ROMS_AGRIF sediment application

[ahocean]

Dear friends,

There are some applications (specially SEDIMENT and BBL) which show errors in compilation. for example after setting "# define SEDIMENT" in the cppdefs.h file, this problem encounters:

Code: Select all

[ah@x Run]$ ./jobcomp 
OPERATING SYSTEM IS: Linux
PROCESSOR IS: i686
  file namelist.trc.sms exists in Run directory
/bin/cp: cannot stat `*.F': No such file or directory
/bin/cp: cannot stat `Make*': No such file or directory
Checking COMPILEAGRIF...
Checking COMPILEMPI...
/bin/rm -rf core *.o AGRIF/AGRIF_YOURFILES/*.o *.i *.s *.f *.trace *.mod 
/bin/rm -rf roms mpc cross_matrix cppcheck srcscheck checkkwds partit ncjoin ncrename ./rii_files
cpp -traditional -DLinux -P -I../../netcdf_ifc/ -IROMSFILES/AGRIF_INC mpc.F > mpc_.f
gfortran -O3 -fdefault-real-8 -fdefault-double-8  -o mpc mpc_.f
.
.
.
cpp -traditional -DLinux -P -I../../netcdf_ifc/ -IROMSFILES/AGRIF_INC analytical.F | ./mpc > analytical_.f
gfortran -c -O3 -fdefault-real-8 -fdefault-double-8 analytical_.f -o analytical.o
analytical_.f:661.35:

     &                      worksed                                     
                                   1
Error: Symbol 'worksed' at (1) has no IMPLICIT type
analytical_.f:1341.35:

     &                      worksed                                     
                                   1
Error: Symbol 'worksed' at (1) has no IMPLICIT type
make: *** [analytical.o] Error 1
mv: cannot stat `roms': No such file or directory

Also for the application BBL, after:
# undef SEDIMENT
# define BBL

Code: Select all

[ah@x Run]$ ./jobcomp 
OPERATING SYSTEM IS: Linux
PROCESSOR IS: i686
  file namelist.trc.sms exists in Run directory
/bin/cp: cannot stat `*.F': No such file or directory
/bin/cp: cannot stat `Make*': No such file or directory
Checking COMPILEAGRIF...
Checking COMPILEMPI...
/bin/rm -rf core *.o AGRIF/AGRIF_YOURFILES/*.o *.i *.s *.f *.trace *.mod 
/bin/rm -rf roms mpc cross_matrix cppcheck srcscheck checkkwds partit ncjoin ncrename ./rii_files
cpp -traditional -DLinux -P -I../../netcdf_ifc/ -IROMSFILES/AGRIF_INC mpc.F > mpc_.f
gfortran -O3 -fdefault-real-8 -fdefault-double-8  -o mpc mpc_.f
.
.
.
cpp -traditional -DLinux -P -I../../netcdf_ifc/ -IROMSFILES/AGRIF_INC def_rst.F | ./mpc > def_rst_.f
gfortran -c -O3 -fdefault-real-8 -fdefault-double-8 def_rst_.f -o def_rst.o
def_rst_.f:937.72:

        call nf_add_attribute(ncid, rstBBL(1), indxHrip, 5,             
                                                                        1
Error: Syntax error in argument list at (1)
def_rst_.f:938.17:

        NF_DOUBLE, ierr)                                                
                 1
Error: Cannot assign to a named constant at (1)
make: *** [def_rst.o] Error 1
mv: cannot stat `roms': No such file or directory

Any suggestion would be appreciated,
Thanks,

Amirhosein

[kate]

Is there a ROMS_AGRIF forum you can ask this on?

Otherwise, the troubles seem straightforward enough that you could try fixing them yourself. Which file has "worksed" in it and can you fix it by adding a definition for it? Is it used as a scalar or an array?

As for def_rst.F, look in there and see how attributes get added for the other variables and get the bbl ones to match.

[ahocean]

I hadn't noticed before that a forum was existed for ROMS_AGRIF :
https://gforge.inria.fr/forum/forum.php?forum_id=3526
Thank you for mentioning it.
But just a point; Is the problem surely due to some flaws in the FORTRAN modules or could it have been caused due to my own faults (e.g. not setting up cppdefs.h correctly, ...)?

[kate]

They are fortran problems. I have a copy of AGRIF from two years ago without either of these issues.