Puppet Community

Never mind, looks like I haven't checked in a while

I have one on RHEL 9 and one on Ubuntu Bionic (working on upgrading that too)

I'm curious why `os.{name,family}` would ever be `nil`?

Also wondering why the downcase is there.. did facter used to return the fact in lower case?

<https://github.com/voxpupuli/puppet-keepalived/actions/runs/6630290840/job/18011428544?pr=293> this should fail in a minute or two

and I've seen this in other modules but can't remember which ones...

Facter has mostly returned capitalized versions <https://github.com/puppetlabs/facter/commit/8f938c18adfa65bf88712f1f648670325b744ae5>

I pinged ewoud in the PR, let's see what he says

I could see it happening if the mocked facts don't include os.name or os.family

yeah, but those should always be presend in facterdb :thinking_face:

I will talk to ewoud and come back to you

ok sounds good, I'll try to see where the `downcase` came from in puppet. It definitely seems unnecessary

Hi <@U2S1MUUQ7> I think it looks good. The one thing we have to be careful of changing `confines`/`defaultfor` can produce a side-effect that changes the suitability for other OS'es

This the "special" code <https://github.com/puppetlabs/puppet/blob/945beb6ef0200b1bc7879b62131282a61fc9a935/lib/puppet/provider.rb#L339-L347>

yeah I think what you have looks good, I'll need to take a look what happens on other OSes though

please. I am not sure how far along with AL2023 you are in general, but this would really help meanwhile

<@U2S1MUUQ7> could you preface the commit summary with `(PUP-11976)`?

Josh, I can't apply the same patch on top of 7.x, I will submit a new PR for it

<https://github.com/puppetlabs/puppet/pull/9132/>

Awesome thanks <@U2S1MUUQ7>! We'll merge the 7.x one and it will "merge up" to main automatically, so you can close PR 9128.

take a look, it's a totally different fix

Ah yeah, we usually resolve conflicts when the auto merge fails. But it's good having both PRs.

Speaking of which, I've been wanting to switch puppet over to a backport model, like most software I know of. It's always a pain not knowing what branch to target.

it's a trivial case here, but I can see how that can be a problem. good thing no need to patch other branches :slightly_smiling_face:

I believe composite namevars was added after parsed file providers, so the lack of support doesn't surprise me

and people using parsed file are rare, people using composite namevars are rare as well

I'll comment on the PR, but I think the changes would introduce breakage, for example `property_hash` is referenced in other parsed file providers like <https://github.com/puppetlabs/puppetlabs-sshkeys_core/blob/fbd0c6dc9ca7314a567b6873cb66e2adf52af30b/lib/puppet/provider/sshkey/parsed.rb#L32C8-L32C21|https://github.com/puppetlabs/puppetlabs-sshkeys_core/blob/fbd0c6dc9ca7314a567b6873cb66e2adf52af30b/lib/puppet/provider/sshkey/parsed.rb#L32C[…]2C21>

This is what we did to the sshkeys_core module to support composite namevars <https://github.com/puppetlabs/puppetlabs-sshkeys_core/commit/92a734a6620fcc51691d60347e344f87ac801bde>

I was hoping we can maybe get it released in puppet 7

TBH we can't merge that PR in puppet 7 or 8 in its current form. I'd recommend trying to fix the issue in the module and not the ParsedFile provider like we did for sshkeys_core

:indeed: Next up is _:teacher: Bolt_ in <https://www.timeanddate.com/worldclock/converter.html?iso=20231024T220000&amp;p1=202&amp;p2=179&amp;p3=919&amp;p4=307&amp;p5=3332&amp;p6=236&amp;p7=240|4 hours>

Ok folks, that's a wrap. Thanks for your feedback and PRs!

:allthethings: _:teacher: Bolt_ is about to start up in 15 minutes

:fry_dancing: Next up is _:teacher:Puppet Forge_ in <https://www.timeanddate.com/worldclock/converter.html?iso=20231026T180000&amp;p1=202&amp;p2=179&amp;p3=919&amp;p4=307&amp;p5=3332&amp;p6=236&amp;p7=240|43 hours>

:goodnews: _:teacher:Puppet Forge_ is about to start up in 15 minutes

we are working on some back end things right now, so nothing super visible. But foundation for some cool things coming up

and doing some UX reviews to make the search experience a little better

I think there was some work on the stability or performance in the past? Is that done?

The particular project to focus on that is done, but it’s an ongoing thing. We’ll dedicate a little time each sprint to improve

Maybe it's ok if I bring up something that is tangentially related since there hasn't been a ton of activity. With 40 seconds left...

sometimes we will have a forge module fail to download while running r10k

it doesn't look concurrent when I look at the live logs