Open josefongit opened 3 years ago
I believe that the flag -fallow-argument-mismatch is the correct one to use.
I was able to compile FAC using macports openmpi gcc version 10 compilers, and I did indeed need to use this flag. The configure option that enables adding fortran compiler flags for MPI compilers is --with-mpifflag.
For compiling with ordinary macports gfortran 10, I tried to set the FFLAGS environment variable, but it is not used in the Makefiles generated by configure. I think there is a bug in configure that is preventing FFLAGS from being correctly passed through.
Since I was able to get the MPI compiler to work (and also gcc9 gfortran for non-parallel work), I gave up on trying to fix the problem.
I obtain the same error in coulcc.f
as josefongit. I am running gfortran 12.1.0 on a Mac with OS Ventura 13.2. Configuring with the -fallow-arguent-mismatch
exits successfully, but make
fails.
Configure command:
./configure FCFLAGS="-fallow-argument-mismatch" FFLAGS="-fallow-argument-mismatch"
Error during make
:
Error: Type mismatch in argument ‘x’ at (1); passed COMPLEX(8) to REAL(8) coulcc.f:237:10:
237 | F = CF1R(X,ETA,ZLL,ACC8,SF ,RK, ETANE0,LIMIT,ERR,NFP, | 1
Compiling issues occurred with release 1.1.4. After reading #604, I downloaded a .zip of the repository and was able to successfully compile that. A note could be included in the installation guide that recommends using the current repository and not the most recent release.
Hi,
compilation aborts with a number of coding error messages. I use gfortran/gcc 10.2 on a Mac (non-Apple versions). The trouble starts in a Coulomb routine:
I can fix the bunch of errors in coulcc.f by hand, but then compilation next stops in the file tn.f
I ran FAC in the past and compilation wasn't a problem. Could it be a stricter interpretation of language with newer compilers? If this is true, such errors are bound to appear more frequently. The flag -fallow-argument-mismatch did not make it work either.
Thanks a lot, Josef