> > Note: Perhaps a solution is to have two variants of the procs, one for
> > code points and one for code units. The code units variants would > > guarantee O(1) and the code point ones wouldn't. > > In R5.92RS, Scheme characters are codepoints (or more precisely scalar > values), not code units. Right, that's kind of the point; the code unit variants wouldn't return characters. The other ideas on the table seem to be UCS-4 strings or O(n) string-ref for UTF-8/UTF-16 strings. O(n) string-ref would make something as simple as searching for a substring in a string rather slow. And there would be no standard way of making it faster. _______________________________________________ r6rs-discuss mailing list [email protected] http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
