Many failures in the test suite with -fdefault-integer-8 come from tests mixing C and Fortran. While definitively not an expert in this area, I find pretty obvious that changing the default integer/real type on one side and not the other is asking for troubles. Apparently gfortran, but not the test suite, is finding some of them as in gfortran.dg/c_loc_tests_2.f03:
[karma] f90/bug% gfc -fdefault-integer-8 /opt/gcc/gcc-4.3-work/gcc/testsuite/gfortran.dg/c_loc_tests_2.f03
if(test_array_address(my_c_ptr_1, 100) .ne. 1) then
Error: Type/rank mismatch in argument 'num_elements' at (1)
Fatal Error: Can't open module file 'c_loc_tests_2.mod' for reading at (1): No such file or directory
Using -fdefault-integer-8 with 'intrinsic :: iso_c_binding' should probably trigger an error or at least a warning.
At this point, the easiest fix is probably going to be to document
that -fdefault-integer-8 should not be used with bind(c) code. A
check of this option can be inserted at various locations during the
Subject: Re: C/Fortran interoperability and
> A check of this option can be inserted at various locations during
> the parsing.
Certainly, but this will give a very interesting exercise to build test cases!
If I apply the following patch:
--- c_loc_tests_2.f03 2009-11-25 18:17:45.000000000 +0100
+++ c_loc_tests_2_db_1.f03 2009-12-06 11:43:31.000000000 +0100
@@ -44,13 +44,13 @@ contains
end subroutine test0
subroutine test1() bind(c)
- integer, target, dimension(100) :: int_array_tar
+ integer(c_int), target, dimension(100) :: int_array_tar
type(c_ptr) :: my_c_ptr_1 = c_null_ptr
type(c_ptr) :: my_c_ptr_2 = c_null_ptr
- int_array_tar = 100
+ int_array_tar = 100_c_int
my_c_ptr_1 = c_loc(int_array_tar)
- if(test_array_address(my_c_ptr_1, 100) .ne. 1) then
+ if(test_array_address(my_c_ptr_1, 100_c_int) .ne. 1) then
end subroutine test1
the test passes with -fdefault-integer-8. Note that
gfc -fdefault-integer-8 c_loc_tests_2_db_1.f03 c_loc_tests_2_funcs.c
gives a warning:
cc1: warning: command line option "-fdefault-integer-8" is valid for Fortran but not for C
Although such warning can be trapped by adding in gcc/testsuite/lib/prune.exp lines such that
regsub -all "(^|\n)cc1: warning: command line option .-fwhole-file. is valid for Fortran but not for C" $text "" text
it would be better to change the gfortran script to not pass gfortran options to other languages.
If there is any interest in updating the interoperability tests to works with -fdefault-*-8, I can comb the testsuite.
Created attachment 26187 [details]
patch to fix some failures with -fdefault-integer-8
I have this patch in my tree for a quite long time. It fixes some failures with -fdefault-integer-8 by replacing some implicit kinds with explicit ones.
(In reply to Dominique d'Humieres from comment #4)
> Created attachment 26187 [details]
> patch to fix some failures with -fdefault-integer-8
> I have this patch in my tree for a quite long time. It fixes some failures
> with -fdefault-integer-8 by replacing some implicit kinds with explicit ones.
patches go to the gcc-patches mailing list