16:01:01 #startmeeting Documentation Working Group aka DaWGs 16:01:01 Meeting started Tue Nov 9 16:01:01 2021 UTC. 16:01:01 This meeting is logged and archived in a public location. 16:01:01 The chair is samccann. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 16:01:01 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:01:01 The meeting name has been set to 'documentation_working_group_aka_dawgs' 16:01:15 #topic Opening Chatter 16:01:15 o/ 16:01:17 o/ 16:01:17 Who's around to talk the docs? 16:01:27 #chair felixfontein acozine 16:01:27 Current chairs: acozine felixfontein samccann 16:01:30 \o 16:01:54 #chair ariordan 16:01:54 Current chairs: acozine ariordan felixfontein samccann 16:02:29 #chair felixfontein 16:02:29 Current chairs: acozine ariordan felixfontein samccann 16:02:55 double felix! 16:03:17 hehe :) 16:03:55 andersson007_: deric.crago dmsimard0 gundalow tadeboro briantist Xaroth - you folks around to talk docs today? 16:04:11 heh yeah I keep messing up and not getting the right pings here in matrix land 16:04:45 meanwhile, let's get started 16:04:51 #topic Action Item review 16:05:16 #link https://github.com/ansible/community/issues/579#issuecomment-957950690 16:05:37 #info core-2.12 docs including japanese translation are all published and done! 16:06:35 I think there's an issue somewhere for adding a version-switcher for the Japanese docs 16:06:40 #info expand/collapse - ariordan added another example. We can put that up on test later to verify and then we need to coordinate merges as there is a matching antsibull WIP PR for this 16:06:51 acozine: yep 16:07:07 which makes sense now that there are multiple versions 16:07:12 the antsibull WIP is unrelated, I only kept it open since it was the only place where a lot of the discussion took place 16:07:20 there is a version switcher there but it doesn't work and I havent' had the time to figure out how to get it to work. so for now, there are three versions on the main language page, yeah 16:07:34 heh 16:07:39 #undo 16:07:39 Removing item from minutes: INFO by samccann at 16:06:40 : expand/collapse - ariordan added another example. We can put that up on test later to verify and then we need to coordinate merges as there is a matching antsibull WIP PR for this 16:08:19 #info expand/collapse - 16:08:19 ariordan added another example. We can put that up on test later to verify before we merge 16:08:23 hooray for expand/collapse 16:08:27 Once we have this functionality, was there something we we going to use it on specifically? 16:08:35 other than the occassional super-long output section etc 16:09:37 we talked about whether we could use it in the module docs, I think 16:09:43 not sure if that's a priority 16:09:57 I guess starting to use it for large output sections is a good start 16:10:09 Not around today 16:10:10 do we have consensus on which versions of everything we're using - in the build and in testing? https://github.com/ansible/ansible/pull/75695/files#r727303757 ? 16:10:14 yeah the thing is, the module docs pages already have a convenient TOC at the top to jump to any section you want 16:10:19 true 16:10:38 acozine: no - we can't get the sphinx version above 3 yet 16:10:50 okay, but 3 is confirmed? 16:10:50 but since it's now at 2.12, I figure we are at least moving in the right direction 16:10:59 heh, fair enough 16:11:12 what confirmation do we need beyond the PR itself is passing CI? 16:11:31 I was thinking confirmation that CI is actually using 3 16:11:38 https://github.com/ansible/ansible/pull/75695/files#diff-418649f5e87f149692d365924e50aeae9f5d28cf2d165c451e3829d107614e81R14 says that sphinx 4.0.2 works fine? 16:12:12 \o (but I'm really distracted/overbooked) 16:12:34 IIRC Sphinx 4 works for the build, but fails CI because of recent changes in core 16:12:38 Here's the link to the file where I added the expand/collapse in the test docs build: http://docs.testing.ansible.com/ansible/devel/user_guide/playbooks_vars_facts.html 16:13:11 yep, docs build works with 4, CI doesn't so the highest I could go and have it pass CI is 3. 16:13:20 > I was thinking confirmation that CI is actually using 3 16:13:20 How do we prove that? 16:13:35 just confirm with matt clay 16:14:50 #action samccann to confirm expand/collapse PR is correctly setting CI to use sphinx 3 with matt clay 16:15:01 @smccann:matrix.org cyb-clock chimes Fifteen exciting, thrilling minutes have slipped by whilst we plot to take over the universe, one doc at a time! 16:15:14 d;oh 16:15:17 heehee 16:15:25 the update is TO the test 16:15:43 I'd still confirm with him that it's okay 16:15:47 but yeah, looks good 16:15:53 heh thanks 16:16:19 meanwhile one last look at it based on ariordan 's link about. Are folks happy with this? 16:16:43 hrm, I see what he means 16:16:50 he wants https://github.com/ansible/ansible/pull/75695/files#diff-418649f5e87f149692d365924e50aeae9f5d28cf2d165c451e3829d107614e81R14 changed back to 3.x 16:17:17 ariordan's change to the Ansible Facts page makes it much more readable 16:17:34 I might do some styling to the bar that expands the code, though 16:17:53 it's easy to skip over it, visually, so if you didn't know the page you might not realize there's anything else there 16:18:05 but that can be a separate PR 16:18:21 @acozine not sure what you mean - change what back to 3.x? the sphinx version we use today for docs? 16:18:29 I agree: it doesn't look like a code block. 16:19:06 I'm not sure what kind of styling is allowed with this extension but we can experiment 16:19:08 yeah, what matt meant in his comment, I htink, was that the known_good list should match what we use in CI . . . and he's right, but it feels wrong to dial backwards . . . 16:19:45 anywya, this change gets us closer 16:20:15 #action samccann ariordan - expermiment with styling of the expand/collapse bar to make it more visually obvious 16:20:35 #action samccann to experiment to see if we can set known-good version of sphinx back to 3.0 16:21:03 I'm not sure we can go all the way back to 3.0. I have a vague feeling something somewhere is dependent on sphinx >3 but that could just be my imagination 16:22:00 meanwhile, the other action item is we don't have docs yet for split controller that landed in 2.12. Still tracking that with the PR owner. 16:22:16 Any other action items to review before we move on? 16:22:16 if we're going to invest time and effort, it's probably better to invest it in bringing CI up to 4.x 16:22:28 true 16:22:38 acozine: alas beyond my capabilities. It's just a wall of.. urf?? what does this mean??? when I bump it up past 3. 16:22:48 heh, yep 16:23:22 I can ask gundalow if one of the community-team folks could have a look. Who did you work with before @acozine? Was it matt clay? 16:24:27 matt and toshio, I think 16:24:39 so yeah, ask gundalow who can help 16:25:38 #action samccann to follow up with matt clay and gundalow on how to get sphinx up to 4.x in CI 16:26:28 it may be fixable now, my recollection was that attributes and CLI options were the blockers 16:26:38 and those are handled in antsibull now 16:26:58 ok I'll try it again locally before I bug others 16:27:12 #topic docs survey issues 16:27:12 meanwhile, new topic 16:27:27 #info we need some help fixing the issues readers told us about in the last survey 16:27:59 #link https://github.com/ansible/ansible/issues/75508 16:28:02 https://github.com/ansible/ansible/issues?q=is%3Aopen+is%3Aissue+survey+in%3Atitle 16:28:10 This one looks like something someone here or in community might be able to help with 16:28:24 Sorry, trying to go with the tips on playbook into a role one 16:28:33 Anyone here up for giving that a try? 16:29:14 I could do a draft that outlines my thinking, that might jumpstart some comments from others who approach the question differently 16:29:38 that would be great, thanks! anything as a start would help 16:29:58 #action acozine to draft some ideas on when to convert playbook to role for #75508 16:30:01 @smccann:matrix.org cyb-clock chimes Fifteen exciting, thrilling minutes have slipped by whilst we plot to take over the universe, one doc at a time! 16:30:03 Next one https://github.com/ansible/ansible/issues/75506 16:30:27 This one looks like it might not be hard.. though brian had some thoughts we'd need to followup on 16:30:31 Any volunteers? 16:31:33 I could do the first bullet point - make a list of all the pages that use `hosts: all` 16:31:39 that would get us a step further on 16:31:50 yeah if you can do some grep magic and stick it in the issue, that would help, thanks! 16:32:26 #action acozine to make a list of all the pages that use hosts: all for issue #75506 16:32:55 That's definitely a good start on some of these! 16:33:33 #topic reviving semantic markup 16:34:00 I'm hoping now that 2.12 is out and Ansible 5 doesn't have too much work left from the docs perspective, we can revive this conversation. 16:34:39 Last I recall, it was waiting on Red Hat to determine any impact on other module docs output (galaxy-ng/AH, ansible-docs). Does that sound familiar? 16:35:32 yeah, that has a familiar ring to it 16:36:38 #action samccann to revive Red Hat conversation on semantic markup and impact on ansible-docs and galaxy-ng/AH etc 16:37:10 \o/ 16:37:33 it might be worth going back to the DaWGs minutes from a year ago or so when we were discussing the topic regularly 16:37:57 yeah I'll do a general review before I start poking folks for input/approval etc 16:38:05 there's already a PR for antsibull and ansible-doc, so it's mostly things like galaxy-ng/AH, ansible-navigator (in case it formats docs?), and maybe other tools that format docs missing 16:38:46 #info there's already a PR for antsibull and ansible-doc, so it's mostly things like galaxy-ng/AH, ansible-navigator (in case it formats docs?), and maybe other tools that format docs missing 16:38:57 cool, thanks for the refresher! 16:39:31 #topic community-contributed examples 16:39:34 #link https://github.com/ansible-community/community-topics/issues/39 16:39:59 Last time we /I paid attention, I think there was general support for the idea. 16:40:32 yeah, agreed, I just can't remember what the counter-arguments were 16:41:01 yeah flying a little blind here on that as well. Maybe we can action all of us to review and we can delve in more next week? 16:41:16 felixfontein: ansible-navigator does it's own doc construction 16:41:39 ooo who sneakin in outa the woodwork today! 16:41:42 #chair bcoca 16:41:42 Current chairs: acozine ariordan bcoca felixfontein samccann 16:41:56 got my conflicting meetings cancled! 16:42:21 #info ansible-navigator does its own docs construction so also needs to review semantic markup for any potential impacts 16:42:43 ok back to community contributed examples - 16:42:45 doesn't network also have its own docs generator? 16:42:53 a static one, yes 16:43:05 not sure how much it is used now 16:43:09 #action all - review https://github.com/ansible-community/community-topics/issues/39 for community contributed examples and let's make some progress on this at our next meeting 16:43:25 What doc generator does network have? 16:43:37 felixfontein: my hope is 'sidecar' will unite them all .. to ansible-doc 16:43:40 not familiar with that one 16:43:50 samccann: a scrip tthey run to prerender some docs on their collections/repos 16:43:57 bcoca: it's used for many ansible-maintained collections, like the files in here: https://github.com/ansible-collections/community.aws/tree/main/docs 16:44:02 bcoca: how many would `sidecar` unite? 16:44:06 samccann: ^ 16:44:09 hope is 'all' 16:44:22 bcoca: isn't sidecar about something else? 16:44:36 #info network/cloud collections also use their own docs generator to render rst files within their collection repos 16:44:39 i.e. not about convering `ansible-doc --json` output to some other format? 16:44:44 felixfontein: ist about adding a few docs features the collections need 16:45:01 @smccann:matrix.org cyb-clock chimes Fifteen exciting, thrilling minutes have slipped by whilst we plot to take over the universe, one doc at a time! 16:45:03 felixfontein: no, but that --json should provide ALL the info that is currently implemented in diff ways 16:45:36 @bcoca so if we want semantic markup, then it could be added to sidecar? 16:45:45 so when we add thing like 'retund doc fragments' they all automatically get em 16:46:04 samccann: no, but same as above, if all use ansible-doc they automatically get the 'common things we do' 16:46:13 vs having to reimplement them in 5 places 16:46:21 is anyone not using ansible-doc? 16:46:37 (I never looked at the navigator code, but the others I know of use ansible-doc) 16:46:47 for many things the script above, ansible-navigator and in some cases the buidbot 16:47:16 felixfontein: they use it and they also implement their own functions, in many cases conflicting and overlaping 16:47:52 sidecar = add optional yml file to document any/all plugins and allow test/filters to be documented in several ways 16:48:29 so for semantic markup, is this something sidecar docs will help with, or is it not directly involved? 16:49:39 I think sidecar and semanatic markup are orthogonal, they shouldn't affect each other 16:49:49 sidecar is supposed to fill in missing features others have developed in parallel, so if they use ansible-doc they get any other changes 'for free', including semantic markup removal 16:49:51 ok thanks! 16:50:14 but yes, directly unrelated 16:50:22 heh ok thanks 16:50:35 we have 10 minutes left so let's open the floor for a bit 16:50:39 #topic Open Floor 16:50:49 This is the time to bring up anything you want related to docs. 16:51:05 Favorite/stagnated PR? Issues? Comments? 16:51:40 https://github.com/ansible/ansible/pull/74963 <= planning on taking it up back next week, also planning to document jinja2 filters/tests in core 16:52:19 I'm looking forward to that! 16:52:54 @bcoca cool! is jinja2 filers/tests docs going to be a separate pr? 16:53:15 #info sidecar docs pr https://github.com/ansible/ansible/pull/74963 16:53:47 #info jinja2 filters/tests docs coming as well 16:54:12 possibly, that pr already contains some docs but mostly for testing , not sure it will contain complete docs for all or if making a followup makes sense 16:54:17 it would be great to see example output from that PR 16:55:19 ok 5 minutes left. any other lingering items to talk about today? 16:55:37 https://paste.debian.net/1218821/ <= ansible-doc -t filter ternary 16:57:28 cool, thanks 16:59:51 ok looks like that's about it for today folks! 16:59:57 Thanks everyone! 17:00:00 #endmeeting