...I remembered that it produced compiler-dependent results.

Added a script to simplifly the generation of tests and modified the testing environment. Documentation still missing. Contributes to #190

The preconditioner already improves the convergence of Cu and Fe files, but the starting distance is higher than for the older fleur version and tests run up to test 7 only (does not fail, but hangs).

...because the other tests involving LOs seem to be not sensitive enough to LO errors.

This is because these test systems have LOs and we realized that
our tests are not sensitive enough for LO related errors. I hope this
change helps. But I cannot guarantee this.

The old test is renamed to SiFilmPlotXML.

+ bugfix for density slice generation in main/cdngen.F90

The tolerance was increased from 10^-7 Htr to 10^-6 Htr.

The change was introduced in commit 5bf8da4d and this temporarily
broke the result comparisons in the Bi2Te3 test. The Bi2Te3XML test was not affect since it
uses an enpara file.