[11:40] <wgrant> cjwatson: I could randomise the sequences on DB setup, but few enough tests were affected that I don't think it's going to be a huge problem.
[11:42]  * cjwatson nods
[12:26] <wgrant> cjwatson: So close.
[12:26] <wgrant> In the most obvious test, too :P
[12:28] <cjwatson> Grr.  I ran all the tests but must have been a little out of date.
[19:41] <blr> morning
[20:40] <blr> wgrant: made that wee change to launchpad-buildd when you have a moment. https://code.launchpad.net/~blr/launchpad-buildd/snap_http_proxy/+merge/273355
[21:11] <wgrant> blr: Your os.environ set is bad.
[21:11] <wgrant> And did you manage to confirm the approach with Martin?
[21:11] <blr> wgrant: no he hasn't responded yet.
[21:13] <blr> wgrant: at the snappy sprint presumably
[21:13] <blr> I'll try pinging him again
[21:14] <wgrant> cjwatson: Huh, no idea how that test slipped through, sorry. I ran the entire test suite with all the sequences set to strange values...
[21:14] <wgrant> I wonder if I missed a case in the fixture, will rerun.
[21:17] <cjwatson> wgrant: I think it must depend on the BPPH sequence starting just the right amount offset from the SPPH sequence
[21:18] <cjwatson> Probably should just test that it doesn't get a thing of the wrong type back, rather than specifically None
[21:18] <blr> wgrant: ah, env=os.environ in _reactor.spawnProcess
[21:24] <wgrant> blr: More that you're setting the http_proxy environment variable to "http_proxy=[...]"
[21:26] <blr> ah, so I am... >.<
[21:27] <wgrant> cjwatson: Yeah, will fix.
[21:27] <wgrant> And try to come up with a more diabolical sequence breaking strategy.
[21:43] <blr> wgrant: updating startBuild in the buildd snap test, however can't find the 'apt
[21:43] <blr> ' module
[21:44] <blr> and there's no setup.py or dependency manifest afaict
[21:45] <blr> is that python-apt?
[21:45] <cjwatson> Yes
[21:45] <blr> thanks cjwatson
[21:45] <wgrant> blr: You are meant to run the buildd tests under LP's bin/py, IIRC.
[21:46] <wgrant> But the snap ones may work without that.
[21:46] <cjwatson> It's in the package dependencies ... and yes, what wgrant says
[21:46] <blr> ah, didn't realise that was possible.
[21:46] <cjwatson> There's a comment in the Makefile about it
[21:47] <cjwatson> It's a bit awkward because they don't necessarily always end up being in version sync, but txfixtures isn't packaged IIRC