------- Comment #4 from tkoenig at gcc dot gnu dot org 2006-05-28 16:10 ------- (In reply to comment #3) > Created an attachment (id=11525) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11525&action=view) [edit] > Proposed patch > > I'm tempted to propose the attached patch. It makes the front-end ultimately > use unsigned chars to do the comparison, and also removes the use of the > reverse ASCII collating table for this, because I don't understand how it is > supposed to be useful :)
What about using struct { int length; unsigned char *string; } character; in gfc_expr? There would be a few fallout issues in other places (I assume some "char *" pointers would have to become "unsigned char*"), but it would be much cleaner overall. As far as the collating sequence goes... should we eventually support locales for sorting etc.? If so, we should leave the code in. in gfc -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27715