=== chihchun_afk is now known as chihchun === cpaelzer_ is now known as cpaelzer === JanC is now known as Guest5365 === JanC_ is now known as JanC === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [10:09] I have lp: git alias setup. However, pushing to lp:~team/project/repo-name does not work [10:10] I had to change the remote to lp:~team/project/+git/repo-name [10:10] * xnox is sad [10:20] also i guess i cannot use private branches for recipes?! [11:02] xnox: Why do you feel you need to specify the repository name? [11:03] xnox: In most cases the correct thing to do is to use the default repository for the team/project combination, i.e. lp:~team/project [11:03] xnox: lp:~team/project/repo-name will never work for git, because that's a namespace stolen en masse by bzr [11:04] xnox: Recipes may indeed only be built from public code at this time. This restriction may be lifted in the future, but it requires some quite complex work. === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [11:34] cjwatson, i feel the need to specify the repository name due to (a) bzr habbits (b) because i'm starting a new sub-components in the project, and don't want to register a separate project. [11:35] also it seems weird that i need to always specify target when creating git branch merge proposals, e.g. refs/heads/master -> i would have thought such a default could be pre-populated? [11:36] also thinking about ubuntu-seeds migration to git, i'm thinking we want to have just [11:36] lp:~ubuntu-core-dev/ubuntu-seeds/+git/platform [11:36] lp:~ubuntu-core-dev/ubuntu-seeds/+git/ubuntu [11:36] lp:~lubuntu-dev/ubuntu-seeds/+git/lubuntu [11:36] etc [11:36] and have branches inside those repositories for all the releases. [11:40] xnox: The need to specify the target ref is a bug, yes - we need better picker logic there [11:40] xnox: Goodness, that's a silly layout [11:40] cjwatson, what layout would be good? given we have different ACLs [11:40] xnox: For ubuntu-seeds, there's no good reason for anything other than a single repository per team. [11:40] oh, so you mean like [11:41] lp:~ubuntu-core-dev/ubuntu-seeds [11:41] lp:~lubuntu-dev/ubuntu-seeds [11:41] lp:~xubuntu-dev/ubuntu-seeds [11:41] I'd do lp:~ubuntu-core-dev/ubuntu-seeds with platform.artful, ubuntu.artful etc. branches. [11:41] and that's it? =) [11:41] ack. [11:41] And also, make lp:~ubuntu-core-dev/ubuntu-seeds in fact be lp:ubuntu-seeds (i.e. the target default) [11:41] That way object sharing works better for the others [11:42] That said, check that that proposed layout works properly with germinate [11:42] sometimes i wish we just had one repository for ubuntu-seeds full stop. [11:42] I'm not convinced I was thinking straight when I did the trivial git support for it, and thus I'm not sure it supports branches in the sensible way. [11:43] ack [11:43] So that may need some thought, *before* you push anything :) [13:35] re:pdiffs i always thought we want them for e.g. security pocket [13:35] not necessoraly for $devel or ppas === chihchun is now known as chihchun_afk [16:37] is there any useful way to debug installability problems inside build chroots like https://launchpadlibrarian.net/317155258/buildlog_ubuntu-artful-ppc64el.libkf5kface_16.12.3-0ubuntu1+mpr1_BUILDING.txt.gz ? [16:37] sbuild-build-depends-libkf5kface-dummy : Depends: libopencv-dev (>= 3.1.0) but it is not going to be installed [16:37] E: Unable to correct problems, you have held broken packages. [16:37] => FAIL and nothing else, ain't so useful :( [16:38] I think new enough sbuild also has a "debug apt problems" flag that could be useful here ^^ [16:51] mapreri: easiest way is typically to use chdist [16:52] :(