On 3/31/2010 7:49 AM, Glenn H Sembroski wrote:
Opps, I left a BIND(C) in my fortran code snippet that shouldn't be
there (I was trying this and that unsuccessfully to fix the problem.
The corrected problem description follows:
*****************************************************************************************
Hi all,
I am having a problem where I get a segfault when my C++ main program
calls a f77 entry point. I am using GCC4.1.2. The C++ is compiled with
g++, the fortran with gfortran. I use the -ffixed-line-length-none
-fno-automatic -fno-underscoring -fbounds-check options with gfortran.
ENTRY RLUXGO(LUX,INS,K1,K2)
I have stepped through this with gdb and the segfault occurs when I
step into the rluxgo call .
These files compiled and built with no problems. I have also compiled
and built these same files using the Intel fortran compiler. That
program ran with no problems.
Do I need a later version of GCC? Is there something else I need to do?
Thanks,
bind(c) would be a better approach, but you would need a more current
gfortran, for sure.
legacy usage of ENTRY incurs many possible abuses, such as an
expectation of all variables default SAVE without so specifying.
-fno-automatic might compensate for missing SAVE, but also may not.
--
Tim Prince
