NumberParser accept and ignore underscore

[privat]

The proposal is to extends number syntax to accept and ignore the underscore character _ (ASCII 95).

Many languages (including Python https://peps.python.org/pep-0515/ , Java https://docs.oracle.com/javase/7/docs/technotes/guides/language/underscores-literals.html or Ruby) accept some forms of numeric literal that ignore _.

The idea is to permit long literals that are still readable, eg. 1_000_000_000 is easier for a human than 100000000 especially since in the previous literal a zero is missing (I'm a tricky deceitful fellow).

Usually, there is not a lot of constraint on the placement of _ to accommodate various locales and usages or radix.

However, duplicate or training _ are usually rejected as syntax error.
The PR just ignore any _ without constrains. We might add them, or add them as rules for case that feel more being a syntax warning than a syntax error.

[svenvc]

I am not so sure it is a good idea to do this by default.
In NeoCSV I have my own number parser that supports this, as an option:

NeoNumberParser>>#digitGroupSeparator: separatorCharacter
	"Set the digit group separator to separatorCharacter.
	The are skipped while parsing digit characters.
	The default is nil (nothing being skipped)"
	
	digitGroupSeparator := separatorCharacter 

If you don't make it an option, you can't do strict checking anymore.

[guillep]

I'd like to have this in Pharo. I think that this can have a lot of value when working with numbers.
And I think it would be nice to to have also a number printer that adds separators...

I think this should become a phep.

[estebanlm]

phep is here: pharo-project/pheps#18

please let's discuss there :)