[01:29] you would think repo.lookup_reference('non-existent-ref') could return None... [01:37] There are valid arguments to both approaches. [01:37] But I tend to prefer returning None, indeed. [04:32] wgrant: would you mind having a look at the awful test test_repo_repack_verify_commits_to_pack in https://code.launchpad.net/~blr/turnip/repack-api/+merge/257312 [04:32] it passes, but with 'fatal: fgets: Input/output error' which is probably a little less than ideal. [04:55] wgrant: oops nvm, that. [04:56] wgrant: but wouldn't mind a quick check over the branch again before I merge it (including that test). [05:20] blr: Hrm, there wasn't a simpler way? [05:24] blr: I'd like to see actual assertions on the numbers of packs, since currently you have a loop over what is presumably one pack. [05:34] wgrant: there are two packs (after pack-objects is called), confirmed if you break and have a look in the repo [05:34] I could certainly add an assertion though [05:35] blr: Yep, just some assertions to that effect would be nice, in case pygit2 decides to become better. [05:35] Don't want it making the test unbreakable. [05:35] it would be nice to have some of this in pygit2 === jamesh_ is now known as jamesh [07:19] wgrant: added a test helper prop @packs and the 2 pack assertion which has cleaned up the test a bit if you want to have a quick look again. [07:29] blr: The "for filename in factory.packs" bit is still weird, because it only uses the last pack that it loops over. [07:29] It overwrites the local on every iteration. [07:45] ah yes, true [07:46] wgrant: added another assertion that there is 1 pack and just used the first element [07:46] that makes more sense :) [07:47] thanks [07:47] Great. [07:47] The test makes sense now! [07:47] yay for sense === jamesh_ is now known as jamesh [12:19] wgrant: Mind having a quick look over my updates to https://code.launchpad.net/~cjwatson/launchpad/git-mp-move-outside-ref/+merge/260105 ? [12:20] And I'm testfixing. [12:20] cjwatson: Amazing. [12:20] Oh, of course, it's keyed on source rather than target branch, so it would never have become a totally huge list. [12:20] Yeah. [12:21] Still. [17:18] hey there, bzr private branches seem to be broken? [17:19] smart server says TypeError _makeBranchTransport() got an unexpected keyword argument 'private' [17:19] er... all bzr branches in fact it seems [17:20] cjwatson: ^if you're still working [17:31] I filed bug 1458948 [17:31] Bug #1458948: bzr smart server operations fail with TypeError [17:32] the world broke..... [17:32] bzr: ERROR: Server sent an unexpected error: ('error', 'TypeError', "_makeBranchTransport() got an unexpected keyword argument 'private'") [17:32] and i'm not pushing to private branch.... [17:32] bug 1458948 [17:32] Bug #1458948: bzr smart server operations fail with TypeError [17:33] Laney: i bet code needs to be pushed to deploy the fix for broken pushes right..... [17:38] could be https://bazaar.launchpad.net/~launchpad-pqm/launchpad/devel/revision/17526 [17:39] ...which is now 503ing