[00:23] <StevenK> wgrant: https://code.launchpad.net/~stevenk/launchpad/hide-build-portlet-archive/+merge/151864
[00:24] <wgrant> StevenK: Have you confirmed that there are no similar cases?
[00:25] <StevenK> On Archive:+index? Yeah
[00:27] <wgrant> StevenK: Great
[00:27] <wgrant> While you're on the PPA pages...
[00:27] <wgrant> Bug #1149544
[00:28] <_mup_> Bug #1149544: PPA +packages timeout when too many package copy errors are shown <Launchpad itself:New> < https://launchpad.net/bugs/1149544 >
[00:28] <wgrant> 15s or so of non-SQL time
[00:28] <wgrant> Rendering PCJ errors
[00:28] <wgrant> Seems a bit odd
[00:29] <StevenK> wgrant: I have a second branch on archive pages already
[00:41] <StevenK> wgrant: I'll look at that bug after I clear my plate. That MP is first on my list.
[00:41] <wgrant> StevenK: I'm not sure it's worth having display_latest_updates, rather than just using context/required:launchpad.View in the template
[00:42] <StevenK> wgrant: Oh, I didn't know about that pattern.
[00:44]  * StevenK tries to swap back in the changes for that branch
[00:45] <StevenK> wgrant: http://pastebin.ubuntu.com/5591925/
[00:49] <StevenK> wgrant: Any other issues?
[00:50] <wgrant> StevenK: Seems fine otherwise
[00:53] <StevenK> Strange that r16519 turns up as an unmerged revision
[00:53] <wgrant> The scan of the last landing probably failed
[00:53] <StevenK> Indeed
[00:53]  * StevenK stabs the branch scanner more
[00:53] <wgrant> Was that the spec notification fix?
[00:53] <StevenK> Yeah
[00:53] <StevenK> name => Name
[00:54] <StevenK> wgrant: MP updated
[00:56] <wgrant> StevenK: Done
[00:56] <wgrant> Thanks
[01:20] <StevenK> wgrant: https://code.launchpad.net/~stevenk/launchpad/preload-subscriptions/+merge/152073
[01:22] <wgrant> StevenK: We do not want to link to Archive:+index
[01:22] <wgrant> As it's completely useless to the subscriber until they've activated their sub to get a token
[01:24] <StevenK> wgrant: Gar.
[01:29] <StevenK> wgrant: So I have to cache a negative permission
[01:31] <wgrant> StevenK: An easy change
[01:31] <wgrant> However
[01:31] <wgrant> You'd need to actually do the view check as well
[01:31] <wgrant> Or you can just fix +archivesubscriptions so it never links :)
[01:31] <StevenK> That sounds good
[01:33] <StevenK> It uses subscription/archive/displayname and subscription/archive/fmt:reference
[01:33] <StevenK> So the PPAFormatterAPI is doing the linking
[01:33] <wgrant> Are you sure that it actually does link?
[01:34] <wgrant> I don't get links
[01:34] <wgrant> And I can see every PPA
[01:34] <StevenK> It looks up View to see if it should link
[01:34] <wgrant> Are you sure?
[01:34] <wgrant> Because it doesn't link even though I do have View
[01:34] <StevenK> The code that's on prod is odd
[01:35] <StevenK> It checks if you're a subscriber if it's private
[01:35] <StevenK> Oh
[01:35] <StevenK> It doesn't link
[01:36] <StevenK> But it should still check permission so it doesn't leak
[01:36] <wgrant> Ah, right
[01:36] <wgrant> So, that bit could be adjusted to SubscriberView
[01:36] <StevenK> Right
[01:41] <StevenK> wgrant: http://pastebin.ubuntu.com/5592006/
[01:43] <wgrant> StevenK: Sounds reasonable.
[02:02] <StevenK> wgrant: Can haz +1?
[02:16] <wgrant> StevenK: Done
[02:39] <StevenK> wgrant: Current plan is to deploy after r16521 is QA'd.
[03:32] <StevenK> wgrant: One failure in tales.txt
[03:33] <StevenK> wgrant: Trying to fmt:link a disabled ppa for anonymous results in <BLANKLINE> now, rather than <span>displayname</span>
[03:35] <wgrant> StevenK: Hm, I guess you'll need to reintroduce the private special case
[05:39] <StevenK> 280 PCJs makes me sad
[05:42] <StevenK> wgrant: I wonder if we want to limit the view method to only return 5 and show somehow that there are X others
[05:43] <wgrant> StevenK: Quite possibly, yes
[05:43] <wgrant> But it's also worth investigating why they take so long to render without queries
[05:43] <StevenK> Because TAL hates life?
[05:44] <StevenK> wgrant: The view method that drags them out of the DB is quite good and preloads everything
[05:45] <wgrant> Sure, but what sort of data do they have?
[05:45] <wgrant> Can't be that huge
[05:45] <StevenK>                   tal:attributes="class python: 'failed' if job.status.title=='Failed' else 'non-failed'"
[05:45] <StevenK> That can't be helping
[05:45] <wgrant> Why not?
[05:47] <StevenK> Let's see. It pulls out job.status, job.package_name, job.package_version, job.source_archive, job.requestor, job.target_distroseries and job.error_message
[22:10] <StevenK> wgrant: OOPS-7ed453c05c38c1f860dcf3c63e47eb6c
[22:56] <StevenK> wgrant: https://code.launchpad.net/~stevenk/launchpad/preload-subscriptions-redux/+merge/152302
[23:01] <wgrant> StevenK: k
[23:33] <cjwatson> wgrant: existing_publishing_status> what would a legal method name be?
[23:34]  * cjwatson digs through style guides
[23:35] <cjwatson> Oh, whoops, you mean existingPublishingStatus?
[23:36] <wgrant> cjwatson: Technically it should be a verb too, but there is precedent for just mixedCasing it, yes
[23:37] <cjwatson> OK.  In an earlier version it was just a class attribute, hence the misnaming
[23:40] <cjwatson> Sod it, I"
[23:40] <cjwatson> 'll be lazy and prepend "get".  There's precedent in the same file
[23:41] <wgrant> Yup
[23:43] <cjwatson> Now, will bzr-pqm work on this new machine ...
[23:49] <cjwatson> Yes.  Excellent.