[07:04] <rog> davecheney: mornin'
[07:09] <davecheney> rog: 'sup!
[07:09] <davecheney> just putting the final touches on the dummy change
[07:09] <rog> davecheney: i looked at the python code yesterday and discovered that provisioner StateInfo problem has a trivial solution...
[07:09] <davecheney> ohh do tell
[07:10] <rog> davecheney: http://paste.ubuntu.com/1053800/
[07:10] <davecheney> that'll wokr
[07:10] <davecheney> work
[07:11] <rog> davecheney: yeah. so obvious now that i know it...
[07:24] <rog> davecheney: this my last day before going on hols for two weeks BTW, so if there's anything you wanna talk about...
[07:39] <davecheney> rog: sorry, was in another screen
[07:39] <davecheney> i tried to do a juju bootstrap today
[07:39] <davecheney> it didn't work (not surprising)
[07:39] <davecheney> what are the steps you think remain between now and having juju bootstrap working ?
[07:44] <davecheney> opp, getting my hurry up call
[07:44] <davecheney> rog, have a good holiday, i'll catch ya in a fortnight
[07:45] <rog> davecheney: i'll try me very best :-)
[07:45] <rog> davecheney: have fun in the meantime, and don't finish everything before i'm back - i don't wanna miss all the fun!
[11:04] <rog> TheMue: hiya
[12:54] <Aram> heyhey.
[13:00] <rog> Aram: hiya
[13:03] <TheMue> Hi Aram, hi rog
[13:05] <Aram> TheMue: rog: does this look right to you? http://go.pkgdoc.org/launchpad.net/juju-core
[13:06] <Aram> I wonder how can we force a refresh
[13:06] <rog> Aram: it only shows subpackages if they've been requested
[13:07] <Aram> I see.
[13:08] <rog> Aram: try it now
[13:09] <Aram> yeah, much better
[13:09] <rog> Aram:  for(i in `{ls  | grep -v '\^.'}) {curl -s http://go.pkgdoc.org/launchpad.net/juju-core/$i > /dev/null &}
[13:09] <rog> :-)
[13:15] <Aram> hi niemeyer
[13:17] <Aram> rog: even better, I did it recursively: http://go.pkgdoc.org/launchpad.net/juju-core
[13:17] <Aram> to refresh everything.
[13:17] <rog> Aram: good call
[13:17] <rog> Aram: we need to add more package-level docs!
[13:17] <niemeyer> Heya!
[13:17] <rog> niemeyer: yo!
[13:20] <Aram> niemeyer: I had the same experience booking my flight, I found a flight on the internet they didn't find, but after I told them about it surely they found it!
[13:21] <niemeyer> Aram: We should bring that kind of stuff with management..
[13:21] <niemeyer> The agency is of limited usefulness if they can't find flights unless we point out that they exist.
[13:31] <TheMue> Aram: Hehe, my flight has been ok so far, but then KLM cancelled one of the flights. Now we look for an alternative.
[13:55] <rog> niemeyer: here's a CL to look at if you fancy it: https://codereview.appspot.com/6336047
[13:56] <rog> niemeyer: one more step towards getting the provisioner actually deployed
[13:57] <rog> niemeyer: i'm very happy about the code move BTW
[13:57] <niemeyer> rog: Sweet!
[13:58] <rog> niemeyer: there's one thing i wasn't entirely sure about
[13:59] <rog> niemeyer: how to handle the distinction between the original StateInfo (as passed on the command line) and the StateInfo as retrieved from the environment
[13:59] <niemeyer> rog: Not sure I get the point
[13:59] <niemeyer> rog: ACtually, I'm sure I don't :)
[13:59] <rog> niemeyer: in the CL it always connects with the original StateInfo, but that might be dubious
[14:00] <niemeyer> rog: Yeah, we'll want to improve that over time
[14:00] <niemeyer> rog: But sounds fine for the moment
[14:00] <rog> niemeyer: ok. i think that perhaps we'll eventually want to connect using the most-recently-retrieved info, but fall back to the original if that fails
[14:01] <niemeyer> rog: If mstate works out, we should be saving the live servers every once in a while
[14:01] <niemeyer> rog: Persistently, so that a restart sees all the servers it knew about
[14:01] <rog> niemeyer: yeah, and if the PA can't get the correct info then everything is broken!
[14:01] <rog> niemeyer: because clients won't be able to connect
[14:02] <niemeyer> rog: Right
[14:02] <rog> niemeyer: so maybe overwriting Provisioner.info is the right thing to do
[14:04] <niemeyer> rog: Sorry, I lost the leap there
[14:04] <niemeyer> rog: You mean replacing it every once in a while with state.StateInfo()?
[14:05] <rog> niemeyer: yes
[14:05] <niemeyer> rog: Ah, agreed
[14:05] <rog> niemeyer: currently we store it in a different field
[14:05] <rog> niemeyer: (and keep the original in origInfo)
[14:19] <niemeyer> rog: Reviewed
[14:19] <rog> niemeyer: thanks!
[14:19] <niemeyer> rog: Currently or in the CL you're proposing?
[14:19] <rog> niemeyer: in the CL i'm proposing
[14:19] <niemeyer> Okay
[14:20] <rog> niemeyer: the "check that instance is started" TODO is pending dave's changes to the dummy provider
[14:20] <rog> niemeyer: which will make it easy to check.
[14:21] <rog> niemeyer: i can't add the test without making conflicting similar changes to dummy
[14:21] <niemeyer> rog: Okay, let's try to get these changes in now then
[14:21] <rog> niemeyer: the dummy changes?
[14:21] <niemeyer> rog: Yeah
[14:21] <rog> niemeyer: i'm not sure it's ready. unless i missed a message.
[14:22] <niemeyer> rog: It's up for review
[14:22] <rog> niemeyer: hmm, i think i must be missing messages
[14:23] <niemeyer> Hmm.. we left the genericOperation private..
[14:23] <niemeyer> s/we/he
[14:24] <rog> niemeyer: weird, i don't seem to have seen *any* of dave's proposal emails
[14:24] <rog> niemeyer: i see your replies to them
[14:24] <niemeyer> rog: https://codereview.appspot.com/6304104/
[14:24] <niemeyer> I think it's good to go in, except for genericOperation
[14:24] <rog> niemeyer: yeah, i've found it
[14:25] <niemeyer> rog: But you can fix that in your branch easily
[14:25] <rog> niemeyer: i don't know that genericOperation needs to be exported actually
[14:25] <niemeyer> rog: Well, how one knows how to build the type otherwise?
[14:25] <rog> ah, except that Env won't show up
[14:25] <rog> yeah
[14:26] <niemeyer> rog: I'll submit it to unblock you
[14:26] <rog> niemeyer: that would be great, thanks
[14:57] <niemeyer> rog: Branch is in
[14:57] <rog> niemeyer: thanks a lot
[14:57] <niemeyer> rog: np!
[14:57] <rog> niemeyer: am just working on live tests for the PA BTW
[14:57] <rog> niemeyer: crossed fingers for the end of the day
[14:58] <niemeyer> rog: Wow, cool
[14:58] <niemeyer> rog: Having it actually firing stuff up would make the week for sure
[14:58] <rog> niemeyer: BTW it would be great to have that resilience stuff in aws. it really slows down testing when 80% of the really slow tests fail...
[14:58] <niemeyer> rog: Yeah, I know
[14:58] <niemeyer> rog: Sorry about that
[14:59] <rog> niemeyer: 's'ok!
[14:59] <rog> niemeyer: just a little nudge every now and again :-)
[15:04] <niemeyer> rog: Hehe :)
[15:04] <niemeyer> Okay, I'll get some lunch.. back soon
[16:16] <rog> niemeyer: https://codereview.appspot.com/6336047 all done, i think. took a little more work than i thought, but i think the tests are better for it.
[16:17] <niemeyer> rog: Looking
[16:28] <niemeyer> rog: Done
[16:29] <rog> niemeyer: lovely. thanks very much for doing it so quickly!
[16:31] <niemeyer> rog: My pleasure
[16:34] <rog> niemeyer: submitted
[16:39] <rog> niemeyer: i've got a preliminary CL that starts the provisioning agent, but i haven't done the tests yet. https://codereview.appspot.com/6333056/
[16:40] <rog> niemeyer: someone might want to take it on while i'm away
[16:44] <niemeyer> rog: Sounds great, thanks a lot
[16:45] <niemeyer> I'll have to step out for a while.. I'm in a bit of trouble here. Have to take some recurrent papers to the bank before the trip tomorrow and we're out of paper
[16:46] <rog> niemeyer: oops. that reminds me - i must submit my expense!
[16:46] <rog> s
[17:06] <rog> niemeyer: ok, there's now one uncompleted test. it's running now... we'll see if it works!
[17:16] <rog> darn, it failed.
[17:16] <rog> oh well, it did actually start the PA, even if it didn't work!
[17:37] <rog> right, i'm off to pack.
[17:37] <rog> catch y'all in two weeks' time
[17:37] <TheMue> rog: Where do you travel?
[17:37] <rog> TheMue: to cornwall
[17:38] <rog> TheMue: in the camper van
[17:38] <rog> TheMue: should be good i hope!
[17:38] <TheMue> rog: And with you bikes.
[17:38] <rog> TheMue: oh yes
[17:38] <rog> TheMue: hope you have a good two weeks
[17:38] <TheMue> rog: *sigh* Sounds very good. Never been there, but I still want.
[17:38] <TheMue> rog: We'll do our best.
[17:38] <Aram> rog: have a great time!
[17:38] <rog> TheMue: don't finish it all before i get back!
[17:38] <rog> Aram: hope so!
[17:39] <TheMue> rog: Out holidays on Ruegen will be in September.
[17:39] <rog> fingers crossed for the weather
[17:39] <rog> gotta go, i hear a voice from below!
[17:39] <TheMue> rog: No, we'll leave the last line of code for you. ;)
[18:39] <niemeyer> rog: Superb, thanks so much
[18:39] <niemeyer> rog: and have a brilliant time there
[20:06] <TheMue> niemeyer: One little CL: https://codereview.appspot.com/6332048. It's needed for the port of Firewall.
[20:06] <niemeyer> TheMue: Looking
[20:06] <TheMue> niemeyer: It's only the ExposedWatcher.
[20:09] <TheMue> niemeyer: The firewall is only used in the PA. So I think the cmd package would be a good home for it.
[20:11] <niemeyer> TheMue: Done
[20:13] <TheMue> niemeyer: Ah, thx, will quickly change it.
[20:20] <TheMue> niemeyer: Is in again.
[20:24] <niemeyer> TheMue: Checking
[20:26] <niemeyer> TheMue: Done
[20:33] <TheMue> niemeyer: And once again in. ;)
[20:34] <niemeyer> TheMue: Man.. :)
[20:34] <niemeyer> TheMue: https://codereview.appspot.com/6332048/diff/8002/state/watcher_test.go
[20:35] <niemeyer> TheMue: Well, nevermind.. it's actually alright there
[20:35] <niemeyer> TheMue: I thought there were more references missing still
[20:35] <niemeyer> TheMue: LGTM.. please just fix the CL description when submitting
[20:35] <niemeyer> TheMue: Btw, the usual convention is
[20:36] <niemeyer> prefix: summary with lowercase first letter
[20:36] <niemeyer> TheMue: For that summary line
[20:36] <niemeyer> TheMue: But that's minor
[20:37] <TheMue> niemeyer: OK, will keep in mind.
[20:37] <niemeyer> TheMue: Thanks!
[20:47] <TheMue> So, time to leave. Have a nice weekend.