relax SVN archive

Significant improvements to the new error/warning system.

The idea of this system was suggested in two threads which commence at:
https://mail.gna.org/public/relax-devel/2006-08/msg00052.html
https://mail.gna.org/public/relax-devel/2006-08/msg00053.html

The improvements includes the addition of the printing of the traceback, the option of saving the
program state (self.relax.data pickling), a new way of calling the RelaxErrors or RelaxWarnings
(for example RelaxBadError('Hello')), addition of the warning functions, and a number of other small
changes.

Start of a new relax error/warning subsystem.

The idea of this system was suggested in two threads which commence at:
https://mail.gna.org/public/relax-devel/2006-08/msg00052.html
https://mail.gna.org/public/relax-devel/2006-08/msg00053.html

Essentially a rudimentary framework for replacing the entire Error and Warning system has been set
up. Many essential elements are missing (for example the warning function, tracebacks, etc)
however it shouldn't take long to add these. This new system significantly simplifies the
RelaxError/RelaxWarning system, allows easy implementation of the --ignore-errors and --pedantic
command line flags (https://mail.gna.org/public/relax-devel/2006-08/msg00052.html), and allows fine
tuning of how RelaxErrors and RelaxWarnings are handled within the different relax user interfaces
(UIs).

Janitorial changes to r2533

As discussed on relax-devel at gna dot org:
https://mail.gna.org/public/relax-devel/2006-08/msg00055.html

Alterations to the debug mode.

These changes reflect the discussion on relax-devel at gna dot org:
https://mail.gna.org/public/relax-devel/2006-08/msg00049.html

Two functions are implimented when in debuging mode:
1) KeyboardInterupt during script execution will print the normal
python traceback
2) The all of the relax exception classes will save the program
state

Created the RelaxWarning subsystem. The standard 'RelaxWarning' and the 'RelaxZeroVectorWarning'
warning classes have been created.

Modification of proposed fix to bug #6503

Based on discussion on relax-devel at gna dot org:
https://mail.gna.org/public/relax-devel/2006-08/msg00033.html

This modification does not raise RelaxError if the amide bond length
is 0, but rather sets the bond vector to None

Fix for the RelaxInfError and RelaxNaNError.

Proposed changes to fix bug #6503 (https://gna.org/bugs/?func=detailitem&item_id=6503) which has
been discussed within the thread starting at
https://mail.gna.org/public/relax-devel/2006-08/msg00000.html.

Made a private branch of the 1.2 line to test proposed NaN catching code.

Modified the release checklist and added an entry for updating the 'Release' field of the bug
tracker.