I'm using a VB test harness to test functions in my FORTRAN DLL. The VB reports errors with certain calculations ('Overflow'), however, on debugging I see that the FORTRAN calculates the correct answer but somehow fails to return it properly all the time.
The VisStudio debugger reports 'inexact result'.
If I manually test the DLL using its original host program, the same combinations of inputs give me no errors.
Anyone seen this before and / or know what to do with it?
This sounds like your classic C <--> Fortran argument/calling convention mismatch. Try searching through the forum for VB content. Here is one of the latest postings where I think this problem was nipped in the bud...
