22:04:24 <gabrielhurley> #startmeeting horizon
22:04:24 <openstack> Meeting started Tue Jun  4 22:04:24 2013 UTC.  The chair is gabrielhurley. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:04:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:04:28 <openstack> The meeting name has been set to 'horizon'
22:04:36 <gabrielhurley> Hello Horizon people!
22:04:46 <gabrielhurley> This one should be pretty short, I think.
22:04:47 <david-lyle> Hello
22:04:48 <jpich> Hello
22:04:51 <bradjones> hey
22:04:56 <gabrielhurley> H1 was a success, H2 is moving along.
22:05:02 <lcheng> hello
22:05:08 <BrooklynChen> hello
22:05:31 <kspear> howdy
22:05:53 <gabrielhurley> I've nominated david-lyle and lcheng for horizon core. Many of you have offered feedback there, and that's all very good. I'll give it another day or two just on principle.
22:06:09 <gabrielhurley> Just wanted to make sure everyone was in the loop in case you missed it on the -dev mailing list
22:06:37 <gabrielhurley> Beyond that nothing's really changed since last week.
22:07:10 <gabrielhurley> I've got a new intern starting here at Nebula who's eating a bunch of my attention currently, but I'm also actively engaged in Heat stuff, so hopefully we can finally get that merged in and close out that blueprint
22:07:23 <gabrielhurley> that's gonna be my focus this week
22:08:19 <gabrielhurley> I can't think of any other general announcements...
22:08:19 <gabrielhurley> #topic blueprints and bugs
22:08:23 <gabrielhurley> is meetbot ignoring me? I dunno... IRC's been flaky today.
22:08:27 <gabrielhurley> ah. there it goes
22:08:35 <gabrielhurley> I do have one note on bugs
22:09:28 <gabrielhurley> I pinged dolph about the keystone regressions that showed up in Grizzly which we've been suffering from and he's gonna make sure those get fixed and backported.
22:09:29 <gabrielhurley> specifically:
22:09:35 <gabrielhurley> https://bugs.launchpad.net/keystone/+bug/1170186
22:09:37 <uvirtbot> Launchpad bug 1170186 in horizon "Unscoped tokens are revoked when assigning a role to a user" [High,Confirmed]
22:09:41 <gabrielhurley> https://bugs.launchpad.net/keystone/+bug/1161963
22:09:44 <uvirtbot> Launchpad bug 1161963 in keystone "Editing User fails when the user already has a Primary Project" [Medium,In progress]
22:09:47 <gabrielhurley> and   https://bugs.launchpad.net/keystone/+bug/1170649
22:09:48 <uvirtbot> Launchpad bug 1170649 in keystone "Removing a user from a project would result to all members of that project to be removed" [Critical,In progress]
22:10:12 <gabrielhurley> so that's good news on those
22:10:23 <gabrielhurley> anyone else have updates they'd like to give on blueprints (or bugs)?
22:11:10 <david-lyle> https://bugs.launchpad.net/horizon/+bug/1186379 question
22:11:11 <uvirtbot> Launchpad bug 1186379 in horizon "url_for() function needs refactoring of admin flag" [Undecided,Confirmed]
22:11:11 <lcheng> gabrielhurley: dolph remove the token revocation call when adding roles to user. That is merged.
22:11:48 <BrooklynChen> the ceilometer bp https://blueprints.launchpad.net/horizon/+spec/ceilometer
22:11:55 <lcheng> However, adding/removing  users to group still revokes the token. https://bugs.launchpad.net/keystone/+bug/1187359
22:11:56 <uvirtbot> Launchpad bug 1187359 in keystone "Tokens are revoked when assigning a role to a group" [Medium,Confirmed]
22:13:29 <gabrielhurley> lemme go one at a time there
22:13:38 <gabrielhurley> david-lyle: what's the question about that bug?
22:13:53 <david-lyle> the submitter would like the option to attempt a fallback endpoint_type, specifically because so many api calls require adminURL, they would like that as the default in settings and then have the option to fallback to public.  Are there any string objections to having a fallback setting as well handled in url_for?
22:14:19 <gabrielhurley> I'd like to maintain the current behavior in the default case, but otherwise I'm fine with refactoring that code
22:14:35 <gabrielhurley> how it arrives at that default behavior can be improved, I'm sure
22:14:51 <david-lyle> yes, purely settings driven changes and some logic around those
22:14:57 <gabrielhurley> yep
22:15:11 <david-lyle> also want to get rid of the admin param as it's not used anywhere and complicates the logic
22:15:13 <gabrielhurley> marked it as H2
22:15:19 <david-lyle> thanks!
22:15:53 <gabrielhurley> lcheng: hopefully that bug will also get fixed and backported. feel free to ping dolph about it. he's a nice guy.
22:16:33 <gabrielhurley> BrooklynChen: I've been meaning to look at the ceilometer code you posted, but I haven't had a chance. perhaps you can try to work with some of the other core folks to get some eyes on it.
22:16:52 <gabrielhurley> kspear, jpich, mrunge: any interest in following up with BrooklynChen about ceilometer?
22:17:06 <BrooklynChen> gabrielhurley:  ok.
22:17:14 <lcheng> Okay. Btw,  I've implemented the v3 auth for keystoneclient. It is up for review: https://review.openstack.org/#/c/21942/ After that get merged, I can go back on domain login support.
22:17:35 <gabrielhurley> lcheng: yep, I actually happened to see that this morning
22:17:52 <BrooklynChen> gabrielhurley:  some questions about the bp. do we need to export charts to pdf? it will introduce other python lib
22:18:30 <gabrielhurley> I'd really rather avoid that, actually
22:19:03 <gabrielhurley> I'm not a fan of reportlab
22:19:13 <jpich> So, is the idea to go with the separate ceilometer dashboard until we can implement the other "integrated" approach that seemed favoured during the summit?
22:19:15 <gabrielhurley> there's another project called PyPDF which is newer that I haven't tried
22:19:21 <gabrielhurley> but I'd like to just leave that out for now
22:19:47 <BrooklynChen> gabrielhurley:  okay. I agree with you.
22:19:50 <gabrielhurley> jpich: to be honest I was hoping we could perhaps go straight to integrating things, but this code is here and it's worth evaluating
22:20:04 <BrooklynChen> that's developed by yolanda.
22:20:56 <BrooklynChen> another question. we need to choose date period for chats, it may need  a datepicker js plugin.
22:21:21 <jpich> gabrielhurley: Fair enough. I'll try to make the time to have a look
22:23:14 <jpich> BrooklynChen: There was an (unmerged) review by vkmc that meant to introduced a bootstrap date picker, maybe worth investigating
22:23:14 <BrooklynChen> jpich:  great
22:23:20 <jpich> BrooklynChen: https://review.openstack.org/#/c/22471/ http://www.eyecon.ro/bootstrap-datepicker/ -- I think vkmc looked at multiple ones and this one the most lightweight that does the job
22:25:37 <BrooklynChen> so we decide to integrate the ceilometer panel into the overview?  I will do that and move the usage tables to the overview page.  any idea about the chart?
22:25:49 <gabrielhurley> jpich: thanks
22:26:39 <gabrielhurley> BrooklynChen: nothing explicit. we've got d3.js in horizon now, so building from there is the way to go. bradjones can probably be of assistance
22:27:13 <bradjones> yeah I can help out with d3
22:27:49 <BrooklynChen> bradjones: thanks. :)
22:28:08 <gabrielhurley> awesome
22:28:09 <gabrielhurley> :-)
22:28:13 <gabrielhurley> I love teamwork
22:28:28 <gabrielhurley> #topic open discussion
22:28:34 <gabrielhurley> questions on anything from other folks?
22:29:00 <ayoung> sorry just saw that I was pinged.  It was my understanding that the Heat team was working on the keystoneclient support for trusts since they needed it
22:29:18 <gabrielhurley> uh... that was the previous meeting ayoung
22:29:22 <ayoung> Hah
22:29:25 <gabrielhurley> lol
22:29:33 <gabrielhurley> but good to know
22:29:38 <gabrielhurley> I *was* actually wondering about that
22:29:48 <david-lyle> there have been a couple cases where the wrong django-compressor version has broken master, does anyone have a good idea as to how to include a javascript sanity test into the gate jobs?
22:30:23 <jpich> Selenium to the rescue?
22:30:48 <gabrielhurley> two options: include a test that runs the django manage.py compress command and verifies the success of it, or write a selenium test that checks it
22:31:01 <gabrielhurley> I prefer the latter (selenium) personally
22:31:34 <david-lyle> it would be a nice addition
22:33:12 <gabrielhurley> yeah, certainly wouldn't hurt
22:33:35 <gabrielhurley> Last call on questions/comments before I wrap the meeting....
22:36:53 <david-lyle> if anyone has time I'd love feedback on https://review.openstack.org/#/c/28855/
22:37:14 <bradjones> BrooklynChen: ping me offline let me know what I can do to help out with celiometer bp
22:37:47 <BrooklynChen> bradjones: I will.
22:46:22 <kspear> #endmeeting
22:46:24 <gabrielhurley> I keep forgetting to officially end these meetings
22:46:27 <gabrielhurley> #endmeeting