> I also ported the same code to VC++ and VB for PC users.

> The Win32 PC application works fine every time.

> The WinCE version simply stalls/freezes during testing and doesn't
generate
> any errors.

> Example of the eVC DLL code

> double APIENTRY DMSToDec(int Degrees, int Minutes, int Seconds)
> {
>     //Implementation code
> }
> double APIENTRY DecToRads(double DecValue)
> {
>     //Implementation code
> }

> Def file created and functions exported.

> VB Definition
> ***********

> Declare Function DMSToDec Lib "Windows\MYDLL.DLL" (ByVal Degrees as
Integer,
> Minutes as Integer, Seconds as Integer) as Double
> Declare Function DecToRads Lib "Windows\MYDLL.DLL" (ByVal DecValue as
> double) as Double

> Now, I believe that the double data type is not valid for declaration
> satatements, amoung others, so I converted the code accordingly.

> Declare Function DMSToDec Lib "Windows\MYDLL.DLL" (ByVal Degrees as
Integer,
> Minutes as Integer, Seconds as Integer) as Variant
> Declare Function DecToRads Lib "Windows\MYDLL.DLL" (ByVal DecValue as
> Variant) as Variant

> I also tried passing the variants ByRef.  Nothing works.   I now read that
a
> function declaration cannot return a variant.  I'm confused...  It appears
> that I cannot use a declaration the returns any variation of a floating
> point variable.

> The lack of any error message except the occassional "fatal error in
> pvbload.exe" and the fact I can't terminate the running app from within
> WinCE has me stumped.

> Any ideas would be appreciated.

> Regards

> Shane Brodie

> I am not of much use to you re C++, but I can tell you that a C++ in
should
> translate to a VB Long, not int - and you also appear to have missed the
> ByVal from the remaining parameters:

> Declare Function DMSToDec Lib "Windows\MYDLL.DLL" (ByVal Degrees as Long,
> ByVal Minutes as Long, ByVal Seconds as Long) as Variant

> I think you are right re the double parameters, so perhaps doing something
> like this:

> MyDoubleVariant = CDbl(0)

> Before passing it "ByVal" may force it to pass as a double given that the
> variant sub-type will now be double... just an idea.

> HTH

> Dino

> > I have developed a simple DLL in eVC that provides some complex math
> > calculations for my eVB client application.

> > I also ported the same code to VC++ and VB for PC users.

> > The Win32 PC application works fine every time.

> > The WinCE version simply stalls/freezes during testing and doesn't
> generate
> > any errors.

> > Example of the eVC DLL code

> > double APIENTRY DMSToDec(int Degrees, int Minutes, int Seconds)
> > {
> >     //Implementation code
> > }
> > double APIENTRY DecToRads(double DecValue)
> > {
> >     //Implementation code
> > }

> > Def file created and functions exported.

> > VB Definition
> > ***********

> > Declare Function DMSToDec Lib "Windows\MYDLL.DLL" (ByVal Degrees as
> Integer,
> > Minutes as Integer, Seconds as Integer) as Double
> > Declare Function DecToRads Lib "Windows\MYDLL.DLL" (ByVal DecValue as
> > double) as Double

> > Now, I believe that the double data type is not valid for declaration
> > satatements, amoung others, so I converted the code accordingly.

> > Declare Function DMSToDec Lib "Windows\MYDLL.DLL" (ByVal Degrees as
> Integer,
> > Minutes as Integer, Seconds as Integer) as Variant
> > Declare Function DecToRads Lib "Windows\MYDLL.DLL" (ByVal DecValue as
> > Variant) as Variant

> > I also tried passing the variants ByRef.  Nothing works.   I now read
that
> a
> > function declaration cannot return a variant.  I'm confused...  It
appears
> > that I cannot use a declaration the returns any variation of a floating
> > point variable.

> > The lack of any error message except the occassional "fatal error in
> > pvbload.exe" and the fact I can't terminate the running app from within
> > WinCE has me stumped.

> > Any ideas would be appreciated.

> > Regards

> > Shane Brodie