17:00:03 #startmeeting Testing Working Group 17:00:03 Meeting started Thu Aug 25 17:00:03 2016 UTC. The chair is gundalow. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:03 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:03 The meeting name has been set to 'testing_working_group' 17:00:12 #chair gundalow mattclay 17:00:12 Current chairs: gundalow mattclay 17:00:19 hi folks 17:00:32 #topic Roll call 17:00:33 Who is here? 17:00:33 Who is new this week? 17:00:40 hey ryansb 17:00:41 ryansb is here 17:00:46 :) 17:00:53 Hi 17:01:00 hey Reaper 17:01:02 Hi gundalow 17:01:03 I'm new to this forum 17:01:06 * mattclay waves 17:01:16 Reaper: Care to introduce yourself? 17:01:57 I'm a network test engineer. 17:02:07 New to the open source community 17:02:11 e.g. How/where you use and what your interest in testing is 17:02:57 We use python as an automation tool to evaluate the modules 17:03:40 Interested to do more python coding and participate in the open source community 17:03:53 Used ansible to write simple playbooks 17:04:10 just to execute very basic testcases 17:04:16 interested to explore more 17:04:52 who else is new/lurking 17:04:59 linuxdynasty: you around? 17:05:02 I am lurking 17:05:15 skg_net: sivel hey :) 17:05:48 #topic Update on internal work 17:05:49 first time to open community,from networking testing background 17:06:01 bruno_cgl: Welcome 17:06:31 gundalow Hi 17:06:51 So quick update on what I've been working on 17:08:37 I'm working on playbooks to build up test VMs for the different network appliances (ansible.com/network). Ansible 2.2 will have a large network module refactor in, so this will help test that 17:09:03 We will also be adding in pep8(ish) checks for these modules once the feature work is done 17:09:41 yeah 17:09:47 im around 17:10:17 mattclay: would you like to give an update 17:11:20 This week I've enabled integration tests for OS X and Python 3 on Shippable. 17:11:52 abadger1999 has been making great progress with Python 3 compatibility, and the tests on Shippable will help prevent regressions. 17:12:32 \o/ 17:12:41 Not so recent additions included enabling integration tests on Shippable for FreeBSD and Windows. 17:14:27 There are still some integration tests for Python 3, FreeBSD and OS X that are disabled, as they're not currently passing. 17:14:43 We'll be working on fixing those over the next few weeks. 17:14:58 Any contributions to help in that area are definitely appreciated. 17:15:02 Which I guess brings us nicely onto asking for help 17:15:41 * gundalow finds the links 17:15:55 For Python3 17:16:00 https://github.com/ansible/ansible/blob/devel/test/utils/shippable/python3-test-tag-blacklist.txt 17:16:03 https://github.com/ansible/ansible/blob/devel/test/utils/shippable/python3-test-target-blacklist.txt 17:16:45 and for freebsd. & osx 17:16:50 https://github.com/ansible/ansible/blob/devel/test/utils/shippable/remote-integration.sh#L17 17:16:57 https://github.com/ansible/ansible/blob/devel/test/utils/shippable/remote-integration.sh#L145 17:17:31 We'd *Really* welcome help on getting those lists reduced 17:17:50 I'm sure there are some modules you use regularly in there and care about 17:18:04 :) 17:18:20 * gundalow looks round at the tumbleweed 17:19:52 we need a topic? 17:20:05 sivel: ? 17:20:20 I have something to talk about, just making sure we are ready to move on 17:20:30 sure 17:20:35 ansible-testing 17:20:57 I received the go ahead from our IP legal team, that I can move that repo 17:21:04 #topic ansible-testing 17:21:19 so assuming that we are still interested, I just need to work with someone to move it to the ansible org 17:21:20 #info < sivel> I received the go ahead from our IP legal team, that I can move that repo 17:21:56 which would require someone with admin access to the ansible org 17:22:33 So would this move to github.com/ansible/ansible-testing / 17:22:50 yes, assuming that everyone is still ok with that 17:23:15 +1 17:23:17 sure, or /ansible-validate-modules 17:23:20 it had been discussed a while ago, but I hadn't gotten the legal go ahead until just before my vacation, so I didn't try to do anything about it then 17:23:38 I had invisioned this going farther than just modules, which is why it had a more generic name 17:23:39 Sure, thanks for progressing with it. We all know legal stuff can take a while 17:23:52 an, OK 17:23:59 I figured we could write other tests for other parts of ansible, and so on 17:24:32 so, assuming you guys tell me who to work with, we can work on migrating it to the ansible org 17:24:43 sure, leave that with me 17:25:36 #action gundalow to work out 1) Do we still want to move sivel/ansible-testing/ (we should) 2) Are we happy with the name 3) Who has github rights 4) Update links 17:25:53 and if needed, I can provide my super awesome confirmation that I am allowed to do this, which is like a 1 line email :) 17:26:02 but I doubt it is needed 17:26:28 "The Rackspace Intellectual Property Committee has approved this request under the terms you identified below" 17:26:30 ;) 17:26:35 haha 17:26:40 haha 17:26:54 excellent 17:27:22 would be cool to have this 'inside' ansible 17:27:22 #topic AnsibleFest & Contributors Summit 17:27:30 (hello btw) 17:27:40 can you point me to a PR which already migrated some test for Python 3 17:27:43 hey jhawkesworth. I think we wil 17:27:57 mattclay: You got a link for skg_net 17:28:32 * mattclay finds a PR 17:28:47 So we will be having our 3rd Contributors Summit - https://www.ansible.com/blog/ansible-contributor-summit-brooklyn-2016 17:28:58 There are agenda and details on the above blog post 17:29:11 This is a great way to help shape ansible 17:29:25 You can attend in person, or online 17:29:43 https://github.com/ansible/ansible-modules-core/pull/4526 17:29:44 If you attend in person you get a massive 50% off your AnsibleFest ticket 17:30:08 That's a PR for actual module fixes for Python 3. 17:30:13 skg_net: example of py3 fixes https://github.com/ansible/ansible-modules-core/pull/4526 17:30:30 Thanks @mattclay @gundalow 17:30:40 I cannot attend in person, but I will likely be able to contribute "online" 17:31:05 skg_net: It would be worth you pinging mattclay and abadger1999 before you work on something incase they are actively working on the same module 17:31:06 skg_net: Here's an example of a PR to enable tests once the modules have been fixed: https://github.com/ansible/ansible/pull/17212 17:31:09 skg_net: great 17:31:35 yeah most likely I'll be on line too 17:31:44 sure will work with them 17:31:55 my wife happens to be out of town during ansiblefest, so no travelling for me 17:32:22 no travel budget for me! 17:32:39 jhawkesworth: Where areyou based? 17:32:45 +1 @jhawkesworth 17:32:47 skg_net: Most changes will be to code. Run the test manually, see what breaks, fix the module code, then submit the PR for the fixed code. 17:33:06 worcestershire, UK 17:33:09 sv--: and say that we can enable the test once the code is merged. 17:33:16 skg_net: ^ 17:33:28 jhawkesworth: See you in London next year then :) 17:34:11 skg_net: Any modules that jump out at you from that list? 17:34:14 that's my plan. Hope I can do a windows workshop or something like if theres enough interest. 17:34:23 ok got it 17:34:38 great 17:35:33 gundalow: will get back to you and others tomorrow 17:35:46 skg_net: Thank you :) 17:36:00 did i miss the majority of the testing group meeting? 17:36:07 caphrim007: hop, still ongoing 17:36:11 k 17:36:18 caphrim007: You can see the IRC logs at the end 17:36:22 k 17:36:25 #topic Improving Documentation 17:36:42 From the voting (and general feedback) https://public.etherpad-mozilla.org/p/ansible-testing-working-group 17:36:54 One of the really good suggestions was 17:37:17 having a decent example of a well written module 17:37:47 So, What are examples, and what are the characterists of a good module 17:38:14 Anyone got any examples of a well written module (it can be your own :P) 17:38:18 did anyone suggest raiding jimic's AF london presentation for the module he wrote to control Hue lights? 17:38:58 his talk was on writing modules so ought to be model 17:39:01 jhawkesworth: nop, I'll make a note 17:39:15 but perhaps a bit 'niche' in purpose 17:39:33 also hard to try out without (expensive in UK) hardware 17:39:33 couple of examples wouldn't hurt 17:39:56 can we codify all of the module checklist in sivil's ansible checker? 17:40:03 or some tool 17:40:10 caphrim007: YES 17:40:14 well, more of it 17:40:27 gotta drop - next time! 17:40:57 can a second ask be that not all the checks be inline in the code, but pluggable so that devs in other orgs can add tests without having to change existing code? 17:40:59 on https://public.etherpad-mozilla.org/p/ansible-testing-working-group (from last weeks meeting) I've said we should split out the checklist and move out the stuff that's checked by ansible-validate-module 17:41:33 should be unittestable -- separate the code that has sideeffects (like anything using module.*) from the code that is purely algorithmic. 17:41:34 for lack of a better example, consider https://github.com/willthames/ansible-lint/tree/master/lib/ansiblelint/rules 17:41:52 i can add more rules if i have something specific i want to check in house 17:41:54 I was about to suggest that 17:42:26 be aware of common parameter names. 17:42:42 Make proper use of parameter types (path, json, bool) 17:43:12 Do more than just wrap a command line tool. 17:43:40 abadger1999: If we can check more with CI it's fewer things that humans need to check/miss :) 17:44:20 Work on python3 and (until March/April of next year) on python2.4 17:45:59 linuxdynasty: How would you feel about us using one of your modules as an example 17:46:43 linuxdynasty: and do you have thick enough skin to deal with us ripping it to shreads to make it bullet proof/solid documentation :P 17:46:53 np 17:46:59 yes i do 17:47:04 excellent 17:47:06 i bite back :P hehe jk 17:47:15 haha 17:47:54 #action linuxdynasty to pick a module to use as an example of a perfect module. Group then to review and ensure that it is 17:48:06 Smaller is better :-) 17:48:22 as no matter the module, we're going to find tons of stuff to change. 17:48:29 I document like a mad man in my functions/methods :( 17:48:43 so I will look for a smallish one :p 17:48:44 linuxdynasty: pick one and ping me 17:48:48 kk 17:48:57 doesn't have to be today 17:49:12 linuxdynasty: no problem -- documentation doesn't count in what I was thinking of. 17:49:16 Mostly about the code. 17:49:23 kk 17:49:34 (although fewer parameters is also god ;-) 17:49:43 I think function/method documentation should also be part of the examples 17:49:49 17:49:56 to motivate module developers to do the same 17:49:57 +1 17:50:00 Do you sphinx markup? 17:50:01 +1000 17:50:11 That's what everything in core and module_utils will eventually use. 17:50:20 10 minutes left 17:50:30 I use this for the past 6 years 17:50:30 http://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html 17:50:34 Excellent. 17:50:42 all of python docs looks like the above 17:51:38 The section style is an extension to sphinx? 17:51:39 Cool. So if we can pick a module, and and get most way through a re-review by next week I'd be really happy with that 17:51:55 cool, I will get you something in the next couple of hours 17:53:11 excellent 17:53:15 ah, I suppose that's what the package that documentation is for is. 17:53:28 linuxdynasty: did you have some mock code you would like reviewing? 17:53:48 nope got to modify it as per jimi's suggestion 17:53:53 cool 17:53:53 I will ping once it is ready 17:53:55 ty for checking though 17:53:57 nod 17:54:04 #topic openfloor 17:54:13 mattclay: abadger1999 anything else? 17:54:20 or anyone else? 17:54:24 Nothing from me. 17:55:11 linuxdynasty: You'll want to talk to people (Not sure if it's a core team decision or doc team decision or just consensus) about whether to use that extension or stick with standard sphinx markup. 17:55:12 cool. I think we can use the example module as a way forward with improving docs.ansible.com/ansible/developing_modules.html 17:55:39 kk 17:56:19 docschick and dharmabumstead are docs team. You know who all the core team members are :-) 17:56:37 yep yep 17:56:42 Cool. 17:56:56 Nothing else from me. 17:57:13 cool 17:57:42 caphrim007: shertel bruno_cgl Reaper anything else? 17:57:52 Nothing from me 17:58:09 cool 17:58:14 Thank you all for your time 17:58:15 nope 17:58:30 If you think of anything else feel free to add it to https://github.com/ansible/community/issues/114 17:58:34 #endmeeting