# New Ticket Created by Adam Thomason
# Please include the string: [perl #29402]
# in the subject line of all future correspondence about this issue.
# <URL: http://rt.perl.org:80/rt3/Ticket/Display.html?id=29402 >
The recently posted sha1 and md5 programs produce incorrect answers on amd64, at least
in part because bitops on 64-bit values are screwy. One particularly odd case:
_main:
set I1, 0xffffffff
shl I1, I1, 32
set I2, 0xffffffff
band I2, I2, I1
new P1, 32
set P1, 2
set P1[0], I1
set P1[1], I2
sprintf S1, "%lx %lx\n", P1
print S1 # prints "ffffffff00000000 ffffffff00000000"
end
Of course, ANDing ffffffff00000000 with 00000000ffffffff should be 0. ops/bit.ops is
implemented as $1 &= $2, which seems straightforward enough as C, pointing towards
gcc, but the following works fine:
int main( )
{
long I1 = 0xffffffff;
long I2 = 0xffffffff;
I1 <<= 32;
I2 = I1 & I2;
printf( "%lx %lx\n", I1, I2 ); # prints "ffffffff00000000 0"
}
I don't see any bitops tests specifically for 64-bit values... has anyone addressed
this?
--
Adam Thomason
[EMAIL PROTECTED]
