ENH: ignore draft PRs

[bsipocz]

Changelog checker should not error out on draft PRs. (but if there is an entry added already, that is great and could probably give a green.

[pllim]

But why? Then your workflow will have to listen to some event that triggers when a PR gets out of draft, if there is such a thing. If you don't, then you get a false positive.

[bsipocz]

Because it's premature to think about changelog when it's not even clear which release a draft will end up in, or if it even comes out of the experimental draft status. Think about packages where it's encouraged to open a PR super early, so discussion can happen on it, but packages where the release cycle is much less defined.
(E.g. in pyvo there are currently a couple of drafts, some will never be merged as they are more like experiments, others are a proof of concept that may eventually end up being merged and others are just a very early version of the work waiting for initial review. None of those currently open have ended up in the release closest to their opening time and for most it's still difficult to say which release they end up in. But at this point having a red failing status is not appropriate, I would rather have it as a skipped one and trigger the changelog once they are out of the draft PR status)

[pllim]

Because it's premature to think about changelog

True but personally failed changelog check in a draft PR does not bother me because it would be green anyway later. Perhaps that is a matter of personal preference that can vary greatly between people.

I would rather have it as a skipped one

Why not use the skip change log label in this case until you are ready for the check to run?

[bsipocz]

draft PR does not bother me

Me neither, but it does contributors. Recently I had ones who went in and started messing with the CI config and packaging to make their draft pass.

[pllim]

ones who went in and started messing with the CI config

Maybe this is a case of just telling them to ignore a red x until their PR is ready? Do you have a link to that PR for full context? Thanks!