On Jul 28, 2007, at 4:41 PM, Lisandro Dalcin wrote:
In the mean time, I would prefer to follow the standard as close as
possible. If not, some external, stupid test suite (like the one I
have for mip4py) would report that OMPI is wrong about this point.
What exactly are you testing for?
BTW,
It looks like there is a check for COMM_SELF in comm_free.c that
explicitly calls the error handler on COMM_WORLD; it looks like this
was introduced in the very first implementation of COMM_FREE way back
in r983.
I don't see any reason why this would be the case; I'm guessing it's
a bug b
It looks like an explicit check for MPI_GROUP_EMPTY was removed from
group_free.c in r2623 by Edgar. I suspect that there is some subtle
reason for this.
Edgar: do you have any recollection why this is? And how we can
otherwise avoid the segv?
On Jul 28, 2007, at 6:40 PM, Lisandro Dalc
