[02:42] <StevenK> wgrant: Pushed a pruner branch up, I fits after switch-bmp and before drop-branchmergeproposal
[02:43] <StevenK> wgrant: I'm waiting to see if the scanner actually loves the branches
[02:44] <wgrant> StevenK: I pushed a couple of branches this morning, so it should be fine.
[03:27] <wgrant> StevenK: How well does BulkPruner handle long queries like the one in PreviewDiffPruner?
[03:31] <wgrant> StevenK: Can you push -r-2 and repush https://code.launchpad.net/~stevenk/launchpad/switch-bmp-to-previewdiff-merge_proposal to get the diff updated?
[03:32] <StevenK> wgrant: From BulkPruner's docstring: "This is designed for the case where calculating the list of items is expensive, and this list may be huge. For this use case, it is impractical to calculate a batch of ids to remove each iteration."
[03:32] <wgrant> StevenK: Aha
[03:32] <StevenK> wgrant: I can't really push from a pipe'd branch
[03:33] <wgrant> StevenK: Why not?
[03:33] <wgrant> It's not that magical.
[03:33] <wgrant> Switch to the relevant pipe
[03:33] <wgrant> Then push
[03:34] <StevenK> The push branch in bzr info still includes .bzr/branches
[03:36] <wgrant> StevenK: Sure, that automatic push branch will be wrong
[03:37] <wgrant> So don't use the automatic push branch :)
[04:03] <StevenK> wgrant: I need --overwrite with -r -2?
[04:04] <wgrant> StevenK: Yes
[04:05] <wgrant> StevenK: Only a push of a superset will work without --overwrite
[04:22] <StevenK> wgrant: Hopefully that fixed switch-bmp
[04:31] <wgrant> StevenK: Indeed
[04:31] <StevenK> wgrant: I've re-sync'd the pipeline to fix your concern with 44-1
[04:31] <wgrant> StevenK: So, apart from the DB branch I think that's fine. Normally now I'd do a full trial migration on DF just to catch any gotchas, given that it won't get a full test run on qastaging until most of it's already deployed to production (and the DB)
[04:33] <wgrant> StevenK: Are you going to move the NOT NULLs into 44-2?
[04:33] <StevenK> wgrant: That was your concern with 44-1
[04:33] <wgrant> StevenK: 44-2 is an fdt patch
[04:33] <wgrant> 44-1 is live
[04:33] <wgrant> a NOT NULL on a non-trivial table must be fdt
[04:33] <StevenK> And yeah, they'll be in 44-2
[04:33] <wgrant> Right
[04:35]  * StevenK reaches for mawson
[04:36] <StevenK> wgrant: Anything surprising in ZTK 1.1.6?
[04:36] <wgrant> StevenK: No, just a few minor upgrades
[04:37] <wgrant> Our stuff also mostly works with ZTK 2.0a1, once I hacked buildout 2 to be sane.
[04:37] <StevenK> Hah
[04:37] <StevenK> Bleh
[04:37] <wgrant> (by "mostly works" I mean RootObject:index.html renders)
[04:37] <StevenK> The pruner stuff is infecting drop-branchmergeproposal-merge_diff
[04:37] <StevenK> Now I have to resubmit it
[04:38] <wgrant> Or you could just pretend that drop-branchmergeproposal-merge_diff isn't really part of the pipe any more
[04:38] <wgrant> Given you're basically at the end anyway
[04:38] <wgrant> uncommit the merge?
[04:38] <StevenK> Not that simple
[04:38] <wgrant> It's not the last branch?
[04:38] <StevenK> Given r16611 is the last revision which changes the DB patch
[04:39] <StevenK> And r16609-r16610 are merges from the prune branch
[04:46]  * StevenK waits for mawson to finish WADLing
[04:57] <StevenK> wgrant: Do you still require the lib/lp/soyuz/scripts/publishdistro.py hack on DF?
[04:59] <wgrant> StevenK: The one to publish just my PPA?
[04:59] <StevenK> wgrant: Yup
[04:59] <wgrant> StevenK: If it's just that, then kill it
[05:25] <StevenK> wgrant: DF populated, 108020 previewdiffs with NULL branch_merge_proposal and date_created
[05:25] <wgrant> StevenK: And how many not null?
[05:26] <StevenK> Oh, sorry, that is NOT NULL
[05:26] <StevenK> 178059 are both NULL
[05:31] <wgrant> StevenK: Does 108020 roughly equal the number of merge proposals?
[05:32] <StevenK> 114550
[05:32] <StevenK> Close enough
[05:34] <wgrant> Right
[05:35] <StevenK> wgrant: Let me slay 178059 previewdiffs
[05:35] <wgrant> StevenK: Perhaps not
[05:35] <StevenK> Oh?
[05:35] <wgrant> StevenK: That part of the testing is difficult to revert
[05:35] <wgrant> Everything else is easy to redo later
[05:35] <wgrant> Deleting them is not :)
[05:36] <wgrant> So I'd skip that piece of testing until it's unavoidable.
[05:36]  * StevenK drags DF up to switch-bmp
[05:47] <nigelb> Hrm, is there any way I can get the api I made for bug status in comments made more public?
[05:49] <wgrant> nigelb: What do you mean "more public"?
[05:49] <nigelb> wgrant: Um, can somethign that's not LP hit it?
[05:49] <nigelb> It's one of those +link urls.
[05:49] <wgrant> +check-links?
[05:50] <wgrant> I don't see why not.
[05:50] <nigelb> Oh.
[05:51] <nigelb> Stupid question - how do I hit it?
[05:51] <nigelb> launchpad.net/+check-links?
[05:51] <wgrant> +check-links should exist on every object, I think
[05:56] <wgrant> https://launchpad.net/+check-links?link_hrefs={%22branch_links%22:%20[%22/+branch/foo%22,%20%22/+branch/launchpad%22]}
Launchpad requires a <code>REFERER</code> header to perform this action. There is no <code>REFERER</code> header present. This can be caused by configuring your browser to block <code>REFERER</code> headers.</p>
[05:58] <nigelb> (hitting it with requests)
[05:59] <wgrant> nigelb: GET should work
[05:59] <wgrant> For POST you'll need to forge a referer
[05:59] <nigelb> aha
[06:02] <StevenK>     raise LookupError(aliasID)
[06:02] <StevenK> LookupError: 51928797<br />
[06:02] <StevenK> BAH
[06:17] <StevenK> wgrant: I guess I need to create a BMP and then run the BMPJ to get a real previewdiff on DF?
[06:17] <wgrant> StevenK: Or create a BMP and fake the BMPJ
[06:17] <wgrant> I'd fake the BMPJ
[06:17] <StevenK> I have a bunch of BMPs
[06:18] <wgrant> DF can in theory talk to parts of prod codehosting, but probably not directly to the branch store
[06:18] <wgrant> So create the PD manually with fake content, however the BMP does it.
[06:31] <StevenK> wgrant: /srv/launchpad.net/dogfood-logs/2013-05-10/OOPS-e4a155a1aec93e2714a3a720edf626f7
[06:38] <wgrant> StevenK: That's why I said not to use a BMPJ
[06:38] <StevenK> Bleh
[06:39] <czajkowski> morning
[06:46] <StevenK> wgrant: https://code.dogfood.launchpad.net/~stevenk/launchpad/skip-private-dev-focus-mishmash/+merge/131124
[06:47] <StevenK> wgrant: http://pastebin.ubuntu.com/5650261/ is relevant too
[06:48] <wgrant> StevenK: And date_created is set?
[06:49] <StevenK> 2013-05-10 06:40:50.0436
[06:49] <wgrant> Excellent
[06:55] <StevenK> wgrant: So I can land 40-0?
[06:56] <wgrant> StevenK: What's your full deployment plan?
[06:56] <wgrant> The sequence of landings, deployments and migrations.
[07:00] <StevenK> wgrant: http://pastebin.ubuntu.com/5650283/
[07:10] <wgrant> StevenK: Sounds good
[07:10] <wgrant> (plus the pruner after that)
[07:11] <wgrant> StevenK: https://code.launchpad.net/~wgrant/launchpad/bug-1124352/+merge/163270
[07:11] <StevenK> wgrant: Pruner is step 5
[07:11] <wgrant> Er yeah
[07:11] <wgrant> That
[07:14] <StevenK> wgrant: You allow setting of buildd_secret in the ZCML
[07:14] <StevenK> For IArchiveAdmin
[07:16] <wgrant> StevenK: That's necessary because setting private sets buildd_secret
[07:16] <wgrant> But no forms exposes it, and it's not on the API
[07:17] <StevenK> Right
[07:17] <wgrant> Locking it down further in ZCML will cause lots of test failures
[07:17] <wgrant> But we should do it eventually
[07:17] <StevenK> wgrant: +98/-97 disappoints me
[07:17] <StevenK> But r=me
[07:17] <wgrant> Huh
[07:17] <wgrant> I assumed it'd be net +100 or so
[07:17] <wgrant> What did I do...
[07:18] <wgrant> All the buildd_secret stuff, I guess
[07:18] <wgrant> This wasn't meant to be even close to negative :)
[07:18] <wgrant> A convenient accident
[07:18] <StevenK> Haha
[07:20] <wgrant> I'll create the celebrities on all four instances.
[07:28] <wgrant> StevenK: Can you QA your changeOverride thing?
[07:28] <wgrant> I hope to deploy tonight