Fix FasterXML#3013: Implement int-to-string coercion config

[Tomasito665]

Description

This pull request implements configurable integer to string coercion.

Issue

#3013

[cowtowncoder]

Note to self: we do have CLA for this.

[cowtowncoder]

Could this be simplified by calling _parseString(...) and removing handling of START_OBJECT, VALUE_EMBEDDED_OBJECT (that are handled there)?

[Tomasito665]

Good one! Yes, it can. I've rewritten this method in terms of _parseString(...), leaving only the START_ARRAY case in the original method and the VALUE_STRING case at the very top as an optimization — I assume it is the most common case, so I figured it would make sense to keep that there even if it is already handled in _parseString(...). If you think this is a premature optimization, let me know and I'll remove it :-)

[cowtowncoder]

Looks good; I have one potential suggestion here, but happy to merge otherwise.

[Tomasito665]

I have implemented your suggestion. For some reason, the ObjectReaderValueOfWithValueTypeTest test case fails in CI for Java 8. The tests pass on my machine, both with Java 8 and 17. Any idea what's happening here?

[cowtowncoder]

@Tomasito665 That is odd and I doubt it has anything to do with your changes. I actually ended up removing that test case since I do not think it adds any value (uses Mock objects to verify something that ... does not seem valuable).

I'll see if I can get PR itself now reviewed.

[cowtowncoder]

Ok, this looks good. I will need to go over it one more time, merge hopefully tomorrow, possibly making minor tweaks.
I think we need to add an overload for old version of _parseString() just because something in other modules might be calling it -- need to default NullValueProvider to, NullsConstantProvider.nuller(). If you get to do it first that'd be great; if not I'll add it after merging.

[cowtowncoder]

@Tomasito665 I just merged this, so if you want to do another PR for coercion from Float etc, should be possible.

Thank you again for contributing this!

[Tomasito665]

Wonderful, thanks for merging, @cowtowncoder. I wanted to add the method overload for _parseString() without the null provider, as you suggested, but you were faster :-) I'll file a new PR for the pending types in the coming days.