jv@mh.nl (Johan Vromans) (01/13/91)
[I was preparing this when patches 42-44 came in. So I quickly built 3.044 to verify that the problem still exists. It did.] System: DECsystem5000 / Ultrix4.0 / Gcc 1.37 [OSF version] 'split' considers the non-ascii character \0351 to be whitespace: @a = split (/\s+/,"abc\351def"); print join(":",@a), "\n"; => abc:def But with \s instead of \s+ : @a = split (/\s/,"abc\351def"); print join(":",@a), "\n"; => abc\351def However, "abc\351def" =~ /\s/ "abc\351def" =~ /\s+/ fail, like they should. Apparently the regexp matching and split use a different approach. The problem does not occur on VAX3100 / Ultrix3.1 / Gcc 1.37 . If it helps: in C: isspace('\351') => 8 char c = '\351'; isspace(c) => 8 unsigned char c = '\351'; isspace(c) => 0 Johan -- Johan Vromans jv@mh.nl via internet backbones Multihouse Automatisering bv uucp: ..!{uunet,hp4nl}!mh.nl!jv Doesburgweg 7, 2803 PL Gouda, The Netherlands phone/fax: +31 1820 62911/62500 ------------------------ "Arms are made for hugging" -------------------------