[11:41] <gary_poster> bac, morning.  did you switch tarmac to look for the well-known file, but we still need to add it for tarmac?
[11:41] <bac> gary_poster, both are updated.  i'm trying to figure out why it isn't finding it
[11:41] <gary_poster> bac, weird, k thanks
[11:42] <bac> and when i google for info about setting up the verify_command, my own wiki page is one of the top hits.  unhelpful.
[11:44] <gary_poster> heh
[11:44] <gary_poster> bac, any bites from rockstar, or hints that he might actually review/merge your branch anytime soon
[12:03] <bac> gary_poster, no.  i forgot to ping him y'day
[12:04] <gary_poster> k, we were busy
[12:04] <bac> gary_poster, him merging the changes would be nice but we do have work-arounds in place
[12:04] <bac> so it is annoying but not impacting us
[12:04] <gary_poster> understood bac.
[12:04] <bac> figured out the pre-commit hook problem
[12:04] <bac> weird
[12:04] <gary_poster> what was it, if it is short to describe?
[12:05] <bac> at 15 and 45 a cronjob pulls new tarmac-puppet branch
[12:05] <bac> but at 0 and 30 a different cronjob actually installs it.
[12:05] <bac> so for 15 minutes you have good code that is not being run
[12:06] <bac> the first is run as tarmac but the second as root.  i think it would be nice if the skew were smaller.
[12:09] <gary_poster> huh
[12:09] <gary_poster> weird
[12:09] <gary_poster> bac benji gmb https://plus.google.com/hangouts/_/f135cc1be8fc9bbfbb3d0128a6f81e6403c1c8f4?authuser=1&hl=en-US
[12:09]  * gmb does the firefox dance
[13:55] <gary_poster> bac, do you want to pair on the next steps of the install thing? (No is fine, just asking)
[13:55] <bac> yes, that would be good
[13:56] <bac> if you start a hangout i'll join in a few minutes
[14:09] <bac> gary_poster, ^^
[14:09] <gary_poster> bac https://plus.google.com/hangouts/_/f82bede9fbd65df1dda02cb90f7324f306fd834e?authuser=1&hl=en-US
[14:10] <gary_poster> invited both accounts
[14:10] <bac> sorry,i'd missed it
[14:23] <gary_poster> benji, bac sez ctrl-space (important for emacs) does not work in tb, fwiw
[14:23] <benji> gary_poster: interesting, I'll file a bug
[15:01] <bac> https://code.launchpad.net/~bac/lpsetup/refactor-install/+merge/113753
[15:11] <gary_poster> bac https://bugs.launchpad.net/lpsetup/+bug/1021771
[15:11] <_mup_> Bug #1021771: Write new "update" command <lpsetup:In Progress by bac> < https://launchpad.net/bugs/1021771 >
[15:41] <benji> bac: you've done some work with getting ssh to shut up about ec2 instances that "change", was any of that in the context of ssh called by juju?  I'm trying to automate some juju calls and the unskippable ssh prompts are killing me
[16:42] <bac> benji, trying to parse that sentence.
[16:43] <bac> benji, which prompts?  i only know about using -o StrictBlahBlah=no
[16:59] <gary_poster> bac, when do you want to hang out again?
[16:59] <bac> yep
[17:00] <bac> gary_poster, putting in my earphones
[17:11] <benji> bac: (I was AFK/lunch.)  Yeah, those prompts were what I was talking about.  I was having a problem with juju calling ssh which would then prompt.  The solution I came up with was to use the "sshpass" program to trick ssh into thinking a real person typed in it's oh-so-precious "yes"
[17:11] <bac> it's not nice to taunt SSH
[18:08] <gary_poster> bac abspath(expanduser('../foo'))
[18:18] <gary_poster> bac
[18:18] <gary_poster>   --use-http     Force bzr to use http to get the sourcecode branches rather
[18:18] <gary_poster>                  than using bzr+ssh.
[20:51] <gary_poster> benji, how goes integration tests?
[20:52] <benji> gary_poster: they have determined that lpsetup init-host is broken
[20:52] <gary_poster> s/goes/go the/
[20:52] <gary_poster> great!
[20:52] <benji> I like that attitude. ;)
[20:52] <gary_poster> in what way, benji?
[20:52] <gary_poster> :-)
[20:53] <benji> one part of lpsetup is generating ssh keys and then another part tries to do the same thing and falls over
[20:53] <benji> I am trying to figure out where the first part is.
[20:54] <benji> oh! I found it
[20:55] <benji> aparently this code path hasn't ever been run
[20:56] <gary_poster> cool benji.  so you have a pattern for us to follow, and it is ready to merge once the test actually passes?
[20:56] <benji> gary_poster: yep
[20:57] <benji> I did have to move away from using unittest though; SSH's deep need to interact with the user foiled my plans
[20:57] <benji> (ssh is so needy)
[20:57] <gary_poster> great news benji, ty.  I sent an email: bac and I have broken lxcinstall and run out of time.  automated tests will be so nice.  what do you run now? custom harness?>
[20:59] <benji> it's just a script; once we want a second test like this I bet the best thing to do will be to just add another function to the script
[20:59] <benji> or at worst, extract a couple of functions into a module so all the test scripts can use them
[20:59] <gary_poster> sounds good benji
[21:00] <benji> I'm going to push what I have and write up a hand-off email so the euro-yellows can work on it if they have time monday
[21:00] <gary_poster> great benji thank you.  I believe frankban is back Wed, so it is just gmb Monday
[21:00] <benji> k