[01:25] blr: Can you have a look at https://bugs.launchpad.net/launchpad/+bug/1471426? [01:25] Bug #1471426: AttributeError: 'BinaryPatch' object has no attribute 'get_header' [01:28] wgrant: sure [01:32] mwhudson: Is go get happy with qastaging now? [01:38] wgrant: seems to be, although it complains about no go source for the repos I've tested [01:38] mwhudson may want to confirm with an actual golang repo [01:56] wgrant: trying to write a test for this.. in what context do we have binary patches in the diff text? [02:01] blr: You can see a couple in https://code.launchpad.net/~blr/launchpad/project-meta-go-import/+merge/262144 [02:06] cjwatson: sorry, which patches are you referring to? [02:06] blr: gitbranch*.png [02:06] blr: "Binary files [...] differ" [02:08] wgrant: ah I see thanks [02:09] given a binarypatch has no header, do we throw them out? there could be a comment there [02:09] we can't throw out anything, a comment might go anywhere [02:41] blr: seems ok to me, fighting the fun that is qastaging codehosting [02:42] mwhudson: great, glad something is working today >.< [02:42] OOPS-ec59fa242226c1c084ff47bddab04e86 [02:42] on trying to register a new series [02:43] well, actually on trying to view the new series [02:43] https://qastaging.launchpad.net/pydoctor/series [02:43] blr: known? [02:45] mwhudson: hmm no, I wonder why I didn't see that on turnip's series page [02:45] blr: Probably because therer's no default branch on that new series. [02:45] i was about to add one :) [02:47] Oh, there's also a correctness bug in that path. [02:47] Ah, the bugs are one and the same. [02:48] wgrant: in https://qastaging.launchpad.net/~mwhudson/pydoctor/new-trunk ? [02:48] blr: ProductSeriesView.golang_import_spec is buggy. [02:49] A template crashing with a LocationError usually means that the expression raised an AttributeError. [02:49] The TALES interrpreter helpfully swallows the AttributeError for you so you can't see what it actually is. [02:49] right, have seen that before. [02:49] * mwhudson twitches [02:51] wgrant: ah, development_focus can be None [02:51] blr: It can, but that's not the issue here. [02:52] Actually, can it? I forget. [02:52] hmm ok [02:52] It's not the cause of this crash, though, as it is set for pydoctor. [03:22] ok have the binary patch handling sorted, just tidying up this test. [03:49] wgrant: ah, that should just be self.context.branch [03:49] waiting for the binary patch branch to scan [04:14] * blr rounds up the goats [04:15] wgrant: could you have a look please https://code.launchpad.net/~blr/launchpad/golang-meta-import-buggy-productseries/+merge/263994 [04:17] made the product series test clearer as well hopefully. [04:25] blr: You pushed the test change to the wrong branch. [04:25] golang-meta-import-https-fix rather than golang-meta-import-buggy-productseries [04:27] * blr sighs [04:28] thanks [04:28] should have made a new branch [04:28] pushed [04:38] blr: Both reviewed, thanks. [04:38] wgrant: thanks, I believe a BinaryPatch will always have a dirty header [04:39] blr: It's still rather confusing for it to be inside that block. It doesn't need to be. [04:40] yes, agreed. [05:41] wgrant: unblocking buildbot, then will get those two branches landed. [05:42] and hopefully can now figure out why elmo's diff was so mangled. [08:29] wgrant: hmm I wonder if blank newlines are not being stripped, bzrlib won't parse subsequent patches otherwise. [10:12] blr: Have you got a test that reproduces it? [10:13] wgrant: I have a failing test yep [10:13] Great. [10:14] will dig into bzrlib again in the morning. [21:34] huh, https://bugs.launchpad.net/launchpad/+bug/1465467 is deployed now, but i don' [21:34] Bug #1465467: put tags on project, series pages

[21:34] t see the meta tags ? :( [21:42] mwhudson: I see one on https://launchpad.net/germinate [21:43] not on code. though [21:43] cjwatson: hmm [21:43] cjwatson: i was looking at https://launchpad.net/pydoctor [21:43] bzr vs git? [21:44] Yeah, I looked at a git example [21:45] blr: ^- looks like you have more to do here :-/ [21:46] this simple sounding feature seems to be cursed :) [21:53] mwhudson: did you set a default vcs? [21:53] blr: oh right, no [21:54] blr: is there going to be a garbo job to set that to bzr when trunk is already bzr? [21:54] it will only render if you've one into code and set bzr (no need to change any of the form fields) [21:54] mwhudson: good question, that would certainly be nice. [21:54] blr: i think its essential to be able to remove the special case code from the go tool [21:55] mwhudson: ok will discuss that with william and colin in 5m (we have our LP meeting) [21:55] or we'd have to ask the maintainers of every go package on lp to visit the code config page... [21:55] blr: ta [21:55] mwhudson: looks like the tag is rendering now at least :) [21:55] blr: can confirm that that fixes it, yeah [21:55] (forgot about that part from qastaging) [21:56] well, less than obvious I think. Hopefully we can run a job (we have an 'inferred_vcs' that we could potentially use to set product.vcs globally) [22:02] ah, you decided not to make it inferred_vcs? I guess that would have performance problems. [22:02] we'd talked about backfilling, yes. [22:02] cjwatson: yes, I think wgrant was concerned about performance there. [22:03] makes sense. [22:13] mwhudson: we'll look at garboing that in the next few days. [22:13] blr: thanks