Dnia czw 28. sierpnia 2003 12:34, Simon Marlow napisał:

The only right way to do this, it seems, is to generate the tables from UnicodeData.txt. However, I'm prepared to live with the current solution as long as we document its shortcomings. After all, it does the right thing on the majority of our installed base.

I have some UnicodeData.txt-generated predicate functions & toUpper/toLower in QForeign. There doesn't seem to be an official mapping from Unicode character categories to various predicates (I've once tried to find one asking on Unicode groups with no success - I was told that for good definitions of some predicates character categories are not precise enough), or even the set of useful predicates, and it's not clear which predicates should recognize only ASCII characters (most probably isDigit). It's all to be designed, perhaps even by changing Haskell 98 a bit. QForeign at least has some machinery to generate tables and functions from UnicodeData and some proposals of predicate definitions. I think these functions should all be locale-indepentent, and ideally Haskell should have a portable definition - perhaps in terms of character categories. Usable wcwidth requires it to be the same on both sides of a remote terminal; <http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c> suggests a definition of wcwidth. wcwidth was removed from the ANSI C addendum which defined other wide character functions, but it's in some POSIX standards. -- __("< Marcin Kowalczyk \__/ qrczak@knm.org.pl ^^ http://qrnik.knm.org.pl/~qrczak/