14:32:02 #startmeeting Docs Working Group aka DaWGs 14:32:02 Meeting started Tue Jan 14 14:32:02 2020 UTC. 14:32:02 This meeting is logged and archived in a public location. 14:32:02 The chair is acozine. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:32:02 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:32:02 The meeting name has been set to 'docs_working_group_aka_dawgs' 14:32:09 who's around? 14:32:14 I am 14:32:15 o/ 14:32:24 #chair cbudz samccann 14:32:24 Current chairs: acozine cbudz samccann 14:32:33 ! 14:32:45 sorry, I'm typing on an actual laptop today 14:32:49 it's gonna get messy 14:32:59 I'm so used to my split keyboard 14:32:59 It's funny how tough that is sometimes 14:33:16 meant to type "welcome cbudz" and apparenlyt I hit some weird key 14:33:21 aaaargh 14:33:25 well, you know what I meant 14:33:42 all is well 14:33:56 any folks around on the far side of the pond? 14:34:10 andersson007_: shaps Xaroth ? 14:34:26 felixfontein: you around? 14:34:48 I see some new nicks I havne't noticed before 14:35:00 PeatTheB33t: welcome to Ansible docs chat 14:36:06 dfed[m]: epereira_ jhawkesworth madonius Pilou tremble xenlo zbr zoredache zyun 14:36:22 I'm going to ping everyone 14:36:45 a ping party! 14:37:21 back in the day, we used to have actual ping parties... when a new piece of hardware was pingable... then there was beer and food and fun 14:37:31 as always, our agenda is the last comment on https://github.com/ansible/community/issues/389 14:37:39 samccann: that sounds fun 14:38:16 it was... tho I do seem to recall walking between buildings at the Cisco campus with a pylon cone on my head... so you know... poor choices etc 14:38:45 heh, was that the result of too much partying, or was that some kind of Cisco hazing ritual? 14:38:52 too much booze 14:39:06 driveby: add new 'advanced' or 'deep dive' page that describes ansbile's inner workings on how we loop through plays/tasks and what happens at each stage .. mostly that will clarify why undefined in loop is not captured by when and other 'mysteries' 14:39:24 bcoca: thanks for dropping in 14:39:53 as always, conflicting meetings, so will not be 100% here 14:39:54 this week's agenda is short: https://github.com/ansible/community/issues/389#issuecomment-571631490 14:39:59 #info add new 'advanced' or 'deep dive' page that describes ansbile's inner workings on how we loop through plays/tasks and what happens at each stage .. mostly that will clarify why undefined in loop is not captured by when and other 'mysteries' (bcoca) 14:40:06 #topic Status Updates 14:40:22 I can start with a couple of nibbly bits 14:40:34 hang on, I tbink we have a page that might be the right home for what bcoca describes 14:40:42 ok 14:40:45 i'll dig for a link, you carry on 14:41:15 #info top left search box on devel is now running swiftype search. Feedback most welcome 14:41:48 we want to see how it works before considering backports etc. a little search experiment :-) 14:42:02 #info Added writing guidelines for search optimization 14:42:05 #link https://docs.ansible.com/ansible/devel/dev_guide/style_guide/search_hints.html 14:42:36 so basically added to the Ansible style guide to help folks understand a few ideas on making their docs content more findable etc. 14:43:44 aha! found it - https://docs.ansible.com/ansible/latest/dev_guide/developing_program_flow_modules.html 14:44:12 that is probably where the "how ansible executes" stuff should go, or at least we should integrate the missing info with that page 14:44:28 samccann: hooray for the SEO guidelines 14:44:43 kudos on those. Saw that PR yesterday 14:44:48 I will look at them every time I look at the headings-syntax page 14:45:06 #topic how ansible executes 14:45:08 so as I regularize the headers markings, I will try to make our docs more findable 14:45:15 ooh, an actual topic 14:45:21 cool! 14:45:24 #link https://docs.ansible.com/ansible/latest/dev_guide/developing_program_flow_modules.html 14:45:33 well trying to keep the meeting minutes making sense :-) 14:45:40 this topic came from a concern on Reddit, I think 14:45:45 the question was . . . 14:46:07 #info above link is a possible home for a deep dive on how ansible executes 14:46:15 https://www.reddit.com/r/ansible/comments/eo6lwq/why_is_ansible_user_not_defined_when_accessed/fe9dcci/ 14:47:11 #link https://www.reddit.com/r/ansible/comments/eo6lwq/why_is_ansible_user_not_defined_when_accessed/fe9dcci/ - where the request came from 14:47:50 and the answer is, as I understand it, that because some users want the values of a variable to vary from one item in the loop to another, Ansible doesn't "carry over" the first-defined value 14:48:16 so . . . loops are powerful, and their behavior is not always well-understood 14:48:34 the behavior of loops is tied to the way ansible executes 14:49:00 so links between thbe loops page and the how-ansible-executes info would also be good 14:49:13 I"m givinfg up on typos in things where I tink it's obvious what I meant 14:49:27 otherwise this keyboard will send me over the edge 14:49:38 only part way through the thread, but this seems to be a user question, not a developer question (that is, not an ansible module/collection developer question). So would somthing in the dev guide be findable by a user? 14:49:44 acozine: welcome to the dark side! 14:50:06 bcoca: you mean the laptop keyboard, or delving into how ansible executes? 14:50:19 all dark... ALLL DARK!!! 14:50:27 ignoring typos 14:50:28 ... 14:50:40 heh, jsut as you typed that, a light went on in the office hallway 14:50:47 lol 14:50:50 I tink someone else is here in the office with me 14:51:00 acozine now sees the light and shall mend her ways 14:51:12 can't hear or see them yet, but . . . 14:51:21 heh, I make no mending-ways promises 14:51:29 REPENT! 14:51:42 i am perhaps digressing the thread here... :-P 14:51:48 anyway, your point about user questions vs developer questions is vlaid 14:51:49 valid 14:52:10 but sometimes user requirements send them off into the weeds 14:52:21 we can have 2 pages, user detail level and then expand to developer detail level 14:52:23 and how else will we recruit the module develoeprs of hte future??? 14:52:44 bcoca: sure, we can try that 14:52:45 * bcoca thinks of unmarked van and strong dark bag as hood 14:52:55 heh 14:53:31 I prefer a glowing box and the thrill of victory as lures 14:53:51 those are not either/or propositions 14:53:54 offer cookies... people always come for cookies 14:53:57 if that's how they're recruited, they generally stay without being chianed u 14:54:01 chained up 14:54:09 but we digress . . . 14:54:20 what is our topic, again? 14:54:30 the reddit thread 14:55:08 https://www.reddit.com/r/ansible/comments/eo6lwq/why_is_ansible_user_not_defined_when_accessed/ 14:55:20 #info consider a user level page to introduce the ansible execution tidbits, then a deeper dive in the dev guide 14:55:26 ^ this and a copule of other issues always come up, mostly due to misunderstanding how the loop works and the 'stages' of a task 14:55:36 loop > when > connection 14:55:57 anybody interested in doing a draft of this topic? 14:56:10 sure 14:56:27 cbudz_ FTW! 14:56:30 cbudz_: great! 14:56:46 #action cbudz to draft a topic to cover this 14:58:07 #topic latest user guide updates 14:58:41 I've done another round of revisions to the User Guide pages 14:58:45 PR: https://github.com/ansible/ansible/pull/66371 14:59:07 New pages are up on the testing site in the `devel` branch ONLY 15:00:16 http://docs.testing.ansible.com/ansible/devel/user_guide/playbooks_reuse.html 15:00:48 that page is the one I'd most like feedback on 15:01:40 big changes to http://docs.testing.ansible.com/ansible/devel/user_guide/playbooks_reuse_roles.html 15:02:11 o/ we do have open topic at the end? 15:02:20 #info looking for feedback on https://github.com/ansible/ansible/pull/66371 15:02:30 and also to http://docs.testing.ansible.com/ansible/devel/user_guide/playbooks_tags.html 15:02:37 #link http://docs.testing.ansible.com/ansible/devel/user_guide/playbooks_reuse_roles.html especially 15:02:39 zbr: yes! 15:02:44 we always have open floor 15:02:52 #link http://docs.testing.ansible.com/ansible/devel/user_guide/playbooks_tags.html as well 15:03:02 also, you can propose a topic ahead of time on the agenda 15:03:11 zbr: ^^^ 15:03:22 also, welcome zbr! 15:04:02 acozine: thanks. now that i registered I will not miss the meetings, maybe only their start. 15:04:24 idea: can we start running ansible-lint on docs examples? so we assure that what we recommend users is always the best way to write it? (obviously that goes both ways - as in updating linter to match recommended style when needed). 15:05:04 zbr: cool, let's discuss that nex 15:05:05 next 15:05:49 to close out the topic of the User Guide changes . . . 15:05:55 all reviews welcome 15:06:07 comments, questions, suggestions 15:06:16 feel free to post them here @ me 15:06:24 or put them directly on the PR in GitHub 15:06:43 #topic linting examples in the documentation 15:07:12 This is an idea we tossed around before. 15:07:18 we have discussed this a few times 15:07:34 as I recall, it requires moving the examples into separate files in a directory, then running ansible-lint against them 15:08:00 reasoning: people copy/paste a lot from docs, any example that is using less than best guidelines, would likely make users write worse code. 15:08:07 one of gthe best solutions I've heard proposed is that we make tests double as docs examples 15:08:07 but, as I think about it now, I wonder if there is a way to run ansible-lint against the module examples ...with a bit of scripting magic? 15:08:18 that way the examples would always be maintained 15:08:30 moving to files may be needed for parsing/linting but I not a goal bit itself. is more a technical aspect I think. 15:08:33 zbr oh definitely agree and it's something we've wanted to do... it's the logistics of it that trap us so to speak 15:08:48 I'm not sure how to make the our-tests-are-also-our-examples approach work from a technical viewpoint 15:09:07 lets create a ticket for it, note what is agreen upon and maybe we can start marking small PRs towards that goal. 15:09:11 zbr: yes, it's a great idea, would be very helpful 15:09:17 now that I think about a scripting solution - could we write a script that just scrapes through the .rst files, searches for code-block yaml and then runs ansible-lint against all of them? 15:09:44 hmmm, would we have to make all the examples full playbooks? 15:09:47 i'm not sure all yaml code-blocks are ansible (inventories etc) 15:09:57 i'm not sure ansible-lint is that specific 15:10:09 mm, htat's true 15:10:19 but I could be wrong. easy enuf to pull out a few examples that aren't full playbooks and see what happens. 15:10:21 well zbr are you willing to open the issue? 15:10:39 it can be hinted, but linting examples does not limit to ansible-lint itself, also applies to other linters like yamllint or flake8. 15:11:24 #info would be great to run ansible-lint against doc examples in .rst and module docstrings. May require some scripts to do this w/o having to move all examples into separate .yml files 15:11:32 I like the idea of "drinking our own champagne" and using the docs examples with ansible-lint to make sure we keep ansible-lint usable 15:11:47 writing now the ticket... 15:11:54 zbr: excellent! 15:13:06 I hope we can do the updates in stages 15:13:22 maybe start with a smaller section of the docs, then work our way up to full coverage 15:13:44 we'll need a proof-of-concept to start with 15:14:04 Are there usage stats to taget higher usage docs with examples to start with? 15:14:16 why yes there is cbudz! 15:14:21 looking for a link... 15:14:33 *target 15:15:00 #link https://github.com/ansible/ansible/issues/58687 - lists top 10 doc pages for Ansible. Good place to start making improvements 15:15:06 samccann is our stats guru 15:15:21 I should run the stats again since that's 6 months old now (my how time flies) 15:15:53 I'll be interested to see if it has changed 15:16:52 zbr: can you add a link from the test-the-examples issue to the top-ten-pages issue samccann just posted? 15:17:58 https://github.com/ansible/ansible/issues/66474 15:18:11 hooray! thanks zbr 15:18:44 Have to drop to prepare for another status meeting, thanks all 15:19:09 thanks cbudz! 15:19:23 #link https://github.com/ansible/ansible/issues/66474 - new issue to track ansible-lint on docs 15:19:59 I put a link across to samccann's top-ten iussue 15:20:01 aaargh 15:20:02 issue 15:20:21 next time I come to the office I am bringing my external keyboard 15:20:21 gonna be a long day for you 15:20:27 no matter how geeky I look 15:20:34 heh, yes, it is 15:20:43 on the ohter hand, my finger muscles are getting a workout! 15:20:50 silver lining 15:20:52 esp. the right pinkie 15:20:56 the one that hits backspace 15:21:01 heh 15:22:00 zbr: if you have ideas for how to implement linting docs examples, please bring them to this channel! 15:22:17 meanwhile, anything else for today? 15:22:22 #topic open floor 15:22:31 I have one PR I think we can close 15:22:35 awesome 15:22:37 which one? 15:22:52 https://github.com/ansible/ansible/pull/66408 15:23:17 now that I read it, it has two folks saying the doc is fine as it is... so gonna just close it unless you see a reason to keep it open for debade? 15:23:25 debate even (and I do have my tented keyboard) 15:25:42 sounds okay to close 15:26:29 samccann: go for it 15:26:43 closed 15:27:10 puts us down to 93 open PRs :-) 15:27:13 does the main ansible-YAML page have a note about "there are different ways to indent" 15:27:20 if not we could add one 15:27:49 something about "there's more than one valid way to indent your YAML, and there are dozens of invalid ways too" 15:28:02 93, eh? that's going in the right direction!@ 15:28:22 anybody else have stuff today? 15:28:35 nothing else here 15:28:47 quiet day in docs-land 15:29:12 all right, thanks bcoca cbudz_ samccann zbr ! 15:29:22 #endmeeting