chore: document upper bound for python lib 'holidays' >= 0.26

[mistercrunch]

Change of plans, adding a note documenting the upper bound

I simply removed the upper bound for the holidays lib and ran pip-compile-multi -P holidays. Note that pip-compile-multi isn't as' deterministic as we'd like for it to be, but things should be fine as "# vias" are just informational here.

For the record: https://github.com/vacanza/python-holidays/blob/dev/CHANGES

[john-bodley]

@michael-s-molina or @villebro do either of you know why we previously had an upper bound of 0.26 here?

[john-bodley]

Should we bump the lower bound and also place an upper bound of 1.0.0 as well?

[john-bodley]

@mistercrunch it would be good to determine if the issue in #25017 has been resolved.

[mistercrunch]

@john-bodley great catch! I totally would have missed that. I think we really need to add comments for all real upper bounds in pyroject.toml, including if it's just "based on semantic versioning and/or stress levels around this specific dependency".

I'm planning on implementing something like supersetbot bump-python --disregard-upper-bound which would do as specified and remove the upper bound in the process. We may need more parsable hints as smart comments so that @supersetbot knows what to do with it say:

"holidays>=0.25, <0.26", #known-upper-bound-issue #couple-with-lib=prophet

or something like

"holidays>=0.25, <1", #semver-upper-bound 

[mistercrunch]

@rusackas closing/cleaning some old PRs if you don't mind stamping this