On Apr 21, 2011, at 10:05 AM, Jed Brown wrote: > On Thu, Apr 21, 2011 at 17:03, Jed Brown <jed at 59a2.org> wrote: > Currently this needs to be called, but is not available from Fortran and is > marked as developer level. Is there a plan to make it somehow automatic (I > thought the function was introduced because it was expensive to do it > automatically), or did someone forget to make Fortran bindings and set the > Level? > > And should this have a push and pop model.
I realized it when I was half done implemented but was too lazy to do it correctly. Also PetscSynchronizedFlush() should be changed to use the PetscViewerFlowControlStart() paradigm allowing bigger chunks then one message at a time for flow control, see for example VecView_MPI_Binary() again I was too lazy to do that correctly. So someone can do these things or add them to http://www.mcs.anl.gov/petsc/petsc-as/developers/projects.html Barry > Otherwise I still need a getter to do a synchronized view within a library > call and leave the viewer in the same state as the user gave it to me in.
