[Bug fortran/28443] New: gfortran does not implement the present intrinsic procedure correctly for optional character strings

2006-07-20 Thread cyan+gcc at compsoc dot nuigalway dot ie 
In file strpresent.F90:

program test_string_present

call test_present(1, "foo")
call test_present(2)

end program

subroutine test_present(a, b)
integer :: a
character*(*), optional :: b

if (present(b)) then
write (0,*) "b is present."
else
write (0,*) "b is not present."
end if

end subroutine test_present

As of SVN revision 115597, this yields the output:

 b is present.
 b is present.

It appears that the present() intrinsic function isn't working correctly for
optional character strings.

Ifort yields the expected output:

 b is present.
 b is not present.

The output of gfortran -v -save-temps -o strpresent strpresent.F90:

Driving: gfortran -v -save-temps -o strpresent strpresent.F90 -lgfortranbegin
-lgfortran -lm -shared-libgcc
Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: /home/pfarrell/redhat/BUILD/gcc4-115597/configure
--prefix=/opt/packages/gcc4 --enable-bootstrap --enable-languages=c,c++,fortran
Thread model: posix
gcc version 4.2.0 20060719 (experimental)
 /opt/packages/gcc4/libexec/gcc/i686-pc-linux-gnu/4.2.0/cc1 -E -lang-fortran
-traditional-cpp -D_LANGUAGE_FORTRAN -quiet -v strpresent.F90 -mtune=generic
-fpch-preprocess -o strpresent.f95
ignoring nonexistent directory
"/opt/packages/gcc4/lib/gcc/i686-pc-linux-gnu/4.2.0/../../../../i686-pc-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:
 /usr/local/include
 /opt/packages/gcc4/include
 /opt/packages/gcc4/lib/gcc/i686-pc-linux-gnu/4.2.0/include
 /usr/include
End of search list.
 /opt/packages/gcc4/libexec/gcc/i686-pc-linux-gnu/4.2.0/f951 strpresent.f95
-ffree-form -quiet -dumpbase strpresent.F90 -mtune=generic -auxbase strpresent
-version -fpreprocessed -I
/opt/packages/gcc4/lib/gcc/i686-pc-linux-gnu/4.2.0/finclude -o strpresent.s
GNU F95 version 4.2.0 20060719 (experimental) (i686-pc-linux-gnu)
compiled by GNU C version 4.2.0 20060719 (experimental).
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
 as -V -Qy -o strpresent.o strpresent.s
GNU assembler version 2.15.92.0.2 (i386-redhat-linux) using BFD version
2.15.92.0.2 20040927
 /opt/packages/gcc4/libexec/gcc/i686-pc-linux-gnu/4.2.0/collect2 --eh-frame-hdr
-m elf_i386 -dynamic-linker /lib/ld-linux.so.2 -o strpresent /usr/lib/crt1.o
/usr/lib/crti.o /opt/packages/gcc4/lib/gcc/i686-pc-linux-gnu/4.2.0/crtbegin.o
-L/opt/packages/gcc4/lib/gcc/i686-pc-linux-gnu/4.2.0
-L/opt/packages/gcc4/lib/gcc/i686-pc-linux-gnu/4.2.0/../../.. strpresent.o
-lgfortranbegin -lgfortran -lm -lgcc_s -lgcc -lc -lgcc_s -lgcc
/opt/packages/gcc4/lib/gcc/i686-pc-linux-gnu/4.2.0/crtend.o /usr/lib/crtn.o


-- 
   Summary: gfortran does not implement the present intrinsic
procedure correctly for optional character strings
   Product: gcc
   Version: 4.2.0
Status: UNCONFIRMED
  Severity: normal
  Priority: P3
 Component: fortran
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: cyan+gcc at compsoc dot nuigalway dot ie
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28443

[Bug fortran/28818] New: C/Fortran interoperability: variable number of arguments passed from fortran to C causes Illegal instruction

2006-08-23 Thread cyan+gcc at compsoc dot nuigalway dot ie 
With svn revision 116324, on x86_64-unknown-linux-gnu:

[EMAIL PROTECTED]:~]$ cat test.c
#include 
#include 

void test_(int *test, ...)
{
printf("test: %d\n", *test);
return;
}
[EMAIL PROTECTED]:~]$ cat forttest.F90
program test_printf

call test(-1, 'd')
call test(-1, 'd')

end program test_printf
[EMAIL PROTECTED]:~]$ make
gcc -c -o test.o test.c
gfortran -c -o forttest.o forttest.F90
gfortran -o test test.o forttest.o
[EMAIL PROTECTED]:~]$ ./test
test: -1
Illegal instruction
[EMAIL PROTECTED]:~]$

This does not happen if

void test_(int *test, ...)

is replaced with

void test_(int *test, char *d)

so it appears there's something wrong with variable numbers of arguments being
passed between fortran and C.

Here's what gdb says:

Breakpoint 1, test_ (test=0x400878) at test.c:6
6   printf("test: %d\n", *test);
(gdb) n
test: -1
8   }
(gdb) nexti
0x0040070e in test_ (test=0x400874) at test.c:8
8   }
(gdb) nexti
MAIN__ () at forttest.F90:4
4   call test(-1, 'd')
Current language:  auto; currently fortran
(gdb) nexti
0x00400741  4   call test(-1, 'd')
(gdb) nexti
0x00400746  4   call test(-1, 'd')
(gdb) nexti
0x0040074b  4   call test(-1, 'd')
(gdb) nexti

Program received signal SIGILL, Illegal instruction.
0x004006ca in test_ (test=0x400878) at test.c:5
5   {
Current language:  auto; currently c
(gdb)

but I don't know enough about assembly to help any more.


-- 
   Summary: C/Fortran interoperability: variable number of arguments
passed from fortran to C causes Illegal instruction
   Product: gcc
   Version: 4.2.0
Status: UNCONFIRMED
  Severity: normal
  Priority: P3
     Component: fortran
    AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: cyan+gcc at compsoc dot nuigalway dot ie
 GCC build triplet: x86_64-unknown-linux-gnu
  GCC host triplet: x86_64-unknown-linux-gnu
GCC target triplet: x86_64-unknown-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28818

[Bug fortran/28818] C/Fortran interoperability: variable number of arguments passed from fortran to C causes Illegal instruction

2006-08-25 Thread cyan+gcc at compsoc dot nuigalway dot ie 


--- Comment #3 from cyan+gcc at compsoc dot nuigalway dot ie  2006-08-25 
11:06 ---
(In reply to comment #2)
> the Fortran program
> is not calling a var-args C function,

I don't understand what you mean. The fortran code calls the test function,
which is written in C (called test_ due to name mangling). test_ is indeed a
variable-args function:
void test_(int *test, ...).

Also, this code works as expected on 32-bit x86 machines.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28818