relax SVN archive

Fix for the Relax_disp.test_korzhnev_2005_15n_sq_data system test for certain MS Windows systems.

This was reported as sr #3142 (http://gna.org/support/?3142).

The problem was simply the lower precision of this system.

The debugging command line option no longer turns on RelaxWarning tracebacks.

These tracebacks can be separately turned on with the --traceback command line option.

Added systemtest to analyse data.

Systemtest for CPMG dataset, (http://dx.doi.org/10.1073/pnas.0907387106)
Kaare Teilum, Melanie H. Smith, Eike Schulz, Lea C. Christensen, Gleb Solomentseva, Mikael Oliveberg, and Mikael Akkea 2009
SOD1-WT at 25 C.

Added dataset for systemtest.

Systemtest for CPMG dataset, (http://dx.doi.org/10.1073/pnas.0907387106)
Kaare Teilum, Melanie H. Smith, Eike Schulz, Lea C. Christensen, Gleb Solomentseva, Mikael Oliveberg, and Mikael Akkea 2009
SOD1-WT at 25 C.

Added the matplotlib module to the Python binary and module seeking script.

Fixes for two relaxation curve-fitting system tests for Python 3.3+.

These are the Relax_fit.test_curve_fitting_height and Relax_fit.test_curve_fitting_volume system
tests which fail due to accuracy differences and a bad call to the UnitTest method assertEqaul()
which should have been an assertAlmostEqual() call.

Changed the backward compatibility hook for old state files for the spectrometer frequency.

The behaviour has been changed so that the data pipe structure spectrometer_frq_list is now sorted.
This simply allows a number of tests to pass on Python 3.3+, a user would not notice any
differences.

Added 4 more relaxation dispersion system tests to the blacklist for when C modules are missing.

This allows these 4 tests to be skipped in the system test rather than failing with "RelaxError: The
exponential curve-fitting C module cannot be found.".

Added Python 3.4 support to some of the development scripts.

Fix for the Test_data.test_count_relax_times_r1rho relaxation dispersion unit test for Python 3.2+.

The '%s' representation of floating point numbers is different on these Python versions - the number
of decimal places used are different by default.