On 05/07/2015 11:34 AM, Lisandro Dalcin wrote:
On 7 May 2015 at 18:44, Rob Latham <[email protected]> wrote:
1) MPI_File_call_errhandler(file, value) should return MPI_SUCCESS instead of value under normal operation (as the MPI spec says), and particularly if the error handler is set to MPI_ERRORS_RETURN. This bug is trivial to fix.
Trivial to fix...if we only had to worry about C and Fortran!
I'm probably missing something here, so I have to ask.
I said this was "trivial to fix" for the particular case of MPI_ERRORS_RETURN. How should C++ bindings behave in case MPI_ERRORS_RETURN is set? In other cases, I agree with you, fixing this beast is not easy. However, you have a point about C++ bindings being deprecated.
It is I who was missing something. I read the code too fast, saw the special C++ case, and thought it mattered for MPI_ERRORS_RETURN. This one is easy. The core dumping one is a little more work but I see an angle of attack now. ==rob
-- Rob Latham Mathematics and Computer Science Division Argonne National Lab, IL USA