[16:19] <EdwinGrubbs> abentley, bac, jml, rockstar, mars: are any of you around to review a one-line buildbot testfix? https://lpbuildbot.canonical.com/builders/lp/builds/1113/steps/shell_7/logs/summary
[16:19] <rockstar> EdwinGrubbs, fire away
[16:19] <bac> EdwinGrubbs: lost by a hair
[16:19] <bac> better reviewer coverage than most weekdays...
[16:20] <abentley> rockstar, enjoy :-P
[16:20] <rockstar> Man, look at all yous guys that ain't out enjoyin' your weekends!  :)
[16:22] <EdwinGrubbs> rockstar: oops, I didn't realize I pasted the wrong url. https://code.edge.launchpad.net/~edwin-grubbs/launchpad/need_from_future_import_with_statement/+merge/31458
[16:23] <rockstar> EdwinGrubbs, r=me, but please file a bug and send out an email about this.  abentley broke the build, but it's not really his fault...
[16:24] <rockstar> I can't believe we're running buildbot in a different version of python than we're running on production.
[16:24] <EdwinGrubbs> rockstar: ok, thanks
[16:24] <abentley> rockstar, oh geez, really?  I thought we were all switched over.
[16:24] <rockstar> abentley, not according to the test failure.  EdwinGrubbs had to do "from future import with_statement"
[16:25] <abentley> rockstar, I couldn't understand what was going on in the test failure.
[16:53] <mars> rockstar, everything in the build system *except* developer boxes is still on Python 2.5
[16:54] <rockstar> mars, expand "developer boxes" because technically, my machine is a developer box.
[16:54] <mars> And it is really really easy to use awesome 2.6 features and forget that "you can't land that"
[16:54] <rockstar> Does that mean all our boxes in the datacenter?
[16:54] <mars> rockstar, nope, just your local machine I believe
[16:55] <mars> In the data center we have one buildbot slave running 2.6/lucid
[16:55] <mars> the others are all 2.5/hardy
[16:55] <mars> devpad is 2.5/hardy
[16:55] <rockstar> mars, but production is running 2.6, right?
[16:56] <mars> rockstar, hmm, I don't know, but I would think not.  The next "big thing" is to get staging ported to Lucid/2.6.  IIRC then the app servers were next in line.
[16:57] <rockstar> mars, okay, so the buildbots prevent us from deploying code that can't be run.  That makes a bit more sense.
[16:57] <rockstar> Still kinda hairy, but makes sense.
[16:57] <mars> hehe
[16:58] <mars> yeah, I thought about putting a Py2.5 pyflakes check in the linter, but the changeover is happening
[16:59] <mars> a pyflakes-2.5 check would have caught this, but I think it is only the second time this has happened
[16:59] <mars> and people aren't using things like namedtuple yet
[17:00] <mars> the 2.5 feature set appears to be very heavily ingrained in our minds