[01:27] <wgrant> StevenK: https://code.launchpad.net/~wgrant/launchpad/builderinteractor-c-b-b/+merge/182290
[01:27] <StevenK> I have fear.
[01:27] <StevenK> Oh, only 700 lines
[01:28] <wgrant> The next one's 1800 lines, but it's just moving code so I'll self-review.
[01:34] <StevenK> wgrant: r=me with one niggle
[01:35] <wgrant> StevenK: Those tests need to be moved and rewritten later in this series
[01:35] <wgrant> I just did enough to make them work here.
[01:35] <wgrant> Thanks.
[01:36] <StevenK> wgrant: Oh, right, they will get burnt down and replaced later. Fine, niggle dropped.
[06:28] <StevenK> wgrant: Hmm, have you added a race?
[06:29] <wgrant> StevenK: Where?
[06:29] <StevenK> wgrant: db-devel buildbot failure
[06:29] <wgrant> No, that's a normal failure.
=== tasdomas_afk is now known as tasdomas
[08:58] <stub> Anyone up for Swift branch review?
[11:53] <cjwatson> wgrant: I still don't quite understand.  What CancellationError do you expect on manager restart with my patch?  AFAICS it will either send abort to a slave that's already aborting, in which case the slave will say "ok then" and carry on, and the master will just have a timeout that extends from the restart rather than the original cancel; or the slave will have died in some other way such that the abort fails, in which ...
[11:53] <cjwatson> ... case my patch will cause the master to immediately resume-or-fail, which seems right.
[12:05] <wgrant> cjwatson: Nevermind, I misunderstood.
[12:06] <wgrant> cjwatson: The slave abort() method should never fail, so you can probably do resetOrFail if it does.
=== tasdomas is now known as tasdomas_afk
=== tasdomas_afk is now known as tasdomas
=== tasdomas is now known as tasdomas_afk
[18:46] <cjwatson> wgrant: I think https://code.launchpad.net/~cjwatson/launchpad/buildmaster-cancel-properly/+merge/177580 is ready now.  Dunno if you care about reviewing e.g. the inlineCallbacks conversion I did after getting lost in a maze of callbacks one too many times.
[18:46] <cjwatson> (Just in checkCancellation, though, not all over)
[18:46] <cjwatson> Need to do some QA on dogfood builders first.
[23:46] <wgrant> cjwatson: I plan to port an awful lot of stuff to inlineCallbacks over the next week or so. Your port looks sane.
[23:47] <cjwatson> Good.  I think we're ready to go on this once we finish fixing recipe builds in launchpad-buildd 115 (see -ops) and deploying everywhere.
[23:48] <cjwatson> Only four hours spent back-and-forthing on that this evening ...