[00:00] <mwhudson> the build fails with proposed enabled too
[00:00] <mwhudson> (not really surprised there)
[16:09] <nacc> cjwatson: so i'm trying to use login_with in the importer and in the snapped version and the normal version, i get a browser prompt to do the oauth authorization and then i get this backtrace if i say yes there: http://paste.ubuntu.com/24448744/
[16:09] <nacc> cjwatson: i didn't use login_with in the older code (this is so we can set the 'default branch' in lp)
[16:17] <cjwatson> nacc: interesting, perhaps the file keyring backend is stricter than other backends.  could you file a bug against launchpadlib with that backtrace?
[16:17] <nacc> cjwatson: will do, thanks
[16:18] <cjwatson> nacc: and include it in your list of stuff that needs to be improved in LP for the git importer when you get round to sending that to us :)
[16:18] <nacc> cjwatson: any suggestion of a workaround for now? I think this must have worked locally (i'll test that now) so maybe it's someting i'm forgetting to setup on my serverstack instance
[16:18] <nacc> cjwatson: and ack :)
[16:18] <cjwatson> I'm not sure, possibly a different keyring backend if you can manage it?  I'm assuming using python2 is too hard
[16:19] <nacc> cjwatson: yeah i think it'd be tricky to do, but i'll look
[16:19] <cjwatson> (and I don't really want to have to recommend that people use python2)
[16:24] <nacc> cjwatson: yep
[16:29] <cjwatson> nacc: I think it's probably as simple as http://paste.ubuntu.com/24448855/ to make sure that we only ever pass unicode to set_password, but that would need some testing in various environments
[16:30] <nacc> cjwatson: ack
[16:36] <mapreri> cjwatson: might I have a tad more space in https://launchpad.net/~mapreri/+archive/ubuntu/opencv-3.1.0-sync-test/+packages ?
[16:36] <mapreri> just 1 or 2 gigs more would do great
[16:38] <cjwatson> mapreri: done
[16:39] <mapreri> cjwatson: great, thank you!
[16:43] <mapreri> btw, is there work ongoing about having virtualized buildds for ppc and s390x? (well, ppc might not be so interesting anymore given it has been dropped…).  anything that can be used for lurking?
[16:45] <cjwatson> mapreri: it's been ongoing for a while but I gather is actually making progress at the moment.  nothing you can use yet though.
[16:45] <mapreri> cjwatson: ic, ack.
[16:45] <cjwatson> (it involves deploying a new scalingstack region)
[16:48] <mapreri> (I figure(d) it's not straightforward)
[16:48] <nacc> cjwatson: fwiw, that does avoid the backtrace
[17:56] <nacc> cjwatson: also, perhaps i'm thinking about it wrong -- but what would your recommendations for running with oauth in the importer in a detached way? I get prompted to either auth the importer in the web ui (on first run) or have to unlock the keyring (on subsequent runs). Which means it can't be trivially automated. Are we better off not using login_with() for now and just manually fixing the branch
[17:56] <nacc> pointers (or pushing master and avoiding it altogether)?
[17:57] <cjwatson> nacc: in the automated Ubuntu archive tools that run on snakefruit, we use login_with, and just authenticate once
[17:58] <cjwatson> nacc: (copying the resulting URL to a local web browser)
[17:58] <cjwatson> nacc: I've never been prompted to unlock the keyring
[17:59] <cjwatson> nacc: perhaps you're doing something that stops it working with wherever it writes the auth details?
[17:59] <nacc> cjwatson: yeah i wonder if the snap is being ... snappy :)
[17:59] <nacc> cjwatson: thanks for the pointers!
[18:06] <nacc> cjwatson: what backend does snakefruit use? file-based as well? or is gnome-keyring or something installed?
[18:18] <cjwatson> just file-based I think; it lands in ~/.local/share/python_keyring/keyring_pass.cfg
[18:23] <nacc> cjwatson: ah, so using PlaintextKeyring maybe?
[18:24] <nacc> cjwatson: i think that's what i was missing before -- changed it now (in keyringrc.cfg) and it no longer prompts to unlock the keyring
[18:25] <cjwatson> nacc: I don't remember us doing anything specific to configure that
[18:29] <nacc> cjwatson: ok, i wonder if something changed
[18:30] <nacc> cjwatson: context is, with the serverstack rebuild, i am recreating the environment from zero, and using the snap
[18:30] <nacc> cjwatson: i think that if you are using that file, something tweaked that, which is good enough info for me
[20:07] <mapreri> cjwatson: wow, making package building made it filled again... could I have some more space?
[20:09] <mapreri> (I didn't anticipate so much space would be needed, btw)
[20:25] <cjwatson> mapreri: done
[20:26] <mapreri> cjwatson: ♥
[20:31] <nacc> cjwatson: thanks for your help, importer is running again from the snap
[20:34] <cjwatson> good good