/srv/irclogs.ubuntu.com/2016/04/22/#ubuntu-devel.txt

cpaelzergood morning05:36
pittiGood morning05:44
=== smb` is now known as smb
=== bigon_ is now known as bigon
=== athairus is now known as afkthairus
rayzinnz'ello08:59
hikikorbasak ping :)09:24
hikikonew questions on the sru09:25
hikikothat's what I did so far: https://ufuntu.wordpress.com/2016/04/21/sru-process/ password:trusty used the tools you suggested, quilt and dch and I think (hope) I'm ready to create the new package09:26
hikiko and go to step 5 here: https://wiki.ubuntu.com/StableReleaseUpdates#Procedure09:27
hikikothe question is09:27
hikikowhich tools I need to create a package?09:27
hikikoand how should I name it?09:28
rbasakhikiko: that looks good so far.09:28
seb128hikiko, did you check with Trevinho? compiz is managed by CI landing09:28
seb128not normal uploads09:28
rbasakseb128: including for an SRU?09:29
seb128rbasak, I was looking at the most recent one, https://launchpad.net/ubuntu/+source/compiz/1:0.9.11.3+14.04.20150313-0ubuntu109:29
seb128the CI bot is the uploader09:29
seb128so I guess so09:29
rbasakWell that makes life quite a bit harder for hikiko :-(09:30
hikikono worries :) I want to learn the full process09:30
hikikoseb128, I didn't I'm now learning the SRU process09:30
rbasakI don't know the CI process.09:31
rbasakseb128: can you help hikiko?09:31
hikikoalso, could you tell me the non-CI process?09:31
seb128hikiko, can you talk to Trevinho?09:31
seb128or maybe do that with him next week?09:31
rbasakOr would it be sufficient for hikiko to put a debdiff in the bug and let a sponsor sort it out?09:31
hikikobtw seb128 I've seen in the changelog that chris townsend had sru some changes09:31
seb128rbasak, hikiko, well for things going through CI landing you just need a merge request with the change09:31
rbasak"just"09:32
hikikoit's not only the CI bot I think09:32
pittiwe do SRUs via the CI train too?09:32
seb128pitti, of course09:32
seb128why not?09:32
hikikoTrevinho, is in Prague :D09:32
seb128well, wait monday09:33
seb128?09:33
hikikoalright :)09:33
pittiI just don't think I've ever seen one09:33
seb128you are coming next week right?09:33
hikikoyes09:33
pittipresumably because most touchy stuff lands in the overlay PPA09:33
seb128pitti, I just gave an url09:33
seb128pitti, all unity landings are done this way09:33
seb128unity709:33
hikikook but just out of curiosity09:35
LaneyThe only difference for the SRU team person is that it's a copy (with binaries) in the queue instead of a source upload09:35
hikikoif I had to SRU it myself09:35
hikikowhat would be the next step?09:35
hikikohow would I create a package and how would I name it?09:36
rbasakhikiko: if not following the CI process, you'd adjust the version number, run "dch -r -D trusty ''" to set the metadata in the debian/changelog file.09:36
rbasakThen run "dpkg-buildpackage -us -uc -S -nc -d" to create a new .dsc and .diff.gz file for the new version.09:36
rbasakAn uploader would then be able to use the generated .changes file to upload.09:37
rbasakIf you need sponsorship, you run "debdiff <old>.dsc <new>.dsc" to create a debdiff, attach that to the bug, and subscribe ~ubuntu-sponsors to the bug.09:37
rbasakI've ignored testing here. You can use sbuild to build a test package locally.09:37
rbasakOr upload to a PPA for testing.09:38
rbasakSo you'd done most of the hard part already for the usual process I think.09:38
hikikoand then everything else is done on launchpad I guess09:39
hikikoupload the package etc09:39
hikikorbasak, thanks a lot!09:41
hikikoI'm going to note these steps just in case and wait for monday to ask Trevinho09:41
rbasak12:29 <rbasak> "Yakkety Yak"09:44
rbasakThursday, April 21st, 2016 "Yakkety yakkety yakkety yakkety yakkety yakkety yakkety yakkety yak. Naturally"09:45
* rbasak isn't sure if that was a coincidence, the obvious choice, or what.09:45
Odd_BlokeALL HAIL RBASAK09:59
Odd_Blokecjwatson: wgrant: FYI, there are 16.04 powerpc images in http://cloud-images.ubuntu.com/daily/streams/v1/com.ubuntu.cloud:daily:download.json now.09:59
wgrantOdd_Bloke: <310:00
Odd_Blokecjwatson: wgrant: I don't believe that infinity has had time to test them since we were building from a PPA, but they shouldn't have changed. since he last told me they were booting.10:00
wgrantBooting is good.10:00
Odd_Blokewgrant: Do you pull from daily streams, or release streams?10:00
Odd_Bloke(Because they aren't currently showing up in release streams, and I'm trying to work out highly to prioritise fixing that :p)10:01
wgrantOdd_Bloke: Do I look like a stable OS sort of person to you?10:01
wgrantWe've been tracking wily dailies for a while.10:01
wgrantNo reason not to use xenial dailies now :)10:01
Odd_BlokeOK, cool.10:01
Odd_BlokeI'll throw fixing that on our backlog then.10:01
wgrantThanks!10:02
caribouLaney: would installing all the reverse deps of a backported package considered adequate testing for a backport request ?10:09
caribouLaney: I'm testing a script that installs each of the rdepends in a container where the binaries to be backported are installed & checking if it installs correctly10:11
cariboucurious to know if this would be considered sufficient10:11
pittirbasak: btw, would you mind having a look at my response in bug 1569434 ? If this is real, it'd drive me nuts :)10:14
ubottubug 1569434 in autopkgtest (Ubuntu) "adt-run fails with "dpkg-genchanges: error: cannot read ../akonadi_15.12.1-0ubuntu4.dsc: No such file or directory"" [Undecided,Incomplete] https://launchpad.net/bugs/156943410:14
Laneycaribou: Just /installing/ doesn't sound enough to make sure that they aren't busted10:25
caribouLaney: apache2 has more than 100 rdepends, you would test all of them manually ?10:30
rbasakpitti: I can try to reproduce again.10:30
Laneycaribou: In this case, does installing enable them?10:33
Laneyand restart apache to get them used?10:33
caribouLaney: I'll check the logs once all of them run but most of them do, yes10:33
caribouLaney: I think tests must be specific to each backport context though10:35
caribouLaney: but in that specific case, that's how I ran over a know issue & fixed it : LP: #128688210:37
ubottuLaunchpad bug 1286882 in mpm-itk (Ubuntu Trusty) "libapache2-mpm-itk postinst failed" [Medium,In progress] https://launchpad.net/bugs/128688210:37
* caribou is just trying to get a better understanding of the backport requirements10:38
=== hikiko is now known as hikiko|ln
Laneypitti: do you need to teach autopkgtest about yakkety?11:33
=== hikiko|ln is now known as hikiko
mgedminwelp, update-manager (15.10 -> 16.04) froze after unpacking gir1.2-atspi-2.0 and "processing triggers for man-db"11:36
mgedminsays it's preparing to configure python-cairo, but it's idle11:36
mgedminpstree shows it has one zombie child (dpkg)11:36
mgedminstrace shows it's bloked trying to write "pmstatus:libde265-0:16.8172:Inst..." to a pipe on fd 1411:37
mgedminah, it's an internal pipe used to communicate between threads?  lsof shows that it's the upgrader itself that has the other end open for reading11:40
mgedminthe GUI window is not responding to events either11:41
mgedmin... I might've hit Ctrl+Shift+I to open the GTK+ inspector11:42
mgedminwhile trying to see if it uses GTK2 or GTK311:42
pittiLaney: yes, in the sense that I need to update distro-info-data12:07
pittiLaney: also, retracers etc.12:07
Son_Gokuah, the joys of using a fresh linux distro release :/12:08
seb128dpm, pitti,  I've the feeling we have been there before but any idea why evolution-3.18.mo are missing from langpacks? it makes evolution untranslated in xenial :-/ ( (bug #1566750)13:15
ubottubug 1566750 in evolution (Ubuntu) "16.04: translations completly missing" [High,In progress] https://launchpad.net/bugs/156675013:15
seb128mdeslaur, do you/security team want to be notified about regressions due to the samba trusty security update? and in which way (bug assign? subscribe?) bug #1571854 is another of those13:17
ubottubug 1571854 in evolution-mapi (Ubuntu) "Recent samba updates broke mapi" [High,Confirmed] https://launchpad.net/bugs/157185413:17
mdeslaurseb128: you can tell me about them13:19
mdeslaurseb128: that particular one is problematic though as I don't believe there's an openchange package that actually works with samba 4.313:20
seb128indeed not13:22
seb128we should at least delete -mapi13:22
mdeslaurseb128: thanks for pointing them out, I'll see if there's any other way to fix it when I get a minute13:22
seb128or turn it into a "sorry, being removed"13:22
seb128thanks13:22
TJ-any ideas why 16.04 "apt-get install --no-install-recommends" breaks Unity desktop, in that there is no 'System' menu-icon, and Dash is unable to find any applications via navigation or via a typed search. Also, All Settings > Universal Access > typing > on-screen keyboard doesn't actually provide the on-screen keyboard, and there's no nm-applet shown13:27
seb128TJ-, that command shouldn't do anything? or do you mean starting from a base system and installing the desktop without recommends ... that might lead to some component missing yes, don't do that13:43
TJ-seb128: Yes. I did a debootstrap install for a 4GiB Flash drive, used that to get the basic desktop, and discovered many issues with missing basic functionality13:49
seb128don't do that13:50
seb128download the iso and dd it to the stick13:50
TJ-seb128: can't; I need an installation on the USB13:51
TJ-my point is "--no-install-recommends" shouldn't leave out vital parts of the interface13:52
dobeythey aren't "vital"13:53
TJ-dobey: I disagree. there being no system icon, no nm-applet, no access to applications via dash are pretty serious deficiencies13:54
TJ-I discovered this whilst trying to workaround the broken bluetooth GUI PIN-entry agent which prevents pairing with external keyboard/trackpad... have to do it via CLI with bluetoothctl13:56
seb128TJ-, you put them as depends and users with no bluetooth complain that they can't remove the bluetooth components13:57
seb128can't win13:57
TJ-seb128: I'm not on about BT with regard to recommends - that is a separate issue13:57
rbasakI think it's reasonable to QA only the default (including recommends case).13:58
TJ-seb128: even with all the BT GUI components installed on a *full* install from ISO, gui bluetooth PIN agent is broken, never works,13:58
rbasakIf there are specific recommends that should be depends, then...patches welcome?13:58
seb128+113:58
TJ-First time I've used Unity in about 6 years, surprised its not more mature13:59
seb128?13:59
rbasakOtherwise you're basically requiring the doubling of QA effort for what?13:59
seb128TJ-, you call it not mature because you did some weirdy install, ignored recommends and then are complaining it's incomplete? or is that an orthogonal statement?14:00
rbasakYou're welcome to hack your system but when you break it you get to keep the pieces.14:00
* dobey confused14:00
rbasakIf you want mature, don't do weird things.14:00
TJ-this isn't hacking this is a regression14:00
dobeywhat is a regression?14:01
TJ-Bluetooth *used* to work fine; was broken several releases ago  bug 149034714:01
ubottubug 1490347 in bluez (Ubuntu Wily) "[Regresision] 15:10 - Cannot pair with devices using PIN codes" [High,Triaged] https://launchpad.net/bugs/149034714:01
dobeythe bt agent has nothing to do with unity14:01
seb128TJ-, that has nothing to do with unity14:01
dobeyand i can pair with devices that use PINs just fine14:01
seb128TJ-, you seem confused?14:01
TJ-dobey: bet you can't!14:01
TJ-dobey: bet you *can* pair with devices that use *passphrase* ... because bluez ripped out the *pin* support14:02
dobeyTJ-: except for the many times i've done it in the many years i've been using ubuntu14:02
dobeyyou must be confused14:02
seb128TJ-, I doubt bluez removed pin support14:02
dobeyif bluez removed pin support you wouldn't be able to use bluetoothctl either14:03
TJ-I was deep in the source-code last year trying to fix it and the related kernel bugs breaking other devices, but had to give up because it was too involved and I had other things to do. the kernel bug is bug 149198814:04
ubottubug 1491988 in linux (Ubuntu) "15.10: BT Command List not created: bluetooth:__hci_req_sync:298: hci1 end: err -ETIMEDOUT" [High,In progress] https://launchpad.net/bugs/149198814:04
dobeywell i just paired a headset in xenial with no problem14:05
seb128TJ-, k, so there is a kernel bug, it's different from being bluez ripping out support14:06
dobeyhad no problem pairing my obd-ii scanner in 14.04, so i expect no problem in 16.04 either14:06
TJ-dobey: yes, headsets and so on seem to work but anything that has to enter the PIN in response to the PC fails. E.g. in the GUI for BT, when you discover a device, there's an option to select the PIN method (0000, 1111, 1234, custom) ... none of those will work with the keyboard even if typing them blind (because there's no GUI prompt for the number as there used to be). Using the bluetoothctl it14:11
TJ-generates a 6-digit PIN. The thing is, although we call it PIN this is actually using the underlying BT 'passphrase' authentication.14:11
TJ-I'll dig my notes out on the debugging I did last year on this, I think I also logged some additional bug reports but will have to track them down14:12
seb128TJ-, anyway on the unity issue, I guess it would be fine to move the files/applications lens from Recommends to Depends, feel free to propose a patch for that or to open a bug14:12
TJ-seb128: is that why I had no applications listed? so I can manually install that to get over the current impass?14:13
seb128TJ-, I guess so14:13
seb128unity-lens-applications14:13
dobeyand indicator-session14:14
dobeyshould probably be Depends of ubuntu-desktop14:15
seb128and onboard and14:15
TJ-seb128: I shall do.. currrently fighting to get everything working on a 2-in-1 Asus T300chi (touchscreen tablet/laptop hybrid with docking BT keyboard. Not being able to have the BT come up at greeter-time is currently a major issue so using a USB connected keyboard, or ssh in (now I've created a wifi profile in nmtui-edit) :)14:15
seb128network-manager-gnome14:15
dobeyi don't think onboard should be depends14:15
TJ-seb128: dobey thanks, I'll install those (already did n-m-g and ubuntu-keyboard - for the on-screen keyboard backup)14:15
dobeyi'm not sure ubuntu-keyboard is the one you want for that14:16
TJ-Full install has all those14:16
TJ-dobey: hmmm, really? it looked the most likely candidate!14:16
seb128TJ-, ubuntu-keyboard is the one used on ubuntu touch, onboard is the desktop one14:16
dobeyTJ-: i think onboard is the one used for ubuntu-desktop14:16
dobeywell, you could just install unity8-desktop-session-mir instead, and use unity8 instead of unity7, then ubuntu-keyboard would make sense perhaps :)14:17
TJ-dobey: ahhh! OK, I think you're correct since enabling on-screen keyboard still didn't show anything!14:17
TJ-dobey: oh, is that what it's for?14:18
dobeyi wonder if indicator-network is a usable replacement for nm-gnome now14:18
dobeyTJ-: yeah, ubuntu-keyboard is the phone/tablet (unity8) osk14:18
dobeyalso, ubuntu-keyboard is in universe14:18
dobeyno way it would be installed on the ubuntu-desktop iso :)14:19
=== cjwatson changed the topic of #ubuntu-devel to: Xenial (16.04) Released! | Archive: closed | Devel of Ubuntu (not support or app devel) | build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of precise-xenial | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots:
TJ-ouch! without --no-install-recommends onboard wants to install tonnes of packages, including a mass of Perl!14:19
dobeylol14:19
TJ-right.. I needed 2 of the onboard recommends: gir1.2* ... now the greeter shows the keyboard, athough as a shrunken image in a sub-area of the bottom of the screen14:24
seb128TJ-, you might want onboard-data14:26
TJ-yes, got that earlier, thanks14:27
TJ-this is what the greeter shows: https://iam.tj/projects/misc/ubuntu-1604-greeter-onboard-keyboard.jpg14:27
TJ-once the user is logged-in the keyboard fills the entire area14:28
seb128unsure why it does that14:31
TJ-and this is after log-in, showing the empty dash 'applications' tab, and the lack of system icon top-right, plus onscreen keyboard14:32
TJ-https://iam.tj/projects/misc/ubuntu-1604-user-onboard-keyboard.jpg14:32
de-factoHey guys im fighting to compile http://packages.ubuntu.com/de/xenial/easystroke via "apt-get source easystroke" then "apt build-dep easystroke" then inside the source dir "dpkg-buildpackage" it gives me this error that i also found in https://bugzilla.redhat.com/show_bug.cgi?id=130291714:33
ubottubugzilla.redhat.com bug 1302917 in easystroke "easystroke FTBFS in rawhide" [Unspecified,Closed: rawhide]14:33
TJ-ahh, and noticed that there's no icon to control the onscreen keyboard14:34
de-factoDoes someone have an idea how the ubuntu repo managed to compile this?14:34
cjwatsonIt probably built at some point in the past, just not with the current state14:34
cjwatsonIt failed in our latest build test14:35
cjwatsonBut that doesn't mean we throw away the existing binaries14:35
cjwatsonhttps://launchpad.net/ubuntu/+source/easystroke indicates that it was built last August14:35
cjwatson(https://launchpad.net/ubuntu/+source/easystroke/0.6.0-0ubuntu5/+build/7771352)14:36
seb128TJ-, you probably lack indicator-application as well14:37
seb128for the onboard indicator14:37
de-factocjwatson yeah that is what i suspected, some incompatible changes in the toolchain, im trying to build it because it has some gfx glitches and also a bug which prevents it from storing its settings correctly. First i want to be able to build the original ubuntu package then upgrade the sources from https://github.com/thjaeger/easystroke to fix those issues (he did not raise the version number unfortunately)14:37
TJ-seb128: thank-you seb128 I'll chase that one down. I was about to do a diff between the (full) SDD install and the USB image to figure it out.14:37
seb128yw14:37
TJ-hehehe and with a 'restart lightdm' greeter doesn't show the onboard keyboard, only the lower blacked-out area :)14:40
TJ-just got to find the system-menu indicator icon now14:40
de-factoany ideas what i could play around with to fix the error ctions.cc: In constructor 'TreeViewMulti::TreeViewMulti()':   ctions.cc:57:39: error: group' is not a member of 'sigc'   get_selection()->set_select_function(sigc::group(&negate, sigc::ref(pending)));14:41
de-factoi guess its something from libsigc++14:41
LocutusOfBorgde-facto, libsigc++ version??14:45
LocutusOfBorgyou might want to test the new upstream release (in debian)14:46
de-factoLocutusOfBorg libsigc++-2.0-dev Version 2.6.2-1 on Xenial release currently14:47
LocutusOfBorgmaybe 2.8.0 works=14:48
LocutusOfBorgdid it worked before=14:48
de-factoi used to compile it on previous versions of ubuntu (dont remember the exact versions though)14:49
LocutusOfBorgseems that the group has been removed between 2.4 and 2.614:50
LocutusOfBorghttps://anonscm.debian.org/cgit/collab-maint/libsigc++-2.0.git/commit/?id=fa41e72a2a68239a84839e59f763e0f49c0ca09814:50
LocutusOfBorg- * @deprecated Use C++11 lambda expressions or %std::bind() instead of libsigc++ lambdas and sigc::group().14:50
LocutusOfBorgI think you did get a lot of warnings with the previous release14:50
de-factocould very well be the case, i dont remember exactly anymore, but i got a working binary on previous versions of ubuntu14:52
dobeyTJ-: indicator-session is the "system menu" indicator15:11
TJ-dobey: thanks... I've been walking the rdepends/depends paths to try to figure out why it was missed15:18
=== juliank is now known as Guest76489
=== chuck__ is now known as zul
=== afkthairus is now known as athairus
Unit193slangasek: Got time for a quick PM?18:35
slangasekUnit193: prime ministers are never quick18:54
=== tedg_ is now known as tedg
Unit193Nope, certainly not.  Thankfully, I'm referring to a private message.18:55
mapreripitti: I'm looking at http://autopkgtest.ubuntu.com/packages/p/pbuilder/ and i just noticed the 'i' sign points to tracker.d.o.  I guess this is fine and cool for the debian instance of debci, but probably shouldn't ubuntu link the launchpad page instead? (actually, I'd love to have both links to tracker.d.o and lp)21:04
=== Guest76489 is now known as juliank
=== juliank is now known as Guest71812
=== Guest71812 is now known as juliank_
=== juliank__ is now known as juliank
bdmurraylamont: Could you have a look at bug 1552801?23:08
ubottuError: Launchpad bug 1552801 could not be found23:08
=== jgrimm is now known as jgrimm-afk

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