oysteinls
Joined: 04 Aug 2004
Posts: 1
|
 Posted: Mon Sep 27, 2004 2:58 pm Post subject: pointer args. to subroutine (different answer in 5.2) |
 |
|
Hi,
I don't know if this qualifies as a compiler bug, but the following little program gives different answer in version 5.2 of the compiler than in 5.0 or 5.1:
----------------
| Code: | program test
real ,dimension(:), pointer :: a
real, target :: e(2)
a => e
a(1) = 1
a(2) = 3
print*, 'a=',a
call sub(a,a)
print*, 'a=',a
end program test
subroutine sub(x,y)
real x(2),y(2)
y = x+1
return
end subroutine sub |
-----------------------------
With the new version of the compiler I get:
[ols@dhcp4154 tmp]$ pgf90 -V -g test.f90
pgf90 5.2-2
Copyright 1989-2000, The Portland Group, Inc. All Rights Reserved.
Copyright 2000-2004, STMicroelectronics, Inc. All Rights Reserved.
PGF90/any Linux/x86 5.2-2
Copyright 1989-2000, The Portland Group, Inc. All Rights Reserved.
Copyright 2000-2004, STMicroelectronics, Inc. All Rights Reserved.
PGF90/x86 Linux/x86 5.2-2
Copyright 1989-2000, The Portland Group, Inc. All Rights Reserved.
Copyright 2000-2004, STMicroelectronics, Inc. All Rights Reserved.
[ols@dhcp4154 tmp]$ ./a.out
a= 1.000000 3.000000
a= 1.000000 3.000000
---------------------------------
while in the 5.1 release I get the expected result (2.0, 4.0) in the second output line. This only happens when I declare A as a pointer; with A an ordinary array I get the expected result.
So am I using pointers incorrectly or is this a compiler bug?
Oystein |
|
mkcolg
Joined: 30 Jun 2004
Posts: 4996
Location: The Portland Group Inc.
|
| Posted: Mon Sep 27, 2004 4:47 pm Post subject: Error only in 5.2-2 |
|
|
The code works in 5.2-1, but fails with 5.2-2, so we obviously have a bug. The good news is that this has been fixed as part of the upcoming 5.2-4 release. Once available, I would recommend upgrading.
Thanks for the great example!
- Mat |
|
Search
Memberlist
Register
Profile
Log in
