FIX: Handling of Invalid Values in Class Wizard for DDX/DDV

ID: Q108584

The information in this article applies to:

Microsoft Visual C++ for Windows, versions 1.0, 1.5
Microsoft Visual C++ 32-bit Edition, versions 1.0, 2.0

SYMPTOMS

Using the Class Wizard included with Visual C++ to associate a member variable with a control on a dialog box causes the minimum and maximum values to be treated differently depending upon different variable types.

RESOLUTION

Currently, there is no workaround for the inconsistent auto-correction. Furthermore, once the auto-correction feature has changed a float or double to use scientific notation Class Wizard will fail when trying to modify the attributes of the control. The code with the scientific notation must be manually changed to a non-scientific floating point value to continue.

STATUS

Microsoft has confirmed this to be a bug in the Microsoft products listed at the beginning of this article. This bug was corrected in Visual C++ 2.1.

MORE INFORMATION

When specifying that a variable will be of type int or uint, providing a value greater than its limit (see below) will cause an error to be displayed. However, setting a value that is over the limit for a long or DWORD is automatically corrected to the maximum value achievable in either direction.

The following table details the behavior of these types for both the 16- and 32-bit editions:


   Min. Limit    Type    Max. Limit   Correction
   ---------------------------------------------------------------------

   -32768        int     32767        Warns user
   0             uint    65535        Warns user
   -2147483647   long    2147483647   Automatic
   0             DWORD   4294967295   Automatic

NOTE: For the 32-bit edition, the maximum negative value allowed for a long is -2147483648.

There is an additional complication for variables of type float and double. Not only do they auto-correct, they also correct to a value given in scientific (exponential) format, which Class Wizard cannot parse. Class Wizard can save a value in that format but gives an error the next time the user tries to load Class Wizard for the same project. The Class Wizard will then appear but with no Class information. For more information, please see the following article in the Microsoft Knowledge Base:

Q107465 FIX: Class Wizard Floating-Point Parsing Errors

The following table details the limits of types float and double for the 16- and 32-bit editions:


   Min. Limit               Type     Max. Limit
   ----------------------------------------------------------------------

   -1.79769e+308            float    1.79769e+308
   -1.79769313486232e+308   double   1.79769313486232e+308

IMPORTANT NOTE: The 16-bit edition auto-corrects these values if the user- specified value exceeds them. However, the 32-bit edition sets the value to 1.#INF if the user-specified value exceeds the above limits for variables of type float or double.

Additional query words: 1.00 1.50 2.00

Keywords : kbwizard kbMFC kbVC
Version : :1.0,1.5,2.0
Platform : NT WINDOWS
Issue type :