20:00:55 <jborean93> #startmeeting Ansible Windows Working Group
20:00:55 <zodbot> Meeting started Tue Feb 19 20:00:55 2019 UTC.
20:00:55 <zodbot> This meeting is logged and archived in a public location.
20:00:55 <zodbot> The chair is jborean93. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:55 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:55 <zodbot> The meeting name has been set to 'ansible_windows_working_group'
20:01:02 <jhawkesworth> hey
20:01:20 <jborean93> #chair it-praktyk[m] jhawkesworth nitzmahone
20:01:20 <zodbot> Current chairs: it-praktyk[m] jborean93 jhawkesworth nitzmahone
20:02:08 <jborean93> will give it a few minutes before starting to see if anyone else joins
20:02:16 <jhawkesworth> sure
20:04:22 * nitzmahone is busy in ansible-meeting
20:05:28 <jborean93> alrighty let's get started
20:05:51 <jborean93> #topic https://github.com/ansible/community/issues/420#issuecomment-464838767 win_get_url checksum
20:06:12 <jborean93> I don't think there's much to discuss on this one, just need some reviews to continue forward
20:06:36 <jhawkesworth> yeah, started reviewing it, but too big to 'live review' i think
20:07:15 <jborean93> yea
20:07:30 <jborean93> not sure if I can get to it this week, quite packed as it is
20:07:50 <jhawkesworth> I will try to give it a go
20:07:55 <jborean93> #topic https://github.com/ansible/community/issues/420#issuecomment-464864737 win_psmodule versions
20:08:13 <jborean93> same thing, looks like it's just ready for review and needs those reviews
20:08:23 <jborean93> anything else you wanted to add to it it-praktyk[m]?
20:08:44 <it-praktyk[m]> nope, I just wait for the final review
20:09:01 <jborean93> cool, will try and get to that this week but honestly it probably won't be until next week
20:09:31 <jborean93> #topic open floor
20:09:43 <jborean93> I'm just working on https://github.com/ansible/ansible/issues/52331 and was playing around with how to deal with it
20:10:04 <jhawkesworth> I don't use psmodules much so I'm not sure how much help I can be with that one.
20:10:06 <jborean93> currently pre-release chocolatey versions use the suffic `-beta-yyyymmdd` and the .NET version cannot parse this
20:10:54 <jborean93> we could either just ignore the version check if it fails to parse, plus an optional warning in those cases
20:11:07 <jborean93> strip out the beta part and parse that
20:11:23 <jborean93> check the file version on the executable which should conform to the Version structure
20:11:44 <jborean93> I was leaning towards 1, just wary of displaying a warning for each invocation when using a beta
20:12:46 <jhawkesworth> as a general principle I'm ok with betas of things spitting out noisy warnings.  not sure if it would generate lots of warnings in practice though?
20:13:11 <jborean93> it just means that each run of win_chocolatey would show the warning
20:13:28 <jborean93> the warning will basically say we failed to parse the version so cannot check the minimum version requirements
20:13:47 <jborean93> I suppose we could potentially combine 1 and 2
20:13:58 <jborean93> so only warn if the version is not in the expected beta form
20:15:51 <jhawkesworth> my first thought is that sounds sensible.
20:16:25 <jborean93> ok, I'll play around with it a bit and open a PR for a review when it's ready
20:16:59 <jhawkesworth> I guess just making it possible is first goal.  If its still difficult and/or noisy then that's less of an issue than actually blocking the user after all
20:17:24 <jhawkesworth> I'm in the middle of trying to reduce some of the limitations of win_xml at the moment.
20:17:53 <jhawkesworth> Trouble is it uses SelectSingleNode which is too restrictive for what I want to do with it.
20:18:18 <jhawkesworth> re-writing it to process all nodes in a list of nodes matched by xpath is proving a big change
20:18:23 <jborean93> is this something you could use template for?
20:18:38 <jhawkesworth> or at least a bigger change than I wanted to do.
20:18:51 <jhawkesworth> using template not desirable in this case.
20:18:56 <jborean93> I can't say I know XML too well so manipulation isn't my strong suite. Although I'm not surprised win_xml only handles certain cases
20:19:28 <jhawkesworth> the actual issue I had was you have to know the exact xml fragment AND the xpath that you want to remove (it has 'state=absent' but its not documented)
20:20:07 * nitzmahone resurfaces
20:20:18 <jhawkesworth> hey nitzmahone.
20:20:58 <jhawkesworth> most xml apis are big and unwieldy.  .net's is no exception
20:20:58 <jborean93> the best bet is to document it and adds tests so it can't be reverted in the future
20:21:10 <jborean93> but  that would have to be reviewed by the module maintainer
20:21:25 <nitzmahone> On previous discussion: seems like having a prerelease version of choco would be warning-worthy anyway; basically "hey, this doesn't look like a legit version of choco, you sure you're cool with that?"
20:22:12 <jhawkesworth> yeah I think I just need to take a good run at it and slather tests all over it
20:22:34 <jhawkesworth> hopefully running beta chocolatey is pretty edge
20:22:47 <jhawkesworth> or are they gearing up for a 1.0 release now?
20:23:19 * jhawkesworth has borked my choco install somehow
20:23:39 <jborean93> the beta is from October last year and is still a point release
20:23:48 <jborean93> no idea if/when 1.x will come out
20:24:12 <jborean93> cool, sounds like an easy fix
20:24:30 <jborean93> Anyone who's interested in SSH, it would be great if you could try out https://github.com/ansible/ansible/pull/47732
20:24:41 <nitzmahone> If people gripe a lot about the warning, we could add another switch for that later, but meh
20:25:12 <jhawkesworth> really want to try ssh stuff.
20:25:14 <jhawkesworth> hey dag
20:25:42 <jborean93> It requires the default shell to be set to PowerShell and the latest Win32-OpenSSH version https://github.com/PowerShell/Win32-OpenSSH/releases `v7.9.0.0p1-Beta`, this can be gotten directly from GitHub or Chocolatey but the Microsoft/Windows capability is lagging a bit too far behind
20:25:52 <jhawkesworth> currently trying to get vmware template built from packer-windoze
20:25:54 * jborean93 yay tests are now failing after a docs change :(
20:26:37 <jborean93> And another shameless plug, reviews on https://github.com/ansible/ansible/pull/52012 would be great. It should be helpful if you have LAPS in your environment
20:26:40 <jhawkesworth> packer-windoze includes your role to install correct version of OpenSSH doesn't it
20:26:57 <jborean93> jhawkesworth: yep it will install the correct version, it just won't set the default shell to PowerShell
20:27:42 * nitzmahone tries to think of a sane way to fix the need for default shell as PS
20:27:47 <jborean93> Once that lookup plugin is in, I should have enough knowledge of LDAP in Python to look at creating an AD inventory plugin
20:27:57 <jhawkesworth> cool.
20:28:03 <jborean93> nitzmahone: I'm hoping to have a `cmd` shell plugin that implements powershell except in the cases it matters for SSH
20:28:12 <jhawkesworth> I did a poc long ago using ldap3 python lib
20:28:15 <jhawkesworth> it worked.
20:28:23 <jhawkesworth> i would have liked it to be faster iirc
20:28:34 <nitzmahone> jborean93: I could actually probably implement it with `ansible_powershell_interpreter: auto`
20:29:25 <jborean93> cmd works in most cases, just need to change the _SHELL_AND and add a special quoting function to not use `'` but `"`
20:29:43 <jborean93> it was just easier to get powershell working for now and then look at supporting cmd sometime in the future
20:30:24 <jborean93> I also didn't want to mix the two, hence why powershell shell plugin requires powershell as the default shell
20:32:13 <jhawkesworth> agreed that cmd can come later
20:32:42 <jborean93> jhawkesworth: I'm sure the python ldap won't be the fastest, getting the LAPS password attribute is ok but that's a simple query compared to what the inventory plugin would use
20:33:50 <jborean93> cool, that's about it for me
20:33:58 <jborean93> is there anything else we would like to discuss?
20:33:59 <jhawkesworth> I think it would probably be ok, especially with some caching (which wasn't an option when I tried)
20:34:52 <tgadiev> Hi
20:34:55 <jhawkesworth> erm. can someone point me at 2.8 freeze dates?
20:34:59 <tgadiev> How about https://github.com/ansible/ansible/pull/51986 ?
20:35:10 <nitzmahone> jhawkesworth: March 14/21
20:35:23 <jhawkesworth> thanks nitzmahone
20:35:41 <nitzmahone> the roadmaps are linked off the README now: https://github.com/ansible/ansible#roadmap
20:35:50 <jhawkesworth> tgadiev: I'm planning on trying that out as soon as I can
20:36:46 <jhawkesworth> thanks that's easier to find than I remembered
20:36:51 <jborean93> tgadiev: hi, we briefly spoke about that PR. Basically we just need to review it but thank you for letting us know it is ready for review
20:36:51 <tgadiev> jhawkesworth: cool, thanks a lot! Let us know if there's anything to be settled
20:37:52 <jhawkesworth> tgadiev: ok I will comment on the PR itself
20:38:00 <jhawkesworth> nothing else from me tonight
20:38:21 <jborean93> cool
20:38:30 <jborean93> will bring it to a close if there's nothing else
20:39:30 <jborean93> in 5
20:39:34 <jborean93> 4
20:39:37 <jborean93> 3
20:39:40 <jborean93> 2
20:39:43 <jborean93> 1
20:39:47 <jborean93> #endmeeting