http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50192
Bug #: 50192
Summary: Wrong character comparision with wide strings
Classification: Unclassified
Product: gcc
Version: 4.7.0
Status: UNCONFIRMED
Keywords: wrong-code
Severity: normal
Priority: P3
Component: libfortran
AssignedTo: unassig...@gcc.gnu.org
ReportedBy: tkoe...@gcc.gnu.org
CC: jvdeli...@gcc.gnu.org
ig25@linux-fd1f:~/Krempel/4> cat compare.f90
program main
character(kind=4,len=2) :: c1, c2
c1 = 4_' '
c2 = 4_' '
c1(1:1) = transfer(257, mold=c1)
c2(1:1) = transfer(64, mold=c2)
if (c1 < c2) print *,"strange..."
end program main
ig25@linux-fd1f:~/Krempel/4> gfortran compare.f90
ig25@linux-fd1f:~/Krempel/4> ./a.out
strange...
The problem is in compare_string. We use
int
compare_string (gfc_charlen_type len1, const CHARTYPE *s1,
gfc_charlen_type len2, const CHARTYPE *s2)
{
const UCHARTYPE *s;
gfc_charlen_type len;
int res;
res = memcmp (s1, s2, ((len1 < len2) ? len1 : len2) * sizeof (CHARTYPE));
which works on big-endian, but not on little-endian, where the memcmp is used
to compare (bytewise) 0x01 0x01 0x00 0x00 vs. 0x40 0x00 0x00 0x00.
