relax SVN archive

Fix for bug #7845 (https://gna.org/bugs/?7845).

This bug was reported by Sébastien Morin (https://gna.org/users/semor).

The fix was to increase the precision of the chi-squared value in the model-free results file.

Ported r2881 and r2882 from the 1.3 line.

The command used was:
svn merge -r2880:2882 svn+ssh://bugman@.../svn/relax/1.3

This fixes two misspellings of the 'params_converged' flag and reformats the print out messages of
the convergence tests.

Improvements to the printed out convergence test messages.

Fixed two misspellings of the 'params_converged' flag.

Ported r2878 and r2879 from the 1.3 line.

The command used was:
svn merge -r2877:2879 svn+ssh://bugman@.../svn/relax/1.3

This fixes two bugs in the 'full_analysis.py' script.

Fixed a bug in the convergence tests where a float was being printed as a string.

Bug fix for the failure of model selection in the 'full_analysis.py' script.

Because the 'run.delete()' user function had been removed the 'previous' run, which is a global
model, was being included in the list of runs (m0, m1, m2, etc.) for model selection cause a
failure.

Ported r2872 from the 1.3 line.

The command used was:
svn merge -r2871:2872 svn+ssh://bugman@.../svn/relax/1.3

This fixes a bug in the 'full_analysis.py' script whereby 'run.delete()' was being called destroying
the data required for the convergence tests.

Ported r2868 from the 1.3 line.

The command used was:
svn merge -r2867:2868 svn+ssh://bugman@.../svn/relax/1.3

This fixes the parameter convergence test in 'full_analysis.py'.

Ported r2864 and r2865 from the 1.3 line.

The command used was:
svn merge -r2863:2865 svn+ssh://bugman@.../svn/relax/1.3

This adds convergence tests to the 'full_analysis.py' script so that the user is told whether
convergence has occured.