[00:01] <thumper> babbageclunk: could I get you to take a look at https://github.com/juju/juju/pull/7766 please?
[00:01]  * thumper goes to walk the dog
[00:01] <babbageclunk> thumper: sure
[00:01] <thumper> thanks
[03:00] <menn0> thumper: easy review pls: https://github.com/juju/description/pull/19
[03:00]  * thumper looks
[03:03] <thumper> hmm...
[03:03] <thumper> I too have a branch that bumps the model type to 4
[03:03] <thumper> model vesion
[03:03] <thumper> perhaps we should get them merged close to gether
[03:03] <menn0> makes sense
[03:04] <menn0> thumper: is yours ready?
[03:04] <thumper> um...
[03:04] <thumper> kinda
[03:04] <menn0> thumper: well if I merge what I have now, and you're not far off, that should be fine
[03:17] <menn0> thumper: would you be able to hit merge on https://github.com/juju/description/pull/19 pls?
[03:20] <thumper> yep
[03:20] <thumper> menn0: done
[03:20] <menn0> thumper: thanks
[03:45] <babbageclunk> axw: thanks for the review!
[03:57] <menn0> thumper: review pls: https://github.com/juju/juju/pull/7770
[03:57] <babbageclunk> thumper: Have you profiled the FullStatus call? Would be interesting to see what hotspots that highlighted.
[05:34] <thumper> babbageclunk: I haven't yet, but I'd like to
=== frankban|afk is now known as frankban
[07:47] <axw> anastasiamac jam: to update the models, list-controllers --refresh could just call the RefreshModels method. we can/should make RefreshModels replace the entire set of models stored in the cache with what's retrieved
[07:47] <axw> anastasiamac jam: still need to get model status to set the machine count though
[07:48] <axw> or we could update the ListModels method to include that in the result
[07:48]  * axw wishes machine count wasn't cached
=== frankban is now known as frankban|afk
=== frankban|afk is now known as frankban
=== frankban is now known as frankban|afk
=== externalreality_ is now known as externalreality
[23:25] <anastasiamac> axw: yes, the idea is similar ecxcept maybe a bit more efficient than RefreshModels. I was going to improve RefreshModels - right now, we make a call to a store to update each model sequentially which also acquires alock for each call..
[23:26] <anastasiamac> axw: i was going to change store to accept a collection of models to update (more efficient when we know that some controllers have a *few* models)
[23:27] <anastasiamac> axw: that call would not be great for 'controllers' because it's controller specific, i.e. we'd end up with an additional api call for each controller in the list.. I think it's best to do what we've agreed on: api call will return desired models info. this will only have 1 api call instead of 1+nXcontrollers
[23:28] <anastasiamac> axw: i'll propose later on today- would appreacite an opinion :D