=== kadams54 is now known as kadams54-away
=== kadams54-away is now known as kadams54
=== kadams54 is now known as kadams54-away
=== kadams54-away is now known as kadams54
=== kadams54 is now known as kadams54-away
=== kadams54 is now known as kadams54-away
=== rogpeppe2 is now known as rogpeppe
[20:05] <mwhudson> menn0, waigani: email standup then?
[20:06] <waigani> mwhudson: no one's on hangout so I'm guessing so
[20:06] <mwhudson> i am!
[20:06] <mwhudson> but ok
[20:06] <mwhudson> suits me to not have the call tbh
[20:06] <waigani> oh, ha I just logged out
[20:06] <waigani> okay cool, let's do email standup
[20:07] <menn0> mwhudson, waigani: sounds good
[21:09] <menn0> waigani: looking at your megawatcher branch now
[21:23] <waigani> menn0: is this your latest: https://github.com/mjs/juju/commit/dca522e6f5bcc369fbf195cff6aa8cd5bdf2db30
[21:23] <menn0> waigani: yep that's the one
[21:24] <waigani> menn0: and are all the changes in that commit?
[21:24] <menn0> yep
[21:31] <menn0> waigani: just finished reviewing your megawatcher branch
[21:32] <waigani> menn0: thanks
[21:32] <menn0> waigani: looks good overall. you're still going to add tests for services, units, relations etc right?
[21:32] <waigani> menn0: yes. I said that in the email. I just needed a sanity check before putting any more time into it.
[21:33] <menn0> waigani: cooll cool.
[21:33] <menn0> waigani: I like the way the tests look now
=== kadams54 is now known as kadams54-away
[21:39] <waigani> menn0: "// XXX do me" - you need to come up with a better placeholder
[21:55] <waigani> menn0: moving txn to it's own file makes sense. using reflection and cycling through the envCollections set on each transaction is going to add overhead, but it looks necessary. And reflection is only needed in the odd case where the doc is not bson.D/M. Overall looks neater and makes sense to me.
[22:00] <menn0> waigani: glad you like it
[22:00] <menn0> waigani: actually, in most cases we insert docs using structs so reflection will be the normal case
[22:00] <menn0> waigani: but the overhead of a little bit of reflection will be nothing compared to the I/O of the actual transaction
[22:01] <menn0> waigani: I'll do some comparisons of the test run over the state package anyway to make sure I haven't made things a whole lot worse
[22:02] <waigani> menn0: considering this touches basically all transactions, it would be worth checking
[22:07]  * menn0 nods