu-server)23:04
naccstgraber: fyi, lxc also ftbfs there23:05
stgrabernacc: yeah, doko_ is looking at it. We'll just revert his change if we can't get this resolved soon.23:15
=== salem_ is now known as _salem

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