On Apr 30, 2013, at 9:41 AM, George Bosilca <bosi...@icl.utk.edu> wrote:
> Ralph, > > The log message is plain wrong. There is nothing non standard compliant in > the way the PML V is using the output, it is just it use an internal storage > for it's stream. My point was that vprotocol isn't a standalone framework like all the others we have - it is tightly coupled to the pml/v component. That isn't the standard way we do things, as you know. The arrangement has caused some issues in the past, and this was just one that bit Nathan as he didn't realize it. > > It should be straightforward to fix it by removing __ALL__ references to the > output variable not only the one from the structure. Yes, but that then requires that the pml/v component directly reference the vprotocol framework object, which seems equally ugly. If that's what you want, then please feel free to do so - in which case, you could have just fixed it the way you want yourself :-) > > George. > > > On Apr 30, 2013, at 18:37 , svn-commit-mai...@open-mpi.org wrote: > >> Author: rhc (Ralph Castain) >> Date: 2013-04-30 12:37:17 EDT (Tue, 30 Apr 2013) >> New Revision: 28418 >> URL: https://svn.open-mpi.org/trac/ompi/changeset/28418 >> >> Log: >> Revert r28417 - given the non-standard way vprotocol is implemented, I see >> no way to use the framework verbosity here. Best to just leave it alone as >> those who use it know what they need to do to get debug output > > > _______________________________________________ > devel mailing list > de...@open-mpi.org > http://www.open-mpi.org/mailman/listinfo.cgi/devel
