FIX: Intrinsic REAL Function Assigned to COMPLEX Fails

The information in this article applies to:

• Microsoft FORTRAN PowerStation for MS-DOS, version 1.0

SYMPTOMS

An application that assigns the result of a REAL intrinsic function to a COMPLEX variable may generate incorrect results.

CAUSE

The compiler generates incorrect code to perform this assignment.

RESOLUTION

There are two methods to avoid this problem, as follows:

• Modify the source code to assign the result of the REAL function to a temporary REAL variable. Then specify the temporary variable in the assignment statement.
• Add zero to the result of the function.

STATUS

Microsoft has confirmed this to be a problem in FORTRAN PowerStation version 1.0 for MS-DOS. This problem has been resolved with FORTRAN PowerStation maintenance release version 1.0a for MS-DOS.

FORTRAN PowerStation version 1.0 can be differentiated from the maintenance release version 1.0a by invoking the linker. Typing "link32 | more" from \F32\BIN directory will show version 2.8 for FORTRAN PowerStation version 1.0, and it will show version 1.0f for the maintenance release version 1.0a.

MORE INFORMATION

Specifying different compiler options does not correct this problem. The various options only change the incorrect value that the code places in the COMPLEX variable.

The following sample code demonstrates this problem:

Sample Code #1

c Compile options needed: None

      complex*8 c
      real*4 x
      x = 1.0
      c = exp(x)
      print *, 'complex c1 = exp(x); c = ', c
      end

Sample Code #2

c Compile options needed: None

      complex*8 c
      real*4 x, tmp
      x = 1.0
      tmp = exp(x)
      c = tmp
      print *, 'complex c1 = exp(x); c = ', c
      end

Sample Code #3

c Compile options needed: None

      complex*8 c
      real*4 x
      x = 1.0
      c = exp(x) + 0
      print *, 'complex c1 = exp(x); c = ', c
      end