John Darrington <[EMAIL PROTECTED]> writes: > That sounds reasonable. It would cause a problem if there is a locale > which considers '@' and '#' to be alphabetic characters, but I'm not > aware of any such locale.
Agreed. > There may also be other characters that we can safely disallow > in any part of a name, for example '=', '+', '/' etc. It occurs to me that some locales might use ASCII = or + (etc.) as the second byte of a multibyte character sequence. UTF-8 is careful to avoid this pitfall but I don't know whether other locales are. That's why I only suggested checking the first byte, because it's guaranteed to be at the beginning of a possibly multibyte character. Internationalization is hard. -- "A computer is a state machine. Threads are for people who cant [sic] program state machines." --Alan Cox _______________________________________________ pspp-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/pspp-dev
