[04:51] <nigelb> Morning
[04:51] <StevenK> O hai
[04:51] <nigelb> \o/ 20 landings
[05:16] <nigelb> StevenK: How do I know when my db patch is ready for landing again?
[06:22] <wgrant> nigelb: You don't, really. I'll try to arrange the remaining deployments on Monday.
[06:22] <nigelb> wgrant: cool, okay.
[07:10] <StevenK> Well, it can be worked it out. But only if you work for us.
[07:17] <nigelb> StevenK: I'm happy to. Except I seem to headed to set records for number of rejections :P
[07:25] <wgrant> Need to arrange cocoplum/germanium/librarian/codehosting/mailman deployments. Shouldn't be too hard.
[08:42] <lifeless> hmm
[08:42] <lifeless> why was registry removed from ~launchpad-translators?
[08:57] <wgrant> lifeless: It looks like it was removed from *everything* :/
[08:58] <wgrant> Hm, it's actually not been a member of much for ages.
[08:59] <wgrant> Ah, ~launchpad is in ~rosetta-admins, so we indirectly own ~launchpad-translators. Not sure we were a member, though.
[08:59] <wgrant> And oh wow that is a lot of mail.
[09:01] <StevenK> Oh?
[09:02] <wgrant> Roughly 10 shitloads of cronspam and wikispam.
[09:05] <StevenK> That much? :-P
[09:06] <nigelb> You never have too much email.
[09:06] <nigelb> Unless you have Nagios
[09:08] <StevenK> Right, Chapter Master finished
[09:11] <LPCIBot> Project devel build #1,082: FAILURE in 4 hr 16 min: https://lpci.wedontsleep.org/job/devel/1082/
[09:11] <wgrant> :(
[09:11] <StevenK> I think I need to disable Jenkins builds, anyway
[09:12] <wgrant> Why?
[09:12] <StevenK> $$$$
[09:12] <wgrant> :(
[09:12] <wgrant> Probably.
[09:12] <StevenK> And I can't get OpenStack working with Jenkins.
[09:13] <StevenK> And the jclouds plugin, which is the new hotness isn't finished and doesn't build.
[10:03] <LPCIBot> Project db-devel build #875: FAILURE in 3 hr 36 min: https://lpci.wedontsleep.org/job/db-devel/875/
[10:22] <lifeless> mtaylor: ^
[11:03] <cjwatson> so how do I upload to dogfood for QA purposes?  it doesn't seem to have a poppy instance running - at any rate, nothing's listening on ftp
[11:05] <wgrant> cjwatson: Try 2121
[11:06] <wgrant> cjwatson: Should also have SFTP on 5022.
[11:36] <cjwatson> wgrant: neither of those is working for me, not even from chinstrap
[11:36] <StevenK> Lesse
[11:37] <StevenK> cjwatson: poppy is now running on mawson.
[11:41] <cjwatson> and on FTP port 21 - thanks, sorted
[11:42] <wgrant> Ah, someone must have fixed the firewall.
[11:44] <cjwatson> what's the rune for running the publisher?  LPCONFIG=dogfood /srv/launchpad.net/codelines/current/cronscripts/publish-ftpmaster.py -d deribuntu -q # ?
[11:59] <wgrant> cjwatson: I normally run publish-distro directly, but yeah, that should work.
[12:01] <wgrant> cjwatson: You're QAing the translations compression change?
[12:10] <cjwatson> yep
[12:10] <cjwatson> if you have the thing you normally run in shell history, that would be ideal, I don't want to be inventing things as I go :)
[12:15] <StevenK> There's a distro-run script somewhere on mawson, too.
[12:16] <cjwatson> not finding it ...
[12:20] <cjwatson> https://dev.launchpad.net/Soyuz/QA was somewhat helpful but only up to a point
[12:21] <cjwatson> (or do I want to use --all-derived?  https://dev.launchpad.net/Soyuz/TechnicalDetails/Publishing)
[12:24]  * cjwatson decides to go ahead with publish-ftpmaster -d deribuntu and see what happens
[12:26] <cjwatson> That seemed to work.  Lots of errors for missing files but they seem unrelated
[12:26] <cjwatson> http://paste.ubuntu.com/691497/
[12:31] <wgrant> cjwatson: Yeah, we don't have the whole pool, because it's slow and not small.
[12:31] <wgrant> cjwatson: That looks fine, then. Thanks for QAing it.
[12:33] <cjwatson> Marked qa-ok.  Thanks for the guidance
[12:36] <StevenK> wgrant: Can you think of any reason why queue-builder shouldn't be removed from the tree and burnt at the stake?
[12:37] <wgrant> StevenK: It's possible we might want to do a rescoring run at some point... maybe.
[12:37] <wgrant> StevenK: add-missing-builds also only works for a particular DAS, I think.
[12:37] <wgrant> If we generalise add-missing-builds to also work over a whole archive, or it does already, then we should rip cMB out of queue-builder, at least.
[12:37] <StevenK> I thought we hadn't run it for years?
[12:38] <wgrant> I think it was running until early last year.
[12:38] <wgrant> In --score-only, at least.
[12:38] <wgrant> I came across its logs a while back.
[12:38] <wgrant> It was a terrible experience.
[12:38] <StevenK> Haha
[12:39] <StevenK> add-missing-builds already takes an archive argument
[12:39] <wgrant> Yes, and I even made it work for the primary archive.
[12:39] <wgrant> But it requires an explicit arch list, which is possibly silly.
[12:39] <wgrant> OK for its original purpose, but that was short-sighted.
[12:41] <StevenK> So I'm free to gleefully rip cMB out of q-b?
[12:41] <wgrant>         sources_published = distroseries.getSourcesPublishedForAllArchives()
[12:41] <wgrant> From q-b.
[12:41] <wgrant> It works over all archives at once.
[12:42] <wgrant> (wow, how did that ever not take a week...)
[12:42] <wgrant> So, q-b still possibly has a purpose.
[12:42] <wgrant> And it's only like 180 reasonably unoffensive lines.
[12:44]  * StevenK is tempted to put the final nail in the coffin of canonical.lp
[12:44] <wgrant> Please don't.
[12:45] <wgrant> I have big plans for initZopeless.
[12:45] <wgrant> But there's more that has to be untangled first.
[12:45] <wgrant> I have a rough idea of how it is going to go down, but need to work out specifics.
[12:45] <StevenK> wgrant: My plan was to move it away so lib/canonical/lp can be deleted.
[12:45] <wgrant> No.
[12:45] <wgrant> That's the easy way out.
[12:45] <wgrant> But initZopeless is pointless now.
[12:45] <wgrant> It should die entirely.
[12:45] <wgrant> Moving it away just hides tech debt.
[12:46]  * StevenK rewrites lib/canonical/lp/ftests/test_zopeless.py, since it's utterly disgusting.
[12:46] <wgrant> I deleted most of that last week.
[12:46] <wgrant> All that's left is that doctest.
[12:46] <StevenK> But it isn't even a doctest!
[12:46] <wgrant> It is a doctest in a unittest!
[12:48] <wgrant> Things that remain:
[12:48] <StevenK> Which is disgusting
[12:48] <wgrant>  - untangling our dbuser change support.
[12:48] <wgrant>  - renaming "Zopeless" mail to non-queued mail
[12:48] <wgrant>    (and detaching it from initZopeless)
[12:49] <wgrant>  - Letting FunctionalLayer users specify a security policy.
[12:49] <wgrant>  - Porting *ZopelessLayer users to *FunctionalLayer, using PermissiveSecurityPolicy
[12:49] <wgrant>  - Delete initZopeless, because it no longer has a purpose.
[12:50] <wgrant> (Zopeless currently means three things: 1) non-queued mail, 2) DB user and isolation level reconfiguration, 3) PermissiveSecurityPolicy)
[12:51] <wgrant> I've got rid of most of the terrible old code, but we still have this awful tangled wrapper.
[12:51] <wgrant> Conflating these three largely unrelated things.
[12:53] <StevenK> wgrant: http://pastebin.ubuntu.com/691519/
[12:54] <wgrant> StevenK: You could do that, but the sole remaining callsite of isZopeless is Zopeless mail, which is next on the list to remove.
[12:54] <wgrant> So I would be surprised if that test wasn't deleted within a fortnight.
[14:58] <nigelb> Gah
[14:58] <nigelb> My nightmare with that bug continues
[15:27] <nigelb> wgrant / StevenK around?
[15:27] <nigelb> Hrm,looking at the time, I guess not.
[15:40] <LPCIBot> Yippie, build fixed!
[15:40] <LPCIBot> Project db-devel build #876: FIXED in 4 hr 50 min: https://lpci.wedontsleep.org/job/db-devel/876/
[16:09] <nigelb> If anyone has a few minutes to review an MP for a regression that'd be great! https://code.launchpad.net/~nigelbabu/launchpad/js-regression-849121/+merge/75858
[20:33] <lifeless> and again, deactivated. wtf
[20:34] <lifeless> should be ban untaintableangel ?