relax SVN archive

The diffusion tensor spheroid_type variable is now only updated if it is never explicitly set.

This is to fix the diffusion tensor setting bug found by Vitaly Vostri <vvostri att gmail dott com>
(Message-id: <AANLkTinRO3ZCNwL6GeqWdx0uBvv=PQ=zw6vpdQQeruBW@...>
https://mail.gna.org/public/relax-users/2010-12/msg00014.html).

The result of this bug is that initialising a prolate diffusion tensor with Da set to zero caused
the spheroid_type to be automatically set to 'oblate'. This bug was introduced at exactly r11678,
and was part of changes for better BMRB support. This affects relax version 1.3.6 and the 1.3
repository line from r11678 to 12174.

The type of spheroid is now being checked in all of the oblate and prolate system tests.

Added a new system test for the prolate diffusion tensor, this time checking the diffusion type.

A spherical tensor is sent in, triggering the bug found by Vitaly Vostri <vvostri att gmail dott
com> (Message-id: <AANLkTinRO3ZCNwL6GeqWdx0uBvv=PQ=zw6vpdQQeruBW@...>
https://mail.gna.org/public/relax-users/2010-12/msg00014.html).

The problem is that the prolate tensor is changed to an oblate one!?!

Added the spin.copy user function window and menu entry.

Fix for the _get_spin_info() and _get_res_info() methods - they now return None, None if not found.

The _get_spin_id() and _get_spin_info() methods can now handle suffixes for the spin names.

This allows self.spin, self.spin_from, etc. to be used.

Added the file sample_scripts/full_analysis.py with a single line pointing to dauvergne_protocol.py.

Renamed the full_analysis.py sample script to dauvergne_protocol.py to make it more obvious what it is.

Fix for subdirectory names.

In scipt mode, the dAuvergne_protocol function created wrong subfolders for the diffusion models (e.g. 'proteinsphere' instead of 'protein/sphere').

Fix for the Mac OS X Python 2.6.6 link at http://www.nmr-relax.com/download.html.