relax SVN archive

A new PDB file with 3 models and a few atoms for testing of the structure.web_of_motion user function.

Created the structure.web_of_motion user function.

This is used to create a special PDB file which represents the atomic motions between different
structural models. Identical atoms of the selected models are concatenated into one model, within
a temporary internal structural object, and linked together using PDB CONECT records.

Fix for the assembly of atomic positions for the frame order analysis.

This is for the _minimise_setup_atomic_pos() method. Previously the warning "RelaxWarning:
Averaging the 1 atomic positions for the PCS for the spin 'x'" was being given if a single atomic
position was found.

Bug fix for the frame order analysis when multiple models are present.

The averaging of the atomic positions was not working and instead a sum was being performed.

Fix for the frame order grid search for when the pivot point is to be optimised.

The construction of the custom grid point array was failing when the pivot point is to be optimised.

The grid search setup in the frame order auto-analysis now supports the pivot point optimisation.

The grid search increments method custom_grid_incs() did not take pivot point optimisation into
account. Now it says that no grid search over these parameters should be performed.

Added HTML pages for the old relax 1.2 and 1.3 manuals to redirect the user correctly.

There are still links to these manuals across the internet. For example at
http://www.sbgrid.org/news/newsletters/2009. These basic HTML pages use the META redirect,
JavaScript redirect, and provide a link if all else fails to the
http://www.nmr-relax.com/manual/index.html page.

Fix for the frame order analysis for a bug introduced at r18566.

The _domain_moving() method has been redefined to return the spin ID string corresponding to the
moving domain rather than the domain ID, as this method is only used for restricting the spin_loop()
and interatomic_loop() methods. The changes allows the loop methods to function prior to defining
the moving domain.

Big set of fixes for the frame order analysis.

RDC and PCS data for the non-moving domain are now not used in the analysis. This was not a problem
during development as the test models had no RDC and PCS data for the non-moving domain. Now the
spin_loop() and interatom_loop() methods are called with the moving domain ID string as the
selection argument. This avoids all non-moving domain data, as should happen for the analysis.

Improved the default RDC and PCS errors for the frame order target functions if these are not loaded.

These now default to 1 Hz and 0.1 ppm respectively. The previous 0.03 ppm PCS error default was far
too small, causing the PCS to have much more weight than the RDC.