Literal concatenation, strings vs. numbers

[greg]

Ben Finney wrote:

> So, different representations of literals are parsed as separate
> literals, then concatenated. To have the behaviour you describe, the
> case needs to be made separately that digit concatenation should not be
> consistent with the established string literal parsing behaviour.

I think it's a pretty easy case to make, since there is no
obvious way of "concatenating" numbers written in different
bases. So if it's to be allowed at all, it pretty much has
to be restricted to a single base.

However, I'd be just as happy with underscores. I don't
see how there could be any great difficulty with implementing
that -- it only affects the scanner.

-- 
Greg