On Wed, 10 Apr 2013, Roy Stogner wrote: > a potentially-controversial-bugfix that I'll post about shortly),
This ominous warning turns out to be overstating the problem. I'm pretty sure that the one-line-deletion in #89d0841 will fix everything. It reminds me that technically our new LibMeshInit behavior is an API change, though. Previously we'd dupe the MPI communicator handed to LibMeshInit, then free our dupe upon exit. This implied that the original communicator could have been freed by the user with no ill effects. Now the user is going to be responsible for not freeing the passed-in communicator until after the LibMeshInit object is destroyed. I don't think anybody's code will break, and the new behavior is probably more intuitive, but it's worth pointing out. --- Roy ------------------------------------------------------------------------------ Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis & visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter _______________________________________________ Libmesh-devel mailing list Libmesh-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-devel
