[00:50] oh shoot. what's the new rest api for replacing ubuntu-sso-client? [06:52] good morning [07:00] Good morning === ljp is now known as lpotter [09:23] @pilot in === udevbot changed the topic of #ubuntu-devel to: Wily (15.10) Released! | Archive: open | Devel of Ubuntu (not support or app devel) | build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of precise-wily | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots: Mirv [09:57] Mirv, you're a star! Thank you very much for sponsoring again :-) [09:59] flexiondotorg: you're welcome again :) I did some sponsoring last week but it turns out today I'm actually on the calendar so continuing a bit, plus it's s good time with lts milestone coming === vrruiz_ is now known as rvr [10:41] hi, nobody is swearing at me for the libsdl2 sync? [10:42] seeded-in-ubuntu returns ubuntu-mate, ubntukylin, lubuntu, going there [10:48] LocutusOfBorg, you dropped mir support, didn't you? [10:48] LocutusOfBorg, in general it's not nice, to forcesync things, effectively stealing a merge, and dropping ubuntu delta. [10:49] LocutusOfBorg, it's ok for things to be ahead in debian, but a debian maintainer does not get to drop ubuntu delta for no reason. [10:50] xnox, the delta is upstream [10:50] lubuntu have been working on mir enablement, and you should have talked to e.g. Artur or Brandon. [10:50] I checked and it is included upstream [10:51] I asked a while ago, and nobody complained [10:51] LocutusOfBorg, you have dropped libmirclient-dev [i386 amd64 armhf], build-dep. [10:52] thus it's not compiled with mirclient support any more, is it? [10:52] let me check, brandon answered me a few seconds ago [10:53] LocutusOfBorg, hm, i see libmirclient-dev installed none the less in the new build. [10:53] and it doesn't link against it. [10:54] maybe it's actually all ok, as you say. [10:54] actually I checked, checked and double checked the delta, it was patching upstream source files [10:54] changing function names [10:54] no [10:54] and I dropped, because upstream actually duplicated them [10:54] checking for Mir support... no [10:55] Laney, yeap, found the delta. [10:55] some mir things get pulled in, evidently transiently, but that seems to not be sufficient [10:56] ok let me fix it [10:56] LocutusOfBorg, see at the end http://paste.ubuntu.com/14662490/ there are chuncks of things from debian/rules [10:56] LocutusOfBorg, that are dropped. Evidentaly, since libmirclient-dev is available anyway, you can add those bits to debian/rules but with extra DEB_VENDOR treatment [10:56] it's auto detected [10:56] to e.g. enable it only on ubuntu, whilst uploading it into debian. [10:56] so I guess you could add that to fail the build if it doesn't get enabled [10:57] but probably not synacble if additional BDs are needed [10:57] then again i don't know how MIR or libsdl thing work... [10:57] LocutusOfBorg, please work out and fix/re-enable mir support if needed =) [10:58] yes, I'll probably upload it in ubuntu in a few moments (I need to check in a ppa) [10:58] and then if the patch is sane enough go in debian too [10:58] merci [10:58] thanks to you both for the help [10:58] but at the end, why isn't it autodetected? [10:59] I think I should patch configure.ac somewhere, to fix the detection, instead of tweaking rules file [11:01] "if $PKG_CONFIG --exists mirclient egl xkbcommon ; then" [11:04] probably just a missing build-dependency [11:04] I suspect mir support has been broken anyway also before [11:04] for some little value of "broken" probably? [11:06] broken how? [11:07] LocutusOfBorg, it was not broken before [11:07] checking for Mir support... yes [11:07] -- dynamic libmirclient -> libmirclient.so.8 [11:07] -- dynamic libxkbcommon -> libxkbcommon.so.0 [11:07] https://launchpadlibrarian.net/196165225/buildlog_ubuntu-vivid-amd64.libsdl2_2.0.2+dfsg1-6ubuntu2_UPLOADING.txt.gz [11:07] yes, true [11:07] I'm actually checking, thanks [11:08] the question is: the autodetect code never worked, but before it was forced in rules file [11:08] instead of forcing it would be better to fix the configure.ac file maybe? [11:09] cjwatson, hi [11:16] ok I found a clie [11:16] clue [11:16] autoconf tries to compile a sample program [11:17] MirMotionToolType tool = mir_motion_tool_type_mouse; [11:17] [11:17] but that enum doesn't exist, even if it is in the documentation === _salem is now known as salem_ [11:20] tkamppeter: hi, hwat's up? [11:20] er, what [11:20] excessively Old English there [11:26] cjwatson, you told me that liblouis-bin would make it automatically into Main when I upload something which depends on liblouis-bin to Main. [11:28] tkamppeter: not automatically, but it's a prerequisite [11:28] tkamppeter: it's on http://people.canonical.com/~ubuntu-archive/component-mismatches.txt , I'll promote it now [11:29] tkamppeter: done, it now just needs a publisher run and mirror sync, so < 1 h [11:29] beat me to it [11:30] indeed, I didn't explicitly say it was automatic; though I wasn't very clear [11:36] LocutusOfBorg: looks like that's only in the check [11:37] you can probably replace it with a version check instead? [11:37] Laney, it uses the check of a common api to know if the code can be compiled [11:37] so probably since somebody removed that part of api, I need to have something better-proof call? [11:37] LocutusOfBorg: yes but it's not like the actual code itself uses API which is removed [11:37] I'm looking at bzr mir code to understand the actual replacement [11:37] can you just pkg-config check for a recent ish version? [11:37] Laney, this is true, I might also look to some calls that libsdl2 uses [11:37] instead of try-compiling? [11:38] since the api seems to change, I would like to keep a try-compile and forward upstream [11:39] I don't understand, you might pick the wrong bit [11:39] it's going to fail to build if the API breaks anyway [11:40] so how do you suggest to fix it? [11:41] "if $PKG_CONFIG --exists mirclient egl xkbcommon ; then" [11:41] enable and live happy? [11:42] If you know a lower bound on the version then you can check for that - there's a PKG_CHECK_MODULES macro [11:42] I don't know a lower bound unfortunately [11:42] it says something about 14.04 [11:43] so that's a good indicator, or maybe brandon knows [11:43] precise isn't supported anymore AFAIK, so no check isn't good? :) [11:44] I'm a cmake guy, autoconf syntax is obscure to me :) [11:52] the fix seems to be working [11:52] but I need to patch it again, because mir changed api [11:57] LocutusOfBorg: actually it looksl ike you can pass >= to pkg-config --exists, if that's helpful [11:57] --exists 'libmirclient >= the-version-in-trusty' [11:57] hi guys, what does the SRU queue look like nowadays? I have this bug for which I uploaded packages but they need review by the sru team: https://bugs.launchpad.net/landscape-client/+bug/1508110 [11:57] Launchpad bug 1508110 in landscape-client (Ubuntu Wily) "Users tab doesn't work as expected" [Undecided,New] [11:58] ahasenack: http://people.canonical.com/~ubuntu-archive/pending-sru.html#upload-queues I've seen worse [12:22] Laney, I'll use the one from there [12:22] https://bugs.launchpad.net/ubuntu/+source/libsdl2/+bug/1513241 [12:22] Launchpad bug 1513241 in libsdl2 (Ubuntu) "libsdl2 No longer works with MIR 0.14 and greater (ABI/API break)" [Medium,In progress] [12:38] cjwatson: I'm looking at bug 1537136; does it sound workable/correct if udev-udeb starts shipping a hook /usr/lib/finish-install.d/copy-persistent-net which copies /etc/udev/rules.d/70-persistent-net.rules to /target/etc/udev/rules.d/ ? [12:38] bug 1537136 in debian-installer (Ubuntu) "ISST-LTE: no network after install due to interface order change" [High,Triaged] https://launchpad.net/bugs/1537136 [12:38] cjwatson: (sorry, not familiar with udebs at all) [12:39] just looked at http://d-i.alioth.debian.org/doc/internals/apb.html [12:39] cjwatson: earlier would be better, so would post-base-installer.d/ be ok? [12:42] cjwatson: unping, this script already exists, it just seems we aren't installing the generator [12:43] pitti: I believe your force for jquery & friends needs some adjustments, since jquery (and friends) are not migrating. [12:44] Laney, xnox actually due to MIR api changes, libsdl2 wasn't working anymore with the new MIR protocol [12:44] https://bugs.launchpad.net/ubuntu/+source/libsdl2/+bug/1513241 [12:44] Launchpad bug 1513241 in libsdl2 (Ubuntu) "libsdl2 No longer works with MIR 0.14 and greater (ABI/API break)" [High,In progress] [12:45] so it isn't actually a regression, but I promise I'll take care of it :) [12:45] pitti: heh. right, if the generator isn't in udev-udeb it probably won't work so well [12:46] cjwatson: my grep apparently didn't catch the base-installer.d/05udev script, so that part should be okay [12:47] pitti, thank you very much. [12:47] cjwatson: all the udebs are bundled in netinst's initrd.gz, right? so I could just modify that to install an updated udeb locally [12:47] mapreri: I can't force the migration as it makes things uninstallable; see "trying: jquery" on http://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt [12:47] mapreri: apparently this needs some more rebuilds or porting to the new version [12:47] * amd64: gitit, libghc-gitit-data, libghc-gitit-dev, libghc-gitit-prof [12:47] pitti: yep [12:48] mapreri: those packages become uninstallable with the new jquery [12:48] cjwatson: ack, thanks [12:48] oh, haskell might be improvable now that aeson is fixed [12:48] (gitit) [12:48] it was stuck for a while [12:48] * cjwatson goes to poke that [12:48] thank you. [12:49] pitti: wasn't force a thing to ignore uninstallability troubles and migrate that anyway? [12:49] mapreri: not force-skiptest [12:50] mapreri: sure, there are hints to force things really hard but they should almost never be used [12:50] pitti: nfs shares fail to get mounted on boot if using the hostname in fstab, because apparently it can't resolve the hostname yet at that point.. do you know if this is filed somewhere already, or if it's a bug at all? [12:51] and would it be in nfs-utils or elsewhere [12:51] ah, then the "forced by pitti" in _excuses means something related to the tests... I thought it was the same "force" hint for britney... [12:51] * mapreri is learning stuff every time he put his nose inside this channel :D [12:51] pitti: if you need more complicated initrd changes, then you can grab the debian-installer source, drop updated udebs into build/localudebs/, and run "make -C build rebuild_netboot" [12:52] cjwatson: oh cool, thanks; that seems useful as a more reliable way to test changes [12:52] also, all the hosts here that have sssd set up to use my local directory service kick me out during dist-upgrade to a new release. annoying as hell and no idea what's wrong and where [12:52] mapreri: well, to be fair it does say "Should wait for tests relating to jquery 1.11.3+dfsg-4, but forced by pitti" :-) [12:52] pitti, I would like to do a no change rebuild against libpodofo0.9.3. for your krename [12:52] mapreri: right, I forced the tests, but we can't/don't want to force over uninstallability [12:53] mapreri, ^^^ [12:53] LocutusOfBorg: oh, sure; I can do that with a single command, hang on [12:53] I have already it pending [12:53] it'll take a couple more days to work through the current ghc transition, assuming nothing else goes wrong with it [12:53] pitti: then please also add scribus and calibre to the list ♥ [12:53] but I'll try to stay on top of it [12:53] (same transition) [12:54] LocutusOfBorg: krename upload [12:54] ed [12:54] $ rebuild-lib-transition libpodofo0.9.3 krename [12:54] mapreri: for the same libpodofo0.9.3 transition? [12:54] pitti: yup. [12:54] yes [12:54] done in debian a cople of days ago [12:55] pitti, I did stole without shame from cjwatson, adding some bits http://paste.ubuntu.com/14662972/ [12:55] done [12:55] cool [12:55] I suspect pitti's scripts are nicer than mine [12:55] LocutusOfBorg: heh, I guess everyone has their own little helpers for this :/ [12:55] That is often the way [12:55] yes, I'm wondering why there isn't a tool in ubuntu-dev-tools for this [12:55] cjwatson: by and large the same: http://paste.ubuntu.com/14662976/ [12:55] (what LocutusOfBorg pasted isn't exactly mine though, I wouldn't have used the gratuitous/unnecessary bashism :P ) [12:56] :p [12:56] pitti: ah yes, close enough [12:56] my script should work also after xenial [12:56] assuming "distro-info --devel" gets updated quickly :) [13:03] bschaefer, :D [13:04] hello! [13:04] * LocutusOfBorg is waiting for a punch [13:04] hello and sorry :) [13:04] * bschaefer wonders what happened [13:04] wrt libsdl2 [13:05] reading the comment umm [13:05] LocutusOfBorg, is your patch base on the branch or the WIP branch? [13:06] * bschaefer should just look [13:06] and i would prefer to have mir enabled [13:06] it is based on the patch on the bug report [13:06] bschaefer, sure, this is why I would like to upload ASAP [13:06] we were also talking about just doing some pkg-config --exists mir or whatever [13:06] without the try-compile [13:07] LocutusOfBorg, the patch needs to be based on this branch: https://code.launchpad.net/~brandontschaefer/+junk/SDL2-new-mir-ABI [13:07] since it fixed quite a few bugs [13:07] umm i made a ppa [13:07] with a patch [13:07] https://launchpad.net/~mir-team/+archive/ubuntu/staging [13:07] just a question: libsdl2 was broken even before my sync, right? [13:07] I mean, with a no-change rebuild I would have triggered a build failure? [13:08] LocutusOfBorg, yeah its been broken since mir 0.13 [13:08] IIRC [13:08] ok, this might explain something :D [13:08] btw, with this patch you know we are enabling mir for everywhere available? [13:08] yeah, just been slow to getting a patch out then saw 2.0.4 being worked [13:08] on soo i waited [13:08] before it was enabled for i386, amd64, armhf [13:09] now we are adding arm64, and maybe more [13:09] it should be i386/amd64/armhf [13:09] hmm [13:09] arm64 should be supported [13:09] ok wonderful [13:09] *should* as in, i think :) [13:09] bregma, ^ [13:12] so can I upload that version? [13:13] you added "libmirclient-dev" as dependency, in my environment it gets installed automatically [13:14] LocutusOfBorg, the version you attached to the bug? (Im also not *super* excellent on debain packing soo there could be a mistake there) [13:14] or the one in the ppa? [13:14] direct dependencies should be explicit even if they're installed automatically via something else [13:14] the one from ppa [13:14] simple rule: if you use it explicitly, depend on it explicitly [13:14] but I'm looking to fix something [13:14] ok [13:15] if we want to enable mir support by default it will be needed [13:15] but its also not *required* to run SDL2 [13:15] but since we want to run mir with sdl2 it should be included explicitly [13:15] LocutusOfBorg, let me ... install that package and test it out [13:16] bschaefer, I'm uploading in a ppa [13:16] please wait :D [13:17] o cool [13:17] LocutusOfBorg, since i pushed that patch to a ppa and i tested the source file and checked mir was built [13:17] but i havent tested the ppa out it self yet [13:17] https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa/ [13:18] I did some little changes, from your package [13:18] (I started from the ppa) [13:18] 1) changelog merged with the debian one [13:18] 2) removed the copyright patch parts (your patch was changing back copyright years, I removed the change) [13:18] 3) rebased the version number [13:21] nice [13:21] oh I also renamed from .patch to .diff, to be the same as the other patch [13:21] and I'm adding some verbosity to changelog [13:22] explaining the ubuntu delta, because I presume I can't upload it on debian [13:23] bschaefer, once the SDL 2.0.4 package is in Ubuntu, can all the changes just go upstream so 2.0.5 can be just a plain synch? [13:24] bregma, that would be ideal, and once 2.0.5 hits the patch *should* be removed and all should be well [13:24] LocutusOfBorg, yup, ... once 16.04 hits ill push these changes upstream [13:25] and debian will have it from upstream :) [13:25] bschaefer, so I'll wait for your ack [13:25] yup adding the ppa now! [13:25] http://paste.ubuntu.com/14663107/ [13:26] this is the debdiff [13:26] I just added some verbosity in changelog [13:26] and as said before rebased the patch [13:27] but no *useful* changes, and please note: we are enabling mir almost everywhere the test code compiles, not as before [13:27] +1 for upstreaming changes [13:28] cool, yeah once you see it builds with mir [13:28] and it all compiles it *usually* worked out well :) [13:28] LocutusOfBorg, hmm your ppa ... is losing to the archive :( [13:28] @pilot out === udevbot changed the topic of #ubuntu-devel to: Wily (15.10) Released! | Archive: open | Devel of Ubuntu (not support or app devel) | build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of precise-wily | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots: [13:29] *** 2.0.4+dfsg1-2 500 [13:29] 500 http://us.archive.ubuntu.com/ubuntu xenial/universe amd64 Packages [13:29] https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa/+build/8893186 [13:29] you can just grab the deb files [13:29] they aren't published yet I guess [13:29] o [13:29] * bschaefer should have checked that :) [13:29] checking for Mir support... yes [13:29] -- dynamic libmirclient -> libmirclient.so.9 [13:30] -- dynamic libxkbcommon -> libxkbcommon.so.0 [13:30] the build seems fine [13:30] cool, sooo now its down to ... did i mess the patch up somehow :) [13:30] this has me scratching my head: [13:30] old binaries left on amd64: libpe-status4, libpengine4, pacemaker-dbg (from 1.1.13-2ubuntu1) [13:31] I'd assumed that it was due to some transitions that needed to be done, which I think I have completed - but I can't figure out what's holding onto those... [13:32] BTW bschaefer bregma if we force "libmirclient-dev" build-dependency we won't be able to sync it from debian [13:32] even if the patch is upstreamed [13:33] LocutusOfBorg, upstream wont have libmirclient in its control [13:33] i would assume we would just have a small patch in ubuntu to have that [13:33] bschaefer, neither debian [13:34] ok, that was the mean [13:34] so, plain sync isn't possible [13:34] unless debian has libmir :) [13:34] I guess not :D [13:37] Mirv, you about? [13:47] pitti, cjwatson, CUPS has built now, thank you for the help. [13:47] LocutusOfBorg, yay terria still works [13:48] pitti, cjwatson, will CUPS now get from -proposed into -release or do you have to give it a kick to do so? [13:48] :| but mouse doesnt seem to work... hmm one sec [13:48] hi, is there a reason we don't have ceres-solver in ubuntu ? https://tracker.debian.org/pkg/ceres-solver I don't see it in the sync blacklist and it was uploaded to unstable at the end of november. [13:50] tkamppeter: the easiest way to find out is to wait and see [13:50] tkamppeter: normally no manual action is needed though [13:52] jamespage: corner case in the handling of -proposed as a partial suite; I've removed those [13:52] cjwatson, ack - thanks [13:52] LocutusOfBorg, ignore my concern! Using an older client [13:52] vs a newer server and i broken the server/client protocol soo no client mouse events [13:53] LocutusOfBorg, patch working correctly for me! +1 to use it [13:53] so can I push on ubuntu? [13:53] uploaded :D [13:53] cjwatson, thanks. [13:54] thanks a lot bschaefer [13:54] LocutusOfBorg, awesome thanks! didnt expect to get that patch pushed so quickly :) [13:54] thank you! [13:54] bschaefer, I actually broke the build, so I had to :D [13:54] for the next time, just ping me, and I'll take care of it [13:54] BTW In debian we are planning an update of all the sdl2-* related libraries (just bugfixes) [13:55] LocutusOfBorg, awesome, thats good to know [13:55] all the sdl2-* shouldnt cause issues for the mir stuff? [13:55] and then a question for cjwatson [13:55] at lease i never had to do anything special with those [13:55] In debian the libpng16 transition is mostly "done" I mean, we patched the sources except for a few build failures, and we are waiting for the release team to proceed [13:56] I did ~30 NMUs and they are pending, and I'm planning to merge ubuntu whenever possible [13:56] do you think we can arrange a transition for xenial? [13:57] LocutusOfBorg: let's not have me be in charge of deciding whether to do transitions - I'm not sufficiently on top of things [13:57] so, moving to -release maybe? [13:57] sure [13:57] thanks :D [13:58] bschaefer, I'm also closing the #1513241 bug [13:58] in changelog [13:58] wow it is building awesomly fine [13:58] LocutusOfBorg, awesome thanks! [13:59] now mir is enabled everywhere, lets see how it goes [14:00] * bschaefer hopes quite well [14:00] thanks to you, without your patch sdl2 would have been in a bad state [14:00] I should have asked before doing the sync [14:03] LocutusOfBorg, :), didnt *know* anyone needed to be poked but ill know next time! (though hopefully this wont happen again!) [14:03] as it'll be upstream [14:05] bschaefer, it was me having to poke you before doing the sync [14:05] but I actually checked the actual delta and it was upstreamed [14:06] I missed the rules part change [14:06] and I missed the new patch not yet uploaded [14:06] but well, now we should be good [14:06] LocutusOfBorg, the reason i held off on the 2.0.3 patch is it didnt cleanly go from working with 2.0.3 --> 2.0.4... [14:06] for some reason [14:06] sooo i figured i should wait [14:07] I hope it is fine now [14:07] mvo_: hey! I see you're doing commits to goget-ubuntu-touch along with Sergio - could you take a look at my quick noob MP by any chance? Nothing too urgent tho! https://code.launchpad.net/~sil2100/goget-ubuntu-touch/touch-devel-warning/+merge/282045 [14:07] building is a good start! [14:11] sil2100: sure, I have a look [14:13] mvo_: thank you! === ginggs_ is now known as ginggs === roaksoax is now known as roaksoax-brb [16:45] pitti, is this temporary? http://autopkgtest.ubuntu.com/packages/r/ruby-redcarpet/xenial/s390x/ ? [16:54] doko: temporary how? It had been uninstallable for a while, but apparently that got resolved now? [16:56] pitti, you only answered after it passed [16:57] doko: ah, ok; so I guess ruby 2.2.4 makes it installable? [16:57] pitti, no, that's unrelated === salem_ is now known as _salem [23:14] having a bit of a problem, trying to branch a Bazaar repo and I keep getting public key errors. But my GPG key is imported correctly in Launchpad...what should I do? [23:17] is the error instead in your ssh config? [23:19] sarnold: what do you mean? [23:19] keep getting: [23:19] Permission denied (publickey). [23:19] ConnectionReset reading response for 'BzrDir.open_2.1', retrying [23:19] Permission denied (publickey). [23:19] bzr: ERROR: Connection closed: Unexpected end of message. Please check connectivity and permissions, and report a bug if problems persist. [23:19] (sorry for the paste, it's only 4 lines) [23:21] tsimonq2: yeah, that looks like an ssh error instead of gnupg -- does launchpad know the ssh public key you're using? [23:22] (at least I think it needs your ssh key.. it's been years since I set that up) [23:22] yes, what you need is ssh key for that [23:23] tsimonq2: need to add your SSH public key (usually .ssh/id_rsa.pub) here: https://launchpad.net/~/+editsshkeys [23:24] and I did [23:24] tsimonq2: oh yes it's there :) [23:24] http://keyserver.ubuntu.com:11371/pks/lookup?fingerprint=on&op=index&search=0x9C6058DF83975841F42717F99F9292DB1C174C6F [23:25] this is the one on THIS computer [23:25] tsimonq2: hm, what you're showing is a gpg key, not ssh key [23:25] that looks like a gpg key, not ssh [23:26] it'd be one of these two https://launchpad.net/%7Etsimonq2/+sshkeys [23:26] and this is my gpg --list-keys output(taking away my private key :P): [23:26] pub 4096R/1C174C6F 2015-12-05 [expires: 2016-06-02] [23:26] uid Simon Quigley [23:27] tsimonq2: what happens if you try ssh bazaar.launchpad.net ? (it should reply "No shells on this server.") [23:27] Permission denied (publickey). [23:28] oh I SEE...I'll look at my ssh keys [23:30] yay thanks guys [23:32] hah, my ssh bazaar.launchpad.net fails too. but my bzr+ssh tree still works.. go figure.