Q167750: BUG: The pow() Function Doesn’t Set Underflow Condition

Article: Q167750
Product(s): Microsoft C Compiler
Version(s): 5.0,6.0
Operating System(s): 
Keyword(s): kbtool kbVC500bug kbVC600bugkbbuglist
Last Modified: 13-FEB-2002

-------------------------------------------------------------------------------
The information in this article applies to:

- The C Run-Time (CRT), used with:
   - Microsoft Visual C++, 32-bit Enterprise Edition, versions 5.0, 6.0 
   - Microsoft Visual C++, 32-bit Professional Edition, versions 5.0, 6.0 
   - Microsoft Visual C++, 32-bit Learning Edition, version 6.0 
   - Microsoft Visual C++.NET (2002) 
-------------------------------------------------------------------------------

SYMPTOMS
========

The pow() function does not set the errno variable to indicate that an underflow
condition has occurred.

RESOLUTION
==========

A potential solution is to check the base of the number that is being raised to
a power and check the result of the pow() function. By customizing the
_matherr() function, the underflow condition can be trapped and you can display
an error message. The program below demonstrates this workaround.

      #include <stdio.h>
      #include <math.h>
      #include <ERRNO.H>
      #include <string.h>
      extern int errno;

      /* The _matherr routine can be customized to handle the specific
          underflow problem that is encountered.  See online help for
          more information on _matherr().
      */ 
      int _matherr( struct _exception *except )
      {
          /* Handle errors caused by pow() function.*/ 
          if(strcmp(except->name,"pow")==0)
          {
              switch(except->type)
              {
                  case _DOMAIN:
                  {
                      printf("Domain Error: Argument not in domain.\n");
                      return 1;
                      break;
                  }
                  case _SING:
                  {
                      printf("Singularity Error: Argument singularity "
                             "error.\n");
                      return 1;
                      break;
        }
                  case _OVERFLOW:
                  {
                      printf("Overflow Error: Overflow range error.\n");
                      return 1;
                      break;
                  }
                  case _PLOSS:
                  {
                      printf("Precision Error: Partial loss of "
                             "significance.\n");
                      return 1;
                      break;
                  }
                  case _TLOSS:
                  {
                      printf("Precision Error: Total loss of "
                             "significance.\n");
                      return 1;
                      break;
        }
                  case _UNDERFLOW:
        {
                      printf("Underflow Error: The result is too small to "
                             "be represented.\n");
                      return 1;
                      break;
                  }
                  default:
                  {
                      printf("Unknown Error Occurred While Performing Math "
                             "Operation.\n");
                      return 1;
                      break;
        }
         }
          }
          else
              return 0;  /* non-error condition */ 
      }

      void main(void)
      {

      double x = -1.0e+307;
      double y = -1.0e+307;
      double z;

      z = pow(x,y);

      /*
      Check for error conditions.  _matherr() is automatically
      called should there be a math related error.
      */ 

      /*Note that errno is not set properly & _matherr() is called.*/ 
      if (errno != 0)
          printf("ERROR! Errno: %i\n",errno);
      else
          printf("pow( %e, %e ) = %f\n", x, y, z);
     }

STATUS
======

Microsoft has confirmed this to be a bug in the Microsoft products listed at the
beginning of this article.

MORE INFORMATION
================

The help topic for pow() states:

  No error message is printed on overflow or underflow.

However, an overflow does set errno to the value of 34 indicating "Result too
large". When an underflow occurs, the returned value is 0.

Additional query words:

======================================================================
Keywords          : kbtool kbVC500bug kbVC600bug kbbuglist
Technology        : kbVCsearch kbAudDeveloper kbCRT
Version           : :5.0,6.0
Issue type        : kbbug

=============================================================================

THE INFORMATION PROVIDED IN THE MICROSOFT KNOWLEDGE BASE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. MICROSOFT DISCLAIMS ALL WARRANTIES, EITHER EXPRESS OR IMPLIED, INCLUDING THE WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL MICROSOFT CORPORATION OR ITS SUPPLIERS BE LIABLE FOR ANY DAMAGES WHATSOEVER INCLUDING DIRECT, INDIRECT, INCIDENTAL, CONSEQUENTIAL, LOSS OF BUSINESS PROFITS OR SPECIAL DAMAGES, EVEN IF MICROSOFT CORPORATION OR ITS SUPPLIERS HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. SOME STATES DO NOT ALLOW THE EXCLUSION OR LIMITATION OF LIABILITY FOR CONSEQUENTIAL OR INCIDENTAL DAMAGES SO THE FOREGOING LIMITATION MAY NOT APPLY.

Copyright Microsoft Corporation 1986-2002.