19:04:24 <bcoca> #startmeeting ansible core irc plublic meeting
19:04:24 <zodbot> Meeting started Tue Jul  9 19:04:24 2019 UTC.
19:04:24 <zodbot> This meeting is logged and archived in a public location.
19:04:24 <zodbot> The chair is bcoca. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:04:24 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:04:24 <zodbot> The meeting name has been set to 'ansible_core_irc_plublic_meeting'
19:04:43 <bcoca> #topic https://github.com/ansible/ansible/issues/53402
19:04:52 <bcoca> @mavit?
19:05:11 <cyberpear> +1 to `regex` vs `regexp`
19:05:26 <cyberpear> it's thrown me off multiple times
19:05:37 <bcoca> its just one of many things that make me hate lineinfile
19:06:04 <bcoca> fyi, it already has regex alias
19:06:08 <cyberpear> hmm
19:06:12 <cyberpear> good to know
19:06:15 <bcoca> fixed that a while back
19:06:31 <tchernomax> +1 to `regex`
19:06:45 <bcoca> +1 also to regex (use regexp as alias for backwards compat)
19:07:11 <maxamillion> .hello2
19:07:12 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
19:07:13 <cyberpear> we should add an alias also to the `replace` module
19:07:52 <bcoca> all those listed in ticket that have regexp
19:08:36 <cyberpear> I'd be in favor of updating all of the above to use `regex` at least as an alias
19:08:54 <shertel> also +1 to regex with alias for backwards compat
19:08:55 <bcoca> i would make it principal and regexp the alias
19:09:11 <jillr> +1 to regex
19:09:15 <cyberpear> (if/when modules are moved out of core, it'd be more difficult to do such module param standardization work)
19:09:41 <bcoca> cyberpear: not that being in core has worked wonders for that
19:10:06 <cyberpear> at least `replace` is `stableinterface`, so the alias should probably stick around for a while
19:10:19 <bcoca> it used to be the case when 'core team' was 3 people and we reviewed all modules, but unless we start using enforcement for 'sounds like' in options, its not scalable
19:10:30 <bcoca> im fine with alias being permanent
19:12:13 <maxamillion> +1 regex w/ backwards compat
19:12:51 <bcoca> so let me put up options a) regexp with regex alias, b) regex with regexp alias c) we stop using regexes
19:12:55 <bcoca> b+1
19:13:25 <cyberpear> b+1
19:13:30 <jillr> b+1
19:13:47 <tchernomax> b+1
19:14:20 <shertel> b+1
19:14:50 * bcoca is suprised no one went for c
19:15:02 <bcoca> k, calling it then
19:15:18 <bcoca> #topic to combine with options or to combine_deep
19:15:28 <bcoca> https://github.com/ansible/ansible/pull/58798
19:15:35 <bcoca> https://github.com/ansible/ansible/pull/57894
19:15:38 <bcoca> both options
19:15:46 <bcoca> a+1
19:16:01 <tchernomax> b+1
19:16:08 <bcoca> ^ as of interface, code review might get more traction, specially b
19:16:41 <bcoca> sorry, im b+1
19:16:50 <bcoca> ^ 'specially a)
19:17:33 <tchernomax> > as of interface, code review might get more traction
19:18:02 <bcoca> a) new filter, b) new option on existing filter
19:18:14 <maxamillion> b+1
19:18:16 <bcoca> ^ for some reason those were reversed in my head
19:18:33 <sivel> sorry, was distracted classifying files
19:18:50 <tchernomax> just to be sure of what you mean : you said the interface will be discuss during code review, not during the meeting
19:19:04 <tchernomax> ?
19:19:30 <bcoca> the 2 options are about which interface to use (most of the code is same, just organization and invasiveness, updating hash_merge makes me uneasy)
19:19:36 <jtanner> hi
19:19:42 <tchernomax> ok
19:20:39 <tchernomax> so for now, it's a:0 b:3
19:23:13 <jtanner> i dunno, i like a for the fact that it doesn't complicate merge_hash
19:25:13 <tchernomax> bcoca: are you ok to discuss implementation after the meeting ? (or after the a/b vote ?)
19:25:21 <cyberpear> w/o a new filter, `list_merge` is meaningless for non-recursive
19:26:16 <tchernomax> cyberpear, yes it is: `{"a": ["toto"]} | combine({a: ["titi"]})`
19:27:03 <cyberpear> okay, if list_merge is meaningful for non-recursive, there's no reason to have a separate filter, in my mind
19:27:04 <bcoca> tchernomax: i'll update ticket
19:27:18 <tchernomax> bcoca: ok
19:27:24 <tchernomax> thanks
19:28:15 <cyberpear> b+1
19:28:25 <tchernomax> jtanner: are you voting "a", or are you undecided ?
19:28:38 <tchernomax> for now, a:0 b:4
19:28:41 <jtanner> i probably don't have grounds to vote
19:29:33 <shertel> looks like there is one task in 'filters' target that use 'combine' and expects success... adding tests for the current behavior as well as the feature seem important
19:29:34 <tchernomax> (some others may hesitate… but indeed I think "b" will win)
19:30:24 <bcoca> +10 for tests
19:30:24 <tchernomax> shertel: you, I didn't implement test and documentation yet. But it's in my TODO.
19:30:38 <bcoca> ^ we needed to decide approach first
19:30:44 <shertel> I did read that, just reiterating
19:31:04 <tchernomax> I wanted to now if I should focus on a or b before going further
19:33:15 <bcoca> k, i think you got your answer
19:33:20 <bcoca> #topic open floor
19:33:35 <cyberpear> #topic ansible_remote_tmp and ansible_async_dir
19:33:43 <cyberpear> are these connection variables?
19:33:48 <cyberpear> they seem to be poorly documented
19:33:54 <bcoca> first is .. 2nd ... mebbe
19:34:19 <cyberpear> defaulting to $HOME/.ansible and $HOME/.ansible_async
19:34:25 <bcoca> any var that is in connection/shell/become plugins + ansible_python_interpreter are 'connection variables'
19:34:39 <cyberpear> I couldn't find very good docs on them, but the latter was in the 2.8 porting guide
19:34:58 <bcoca> ansible_remote_tmp is in shell plugins
19:35:02 <jborean93> ansible_remote_tmp is where remote temp files are placed
19:35:17 <jborean93> ansible_async_dir is the directory the async result files are placed
19:35:20 <cyberpear> shell plugins++
19:35:22 <jborean93> IIRC both are documented in the shell plugins
19:35:28 * cyberpear looks
19:35:32 <bcoca> i dont think async is  in shell plugins
19:35:52 <bcoca> async is 'action plugin/module' based
19:35:56 <jborean93> I'm 95% sure I did that there for powershell
19:35:59 <bcoca> its currently in weird state
19:36:01 * jborean93 checks
19:36:05 <bcoca> jborean93: that is the exception then ...
19:36:09 <cyberpear> bingo! https://docs.ansible.com/ansible/latest/plugins/shell/sh.html
19:36:19 <cyberpear> thanks!
19:36:25 <cyberpear> #topic open floor
19:36:30 <jborean93> https://github.com/ansible/ansible/blob/devel/lib/ansible/plugins/doc_fragments/shell_common.py#L33
19:36:30 <bcoca> lib/ansible/plugins/doc_fragments/shell_common.py <= aysnc_dir is here ... i sit corrected
19:42:44 <shertel> https://github.com/ansible/ansible/tree/devel/contrib#contributions-welcome <- can I update it to welcome bug fixes and direct new scripts/features to plugins? That's been done on an individual basis for a while, but I didn't realize there was out of date documentation.
19:42:58 <bcoca> +10K
19:43:10 <bcoca> even pushing to new repo ansible/inventory_scripts
19:49:43 <bcoca> nothing else?
19:50:04 * jborean93 cricket
19:50:08 <bcoca> #endmeeting