[00:13] <cjwatson> wgrant: Yeah, I did it.
[00:13] <cjwatson> Mostly to avoid future accidents.
[00:15] <wgrant> Thanks.
[00:16] <cjwatson> I spent most of the day landing stuff ...
[00:17] <wgrant> I hopefully won't spend most of the day fixing obscure event failures...
[00:17] <wgrant> Events cause no end of trouble.
[01:17] <cjwatson> wgrant: Re bzr-webhooks, my feeling is that git:push:0.1 should probably use the unique name too, to avoid races.
[01:17] <cjwatson> Although I guess some degree of raciness is inevitable, since you could equally change the repository name.
[01:18] <cjwatson> So we could use the shortest path for bzr too if you prefer, though I think it will probably have to gain the lp: prefix in that case (Branch doesn't really seem to want to expose the shortened path without it).
[01:40] <wgrant> cjwatson: Yeah, I'm not really sure which way to go.
[01:40] <wgrant> If we use the full name it gets slightly hideous.
[01:40] <wgrant> But only a little.
[01:44] <cjwatson> To argue against myself, I kind of don't want to encourage people to use unique names unnecessarily
[01:45] <cjwatson> Since they're kind of an implementation detail of "the default repository for <project>"
[01:46] <wgrant> Right, exactly.
[01:46] <wgrant> And it can already be wrong due to renames.
[01:46] <wgrant> Ah, which you already pointed out.
[01:47] <cjwatson> So there remains the question of whether it's OK to add the lp: prefix, or whether we need to add a shortened_path interface to Branch that omits it.
[01:47] <wgrant> Hm
[01:47] <wgrant> BranchSet doesn't have getByPath
[01:47] <wgrant> Just getByUniqueName and getByUrl
[01:48] <wgrant> Possibly we should add lp: to both.
[01:48] <cjwatson> The reason I didn't for GitRepository was that it made GitRef look weird.
[01:48] <wgrant> Ah, yes.
[02:12] <wgrant> cjwatson: When generating keys on DF are you being careful to not upload them to prod keyservers?
[02:15] <cjwatson> Err
[02:15] <cjwatson> Possibly not
[02:15] <cjwatson> http://keyserver.ubuntu.com:11371/pks/lookup?fingerprint=on&op=index&search=0x19D44D6EF975FEC0E91FDEC87B7BBD5205DDBBB2
[02:16] <cjwatson> Wonder if there's some way I can get rid of that ...
[02:16] <wgrant> Nope, it's not possible.
[02:16] <cjwatson> Revoke, I guess
[02:16] <wgrant> Which is why I always disable that code when I test the script every couple of years.
[02:16] <wgrant> Normally we just symlink in one of the existing keys on labbu.
[02:16] <cjwatson> Can we configure DF to not upload the key?
[02:16] <wgrant> That would be a useful option to have.
[03:15] <wgrant> cjwatson: Thanks.