=== deryck is now known as deryck[lunch]
_mup_ensemble/new-hook-semantics-departed r181 committed by jim.baker@canonical.com19:09
_mup_Add -departed hook19:09
kapilbcsaller, jimbaker standup?19:23
bcsalleris there something we need to cover today? I'm finishing up stuff I didn't get to this weekend 19:23
kapilbcsaller, nothing in particular.. i was just wanted to touch base and where we are on our tasks from last week, and what we're working on this week19:25
bcsalleralmost done with this darn spec, but we can do a quick call to cover it 19:26
kapilbcsaller, great19:28
kapiljimbaker, skype?19:28
jimbakerkapil,  sounds good re skype19:32
=== deryck[lunch] is now known as deryck
kapiljimbaker, just see your cell phone on skype19:33
jimbakerkapil, that's probably because i'm not logged in :)19:34
jimbakerhold on, super noisy here - i'm at coffee shop while car is being worked on19:34
jimbakerkapil, i see bcsaller on skype but not yet you. split brain?19:36
kapiljimbaker, cloud :-)19:37
kapiljimbaker, not seeing you on skype19:42
kapiljimbaker, or it fails constantly19:42
kapiljimbaker, you should just be able to just modify unit relation lifecycle _execute change hook19:43
kapiland switch the hook path based on change19:43
jimbakerkapil, back now19:53
kapiljimbaker, did you see my comments above about _execute_change_hook?19:53
jimbakerkapil, yes19:53
kapilcool, afaics that should be all that needs modification to support additional separate hooks for change, depart, join19:54
jimbakerthat's actually where i made the change19:54
jimbakerso maybe i'm overanlyzing here then... i just want to ensure that niemeyer's plan of having the existing semantics co-exist with the new semantics can work19:55
kapila change there should do that19:55
kapiland ditching the removal of the CHANGE env var19:56
kapilalso within that method could be done the expansion of join to include modify19:56
jimbakerkapil, agreed on that19:56
jimbakerso basically we can simply execute an arbitrary number of hook executions (seen with yield self._executor(...))19:57
jimbakeralthough in this case, arbitrary means up to two19:57
kapiljimbaker, generically.. yes. specifically we want two19:57
kapiljimbaker, sounds good19:57
jimbakerkapil, sounds good. i was getting worried this was getting too complicated. and i didn't want to change more unit tests than necessary for all the obvious reasons19:58
kapiljimbaker, np19:58
kapili'm going to meander and see i can find some lunch19:58
kapilhmm.. looks like nova is almost done adding lxc support20:10
_mup_ensemble/new-hook-semantics-departed r182 committed by jim.baker@canonical.com20:23
_mup_Keep existing semantics for departed; make it possible to invoke multiple hooks for one change20:23
_mup_ensemble/new-hook-semantics-remove-ensemble-change-env-var r186 committed by jim.baker@canonical.com23:52
_mup_Modified examples for new hook semantics23:52

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!