WIP: fix standalone output for vector.inc production and usage

[roiser]

This fix applies to the "gpucpp" branch of the repo. When running e.g.

generate e+ e- > mu+ mu-
output standalone test --vector_size=32

there this will produce the error

The compilation fails with the following output message:
	    cd MODEL; make
	    make[1]: Entering directory '/afs/cern.ch/work/r/roiser/sw/madgraph4gpu/MG5aMC/mg5amcnlo/guttu2.sa/Source/MODEL'
	    /opt/rh/gcc-toolset-11/root/usr/bin/gfortran -w -fPIC  -ffixed-line-length-132  -c -o couplings.o couplings.f
	    couplings.f:16: Error: Can't open included file '../vector.inc'
	    make[1]: *** [<builtin>: couplings.o] Error 1
	    make[1]: Leaving directory '/afs/cern.ch/work/r/roiser/sw/madgraph4gpu/MG5aMC/mg5amcnlo/guttu2.sa/Source/MODEL'
	    make: *** [makefile:59: ../lib/libmodel.a] Error 2

the reason being that the vector.inc file containing the vector size for the parallel execution is not being produced. This PR changes the relevant madgraph/iolibs/export_v4.py in three places

• it calls self.write_vector_size also for the SA output
• it moves the write_vector_size from the ProcessExporterFortranME class to its base ProcessExporterFortran so that it can be used also in ProcessExporterFortranSA
• it adds one missing include of vector.inc in couplings.f which maybe was overlooked

[oliviermattelaer]

Thanks a lot Stefan,

Sorry to be nitpicking on this.
But I do not think that you need to move the function "write_vector_size"
Can you revert it to his original place (this is just to avoid spurious conflict).

If you did try and this was not working then no problem.
Consider this has an approval of merging and after this check (and revert if possible) you can merge yourself (if you can, otherwise will do it).

Thanks,

Olivier