[00:38] <StevenK> wgrant: Did you spy bug 1186050?
[00:38] <_mup_> Bug #1186050: Remove Packages-arch-specific use from Soyuz <lp-soyuz> <soyuz-build> <soyuz-core> <Launchpad itself:New> <https://launchpad.net/bugs/1186050>
[00:52] <wgrant> Yeah
[00:53] <StevenK> I think I have a branch that destroys it
[00:55] <StevenK> Just deciding if I want to completly destroy lp.soyuz.pas, and if so where determineArchitecturesToBuild will live.
[01:23] <wgrant> StevenK: It even fixes a critical
[01:23] <wgrant> Ah, no
[01:23] <wgrant> We still use dpkg-architecture for parsing the hint list, of course
[01:24] <StevenK> Yeah
[01:26] <StevenK> wgrant: So I want to destroy lp.soyuz.pas (and I think I do, since the name gives me nightmares), I'm still unsure where to stuff it
[01:31] <wgrant> StevenK: The new code is only 80 lines. I'd do one of two things: a) move it to model.publishing, or b) switch the pubrec arg to an archive and an SPR, and move the whole thing to model.bpb or similar.
[01:32] <wgrant> Or c) switch the pubrec argh and move it to a new small module in adapters
[01:32] <wgrant> Probably c.
[01:32] <StevenK> Yeah, adapters.arch or dpkgarch was what I thinking
[01:33] <wgrant> adapters.buildarch, perhaps
[01:33] <StevenK> Right
[01:33] <StevenK> wgrant: 'Switch the pubrec argument' ?
[01:33] <StevenK> To spph?
[01:33] <wgrant> StevenK: Replace it with an archive and an spr
[01:33] <wgrant> There's no point having it depend on a publication
[01:33] <wgrant> When all it needs is an archive and an SPR
[01:34] <StevenK> Right, I see that.
[01:34] <wgrant> You could even give it a hintlist rather than an SPR
[01:34] <wgrant> To simplify tests further
[01:34] <wgrant> Now that we no longer need to do source/binary name matching
[01:42] <StevenK> steven@undermined:~/launchpad/lp-branches/destroy-pas-yay% bzr di | diffstat -s
[01:42] <StevenK>  6 files changed, 41 insertions(+), 371 deletions(-)
[04:07] <StevenK> wgrant: https://code.launchpad.net/~stevenk/launchpad/destroy-pas-yay/+merge/166639
[04:10] <wgrant> StevenK: Could I convince you to name determineArchitecturesToBuild legally?
[04:11] <wgrant> Also, order should be archive, distroseries, legal_archseries, I think
[04:13] <StevenK> wgrant: Sure, rename to what, since there's like two callsites
[04:13] <wgrant> StevenK: determine_architectures_to_build
[04:13] <wgrant> It's currently disguised as a method
[04:13] <wgrant> Also, I thought there was only one non-test callsite
[04:14] <StevenK> Right, the second callsite is the test
[04:16] <StevenK> wgrant: http://pastebin.ubuntu.com/5718806/
[04:16] <wgrant> This will mean that things that go through unapproved or new might actually get the same sort of builds as autoapproved ones :)
[04:16] <wgrant> And we can remove the stupid builddmaster dirs from haetae and pepo
[04:16] <StevenK> Yeah
[04:17] <wgrant> :)
[04:18] <StevenK> wgrant: Any other issues?
[04:19] <wgrant> I think that's about it
[04:19] <wgrant> It was just a matter of deleting code until I convinced you to fix issues that had been bothering me for years :)
[04:19] <StevenK> Well.
[04:20] <StevenK> I made the mistake of telling infinity I was looking for stuff to do, and he suggested that since I like deleting stuff ...
[04:21] <StevenK> wgrant: Just making sure test_buildarch and test_publishing are happy before I push.
[04:40] <StevenK> wgrant: Can haz +1?
[04:58] <wgrant> StevenK: Sec, cats needed attention
[04:58] <StevenK> wgrant: Duh, that's their default state.
[04:59] <wgrant> StevenK: r=me
[05:41] <StevenK> wgrant: Is that your QA I see?
[05:42] <StevenK> Since Thunderbird is convinced it must keep reading the staging mailbox
[05:45] <wgrant> StevenK: I just rejected lots of stuff
[05:45] <wgrant> To clear out raring's queue