[petsc-dev] Implementing the Fortran interface for DMSHELL

[Barry Smith]

On Oct 26, 2012, at 3:42 PM, Richard Tran Mills <rtm at eecs.utk.edu> wrote:

> Hi Guys,
> 
> I want to take a stab at implementing the necessary custom Fortran stubs so that I can use DMSHELL from inside PFLOTRAN.  I have a couple of dumb questions:
> 
> 1) I used to know this: How do I mark a routine in the code as one that requires a custom stub vs. one that can use an auto-generated one?  Is this just a matter of having "/*@C" vs "/*@" in the comment before the function definition?  Do I need to do something else?

   No C means create fortran stub. C means skip creating fortran stubs. In general any function that has a character string argument or a function argument needs to have a manual stub; manual stubs should go in the ftn-custom directory.

> 
> 2) Should I create a test case for this that does something like, say, create "shell" routines that actually just do something like call the DMDA ones?  Where should this test go?

   dm/examples/tests  or dm/examples/tutorials if you make it nice :-)

   Barry

> 
> Thanks,
> Richard
> 
> -- 
> Richard Tran Mills, Ph.D.
> Computational Earth Scientist      | Joint Assistant Professor
> Hydrogeochemical Dynamics Team     | EECS and Earth & Planetary Sciences
> Oak Ridge National Laboratory      | University of Tennessee, Knoxville
> E-mail: rmills at ornl.gov  V: 865-241-3198 http://climate.ornl.gov/~rmills
>