itaps-parallel IMeshP_protos.h

[Devine, Karen D]

Thanks for your reply, Jason.  Since iMeshP.h includes iMeshP_protos.h, we
should have iMeshP_protos.h in the repository; otherwise, builds that
include iMeshP.h will fail when they can't find iMeshP_protos.h.  If
iMeshP_protos.h is for fortran name mangling only, can we name the file
something like iMeshP_fortran_protos.h?   Also, I like your decision to have
these definitions in a separate file (one for each interface), as users
don't really need to worry about them and having them in the main interface
file would be distracting.

Karen



On 10/22/08 3:45 PM, "Jason Kraftcheck" <kraftche at cae.wisc.edu> wrote:

> Devine, Karen D wrote:
>> Jason:
>>
>> What is iMeshP_protos.h?  It is #included in iMeshP.h, but it is not in the
>> SVN repository.  Similarly, iMesh.h includes iMesh_protos.h, but
>> iMesh_protos.h is not in the repository.  Are these include files
>> implementation-specific?  If so, what do you expect an implementation would
>> include in these files?
>>
>
> This is where the Fortran name-mangling gunk is specified.  If using GNU
> autotools for configuration, the file will typically be composed of a whole
> lot of entries of the form:
> #ifdef FC_FUNC_
>   #define iMesh_newMesh FC_FUNC_(imesh_newmesh, IMESH_NEWMESH)
>   #define iMesh_dtor FC_FUNC_(imesh_dtor, IMESH_DTOR)
>   #define iMesh_load FC_FUNC_(imesh_load, IMESH_LOAD)
>   #define iMesh_save FC_FUNC_(imesh_save, IMESH_SAVE)
>   #define iMesh_getRootSet FC_FUNC_(imesh_getrootset, IMESH_GETROOTSET)
>   ...
>
> If no one objects to the use of autotools 'FC_FUNC_' macro name, we could
> put these in the repository (or the contents directly in the respective
> interface header).  One could use a mechanism other than autotools to define
> FC_FUNC_, or not define it at all.
>
> - jason
>
>