[12:10] <kyrofa> cjwatson, I'm trying to get LP building dailies and publish to edge. In such a situation, the snapcraft.yaml holds the link to https://example.com/latest.tar.gz, and the branch will be in LP (with an associated snap setup to publish to edge). I just need to use the LP API to spin up a build
[12:10] <kyrofa> cjwatson, it would run in a daily cron job
[12:11] <kyrofa> cjwatson, how does one typically go about that? I assume it can only be done by a user who owns the snap. Do I just generate a token and use that? Can I protect it at rest in some way?
[12:16] <cjwatson> https://help.launchpad.net/API/launchpadlib basically.  It (regrettably) can't be usefully constrained at this point; I expect eventually we'll support macaroons so that it can only do that one action, but for the time being it would be a full token for that user.
[12:17] <kyrofa> cjwatson, alright, just wanted to check before I went down a bad path-- thanks!
[12:17] <cjwatson> (Or you could hackily push to the branch containing snapcraft.yaml and use LP's auto build handling ...)
[12:17] <kyrofa> cjwatson, meh, then they're SSH keys :P
[12:45] <kyrofa> cjwatson, the only way to login that I've seen is login_with, which opens firefox to get a token. How might I pre-generate a token to put on a headless server?
[12:45] <kyrofa> Rather, how do I provide it to LPlib
[12:53] <kyrofa> Maybe use credentials_file to generate the token initially and then transfer that file to the server in question? Then continue using credentials_file on the server?
[13:04] <kyrofa> Ah ha, that does work
[13:30] <cjwatson> kyrofa: Yeah, that's typically the one
[13:30] <kyrofa> cjwatson, I found Credentials as well. Seems the access is more fine-grained with that one
[13:32] <cjwatson> I think we typically just use login_with and rely on python-keyring to do the storage.
[13:33] <cjwatson> And when login_with prompts you for authentication, you copy the URL and paste it into your desktop browser.
[13:33] <cjwatson> Ignoring the fact that it tries to pop up a browser on the server.
[13:43] <clivejo> can a comment be deleted from a bug report?
[13:44] <sarnold> there's often a 'hide' button users can hit
[13:44] <sarnold> I've filed launchpad 'answers' questions before when asking for spam comments to be deleted
[13:44] <cjwatson> Yep.  The person who wrote the comment or Launchpad staff can hide comments.
[13:44] <clivejo> I posted to the wrong bug by mistake
[13:44] <cjwatson> You can hide that yourself, then.
[13:45] <clivejo> but cant delete?
[13:45] <cjwatson> You might need to reload the bug page, but you should see a "Hide" link below the comment.
[13:45] <cjwatson> No.
[13:45] <clivejo> ah yes, I see it
[13:45] <cjwatson> "Hide" means that most people can't see it.
[13:45] <clivejo> I thought that was just to hide long comments!
[13:45] <cjwatson> Only if you wanted to stop anyone else seeing them at all ...
[13:46] <clivejo> ah
[13:46] <clivejo> good to know!
[15:06] <rbasak> How should I tackle bug 1523051? I don't see anything I can edit - does it need to be passed to Launchpad admins to fix?
[15:06] <ubot5`> bug 1523051 in net-tools (Ubuntu) "Upstream link needs changing" [Undecided,New] https://launchpad.net/bugs/1523051
[15:15] <cjwatson> rbasak: turn it into a question and move it to Launchpad
[15:15] <cjwatson> (can't deal right now)
[15:19] <rbasak> Will do, thanks.
[18:56] <sforshee> hi - one of the kernel team scripts fetches https://git.launchpad.net/~canonical-kernel/+git/kteam-tools/plain/ktl/kernel-series-info.yaml
[18:56] <sforshee> and that started failing for me sometime today with http code 500. http access to the repository overall seems to be broken, though git access still works.
[19:00] <acheronuk> I've been getting OOPS- messages browsing the kubuntu git repos over http since late afternoon as well
[19:12] <apw> sforshee, that looks to be oopsing:   TypeError: translatePath() takes at most 4 arguments (5 given)
[20:49] <wgrant> apw, sforshee, acheronuk: Fix is on the way.
[20:51] <apw> wgrant, hey thanks
[21:09] <wgrant> Should all be happy now. Let me know if something else looks awry.
=== heroux_ is now known as heroux
[23:47] <mpontillo> wgrant: hey, I'm seeing a weird issue with https://code.launchpad.net/~mpontillo/maas/discovery-view-rdns--bug-1633822--part3/+merge/308980 -- it looks like it's not properly considering the prerequisite branch. the diff should only contain src/maasserver/dbviews.py and src/maasserver/models/tests/test_discovery.py
[23:49] <mpontillo> for what it's worth, --part2 looks happy. maybe because it's has a prereq that has a prereq?