Binding to C - Arrays and Access Violation

[biozic via Digitalmars-d-learn]

On Tuesday, 2 February 2016 at 22:56:28 UTC, jmh530 wrote:
> I'm working on generating a binding to a C library. I've got 
> the .h file converted and can call some parts of the library 
> with no errors. However, I have reached a stumbling block in a 
> critical part.
>
> The library requires passing function pointers to various 
> functions in the library. When I try to run these functions, I 
> get an Access Violation error. I enabled additional DMD 
> warnings, which helped pinpoint the issue.
>
> My D code calls a C function. One of the parameters to the C 
> function is a function pointer to a D function. This D function 
> (below) is one that I copied  from the C library's tutorial. I 
> only slightly changed the signature. This function is 
> eventually called in other functions in the C library.
>
> double myfunc(uint n, const double* x, double* grad, void* 
> my_func_data)
> {
>     if (grad)
>     {
>         grad[0] = 0.0;
>         grad[1] = 0.5 / sqrt(x[1]);
>     }
>     return sqrt(x[1]);
> }
>
> The line (though likely the next will too) that causes a 
> problem is
>
> grad[0] = 0.0;
>
> Thus, as it is an Access Violation, I'm guessing the issue is 
> with accessing elements of arrays in the D function from the C 
> function. I don't know. When I try to call the D function in D, 
> it works, but I have to refer to x and grad as x.ptr and 
> grad.ptr.
>
> I'm not sure how to go about fixing this...

Is grad allocated in the D code? If so, it could have been 
collected because the GC lost track of its use when passing to 
and from the C code. Or is grad owned by the C code? If so, 
either there is a bug in the library or it's misused, because its 
memory has been freed/has never been allocated/has gone out of 
scope.