Improve JavaDoc for config ACCEPT_EMPTY_STRING_AS_NULL_OBJECT wrt special cases

[JooHyukKim]

relates to #4009

[cowtowncoder]

Does not need to be set on mapper; could set on ObjectReader instead?

[cowtowncoder]

Hmmh. I think that this behavior should be changed for 2.16, see my notes on #4009.

What could be done here would be:

1. Change to target 2.15, to verify then-current behavior of coercion
2. Remove Javadoc comment, or make it suggest that there are gray areas; but NOT establish behavior wrt java.util.Local given change for 2.16.

If test was added, it would then be modified as part of #4009 for 2.16.

[JooHyukKim]

Hmmh. I think that this behavior should be changed for 2.16, see my notes on #4009.

Oh yes I just read the comment 👍🏻.

1. Change to target 2.15, to verify then-current behavior of coercion

WIll do.

2. Remove Javadoc comment, or make it suggest that there are gray areas; but NOT establish behavior wrt java.util.Local given change for 2.16.

Will remove reference to Locale.

[JooHyukKim]

For safe verion control, I will not test my git skills rebasing from 2.16 to.2.15 😅 and close this one and cherry pick the changes back.

[cowtowncoder]

Sorry about this. I have not figure out how to use rebase properly with Github PRs in cases like this. :-(

[JooHyukKim]

Np! I don't think there is legit clean way to do such thing, speaking from logical perspective 😆

[cowtowncoder]

Yeah I guess at the point PR is created it is bound to specified branch as... base?
It's just unfortunate that while creating PR it is simple enough to see what diffs were merged, based on target, and this is no longer available once created.