Some fixes and stuff for resonance fit

Fixes:

fix partial-wave fitters in legacy, on Mac a missing colon after a parameter definition for getopt does apparently not matter, for Linux it definitively does
explicitly load the std::comples dictionaries in calcIntegrals

Cosmetics and consistency:

remove the option to write ASCII integral matrices with calcIntegrals from legacy
replace norm.int by norm.root in a couple of places (help texts and default values) (also at places where the ROOT version had been required before)
remove all include statements from the header files in pyInterface only declaring a single export... function, those includes are not required to parse and understand the header files, but are only required in the implementation
rename utils.* to reportingUtilsEnvironment.*
add breakupMomentum and barrierFactorSquared to the pyInterface (those functions have been chosen because I needed them)
update wavelists and scripts generating the keyfiles for pi- pi- pi+, pi- pi0 pi0, and pi- eta eta to the new naming scheme for waves
fix the pyRootPwa test executed by make test, the test still expected a wave name according to the old naming scheme, the expectation has been updated to the new naming scheme

Resonance fit:

remove the parameter to shift the mass used to calculate the break-up momentum in exponentialBackground
add option to set the exponent of the break-up momentum in the background parameterisations
add option to add a barrier factor using an angular momentum from the options file to the break-up momentum (q*F^2(l))^exponent
add new background parametrisations using the phase-space integral instead of the break-up momentum (and the barrier factors)
normalise q*F^2 and the phase-space integral at the highest mass-bin, this should make the parameters for the usual backgrounds and the backgrounds using the phase-space integrals somewhat comparable, but this will render previous fit results to be no longer correct

ROOTPWA-Maintainers / ROOTPWA