Wrong code in pythia used by the matching module

[Juergen Reuter]

There are a plethora of code statements in the matching routines which do not compile with the NAG compiler, i.e. already the make process fails. (cf. the HUDSON messages)

[dwiesler]

Solved the three bugs within the interface code in r2734.

The problem now is faulty pythia code. We should decide wether to fix this (and probably loose the possibility of automated svn updating to the recent pythia versions) or keep it this way and document it in the manual to only use gfortran for compiling pythia.

[dwiesler]

I tried to correct some obvious errors in pythia and finally found myself digging much to deep into a hell of bad code: [NAG Fortran Compiler error termination, 99 errors, 1259 warnings]

Therefore I opt for documenting this in the manual and enabling pythia only for gfortran compiler versions.

[dwiesler]

In r2736 I changed a few culprits in pythia, this should in principle solve the ticket issue. Waiting for a senior to confirm before closing ...

[Juergen Reuter]

There was one other issue, namely that (re-)compilation of PYTHIA from the MIISR directory had the same problem. I transferred Daniel's solution. But this is hacking here, MIISR should use, but not re-compile PYTHIA. We shall keep this in mind for 2.0.4. or so. For now, closing this ticket. (ThO, if you have a good and quick idea, please step in.)