[ET Trac] [Einstein Toolkit] #849: Drop explicit support for Fortran 77 in Cactus

[Einstein Toolkit]

#849: Drop explicit support for Fortran 77 in Cactus
--------------------------+-------------------------------------------------
  Reporter:  eschnett     |       Owner:             
      Type:  enhancement  |      Status:  reviewed_ok
  Priority:  minor        |   Milestone:             
 Component:  Cactus       |     Version:             
Resolution:               |    Keywords:             
--------------------------+-------------------------------------------------

Comment (by rhaas):

 Would that not only apply to source files in a thorn's src directory
 (since it is tied one of Cactus' makefile fragments in Cactus/libs/make)?
 An ExternalLibrary's configure.sh could do whatever it wants to do. BLAS
 in particular contains:
 {{{
         cat > make.cactus <<EOF
 SRCS = $(echo *.f)
 libblas.a: \$(SRCS:%.f=%.o)
         ${AR} ${ARFLAGS} \$@ \$^
         ${RANLIB} ${RANLIBFLAGS} \$@
 %.o: %.f
         ${F77} ${F77FLAGS} ${FIXEDF77FLAGS} -c \$*.f -o \$*.o
 EOF
         ${MAKE} -f make.cactus
 }}}
 so it uses the F77 compiler for the .f files inside of BLAS. It does
 compile though when I comment out F77 in my option list (though there are
 lots of warnings), which given that there is no difference between the F77
 options and the F90 option in my optionlist is not that surprising I
 guess.

 The reason I bring this up that I know of some old code (spherepack
 https://www2.cisl.ucar.edu/resources/legacy/spherepack) that does actually
 use arithmetic if statements that are (I think) no longer allowed in new
 Fortran versions. Since we need to support Fortran 2003 (LoopControl uses
 it) we need to be a bit more careful than supporting only F90 (which I
 think is indeed a proper superset of F77).

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/849#comment:4>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit