18:01:27 #startmeeting trove 18:01:27 Meeting started Wed Apr 22 18:01:27 2015 UTC and is due to finish in 60 minutes. The chair is SlickNik. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:28 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:01:31 The meeting name has been set to 'trove' 18:01:50 o/ 18:01:52 Giving folks a few minutes to trickle in. 18:01:56 o/ 18:01:58 Hello all 18:02:00 o/ 18:02:17 (o_o)/ 18:02:21 Meeting agenda at: 18:02:23 #link https://wiki.openstack.org/Meetings/TroveMeeting 18:02:30 o/ 18:02:37 o/ 18:03:01 o/ 18:03:34 #topic Trove pulse update 18:03:47 #link https://etherpad.openstack.org/p/trove-pulse-update 18:04:47 o/ 18:05:32 o/ 18:05:43 almost double the amount of reviews compared to two weeks ago 18:05:50 \0/ 18:06:21 Yeah — doing pretty well on momentum. 18:06:59 Lots more reviews happening :) 18:07:13 And across the board, not just core — which is good to see. 18:07:18 yeah thats awesome, i first thought if it was 2 weeks ;) 18:07:29 o/ 18:07:42 cores also went from 33 to 70 18:07:50 slicknik - why did the "**" disappear from cores? 18:07:50 to 77 I mean 18:08:05 seems to have stopped after March 11th 18:08:11 yeah, awesome — keep up the good work all around! 18:08:54 new changes : changes merged+abandoned is still ~ 2:1 though 18:09:35 dougshelley66: I've been cleaning that up, I can leave that in there if folks find it interesting. 18:09:56 SlickNik, no worries - just wondered why that changed 18:10:16 dougshelley66: I missed doing that week of March 11. 18:11:42 mvandijk: good point, but given that development for liberty just opened - that might not necessarily be a bad thing. 18:12:40 Okay — let's move on. 18:12:49 #topic Liberty Design Summit sessions 18:12:50 As long as the 'merged' number doesn't dip too much, we're doing ok ;) 18:13:55 So this is _the_ week where I need input from folks as to what summit sessions they would want to drive / participate in for the design summit. 18:14:25 I've started adding a few sessions that I think will be important / interesting at https://etherpad.openstack.org/p/trove-liberty-proposed-sessions 18:14:57 But I need others to step up with suggestions and options too. 18:15:32 was going to say the cluster one 18:15:33 but you already ahve it 18:16:25 edmondk: Yes, I put that one in already based on the discussions we've had so far. 18:17:40 In fact, maybe we should take a few minutes time after this meeting to go flesh that out. 18:18:18 If you'd like to discuss / participate in a session feel free to put in just a line / two for the title and goal. 18:18:37 Others and myself can add in discussion points. 18:20:51 Any ideas folks want to propose? Or any other questions regarding this? 18:22:04 . 18:22:28 Okay, let's move on. 18:22:39 #topic Implement dangling mock detector for unittests (and recommended mocking guidelines - see the below spec for more details) 18:22:46 I am looking for some feedback on the proposed mocking strategy. 18:22:46 I think it would be helpful to have a set of general guidelines on which 18:22:46 approach (see 'Recommended Mocking Patterns') to use when. 18:22:46 If we got an agreement I would submit my detector code together with a 18:22:46 sample usage and fix for review and set up an Etherpad page where people 18:22:47 could sign for modules they'd like to cleanup. 18:23:03 #link https://review.openstack.org/#/c/176158/ 18:24:07 pmalik: I like the idea — we've had a lot of random failures in unit tests because of incorrect mocking. 18:24:25 I also think that we should cleanup all test cases as a part of this project as oposed to just fixing the leaked mocks. 18:24:46 I agree, it's a good excuse to do some overdue housekeeping 18:24:47 yeah, and many-a-times it creates some problem while writing other tests too, have been a victim 18:25:13 It would be good if we all followe the same mocking style. 18:25:41 I tried to propose something in the spec, but of course it is not hard solid... 18:26:41 Have people had a chance to review the suggestions? 18:26:43 we should prefer to inject items into classes vs mocking out globals as well 18:26:44 pmalik: "fixing all tests" doesn't seem specifically actionable. 18:26:51 (I have and tend to agree with them) 18:27:11 pmalik: without having a plan as to what we are actually fixing. 18:27:53 The tests are in packages. Many packages contain only a couple of files. 18:28:02 SlickNik: I think he means to do away with the 'orig=method method=mock method=orig' type used in the past 18:28:06 We need to fix leaked mocks for sure. 18:28:16 yeah we do need to fix them 18:28:19 SlickNik: as part of the cleanup 18:28:42 but I was also thinking about cleanup existing mocking - see peter's comment. 18:29:20 We should really be using the .start(), .stop() methods for those. 18:29:40 Or context manager/decorators for smaller scopes. 18:30:11 pmalik: Do you have a link for the .start() and .stop() methods? 18:30:21 (some sort of doc) 18:30:30 SlickNik: I think there's an example in the spec 18:30:54 Under 'Recommended Mocking Patterns' 18:31:05 see 26.4.3.5. in https://docs.python.org/3/library/unittest.mock.html 18:32:25 Does this exist in 2.7 as well? 18:32:48 (That doc link is for version 3.3) 18:32:58 #link https://docs.python.org/3/library/unittest.mock.html 18:33:36 I've been using it in 2.7. I believe it's the same mock library... 18:34:08 Frankly, I'm less excited about going back and fixing all of the 'orig=method  method=mock  method=orig' style used in the past. 18:34:26 https://pypi.python.org/pypi/mock 18:35:57 Because it's unit tests that already work, and there is a lot of other feature work in Trove that could use effort. 18:36:36 +1 18:36:51 there are better areas to use the effort 18:37:03 ok, so, you suggest using the proposed guideline for new ones only? 18:37:40 I tend to agree with SlickNik and sushilkm - we should fix the broken ones and move on. Establish the new standard going forward 18:37:50 And I would suggest that reviews require modified tests to be converted as well. 18:38:00 pmalik: yes that would be my preference. 18:38:32 i wud agree on using the guideline for new-comers in testing 18:38:50 new-comers=new-unit-tests 18:38:59 pmalik: can we get the info that's in the spec on to a wiki page? 18:39:15 Or into the trove dev docs. 18:39:23 Actually the latter would be better. 18:39:33 sure, why not 18:40:42 Sounds good. 18:40:43 ok, I will submit a review with the proposed implementation then. 18:41:13 pmalik: want to action yourself? 18:41:52 #action pmalik submit a review with the proposed mocking implementation. 18:41:58 ok thanks 18:42:12 Thank you! 18:42:25 #topic Open Discussion 18:42:54 Any items for open discussion? 18:44:12 Looks like we can end a bit early. 18:45:04 #action Everyone please fill out Design summit session suggestions at https://etherpad.openstack.org/p/trove-liberty-proposed-sessions 18:45:26 ^^ reminder — please take a few minutes to do that today. 18:45:31 #endmeeting