fix model_test

[kilian]

After fixing #578, there are still issues with model_test:

• few lines missing in the ref file (trivial)
• problem that shows up with the NAG compiler. Output:

  /afs/tp1.physik.uni-siegen.de/user/kilian/whizard/build/nagfor/libtool --tag=FC --mode=compile nagfor -c -I/afs/tp1.physik.uni-siegen.de/user/kilian/whizard/build/nagfor/src/models -I/afs/tp1.physik.uni-siegen.de/user/kilian/whizard/build/nagfor/src/omega/src -I/afs/tp1.physik.uni-siegen.de/user/kilian/whizard/build/nagfor/src/whizard-core -I/afs/tp1.physik.uni-siegen.de/user/kilian/whizard/build/nagfor/src/misc -C=all -f2003 -gline -nan -PIC models_ccww_threeshl_i1.f90
  libtool: compile:  nagfor -c -I/afs/tp1.physik.uni-siegen.de/user/kilian/whizard/build/nagfor/src/models -I/afs/tp1.physik.uni-siegen.de/user/kilian/whizard/build/nagfor/src/omega/src -I/afs/tp1.physik.uni-siegen.de/user/kilian/whizard/build/nagfor/src/whizard-core -I/afs/tp1.physik.uni-siegen.de/user/kilian/whizard/build/nagfor/src/misc -C=all -f2003 -gline -nan -PIC models_ccww_threeshl_i1.f90  -PIC -o .libs/models_ccww_threeshl_i1.o
  NAG Fortran Compiler Release 5.3.1(907)
  Extension: models_ccww_threeshl_i1.f90, line 3: Line longer than 132 characters
  Fatal Error: models_ccww_threeshl_i1.f90, line 32: Cannot find module TGLUE
               detected at PARAMETERS_THREESHL@<end-of-statement>
  make: *** [models_ccww_threeshl_i1.lo] Error 1
  | command: make compile -j1 -f models.makefile
  | Return code = 512
  ******************************************************************************
  ******************************************************************************
  *** FATAL ERROR: System command returned with nonzero status code
  ******************************************************************************
  ******************************************************************************
  WHIZARD run aborted.
  

[Juergen Reuter]

I fixed the additional lines the .ref file (Reading model file ...). This does not appear any more because now a model_list object is used, and hence the corresponding SLHA model is known in that model list. No model reading necessary. The other thing at the moment is mysterious to me, could be sort of a race condition or a clash of variables. Investigating.

[kilian]

Since the bug occurs in a rather obscure model, would it be an option to disable this until a solution is found? We might postpone the solution, if it is nontrivial. Currently, I can't run the testsuite because it aborts completely.

[Juergen Reuter]

I do have a solution, testing at the moment. But since it is at make distcheck, it takes its time. Sorry.

[kilian]

No problem. I was wondering just in case there is no simple solution.

[Juergen Reuter]

Jesus, forgot to close that one :P Done in r5063.