relax SVN archive

svn+ssh://bugman@.../svn/relax/1.3

........
r8191 | bugman | 2008-12-11 17:24:46 +0100 (Thu, 11 Dec 2008) | 8 lines

Fixes for bug #12573 (https://gna.org/bugs/index.php?12573).

This bug was reported by Jack Howarth (https://gna.org/users/jhowarth).

The problem was simply that the optimisation checks are exact and had to be slightly modified for
the Darwin kernel on 32 bit PowerPC.
........
r8192 | bugman | 2008-12-11 17:48:43 +0100 (Thu, 11 Dec 2008) | 11 lines

Fixes for the 2 test_atom_loop_spin_selection2() unit tests.

This problem was reported by Jack Howarth (https://gna.org/users/jhowarth) in the post at
https://mail.gna.org/public/relax-users/2008-10/msg00089.html (Message-id:
<20081024141150.GA879@...>).

The problem was with the position array comparison as this was being converted to a string and then
compared. The string representations of the numpy arrays is obviously different on the intel vs.
powerpc chips.
........
r8193 | bugman | 2008-12-11 18:09:51 +0100 (Thu, 11 Dec 2008) | 10 lines

Another PowerPC system test bug fix.

This problem was reported by Jack Howarth (https://gna.org/users/jhowarth) in the post at
https://mail.gna.org/public/relax-users/2008-10/msg00089.html (Message-id:
<20081024141150.GA879@...>).

The string.fromstring() function appears to be useless when switching between big and little endian
machines.
........