DEPR: implicit resample in plotting

[jbrockmendel]

• closes #xxxx (Replace xxxx with the GitHub issue number)
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

The use cases for this are all along the lines of

dti1 = pd.date_range(..., freq=high)
dti2 = pd.date_range(..., freq=low)

ser1 = pd.Series(..., index=dti1)
ser2 = pd.Series(..., index=dti2)

_, ax = mpl.pyplot.subplots()

ser1.plot(ax=ax)
ser2.plot(ax=ax)

In this case, ser2 gets silently resampled (period-style, which is different from regular resample i think) so that the x-axis matches the plot for ser1. (I think. We don't have any comments/docstrings explaining "why" for any of the plotting code).

In these cases, "reindex yourself" is a totally valid workaround. Since this is a big use case for kludged state information, I'm hopeful it will move us in the direction of #54485 too.

[jbrockmendel]

this bit is unrelated, can revert and do it in a separate branch

[mroeschke]

Could go as far to suggest f"Call series.resample('{freq}') ..."?

[jbrockmendel]

Sure.

But now im wondering how much this will even help in moving towards #54485. In order to check if we're in this situation we still need the state information attached somehow.

I know "nuke it from space" isn't a viable approach to the plotting code, but it is really really bad.

[jbrockmendel]

it isn't clear to me why the warning is emitted here since ax isn't passed. if that keyword isn't the relevant factor in deciding if the statefulness is needed, then #62442 may not achieve its goals