[petsc-dev] Arguments that are logically void**

[Ethan Coon]

On Mon, 2011-06-06 at 14:34 -0400, Kai Germaschewski wrote:
> Wouldn't it make sense to  have DMDAVecGetArray[23]d() and have the
> argument be PetscScalar ***, and PetscScalar ****, respectively? That
> would actually check that one passes in a reference to the right type,
> making it both safe and avoid the need for casting? 

No.  At least my application uses both 2 and 3 dimensional DAs,
dependent upon whether the code is working in 2D or 3D modes.
VecGetArray is given the argument (PetscScalar **), and the resulting
pointer is passed in to a physical routine which knows the correct
dimension.  

Ethan


> That's also more in line with, e.g., VecGetArray2d().
> 
> I'm not sure where else void ** is used, and whether similar solutions
> would apply. In general, I agree with Jed that I'd rather not have the
> explicit cast in the user code.
> 
> --Kai
> 
> 
> On Mon, Jun 6, 2011 at 2:20 PM, Jed Brown <jed at 59a2.org> wrote:
>         On Mon, Jun 6, 2011 at 20:00, Barry Smith <bsmith at mcs.anl.gov>
>         wrote:
>                 So is the only benefit that one does not need the
>                 extra characters (void**) in front of the argument?
>                 
>         
>         
>         Yes.
>         
>         
>         Rationale.
>         
>         
>         The use of a formal parameter attribute_val or type void*
>         (rather than void**) avoids the messy type casting that would
>         be needed if the attribute value is declared with a type other
>         than void*. ( End of rationale.) 
>          
>                 
>                   Doesn't seem like much of a benefit? At this point I
>                 think doing it either way is fine but if we are
>                 inconsistent we should fix some of them to be
>                 consistent. I am leaning to vote for the (void**)
>                 approach; it is clearer when one looks at code.
>         
>         
>         I think the cast is ugly if it appears in places that should
>         not be "scary". For example, I think DMDAVecGetArray(da,X,&x)
>         is significantly nicer than DMDAVecGetArray(da,X,(void**)&x).
>         
>         
>         Needing an explicit cast makes the reader think about the
>         safety of the cast. This is a good thing for calls that should
>         look a bit scary.
> 
> 
> 
> -- 
> Kai Germaschewski
> Assistant Professor, Dept of Physics / Space Science Center
> University of New Hampshire, Durham, NH 03824
> office: Morse Hall 245E
> phone:  +1-603-862-2912
> fax: +1-603-862-2771
> 

-- 
------------------------------------
Ethan Coon
Post-Doctoral Researcher
Applied Mathematics - T-5
Los Alamos National Laboratory
505-665-8289

http://www.ldeo.columbia.edu/~ecoon/
------------------------------------