On Jun 14, 2012, at 3:53 PM, Ralph Castain wrote:
> I believe we use that constant in several places to define a static array size - you might check to be safe.
I can't find it used anywhere in the code base other than mpi.h.in.
It's a non-standard name (i.e., it's not in the MPI spec). I believe the standard name is MPI_MAX_PORT_NAME (which is OPAL_MAX_PORT_NAME).
> On Jun 14, 2012, at 11:52 AM, Jeff Squyres wrote:
>> WHAT: Remove non-standard MPI_MAX_NAME_LEN from mpi.h.
>> WHY: It looks like this was a carryover from LAM/MPI, but it's not in any MPI spec.
>> WHERE: mpi.h
>> TIMEOUT: This seems non-controversial, so I'll set the timeout to the teleconf next Tuesday: June 19, 2012
>> More details:
>> MPI_MAX_NAME_LEN is in mpi.h, but *not* in mpif.h, nor the C++ bindings. It looks like this is some kind of hold over from LAM/MPI, per the comment in mpi.h:
>> #define MPI_MAX_NAME_LEN MPI_MAX_PORT_NAME /* max port name length, non-std. (LAM < 6.3b1) */
>> This really should be removed to avoid confusion.
>> If there's any discussion needed, I'm happy to push back the timeout -- I'm just assuming that there won't need to be any.
>> Jeff Squyres
>> For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/
>> devel mailing list
> devel mailing list
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/