>>>>> "R" == Rudi Cilibrasi <cilibrar / ofb.net> writes:

R> on the same file on a different machine.  Note that Guy Decoux was probably
R> trying to demonstrate this effect using his test program, but in fact
R> I believe he demonstrated instead that the standard floating point math
R> functions are not 100% consistent across platforms (as the spec allows
R> some latitude), and in fact indirectly proved that the patch is working
R> in an endian-independent fashion.

 I was trying to show that when you work with float you'll never test for
 equality but always compare the 2 floats with some precisions.

 If I have switched from x86 to sparc, this was in reality to have 2
 differents implementations of IEEE754. The difference that you have seen
 is in the implementation of IEEE754.

 You can try to store the exact representation of IEEE754 in marshal, but
 who really care when you know how is defined this standard.


Guy Decoux