cpplib: Preliminary implementation of UCNs

Neil Booth neil@daikokuya.co.uk
Sun Apr 20 07:05:00 GMT 2003


Zack Weinberg wrote:-

> Neil Booth <neil@daikokuya.co.uk> writes:
> 
> ...
> >I don't think it's a good idea to make \u00aa and \u00AA the same
> > identifier, never mind making it the same as the character itself.
> 
> I disagree strongly - please give a rationale for your position.

It breaks the fundamental point that a unique sequence of bytes that
is lexically an identifier is a unique identifier.

In terms of future enhancements, it makes some of what I was planning
to do with cpplib harder, and possibly impossible.

> I don't want support for UCNs in identifiers to be enabled until we
> have both consensus on how it should work, and a complete
> implementation.  This isn't to say that your code shouldn't go in;
> only that it should be disabled for identifiers, requiring one to edit
> the source code to turn it back on.  That way it's there for
> experimentation, but no one comes to rely on any behavior that we
> are likely to change later.

OK, I'll #if-0 out the check in forms_identifier_p, and xfail the
testsuite for ucn-1.c I guess.

Neil.



More information about the Gcc-patches mailing list