Code doesn't compile with recent 5.0

[Juergen Reuter]

This is the problem

Making all in qft
/bin/sh ../../libtool  --tag=FC   --mode=compile gfortran -I../basics -I../utilities -I../testing -I../system -I../combinatorics -I../expr_base -I../physics  -fopenmp -g -O2 -c -o model_data.lo  model_data.f90
libtool: compile:  gfortran -I../basics -I../utilities -I../testing -I../system -I../combinatorics -I../expr_base -I../physics -fopenmp -g -O2 -c model_data.f90  -fno-common -o .libs/model_data.o
model_data.f90:329:49:

   pure subroutine modelpar_data_init_complex (par, name, value)
                                                 1
Error: INTENT(OUT) argument 'par' of pure procedure ‘modelpar_data_init_complex’ at (1) may not be polymorphic
model_data.f90:321:46:

   pure subroutine modelpar_data_init_real (par, name, value)
                                              1
Error: INTENT(OUT) argument 'par' of pure procedure ‘modelpar_data_init_real’ at (1) may not be polymorphic
model_data.f90:700:13:

     buffer = prt%get_name (.false.)
             1
Error: ELEMENTAL procedure pointer component ‘get_name’ is not allowed as an actual argument at (1)
model_data.f90:716:16:

        buffer = prt%get_name (.true.)
                1
Error: ELEMENTAL procedure pointer component ‘get_name’ is not allowed as an actual argument at (1)
make[2]: *** [model_data.lo] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all-recursive] Error 1

The svn revision is 220505. We have to check whether this is a problem with the compiler or our code

[Juergen Reuter]

Bazinga: the first thing is valid in F2003, but _not_ in F2008, confer ​https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54756. Simplest solution is to impure these two subroutines.

[Juergen Reuter]

The first one was revision r219085 of gcc, the second one also happened close to Christmas last year. My last update (before the one now) was for October or so.

[Juergen Reuter]

What Wolfgang wrote seems to violate the Fortran Handbook paragraph 7.5.5.2, subparagraph 3, last sentence :D

[Juergen Reuter]

As of r6550, all problems with gcc 5.0.0 have been resolved. Closing. Commit stop remains until 5.0 is tested on Jenkins.