[petsc-dev] PETSc goes Doxygen!?

[Karl Rupp]

Hi,

>     Oh, we use different definitions of 'documented'. I'm referring to
>     documentation right at the place of declaration/definition.
>     PetscScalar.html is generated from some comment block at some other
>     spot in the sources, hence Doxygen fails to associated it with the
>     correct typedef (this could be resolved by an injection of suitable
>     Doxygen commands).
>
>
> This is used in many places. For example, all the implementations
> (KSPGMRES, MATMPIAIJ, etc) are documented next to the associated create
> function (KSPCreate_GMRES, MatCreate_MPIAIJ), but the macros are in the
> header. I don't see sed being sufficient for this editing, but a Python
> (for example) input filter should be able to convert the sowing comment
> blocks to Doxygen format.

Yes, you're right, this is where things get a bit complicated, but are 
still manageable as long as it's just typedefs, enums, or members of 
structs.

Best regards,
Karli