[01:15] <lifeless> beuno: parallel ftw :)
[01:24] <blr> wgrant: better to add a new property for branch/repo view permission on the views, or just check in the method?
[01:24] <wgrant> blr: Check in the method, or guard in the template using the "required" TALES namespace.
[01:24] <blr> ah actually we already have user_branch_visible in the branch case
[01:25] <wgrant> Hm, I wonder why.
[01:25] <blr> but only on productseries
[01:25] <blr> that seems a little odd
[01:26] <blr> although maybe it is unreasonable to expect too much symmetry between product and productseries.
[01:30] <wgrant> Feel free to increase symmetry where it makes sense.
[02:56] <blr> wgrant: is the best way to restrict a branch by setting information_type=InformationType.USERDATA?
[02:57] <blr> that appears to be what the tests in test_branchvisibility use, but it doesn't seem particularly clear.
[03:36] <wgrant> blr: PRIVATESECURITY is technically the common case, but USERDATA probably works.
[03:38] <blr> thanks, I think that's slightly more obvious
[05:09] <blr> wgrant: buildbot is busy, but need to head out - will land that fix when I get back.
[05:10] <blr> and yay for non-wonky comment mail.
[05:11] <wgrant> blr: You can land it now.
[05:12] <wgrant> It'll fail spuriously in a few minutes, though.
[05:12] <blr> wgrant: oh I see, incoming failure :P
[10:43] <cjwatson> Ah, reproduced the launchpad-buildd 131 crash.
[10:43] <cjwatson> dscpath misunderstanding
[11:38] <beuno> lifeless, indeed!   o/
[11:39] <cjwatson> (launchpad-buildd 132 working its way through the system)
[11:45] <cjwatson> wgrant: Do you know of a reason why BaseMailer specifically avoids producing a decent OOPS, and instead just does an ugly logger.warning thing?
[11:45] <cjwatson> It has a comment saying "Don't want an entire stack trace, just some details", but no rationale for why that should be.
[11:46] <cjwatson> I'm tempted to fix that when consolidating with bugs.
[11:48] <stub> cjwatson: Quite possibly to avoid mail spam infinite loops, from when OOPSes were mailed out directly (or am I misremembering?)
[11:50] <cjwatson> Could be.  At the moment, logger.warning turns into an OOPS in scripts (but in this case probably a rubbish one).
[11:50] <wgrant> cjwatson: logger.warning only started doing that recently.
[11:50] <cjwatson> self.logger.warning("send failed for %s, %s" % (email, e)) - right, let's substitute the entire message text of an e-mail into a clause
[11:51] <wgrant> OOPSes were never mailed out directly, AFAIK.
[11:51] <wgrant> But they were stored in the librarian.
[11:52] <cjwatson> Ah, yes, where "recently" -> 5ya.  But more recent than the BaseMailer change.
[11:52] <wgrant> Relatively recent in Launchpad history :)
[11:52] <cjwatson> So I think OOPS with the message as an attachment, logger.info(request.oopsid)
[12:14] <cjwatson> Not entirely sure how to get a suitable request though - we send mail from both the webapp and scripts
[12:21] <cjwatson> Optional request passed to the BaseMailer constructor, maybe
[16:19] <rpadovani> I was thinking... since in my branch I added a new property in the BranchMergeCandidateView class, maybe test should go in TestBranchMergeCandidateView instead of TestBranchMergeProposalMergedViewMixin?
[16:19] <rpadovani> (wrt https://code.launchpad.net/~rpadovani/launchpad/link-revision-merged-mp/+merge/264654)
[16:20] <cjwatson> You could justify putting a test in the former, though I'd still keep the latter test as well since you're changing the template too.
[16:21] <cjwatson> Your link_to_branch_target_commit implementation is broken for git, BTW
[16:22] <cjwatson> And indeed the current tests are broken too - missing getCodebrowseUrl in various places
[16:28] <rpadovani> cjwatson, indeed, and how could I fix the test? I'm not able to understand how I can retrieve the git url... Also the bzr one is broken, cause it returnes the address of the repo without http://bazaar.launchpad.net at start....
[16:30] <rpadovani> I tried some of the strategies you suggested me, but I'm not able to access to context, so I can't take git_repository or other usefuls properties
[16:30] <cjwatson> I don't understand your comment about the context
[16:31] <cjwatson> In the test, bmp is the merge proposal, which is the context for the view under test
[16:31] <cjwatson> So bmp in the test is equivalent to self.context in the view
[16:32] <cjwatson> Let me go write a bunch of inline comments :)
[16:32] <rpadovani> thanks, I'm quite lost :-)
[16:33] <rpadovani> if we will never meet, I will offer a couple of beers ;-)
[16:51] <cjwatson> rpadovani: does that help now?
[16:52] <rpadovani> cjwatson, definitely, I'll take care of them asap, thank you very much :-)
[16:52] <cjwatson> No problem.
[17:05]  * cjwatson decides he isn't likely to make much more progress on refactoring bug notifications today.  EOD time.
[20:58] <blr> morning
[21:16] <rpadovani> o/