From: David Gingold <david.gingold@XXXXXXXXXXXX>
Date: January 11, 2008 3:58:08 PM CST
To: mpi-21@XXXXXXXXXXXXX
Subject: [mpi-21] Proposal EH2: add const keyword to the C bindings
Reply-To: mpi-21@XXXXXXXXXXXXX

[I'm still trying to catch up on all of the e-mail messages before next week, so please excuse me if this has come up before.]

Here is a way in which adding "const" to the C bindings could break an existing MPI program.  Suppose I write this:

	#include <mpi.h>

	void foo(void)
	{
	    typedef int (* sendfunc_t)(void *, int, MPI_Datatype, int, int, MPI_Comm);
	    sendfunc_t s;
	    s = MPI_Send;
	}

If we re-define MPI_Send() so that its first argument has type "const void *", then this code becomes incorrect.

This problem notwithstanding, I do believe that Erez's proposal to add "const" to the bindings has merit.  My suspicion is that there are very few MPI applications that do something like this.

Perhaps the upshot is that we need to get our heads around moving the standard forward in ways that occasionally break back-compatibility.  Maybe we should come up with a mechanism (e.g. compile with "-DMPI_2.1" to get the old bindings) that provides a migration path for codes.

-dg