/srv/irclogs.ubuntu.com/2018/02/19/#ubuntu-desktop.txt

=== jamesh_ is now known as jamesh
oSoMoNgood morning desktoppers07:25
jibelgood morning oSoMoN07:28
oSoMoNsalut jibel07:29
seb128good morning desktopers08:08
alatieraAre the 18.04 packages in freeze yet? Cause the rustc is kinda outdated08:23
seb128alatiera, hey, no there is no freeze yet, it's just a complex package to update08:26
seb128alatiera, https://launchpad.net/ubuntu/+source/rustc/1.23.0+dfsg1+llvm-0ubuntu2 ... 1.23 seems the current version from a google query, what newest one would you expect?08:27
alatieraseb128:  great, I am writing docs for librsvg and wouldn't want to point newcomers to use rustup instead of the system package08:27
alatiera1.24 but it's like a week old, anything above 1.20 would be fine for my usecase08:28
seb128alatiera, it's blocked in bionic-proposed currently though08:28
alatieraI see, thanks alot!08:30
seb128it's not migrating because it didn't build on s390x08:30
seb128yw08:30
seb128which is weird, it's build-depending on itself it seems08:30
seb128well I guess chrisccoulson knows the situation better than me08:30
oSoMoNgood morning seb12808:32
seb128lut oSoMoN , ça va ? bon w.e?08:32
oSoMoNseb128, pas mal, et toi?08:38
seb128bien! 3j de w.e et une journée de repos au spa :)08:38
oSoMoNsympa08:40
oSoMoNça fait du bien j’imagine08:40
seb128oui :)08:40
Nafallosalut o/08:44
oSoMoNhey Nafallo08:45
Laneyhi there09:03
dufluHi Laney, oSoMoN, Nafallo, seb128, jibel, world09:05
oSoMoNhey Laney, duflu09:06
Laneyhey duflu & oSoMoN09:06
jibelHi duflu09:06
Laneymoin jibel09:06
jibeland all09:06
Laneygood weekends?09:06
jibelwet09:06
jibelbut quiet and restful09:07
seb128hey everyone09:08
seb128Laney, good! spa on friday; nice weather (cold&sunny) on saturday/sunday so quite some walking around ... you?09:09
Laneyhey seb12809:11
* Laney is jealous of this spa experience09:11
Laneywe went walking in the peak district09:11
Laneymostly inside a cloud09:12
dokooSoMoN: https://bugs.launchpad.net/ubuntu/+source/libepubgen/+bug/174992009:12
ubot5Ubuntu bug 1749920 in libepubgen (Ubuntu) "[MIR] libepubgen" [Undecided,Incomplete]09:12
Laneyalthough it did lift a bit once we got to the top on the first day09:13
Laneyhttps://photos.google.com/share/AF1QipN7hT2C3HZMSSrXvCgfne8FNDnkXDzDLk5bMY5vFCpADMNMQ0pn5F7OavjmToJ1MQ/photo/AF1QipNyJT6IMAZfwLy21ADIxk_HCYaQstzwf8lGtIPH?key=V20tV0NOdXY2ZGwyaG1JZk5qd19JNHJKNl9uTU9R09:13
seb128doko, "no bug subscriber" as only feedback, does it mean it's good to be approved outside of that?09:14
dokoseb128: no, looking there's more09:15
seb128k09:15
dokobut I don't understand why even that one is missing09:16
oSoMoNI suppose it's fine if we subscribe ~libreoffice09:20
oSoMoNseb128, ^ ?09:20
seb128oSoMoN, wfm, dunno if that's a team doko is happy with though09:20
seb128speaking of MIR09:23
seb128doko, do you know what's the status of the libblockdev one? somebody posted a comment about a security review, is that an official MIR team bounce? should it be assigned to security team in that case? or is that a random non-MIR-member comment?.09:24
dokoseb128, oSoMoN: is it the same as lo itself? e.g. https://launchpad.net/ubuntu/+source/libixion has both lo and desktop09:24
oSoMoNnot sure about ~desktop-bugs, that's a lot of people potentially not interested in libreoffice and its deps09:31
oSoMoNseb128, wdyt?09:31
seb128oSoMoN, if ~libreoffice is an acceptable team then +1 for that one, otherwise we can do desktop, it's just to define assignment09:33
oSoMoNseb128, doko: I've subscribed ~libreoffice09:33
seb128thx09:36
seb128doko, do we need desktop as well?09:37
oSoMoNdoko, is the lack of a symbols file a blocker?09:42
chrisccoulsonalatiera, how come you wouldn't want to point people to use rustup? The rustc package in the archive is really only there for building distro packages. People hacking on rust code would probably be better off using the rust community's tools, like rustup09:48
alatierachrisccoulson: indeed, I use it myself and prefer it but it's an extra tool newcomers would have to manager and can go wrong09:49
seb128pitti, hey. is python-dbusmock known to have issue with nm 1.10? or was https://github.com/martinpitt/python-dbusmock/issues/37 the first mention of it (it comes from autopkgtest errors since the update)09:53
dokooSoMoN: yes, I think we required those in the past. it shouldn't be that difficult09:59
dokoseb128: could you point me to the list of subscribers which the desktop team is monitoring for bug reports?10:00
oSoMoNdoko, ack. what's the preferred course of action here? submit improvements to debian and sync back, or do it in ubuntu first and submit changes to debian from there?10:03
oSoMoNthe latter might be faster10:03
seb128doko, I'm not sure to understand the question, you want to know the team used for reports? it's desktop-packages10:03
dokoI'm just saying that you should be sure which teams you use to monitor bug reports.10:05
dokoseb128: "somebody posted a comment about a security review, is that an official MIR team bounce?" well I asked for a security review before. now assigned to the security team10:11
seb128doko, thanks, that was not obvious from the current comment/status10:11
seb128doko, right, desktop team uses ~desktop-packages, I don't  know how ~libreoffice is being used since I'm not working on that package10:12
dokowell, but the whole point is that the team knows about these bugs10:14
dokoseb128, jbicha: the s390 plugin should be packaged according to xnox, and symbols files should be added10:16
dokogetting late lunch now ...10:16
seb128doko, enjoy10:16
oSoMoNenjoy10:17
oSoMoNat that time of the day that would be a late breakfast for me :)10:17
dokooSoMoN: I forgot, the lo uitest autopkg test fails10:22
oSoMoNdoko, I know, I filed https://launchpad.net/bugs/175033510:22
ubot5Ubuntu bug 1750335 in libreoffice (Ubuntu) "LO 6.0 autopkgtest failures (uicheck: test_text_direction)" [High,Confirmed]10:22
pittiseb128: it's the first time I hear about it; I just followed up to the issue; should be easy enough to fix10:23
seb128pitti, thx10:24
seb128pitti, just curious, on what n-m version are you? (or is fedora, since I guess that's what you use) is that issue somewhat specific to Ubuntu?10:25
pittiseb128: I'm on Fedora 27, NM is  at 1.8.610:25
seb128oh ok10:25
pittithose were the days when I've always used the current devel series :)10:25
seb128no issue with that version then :)10:25
seb128hehe10:25
pittirawhide isn't nearly as stable as ubuntu devel, so I'm using the current stable10:26
pittihttps://apps.fedoraproject.org/packages/NetworkManager → rawhide indeed has 1.1010:27
pittibut nobody complained so far10:27
pittihah, it failed to build10:28
seb128pitti, https://koji.fedoraproject.org/koji/buildinfo?buildID=1009625 worked though10:31
dufluseb128, I don't know yet. Will need to try bisecting bluez tomorrow. It also depends if upstream declares it a bug or a feature10:40
seb128duflu, what changed exactly?10:40
seb128or what is behaving differently?10:40
dufluseb128, they rewrote the "readline" line editing stuff used in bluetootctl and other interactive tools10:41
dufluSuch a stupid problem to have. It has nothing to do with bluetooth really10:42
seb128right, well if it's a bug in their command line utilies which might create issues in scripts&such it's still a bug worth fixing10:45
seb128duflu, if we believe it's non important we can also skip that test result for now to get the new version in, would still be good to at least report upstream10:46
dufluseb128, yes already reported upstream. AFAICT python-dbusmock tests may be the only thing impacted10:47
dufluAnd that's another day I didn't get time to work on mutter10:48
dufluNight all10:49
seb128night duflu10:49
Laneyseb128: kenvandine[m][m] (dunno if you'll see that): https://code.launchpad.net/~vorlon/livecd-rootfs/ubuntu-channels-for-snaps/+merge/337897 <- will need to publish gnome-calculator to a new channel once this is uploaded11:19
seb128Laney, thanks for the head up11:39
Laneynp11:44
jbichaxnox: doko: enabling the s390 plugin causes libblockdev to FTBFS on s390x, that's why it's disabled11:56
jbichagood morning11:56
gsilvapthello all12:01
gsilvaptjibel, you around=12:01
gsilvapts/=/?12:01
kenvandine[m][m]Laney: yeah, we talked about it in the meeting on Friday.  I've published gnome-calculator to stable/ubuntu-18.04 and then closed the branch12:02
kenvandine[m][m]Per his request12:02
Laneykenvandine[m][m]: ok cool12:03
seb128hey jbicha12:30
jbichahey12:30
seb128jbicha, are you looking at the autopkgtest issues from your uploads? like the n-m urfk one?12:30
seb128and the glib d-t-r one?12:30
jbichawhat's urfk?12:31
jbichabasically no12:32
jbichaNM's autopkgtest issue might be relatively easy based on the comments from LP: #173458612:33
ubot5Launchpad bug 1734586 in network-manager (Ubuntu) "Merge NetworkManager with Debian 1.10.0-1" [Wishlist,Fix committed] https://launchpad.net/bugs/173458612:33
seb128jbicha, https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/1734586/comments/6 the first part12:33
seb128right, which is why I was asking12:33
seb128the python-dbusmock part is being handled12:33
jbichaI've never worked on dbus-test-runner before so it would be cool if someone else could work on that12:34
seb128L_aney said he could have a look12:34
jbichagreat12:34
seb128:)12:34
seb128jbicha, also what's the status of the gnome-desktop transition?12:34
LaneyYou don't need to have worked on dbus-test-runner12:34
seb128it's not clear to me from a quick look what is missing/blocking12:34
jbichaat least with the glib one, there was no easy way to know that a rdepends's autopkgtest would fail until we uploaded to Ubuntu and found out :(12:34
LaneyOnce you start looking at the failure it's clear what it is12:34
jbichait's clear to Laney but not to the rest of us ;)12:35
LaneyNo, it'd be clear to you too12:36
Laneynew glib:12:36
Laneyroot@autopkgtest-lxd-iiuugp:/tmp/autopkgtest.z8x60r/build.4XM/src/tests# dbus-run-session gdbus emit --session --object-path /test/dbustestrunner/signal --signal com.launchpad.dbustestrunner.signal12:36
LaneyError: Destination is not specified12:36
Laneyroot@autopkgtest-lxd-iiuugp:/tmp/autopkgtest.z8x60r/build.4XM/src/tests#12:36
jbichaI mean I did spend a couple minutes looking at it but I didn't understand what I was looking at12:36
Laneyold glib:12:36
Laneylaney@nightingale> gdbus emit --session --object-path /test/dbustestrunner/signal --signal com.launchpad.dbustestrunner.signal12:36
Laneylaney@nightingale>                                                                                             ~12:36
Laneyall I did was reproduce the failure and then look into what it was doing12:38
seb128jbicha, http://packaging.ubuntu.com/html/auto-pkg-test.html has details on how to try things locally12:38
Laneyrunning smaller bits of it until I got to the issue12:38
Laneywhich is this regression in the gdbus tool12:38
Laneyanyways, nearly there now12:39
jbichaseb128: if you're talking about the rdepends issue, it's not obvious (to me) what rdepends autopkgtests will be triggered12:40
seb128jbicha, no, I meant to reproduce a test failing and be able to look at the problem/debug12:41
seb128or did you mean the gnome-desktop question? ;)12:41
jbichaanyway, once glib is fixed, there's only one issue holding up the gnome-desktop3 transition.12:42
seb128are you sure?12:43
seb128which one?12:43
jbichaI'm rebuilding https://bileto.ubuntu.com/#/ticket/3138 (Unity) and if that works, I'm publishing it12:44
seb128I see at least gnome-control-center wanting network-manager to be cleared (not sure why)12:44
xnoxjbicha, well typo, not in v2.14, but fixed in v2.13 so even earlier.12:45
seb128"Depends: gnome-control-center network-manager (not considered) "12:45
jbichaseb128: oh :(12:45
jbichaxnox: ?12:46
seb128xnox, hey, Chris mentioned that you might be interested in fixing cargo/rust on s390x rather than seing it removed on those arches, is that true?12:46
xnoxseb128, i've asked chrisccoulson to fix cargo/rust on s390x; and there it seems strange that he saw bootstrap bugs on s390x on ubuntu, when it is fully bootstrapped to latest releases in Fedora and Suse.12:47
jbichaxnox: ok, I'll try re-enabling the s390 plugin for libblockdev, thanks12:47
xnoxseb128, as far as I can tell, foundations does not maintain cargo/rust by mutual agreement.12:47
seb128xnox, nobody maintains those :)12:47
xnoxseb128, why remove it, if it simply lacks a bootstrap?12:48
seb128xnox, is it?12:48
xnoxseb128, as far as i can tell, it requires no fixing, just bootstrapping -> and like not deleting binaries and making it impossible to bootstrap again.12:48
xnoxseb128, and by using PPAs, instead of the archive, we got ourselfs into the situation of expiring binaries.12:48
xnoxBTW debian raised ISA instruction set to z196 and they will have it bootstrapped too, on s390x, soon.12:49
seb128xnox, chris said that cargo doesn't work properly on s390x and that rust will most probably not build12:50
jbichaoops, Unity still doesn't build12:51
seb128xnox, seems you guys disagree on what is needed then12:51
xnoxrust built up to 1.23 at http://download.sinenomine.net/epel/epel-7/s390x/ and cargo up to 0.24.012:52
xnoxditto in fedora12:52
xnoxhttps://koji.fedoraproject.org/koji/buildinfo?buildID=104542712:52
xnoxhttps://koji.fedoraproject.org/koji/buildinfo?buildID=102584112:52
xnoxsorry second is bad link, this one is the right one https://koji.fedoraproject.org/koji/buildinfo?buildID=102689712:52
seb128ok, good12:53
seb128can you help bootstrapping then?12:53
xnoxall built on s390x.... if it works there, why does it not work for us? well, hard to test, since bootstrapping got dropped.12:53
chrisccoulsonrust stopped building because of what looked like a cargo build, and unless something has changed very recently, our cargo build looks like it's broken on s390 (have a look at the test failures)12:54
xnoxseb128, i am full for the current two week iteration, ping gaughen to schedule things into foundations trello?12:54
chrisccoulsonerr, cargo bug12:54
xnoxchrisccoulson, well, i cannot rebuild cargo at all anymore, since the binaries in the PPAs have expired.12:54
xnoxchrisccoulson, do you have a ppa per-release of rust/cargo, such that you have stashed builds of the last good cargo/rust 21? and you started to have failures with 22 and up?12:55
xnoxchrisccoulson, since each subsequent one, requires previous one to build, you should always make a new PPA to stack each series. Or package each one as a new source/binary package names, such that binaries are not deleted...12:55
xnoxfrom fedora build log, there are some test failures - but they are localhost networking and "assume building from git repository" assertion failures. I don't see any serious / arch specific failures. Most tests are passing.12:58
seb128gnome-session is also an issue for the gnome-desktop transition12:59
seb128gnome-session-bin/amd64 unsatisfiable Depends: libegl112:59
xnoxchrisccoulson, note creating PPAs with all architectures enabled, is completely self-service tick the boxes in settings now. As in, s390x is no longer restricted.13:00
xnoxand then just need to setup build-dependencies between PPAs for each subsequent one to depend on the previous one.13:00
jbichaseb128: oh right, see https://irclogs.ubuntu.com/2018/02/16/%23ubuntu-desktop.html#t12:2913:02
seb128jbicha, there are no easy transitions in Ubuntu are they? ;)13:03
jbichalibgweather was ok, the gnome-desktop3 one was a lot more complicated than I hoped13:04
jbichafor instance, Unity built fine in my PPA but it failed to build with glib which was updated in the mean time13:04
jbichathat mesa transition was surprising too :)13:05
seb128the autopkgtest issues come as a surprise on top then :)13:06
jbichaand LibreOffice's autopktests are entangled with it too!13:07
seb128right, looks like oSoMoN is looking at those though13:07
seb128so yeah, that transition is probably to take the week13:08
oSoMoNI am13:08
seb128jbicha, can you look at doing the tweaks Brian suggested to the n-m autopkgtest?13:08
jbichaseb128: it may take me a few days for NM so anyone else is welcome to work on it earlier13:11
xnoxchrisccoulson, do you have the LP bug # for the cargo issue? location of source packages / build records, and the last built s390x binaries?13:11
seb128jbicha, k, thx13:11
jbichabtw, I'm waiting for gnome-desktop3 to complete before starting https://people.canonical.com/~ubuntu-archive/transitions/html/evolution-data-server.html13:11
jbicha(but I may start it anyway next week because of Feature Freeze)13:12
seb128k13:13
seb128did you identify risky updates13:13
seb128we should build a list before the meeting so we know what we are going to discuss and can prepare13:13
jbichaseb128: do you want me to open a Community hub topic for it?13:15
seb128jbicha, that would be nice, thx13:15
Trevinhojbicha: I think the last mesa has some problems16:03
Trevinhowe get this crash at the end of a test16:03
Trevinhohttps://www.irccloud.com/pastebin/ieALagQN/16:03
jbichatjaalton: ^16:03
jbichatalking about the Unity FTBFS, right?16:04
Trevinhoyeah...16:17
Trevinhoor actually it's just nux doing things wrong16:17
Trevinhotjaalton: well, actually this is happening when doing a glXDestroyContext for a context that has never been free'd17:00
Laneybiab17:28
Trevinhotjaalton: could be related https://bugzilla.redhat.com/show_bug.cgi?id=1526848 ?17:51
ubot5Error: Could not parse XML returned by bugzilla.redhat.com: HTTP Error 403: Forbidden (https://bugzilla.redhat.com/show_bug.cgi?id=1526848&ctype=xml)17:51
popeymorning robert_ancell :)19:28
robert_ancellpopey, hi19:28
robert_ancellkenvandine[m][m], who owns the hangout? I'm sitting on "requesting to join"19:30
popeyrobert_ancell: joining the meeting :)19:32
seb128robert_ancell, he might be off, today is an U.S holiday (president's day)19:44
xnoxhas something changed in colord packaging?20:32
xnoxi'm getting this in systemd autopkgtests20:32
xnoxFeb 19 17:08:36 autopkgtest systemd-udevd[385]: Error resolving group 'colord': Connection timed out20:32
xnoxand that's relatively new.20:32
xnoxIs something supposed to create colord group? and/or did it get dropped, but still referenced in udev rules?20:33
robert_ancellflexiondotorg, do you think you might have something like https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/1648534 ?20:35
ubot5Ubuntu bug 1648534 in gnome-software (Ubuntu Xenial) "gnome-software crashed with SIGTRAP in g_wakeup_new from g_main_context_new from g_dbus_connection_send_message_with_reply_sync from g_dbus_connection_call_sync_internal from g_dbus_connection_call_sync" [Critical,Confirmed]20:35
robert_ancellParticularly comment 320:35
robert_ancellor perhaps https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/165985820:36
ubot5Ubuntu bug 1659858 in gnome-software (Ubuntu) "gnome-software (5) g_wakeup_new → g_main_context_new → g_dbus_connection_send_message_with_reply_sync → g_dbus_connection_call_sync_internal → g_dbus_connection_call_with_unix_fd_list_sync" [High,Confirmed]20:36
flexiondotorgrobert_ancell: I have a .crash in /var/crash but apport didn't want to upload because libssl was out of date.20:41
flexiondotorgI'm running 18.04 daily.20:41
robert_ancellflexiondotorg, if you could try checking the file descriptors if you think it's happening again that would be helpful20:42
flexiondotorgSure thing.20:43
flexiondotorgI'll try and provoke it tomorrow.20:43
ochosihey guys20:44
ochosiquick general question (from someone who has never submitted a new package before): is it too late to submit a new package (that is not in debian) to ubuntu?20:44
ochosiit's a fairly tiny a11y helper (just a few lines of C) that draws expanding circles where the mouse pointer is20:45
ochosii'd like to get that included in xubuntu by default, because we've had that request a few times and our WM doesn't support it (and the maintainer of xfwm4 doesn't really want to have this kind of functionality in the window manager)20:45
jbichaochosi: it's not too late. There are no guarantees it will make bionic, but for best results, get it in the new queue before Feature Freeze20:46
ochosiyeah, ofc20:47
ochosithe alternative is to push it to an existing xubuntu-maintained package20:47
ochosithere are probably some packages that would be okayish20:47
ochosithen again, maybe others want to use this too (lubuntu?)20:47
robert_ancellkenvandine[m][m], I forgot to ask, what was the conclusion on the gnome-software version in bionic? Should I take the work jbicha did in Debian and merge it into Ubuntu?20:56
kenvandine[m][m]robert_ancell: we're going to go with 3.28 for anything that doesn't require a transition21:09
kenvandine[m][m]So gnome-software should be safe21:10
robert_ancellkenvandine[m][m], ta21:12

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