On Mon, 2 Jun 2008, David Knezevic wrote: > I'm running the SVN head version of libmesh, and I was hitting the following > assert in an application code: > > Assertion `neigh_has_remote_children' failed. > [0] src/mesh/unstructured_mesh.C, line 393, compiled May 30 2008 at 17:51:12 > terminate called after throwing an instance of 'libMesh::LogicError' > > I can't see why there should be issues with remote elements, since I'm > running on a single processor and I'm not using ParallelMesh? I've attached a > simple adaptive code which hits this assertion on my machine (interestingly, > it runs fine if max_r_steps<=4).
It's not an issue with remote elements, it's an issue where we had an existing assertion that was correct for a SerialMesh but would occasionally give a false positive when running with a distributed ParallelMesh. So now if we get to a point where that old assertion would have been made, we assume that it's just the remote elements case, but with METHOD=dbg we double-check to make sure. I'll add a "libmesh_assert(!this->is_serial())" first if you think that'll be a less misleading error message, but unless there's been a regression this error shouldn't have come up in the first place. I didn't get your attachment either; go ahead and email me a copy directly if SourceForge is giving us trouble. It sounds like my code is most likely to be the offender here, so I'll certainly do my best to try and fix things. --- Roy ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
