[12:12] sigh [12:13] tests don't work so well when you run them on the wrong branch [12:20] I know the feeling ... [14:24] re earlier conversation, I've been meaning to look at https://github.com/asottile/pyupgrade [14:26] Neat for stripping out 2-isms from pure-3 code [14:27] ...it will also auto-f-string, about which I have Feelings (TM), but hey ho [14:48] ... no [14:57] https://code.launchpad.net/~twom/launchpad-buildd/initial-docker-build-support/+merge/369775 is ready for re-review I think [14:58] tomwardill: queued up. (you have at least one conflict marker in there ... maybe worth going over the diff yourself) [14:58] ergh, thought I'd fixed that [14:58] * tomwardill fixes again [15:01] although it has just occurred to me that I've not tried it in an environment that needs a HTTP proxy for outgoing connections [15:15] * tomwardill diverts into 'how to make a proxy' [16:31] Should I expect --force-with-lease to Just Work when pushing to lp:~doismellburning/turnip? I'm getting rejected with not much context - https://pastebin.canonical.com/p/FDvkYCvpMY/ [16:37] I think that means that your local ref is stale for some reason [16:37] Oh err that invocation is weird [16:38] Normally you'd use --force-with-lease with a proper remote name, since it wants to know what the expected contents of that ref are [16:38] Do you have something like a "doismellburning" remote configured to point to lp:~doismellburning/turnip ? [16:38] Then you'd use git push --force-with-lease doismellburning feature/readme-rst [16:38] Liberal use of tab-completion recommended there [16:45] Ahhh I hadn't appreciated that you could pass a refname to fwl - that'll do it [16:46] I've also managed today to inadvertently push two branches to lp:turnip - I didn't realise I could, or expect to be able to, do that, and I can't figure out how to delete them [16:46] but after that I started explicitly specifying where I was pushing [16:48] I've got the GH-y fork-model embedded deep in my muscle memory so I habitually just "git push" with no further qualification [16:51] git push --delete origin [16:51] Even with GH, I always give the upstream repo a remote name of "origin" and name my fork "cjwatson", so I can "git push cjwatson " [16:52] We could protect everything under lp:turnip but haven't yet/currently done so [16:53] If you delete those branches then I'll apply some more protection [16:55] cjwatson: deleted, thanks [16:56] SpecialK|Canon: And protected further (https://code.launchpad.net/~launchpad/turnip/+git/turnip/+permissions) [16:57] And I've done likewise for lp:launchpad [16:57] cjwatson: Yay, thank you [16:58] And you were right about stale info - I'd missed that I hadn't actually ever fetched lp:~doismellburning/turnip in that clone [17:29] https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/378344 - buildbot fix, if somebody could review that [17:58] Done! [18:11] thanks