Re: New errhandler text

William Saphir (wcs@nsgi3.lbl.gov)
Tue, 6 May 1997 14:33:33 -0700

On May 6, 3:46pm, Jeff Squyres wrote:
> *** THE BOTTOM LINE: I want one callback typedef (of each type -- copy,
> delete, etc.), that takes a (void *) as its first argument, not an MPI
> opaque handle.

I support this. There is no change in functionality from what
has been voted on in the final meeting - it just cleans up
syntax.

The only question I had when reading over your note was whether
there would be conflicts with things in MPI-1. I believe there
are none. Am I correct that the situation is:
1. We are already proposing functions with new names that
reproduce MPI-1 functionality for communicators and extend
it for other types. (The old functions remain, but are effectively
deprecated, as with the datatype constructors with MPI_Aint arguments).
2. The names of your new proposed typedefs have been carefully
chosen to also not conflict with the MPI-1 names.

Bill