[01:47] menn0: plz review! :) https://codereview.appspot.com/94900043/ === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === murga__ is now known as psivaa__ [16:51] why do documents in the state/state.go State struct need a handle back to the State? [16:52] er.... not documents, types. such as state.Settings [16:52] which wrap documents === negronjl is now known as negronjl-afk [18:46] is it necessary to use the state runner.txn stuff if I'm updating a document that no other code will touch? [18:46] (state/state.go) [18:47] I guess it's probably necessary to avoid collision with other transactions [19:18] dimitern: https://codereview.appspot.com/93990043 [21:06] REMOVE ALL THE SAFTY CHECKS! [21:06] ) % juju destroy-environment -y $(juju switch) [21:13] davecheney: Good. ⌤ [22:03] https://bugs.launchpad.net/juju-core/+bug/1305386 [22:03] <_mup_> Bug #1305386: state/apiserver: multiple data races [22:22] Bug: https://bugs.launchpad.net/bugs/1305386 [22:22] error: ERROR: Failed to update bug task: Server returned 400 and body: milestone_link: Constraint not satisfied [22:22] Wheee! [23:09] perrito666: nova/api/openstack/wsgi.py from https://github.com/openstack/nova.git [23:11] menn0: is your branch with more comment fixes up somewhere? [23:12] mgz: not yet. i'm seeing a unit test failing consistently on my machine. Unlikely to be related to my change but I'd like to see a clean run before submitting for review just in case. [23:15] menn0: what's the failure? [23:15] mgz: it's a uniter test: TestUniterUpgradeConflicts [23:19] mgz: just to recap, you need a function that can give you the json encoded version of an error, shouldn't this be a method from the errtype? [23:20] menn0: that one does pass for me: [23:20] $ (cd worker/uniter&&go test -gocheck.v -gocheck.f=TestUniterUpgradeConflicts) [23:20] PASS: uniter_test.go:998: UniterSuite.TestUniterUpgradeConflicts 6.855s [23:20] OK: 1 passed [23:20] PASS [23:20] ok launchpad.net/juju-core/worker/uniter 11.379s [23:21] perrito666: the argument for it not being a method, is we want it to handle errors of other types [23:21] as in, try casting to our type of error, and use the code/Name()/message if so [23:21] mgz: i've just tried with trunk and it still fails consistently for me [23:21] if not, the code is 500 and the json we generate with the error string [23:21] mgz: I see, where do you want that ? [23:22] menn0: it's likely something environmental, can you `... 2>&1|pastebinit`? [23:22] perrito666: probably novaservice for now [23:24] mgz: the output is pretty huge: http://paste.ubuntu.com/7369606/ [23:26] menn0: it's a lovely test [23:26] failure is here: uniter_test.go:1813 [23:27] mgz: that's one word for it [23:29] menn0: failure is from l969 in the file really [23:30] why that charm is on rev1 rather than rev2 is an open question [23:34] mgz: looks like Ian and Will have their names all over this file. Ian said he'll help me debug this later on.