[mpiwg-tools] Too strict for flags in MPI_T_pvar_get_info?

Jeff Squyres (jsquyres) jsquyres at cisco.com
Sat Aug 16 05:37:16 CDT 2014

On Aug 15, 2014, at 11:45 AM, Kathryn Mohror <kathryn at llnl.gov> wrote:

> I suppose it doesn’t change semantics, but possibly implementation of
> tools and MPI libraries. I was under the impression that anything that
> could change how a program that uses MPI is written couldn’t be ticket 0.

Mmm.  Ok, fair enough.

> For my part, I don’t really think it’s wrong to not use the same
> convention as the rest of the MPI library for true and false. We don’t use the same string convention, after all.

Strings are one thing, but booleans are a different (smaller) thing.

I would think that now is the time to actually nip this inconsistency in the bud -- *while* people are initially writing all the MPI_T-based tools, etc.  Vs. someone else in the Forum (outside the Tools WG) noticing the inconsistency, insisting that we fix the inconsistency, and then we have the problem of lots of existing MPI_T tools that need to be updated / backwards-compatibility woes.

It's a small thing; you've convinced me that it should be an errata ticket (vs. a ticket 0 change).  But it's a small/easy one.  And now is the time to do it.

Jeff Squyres
jsquyres at cisco.com
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/

More information about the mpiwg-tools mailing list