[00:30] <hazmat>  niemeyer pong
[00:30] <niemeyer> hazmat: Heya
[00:30] <niemeyer> hazmat: Was just wondering about the boolean stuff, but Clint already mentioned it's still on going
[00:32] <hazmat> hmm. wasn't appearing in the kanban view the other branch masked it
[00:34] <niemeyer> hazmat: Most of the remaining issues seem related to it
[00:34] <niemeyer> hazmat: Also a couple of new cases related to symlinks
[00:47] <SpamapS> niemeyer: I've only seen basic symlinks in charms.. the kind where they link  foo -> ./common.sh
[00:47] <niemeyer> SpamapS: Yeah, these are still fine
[00:48] <niemeyer> SpamapS: We'll prevent only the symlinks pointing out of the charm
[00:48] <niemeyer> SpamapS: The are only two occurrences in the entire repo
[00:48] <niemeyer> There are
[00:49] <SpamapS> I'd bet those are not in the reviewed charms
[00:49] <niemeyer> SpamapS: I've sent the output of the import round to the list
[00:49] <niemeyer> SpamapS: It's
[00:49] <niemeyer> ----- lp:~yellow/charms/oneiric/buildbot-slave/trunk
[00:49] <niemeyer> and
[00:50] <niemeyer> ----- lp:~openstack-ubuntu-testing/charms/precise/nova-cloud-controller/trunk
[01:21] <niemeyer> It's really unfortunate that we didn't start revisions on 1'
[01:41] <jimbaker> finally, reproduced bug 912812 in a simple test
[01:41] <_mup_> Bug #912812: Error condition on relation hooks locks events processing <juju:In Progress by jimbaker> < https://launchpad.net/bugs/912812 >
[01:41] <jimbaker> (about time ;) )
[01:44] <_mup_> juju/purge-queued-hooks r454 committed by jim.baker@canonical.com
[01:44] <_mup_> Added test to verify queued hooks are purged after error
[01:50] <niemeyer> jimbaker: What was it about?
[01:53] <jimbaker> niemeyer, per the above bug, there may be queued hooks (say -changed after a -joined) that are waiting to run. but what if -joined exits with an error? the workflow (and its underlying state machine) at this point is in an error and it cannot run any such hooks
[01:55] <niemeyer> jimbaker: Hmm, right.. but isn't that expected?
[01:56] <jimbaker> niemeyer, sure. hence the bug
[01:56] <niemeyer> jimbaker: Sorry, I mean.. isn't the fact that it stops event processing expected?
[01:56] <niemeyer> jimbaker: What's the bug?
[01:56]  * niemeyer reads
[01:56] <jimbaker> niemeyer, the state machine freezes
[01:56] <jimbaker> and the unit is consequently dead
[01:57] <niemeyer> jimbaker: Ah, ok.. the problem isn't that it stops processing events, but that it stops processing *EVERYTHING* ;-)
[01:57] <jimbaker> niemeyer, correct
[02:24] <_mup_> juju/purge-queued-hooks r455 committed by jim.baker@canonical.com
[02:24] <_mup_> Ensure accompanying -changed hook does not execute if -joined fails
[14:19] <jorge> could anyone help me with a juju test to deploy hadoop-master charm? the detailed scenario is pasted here http://pastebin.com/CrBab8RY ... Thanks! :)
[15:13] <m_3> jorge: hi, I'll give it a try on my end... it looks like it can't see the ppa
[15:30] <jcastro> http://askubuntu.com/questions/102698/juju-wont-run-on-orchestra-due-to-ssh-error
[15:30] <jcastro> new question if someone has time to check it out
[15:43] <jorge> m_3: hum... but, when I run the same install script manually, from the same instance, it works. the instance has the proxy env variables set up (http_proxy no_proxy ....). so, i would like to understand how juju run the script. maybe this is the problem.
[15:47] <m_3> jorge: yeah, I just got it to 'started' state
[15:48] <m_3> and the tests have been passing for hadoop-master
[15:48] <m_3> try it again if you would... destroy-service, then terminate-machine, then deploy it again
[15:49] <m_3> trying to determine if it was just spurious dns fail or something else
[15:54] <jorge> ok
[18:20] <jcastro> niemeyer: heya, I'm not clear about something when you do these imports.
[18:20] <jcastro> do we need to have those 22 bugs fixed before the store can go live?
[18:21] <jcastro> basically, should I be making an effort to find people willing to help fix those?
[18:21] <m_3> jcastro: I think it's more a warning that these won't be visible to the charm-store
[18:22] <jcastro> ok, so they're not like "we need to fix these now or we are doomed"
[18:22] <m_3> jcastro: I'm watching the official "~charmers" ones closely... the rest shouldn't matter, they're just personal branches
[18:22] <m_3> afaik
[18:22] <jcastro> ok, that clears it up for me, thanks
[18:24] <m_3> I think the boolean may've landed recently... I'll verify, but once it does the remaining `~charmers` branches can be fixed
[18:28] <jcastro> we have a guy setting up juju with orchestra if someone wants to answer these: http://askubuntu.com/questions/tagged/juju?sort=unanswered&pagesize=50
[18:30] <SpamapS> niemeyer: will there be somewhere that people can go to see the reason their charm failed to import?
[18:33] <m_3> SpamapS: good question... rather than waiting for reports thru email
[18:33] <SpamapS> even email is dicey IMO.. we don't have a contact field yet
[19:37] <hazmat> SpamapS, we do indirectly
[19:37] <hazmat> if its from a ppa, the owner is known
[19:38] <hazmat> and if its in charmers, the responsible group is known
[19:40] <SpamapS> hazmat: true. I think I'm just anti-email because I have 400 unread
[20:03] <niemeyer> SpamapS: Yo
[20:03] <niemeyer> SpamapS: Sorry, was off for a while
[20:04] <niemeyer> SpamapS: Absolutely.. the only reason I've been sending those reports to the list is because I've been working on the importer and the store itself
[20:04] <niemeyer> SpamapS: Tailoring restrictions and fixing bugs as I go
[20:04] <niemeyer> SpamapS: So it feels like a good idea to share the whole-repo intermediate results so that the repository may be improved while we don't have the final thing
[20:05] <niemeyer> SpamapS: In the future, juju publish will give immediate feedback
[20:06] <m_3> jcastro: post sounds good.. ship it
[20:06] <jcastro> <3 thanks
[20:06] <m_3> I think it's asking for help in exactly the right way... getting peeps to think about what sort of customizations they might already do
[20:06] <jcastro> right
[20:07] <m_3> ... and sharing them
[20:07] <jcastro> so if like everyone is saying "everyone knows you use nginx with wordpress"
[20:07] <jcastro> then we should ship that!
[20:07] <m_3> right
[20:07] <jcastro> of course people will argue about the nitnoid details, but even if you follow most of the things people agree on it's still better than what we have now
[20:07] <m_3> and other choices that people like can just be config options
[20:07] <jcastro> right!
[20:08] <m_3> or fork
[20:08] <jcastro> I should flesh out that detail better.
[20:08] <m_3> that's another post :)