relax SVN archive

Fix for bug #5698 and related issues.

Adds several missing functions to specific_fns/jw_mapping.py, and some
missing pointers to jw_mapping functions in specific_fns/specific_setup.py

Also adds jw_mapping tests to the test suite

fix for missing files from maths_fns in r2455

edward at nmr-relax.com

https://gna.org/bugs/index.php?func=detailitem&item_id=5602

removed memory leak in fitting functions c module

edward at nmr-relax.com

https://gna.org/bugs/index.php?func=detailitem&item_id=5602
https://gna.org/bugs/index.php?func=detailitem&item_id=5473

The creation of a framework for the generation of Molmol macros.

A large number of changes and additions have been made including:

The user function 'molmol.write()' has been created.

A number of functions in the generic molmol code have been renamed. These are 'open_pipe' to
'pipe_open', 'pipe_open' to 'pipe_open_test', and 'write' to 'pipe_write'.

The functions 'self.write()' and 'self.create_macro()' have been added to 'generic_fns/molmol.py'.
The first function calls the second which returns an array of Molmol commands. The point of
splitting the functions is so in the future 'self.exec_macro()' can be added to pipe the command
directly into Molmol.

The two functions 'self.molmol_macro()' and 'self.molmol_macro_classic()' have been added to the
model-free specific code. The second function currently returns a dummy array so that the resultant
macro is useless. The first function is returned by the specific_setup.py code.

A new error class called 'RelaxStyleError' has been created. This error is to be called anytime an
argument called 'style' is unrecognised.

Added Gary Thompson to the list of full access committers.

The run type 'hybrid' model-statistic returning function now handles numerous instances.

Fix for bug introduced in r2450. overfit_deselect() evaluated
self.relax.data.diff[run], and this raised a KeyError in the event the
run was of the local_tm modelfree type.

Fix for bug #5501. Removes all residue selection on read operations, and
impliments protection from underfitting in minimise(), grid_search() and
calc()

Model selection is now enabled for hybrid models. This completes task #3122.

The location of the task is https://gna.org/task/?func=detailitem&item_id=3122.

Fixed two bugs in the hybridisation code.