I've also had problems with linking in variables through f90
modules in a larger code I've been trying to port to ifort on my
MacBook Pro. I reported this to Intel Premier Support in late April
and am still pursuing it with them. In this case I'm using both the
-I<dirname> switch for the *.mod files (which reside in a separate
directory) and linking in a library containing the associated *.o
files. With smaller test codes I've had some success by playing
around with the order that I list things on the load line. E.g.,
ifort -o xrun main.o -Imodpath lib1.a lib2.a
seemed to work for at least one of my codes.
-c Include common symbols as definitions with respect to
of contents. This is seldom the intended behavior
from a library, as it forces the linking of a
just because it uses an uninitialized global that is
at that point in the linking. This option is
because this was the original behavior of ranlib.
is not the default.
Do not post admin requests to the list. They will be ignored.
Fortran-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden