OpenBSD 3.4 isspace() b0rked (was: Problem compacting databases (again!))

[Matthias Andree]

Matthias Andree <matthias.andree at gmx.de> writes:

> That has no relevance. isspace() is only valid for EOF and the values
> that can be represented in an "unsigned char". 0xffffffe0 does not fall
> into this category, hence the isspace() behavior is undefined.¹

Bingo - I tried this on SourceForge's compile farm. The second line
should have been 0xa0 0 (zero as the 2nd number) according to POSIX and
OpenBSD documentation, but we got 8 (eight) instead.

x86-openbsd1:~$ ./try
0xa0 8
x86-openbsd1:~$ uname -a
OpenBSD x86-openbsd1.cf.sourceforge.net 3.4 GENERIC#18 i386
x86-openbsd1:~$ shar try.c
# This is a shell archive.  Save it in a file, remove anything before
# this line, and then unpack it by entering "sh file".  Note, it may
# create directories; files and directories will be owned by you and
# have default permissions.
#
# This archive contains:
#
#       try.c
#
echo x - try.c
sed 's/^X//' >try.c << 'END-of-try.c'
X#include <stdio.h>
X#include <ctype.h>
X
Xint main(int argc, char **argv)
X{
X        unsigned char x = 0xa0;
X        printf("0x%02x %d\n", x, isspace(x));
X        return (0);
X}
END-of-try.c
exit

OpenBSD claims compliance, in the isspace(3) manual page.

     The isspace() function tests for the standard whitespace characters for
     which isalnum(3) is false.  The standard whitespace characters are the
     following:

           ` '    Space character.
           \f     Form feed.
           \n     New-line.
           \r     Carriage return.
           \t     Horizontal tab.
           \v     And vertical tab.

     In the C locale, isspace() returns true only for the standard whitespace
     characters.

But fails to meets its own specification. Cc'd to bugs at openbsd.org

Note that bogofilter does not switch locale, so it is in the default C
or POSIX locale (this is synonymous).

-- 
Matthias Andree