14:31:35 <samccann> #startmeeting DaWGs aka Docs Working Group 14:31:35 <zodbot> Meeting started Tue Oct 15 14:31:35 2019 UTC. 14:31:35 <zodbot> This meeting is logged and archived in a public location. 14:31:35 <zodbot> The chair is samccann. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:31:35 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:31:35 <zodbot> The meeting name has been set to 'dawgs_aka_docs_working_group' 14:31:46 <samccann> Who's around to talk the docs?? 14:32:13 * gundalow is only half around, so feel free to ping me if there is something specific 14:32:18 <samccann> abadger1999 gundalow alongchamps cyberpear ? 14:32:57 * samccann turns gundalow into comfy pillow... not quite a chair 14:33:15 <gundalow> :) 14:33:15 <acozine> o/ 14:33:26 <samccann> #chair acozine 14:33:26 <zodbot> Current chairs: acozine samccann 14:34:13 * acozine told myself to "sit down" and immediately started singing "Sit Down You're Rocking the Boat" 14:34:20 <alongchamps> I'm here 14:34:25 * acozine it's a damnably catchy tune 14:34:29 <samccann> hellooooooo alongchamps !! 14:34:34 <samccann> #chair alongchamps 14:34:34 <zodbot> Current chairs: acozine alongchamps samccann 14:35:13 <samccann> our agenda is here - https://github.com/ansible/community/issues/389 14:35:30 <samccann> ooph this is better - https://github.com/ansible/community/issues/389#issuecomment-542220871 14:35:43 <thedoubl3j> I am here if needed as well 14:35:47 <acozine> the scrolling, the scrolling! 14:35:59 <samccann> kewl!! #chair thedoubl3j 14:36:04 <samccann> hmph. 14:36:15 <samccann> #chair thedoubl3j 14:36:15 <zodbot> Current chairs: acozine alongchamps samccann thedoubl3j 14:36:24 <samccann> there ya go! Furniture. 14:36:30 <acozine> ah, TIL, the command must be the beginning of the line 14:36:30 <samccann> #Topic Roles documentation 14:37:03 <samccann> So to start, we are moving Galaxy docs onto docs.ansible.com... one bit at a time, so that we can eventually retire the docsite associated with galaxy 14:37:39 <acozine> hooray! the Galaxy docs are a backwater, no links to the rest of the documentation 14:38:03 <thedoubl3j> one docs site to rule them all 14:38:04 <acozine> integrating those topics will be a much better user experience 14:38:06 <samccann> Yep. In that effort, we created a Galaxy section on docs.ansible.com, and moved up the galaxy info that was in an appendix - https://docs.ansible.com/ansible/devel/galaxy/user_guide/index.html 14:38:55 <samccann> heh yep, acozine == sauron... and that makes me the guy with the long white hair then has a sticky end in his lonely flooded tower 14:39:04 <felixfontein> hi! 14:39:09 <samccann> helloo!! 14:39:13 <acozine> hey, I object to both of those characterizations! 14:39:22 <samccann> #chair felixfontein 14:39:22 <zodbot> Current chairs: acozine alongchamps felixfontein samccann thedoubl3j 14:39:47 <samccann> info - acozine objects to being the Ruler of All Docs and associated evil empire ambitions 14:40:00 <acozine> I'm a Benevolent Overlord, really 14:40:06 <samccann> Meanwile... roles 14:40:10 <felixfontein> lol 14:40:38 <samccann> Those familiar, please do read that galaxy section and open PRs or issues against it. It probably hasn't seen the docs love it needs yet, so I'm sure there are issues with it. 14:40:45 <samccann> #link https://docs.ansible.com/ansible/devel/galaxy/user_guide/index.html 14:41:03 <samccann> #info looking for experienced roles folks to review that galaxy docs on docs.ansible.com 14:41:28 <samccann> And there is an open PR for installing docs - https://github.com/ansible/ansible/pull/63486 14:41:38 <acozine> do folks think we should promote the Galaxy docs to docs.ansible.com/galaxy? or keep them as a subsidiary to docs.ansible.com/ansible ? 14:41:39 <samccann> Please have a look 14:41:51 <samccann> oo good question 14:41:52 <acozine> I can't remember if we discussed that or not 14:41:59 <samccann> We haven't 14:42:21 <samccann> my initial thought - keep it on docs.ansible.com because people can't work galaxy "in a vacuum" so to speak 14:42:54 <acozine> yeah, i don't know how well we can integrate content at the docs.ansible.com/product level 14:43:14 <acozine> if it doesn't work well, then keeping galaxy under "core" seems smart 14:43:21 <samccann> it's part of Ansible itself if that makes sense. So if I'm creating a collection or a role, I still need to know how to write modules, document modules, etc etc and that's all on docs.ansible.com/ansible 14:43:35 <acozine> yeah, that makes sense 14:43:39 <alongchamps> if it's big enough, and it seems like it is, I like docs.ansible.com/galaxy 14:44:02 <alongchamps> but then there's the version part that ties to ansible 14:44:07 <thedoubl3j> it is almost half and half. 14:44:15 <samccann> imo it won't be that big. There is a user guide and a dev guide, but again, doesn't cover the ansible basics 14:44:29 <thedoubl3j> cli tool comes with ansible, but the site and finding content is a site. 14:44:48 <samccann> There may be a community guide, but I'd rather discuss that with the galaxy folks and see if they want to use the same community guidelines as ansible has 14:45:08 <thedoubl3j> I am fine with keeping it the way it is. 14:45:18 <samccann> thedouble3j - yeah the gui part will be added to the galaxy user guide etc over time. Inching our way there 14:45:51 <thedoubl3j> +1 14:45:57 <samccann> ooo alongchamps++ very good point - galaxy has a different version scheme than ansible core 14:46:16 * samccann ponders the ponderable about versions 14:47:30 <samccann> That has implications I didn't even consider. Right now I'm writing on a 2.9 version of ansible, but galaxy has its own release cycle (and numbering). gundalow - do you know if galaxy releases in the same timeframe as ansible engine? or are they entirely separate on when releases go out? 14:47:41 <acozine> yeah, good point, we may change the Galaxy UI on a different schedule than Ansible itself 14:47:41 <samccann> acozine: same question 14:47:56 <gundalow> Galaxy has it's own release train 14:48:14 <gundalow> ie releases independent to Ansible Core 14:48:18 <samccann> ok then that might push us to moving galaxy 'up' a level as acozine suggested 14:48:40 <samccann> so it can update separate from docs.ansible.com/ansible 14:48:54 <samccann> off the top of my head, that's a bit of work for sure 14:49:23 <samccann> and makes me wonder - if galaxy is its own tab, it's not much different than having it on galaxy website itself. It still may not get the docs love it needs 14:51:05 <samccann> might need to table this discussion for now, gonna take Deeper Thoughts (tm) 14:51:55 <samccann> #info - galaxy has an independent release cycle not tied to ansible engine - it may need its own landing page to reflect different versions etc 14:52:05 <samccann> ok moving on 14:52:33 <samccann> #topic Fix return value PRs 14:52:44 <samccann> These are a pair from felixfontein 14:52:56 <samccann> #link https://github.com/ansible/ansible/pull/63478 14:53:14 <samccann> #link https://github.com/ansible/ansible/pull/63477 14:53:36 <acozine> thank you to felixfontein for working to make the return values documentation more useful 14:53:44 <samccann> \o/ 14:54:06 <samccann> they are both passing CI and have approvals... is there anything holding us up from hitting the merge button on these two? 14:54:47 <felixfontein> these two PRs are preparations (which should be safe to backport) for https://github.com/ansible/ansible/pull/63411 14:54:59 <felixfontein> maybe one more review for https://github.com/ansible/ansible/pull/63411 so that there are at least two reviews for each? 14:55:35 <felixfontein> ah wait one more for https://github.com/ansible/ansible/pull/63478 I meant 14:55:44 <felixfontein> sorry, my network connection is pretty crappy, so I'm slow at writing :) 14:56:19 <samccann> ok so one more review for #63478 and then both are safe for backporting? 14:56:31 <felixfontein> I would say so 14:56:34 <samccann> and once they are done, #63411 can be rebased and merged? 14:57:06 <felixfontein> I'with a bit of review, perhaps :) 14:57:06 <felixfontein> yes 14:57:07 <samccann> acozine and I are doing a batch of docs backporting later today. Do those two earlier ones need #63411 merged and backported as well? 14:57:18 <samccann> (with a review) :-) 14:57:41 <felixfontein> I'm not sure whether 63411 should be backported. also, it's using `elements:` for return values, which currently only exists in devel 14:57:47 <thedoubl3j> _looks at 63478_ 14:57:56 <samccann> my point - the first two can easily be merged and backported today. #63411 requires more work, so could miss the 'backport window' so to speak 14:57:57 <felixfontein> (see last item of docs agenda for today ;) ) 14:58:25 <samccann> ah ok. so the first two are safe to backport, the last one needs discussion 14:58:52 <felixfontein> I'm fine if new validation is only in devel (for 2.10), that probably shouldn't be beackported 14:58:55 <felixfontein> the other fixes from 63411, that's anohter story 14:58:58 <samccann> #agreed can review then merge #63478 and #63477 and backport to 2.9 14:58:59 <felixfontein> (after rebasing we'll see what's left) 14:59:21 <felixfontein> I would also rbackport them to 2.8, since the errors mostly exist there as well 14:59:38 <felixfontein> (63478 and 63477 that is, not 63411) 15:00:01 <samccann> #agreed 63478 and 63477 will be safe to backport to 2.8 as well 15:00:07 <acozine> felixfontein: if parts of 63411 should be backported and parts should not, should it be split again into 2 PRs? 15:00:32 * acozine apologizes if she missed part of the conversation 15:00:47 <acozine> my other meeting just ended, so I am 100% present now 15:01:11 <felixfontein> acozine, I can split it up, but first the others need to be merged, so I'll see what's left of all the changes :) 15:01:16 <acozine> gotcha 15:01:21 * acozine looks at the two PRs 15:01:47 <felixfontein> thedoubl3j: thanks for reviewing! 15:02:09 <acozine> 63477 has had plenty of eyes on it 15:02:27 <acozine> I'm happy to merge, unless someone is still reviewing it 15:02:33 <felixfontein> I think so too, but I didn't want to merge it before the meeting ;) 15:02:49 <acozine> heh, thanks 15:03:15 <acozine> merged 15:03:22 <felixfontein> \o/ 15:03:26 <felixfontein> thanks :) 15:04:13 <acozine> wow, those vultr modules all did the same thing 15:04:40 <samccann> felixfontein - the pr you wanted to test first, is that related to this discussion? 15:04:50 <samccann> https://github.com/ansible/ansible/pull/63409 15:05:00 <samccann> wondering if we should jump to that topic next? 15:05:01 <acozine> so lib/ansible/modules/cloud/amazon/ec2_vpc_nacl_info.py now uses `elements` . . . is that dependent on a change we don't want to backport? 15:05:17 <felixfontein> samccann: yes partially, since 63411 is using the feature added there 15:06:30 <felixfontein> acozine: do you mean because of 63411? 15:06:52 <felixfontein> acozine: it should not be contained in 63477 and 63478 15:07:57 <acozine> felixfontein: as part of 63478 I see `elements` being added to that module 15:08:14 <acozine> I think . . . let me double-check that I wasn't looking at the wrong thing 15:08:54 <acozine> ah, no, my mistake 15:08:56 <thedoubl3j> I am not seeing anything touching elements there. _goes to double check_ 15:09:00 <acozine> ETOOMANYTABS 15:09:34 <acozine> that's in 63411 15:09:51 <felixfontein> ok, then it's fine! 15:09:55 <samccann> phew 15:10:04 <samccann> are we okay to move to the next topic? 15:10:24 <acozine> sure 15:10:25 <felixfontein> abadger just merged https://github.com/ansible/ansible/pull/63409 15:10:33 <samccann> heh 15:10:48 <samccann> #topic should #63409 be tested? 15:10:51 <acozine> felixfontein: should https://github.com/ansible/ansible/pull/63478/files#diff-b8cfe9953f1b6756e8d4da425dd3232aR204 be 'sample' instead of 'example'? 15:11:01 <felixfontein> can someone from you rebuild the test docs site with that branch ()?ODODtemporary-2.9.1-branch-releng-only 15:11:04 <samccann> well, we can still put it on the docs test site if you want to see it in action 15:11:40 <acozine> samccann: you want to kick off a build to the test site, or shall I? 15:12:30 <samccann> hmm well abadger1999 merged it to a temp branch. ansible:temporary-2.9.1-branch-releng-only not sure how to find that? 15:12:47 <samccann> or is it just that - ansible/temporary-2.9.1-branch-releng-only 15:13:10 <abadger1999> yeah... you might want to build the temporary-2.9.1-branch-releng-only to the testing site? 15:13:12 <abadger1999> It is 15:13:27 <abadger1999> https://github.com/ansible/ansible/tree/temporary-2.9.1-branch-releng-only 15:14:31 <samccann> hmm what do I set as the GIT_BRANCH in the jenkins build abadger1999 ? 15:14:46 <abadger1999> I believe: temporary-2.9.1-branch-releng-only 15:15:00 <samccann> ok I'll give that a try, thanks! 15:15:18 <acozine> which fork is that branch on? 15:15:43 <samccann> oh I tried it as orgin/temporary-2.9.1-branch-releng-only 15:15:44 <abadger1999> It's in the main repository 15:15:46 <abadger1999> ansible/ansible 15:15:59 <abadger1999> I believe origin maps to ansible/ansible in jenkins 15:16:08 <felixfontein> should I create backports for #63477, or does anyone else wants to do that? the 2.8 backport will be annoying, since ignore.txt format changed and some new modules and new features are affected 15:16:39 <samccann> acozine and I are doing 2.9 backports for docs later today so feel free to handle that 2.8 fun if you want :-) 15:16:41 <abadger1999> @samccann Do you normally use origin/stable-2.8 or stable-2.8? 15:16:47 <felixfontein> also, how about https://github.com/ansible/ansible/pull/63478, any more reviews, or should we merge? 15:17:03 <felixfontein> samccann: ok, I'll enjoy that then ;) 15:17:15 <felixfontein> I'll backport both PRs (once the second is merged) to stable-2.8 15:17:16 <samccann> @abadger it defaults to origin/devel so i figured origin/<fillintheblank> would work. 12 minutes from now, we should know :-) 15:17:32 <samccann> thanks @felixfontein !! 15:18:03 <abadger1999> @samccann Okay, then origin/temporary-2.9.1-branch-releng-only will probably work.... origin is usually the default so it might work without the origin/ as well. 15:18:07 <samccann> #info putting #63409 onto the docs test site for review 15:18:20 <samccann> good to know abadger1999 15:18:36 <acozine> felixfontein: review added to 63478 15:19:11 <samccann> ooo felixfontein added a review already for the next topic (thanks!) so gonna skip that one 15:19:52 <samccann> The next agenda topic is issues triage, but before hopping to that. gonna do an Open Floor 15:19:58 <samccann> #topic Open Floor 15:20:10 <samccann> anyone have anything to discuss before we attempt a little docs issues triage? 15:21:34 <acozine> I have a quick topic 15:21:38 <samccann> go for it 15:21:53 <felixfontein> acozine: validation allows both `sample` and `example`, I guess we should normalize that (I think also that docs generation only uses one of them, probably `sample`) 15:22:19 <acozine> felixfontein: cool, i put suggestions for the two `example` ones in my review 15:22:41 <acozine> so I'm working on a fairly deep edit of the User Guide pages 15:22:58 <samccann> #topic User Guide revamp 15:23:11 <acozine> the first round was https://github.com/ansible/ansible/pull/63056 15:23:33 <acozine> the next round is going to touch a bunch of the playbooks pages 15:23:58 <acozine> and I'm thinking of reorganizing some of them around a theme of "working with data" 15:24:43 <acozine> I noticed that the mailing lists get a ton of questions about "how do I get the fact/variable value I need in my next task?" 15:24:59 <acozine> so . . . I'm looking for good examples 15:25:08 <acozine> real-world examples of complex return data 15:25:16 <acozine> suitably sanitized, of course 15:26:03 <acozine> ideally it would be a data set I could pull several different examples from 15:26:35 <samccann> #info - revamping the playbooks pages and loking for good examples of complex return data (sanitized) to pull several examples from for getting fact/variable values in tasks etc 15:26:49 <acozine> thanks samccann! 15:27:42 <acozine> if anyone here in DaWGs has suggestions of things in the User Guide that need updating, please ping me! 15:28:19 <samccann> #info if anyone has suggestions on what needs clarification/updating in the Ansible User Guide - ping acozine :-) 15:28:21 <acozine> or comments on the pages that have been edited/updated - all feedback/comments/suggestions welcome 15:28:25 <acozine> that's it 15:28:32 <samccann> Ok we have time to triage maybe one issue - 15:28:38 <samccann> #topic Issues triage 15:28:39 <samccann> https://github.com/ansible/ansible/issues/63495 15:28:53 <samccann> yum module can also run Python 3 `This module only works on Python 2. 15:29:11 <acozine> hmmm 15:29:18 <acozine> I thought I'd seen a bunch of other related issues 15:29:27 <acozine> IIRC the problem is that the module itself runs fine 15:29:28 <samccann> So seems simple, but do we know for sure yum works on python 3? I'm thinking we could label this an `easyfix` item if all that's needed is to remove that line 15:29:51 <acozine> but yum is not available for Python 3 15:30:00 <acozine> so it's not an Ansible restriction 15:30:06 * acozine searches for the other issues 15:30:17 <samccann> not at all? or is it working for the OP because RHEL 8 does a `dnf` under the covers? 15:30:34 <samccann> I seem to recall Fedora did that 'way back when' so maybe that's part of RHEL 8 now? 15:30:38 <acozine> https://github.com/ansible/ansible/issues/62722#issuecomment-540618529 15:31:18 <samccann> #info see https://github.com/ansible/ansible/issues/62722#issuecomment-540618529 for details 15:31:34 <acozine> As you can see above, the yum module is not available to the python3 runtime. In Ansible this is typically not an issue because we can allow the control system to use whatever python version is available and the remote hosts in the inventory to define their own python interpreter via ansible_python_interpreter, as @sivel mentioned and is documented here. This allows the controller and the remote hosts to use differing python versions 15:31:35 <acozine> without issue because Ansible abstracts away that concern. Therefore, Ansible can run in python3 "mode" on the control host and not worry about the remote hosts python version so long as the inventory takes this into account properly for the hosts. 15:31:47 <samccann> ok we are basically out of time, but thanks everyone for participating today and enjoying your time as a piece of docs furniture! 15:31:51 <acozine> ^^^ that's a quotation, just for the record 15:32:07 * acozine looks forward to being an armoire next week! 15:32:08 <samccann> thanks acozine we can review that and decide what to do with that new issue 15:32:24 <samccann> ok gonna close unless someone peeps... 15:32:30 <acozine> thanks samccann felixfontein alongchamps gundalow 15:32:39 <alongchamps> see ya 15:32:45 <acozine> ciao! 15:32:56 <felixfontein> thanks everyone! 15:32:59 <samccann> o/ and thanks again for the version comment on galaxy! that was a real blindspot for me 15:33:03 <samccann> #endmeeting