On Thu, Oct 6, 2011 at 1:58 PM, Paul T. Bauman <[email protected]> wrote: > On Thu, Oct 6, 2011 at 2:23 PM, Paul T. Bauman <[email protected]> wrote: >> >> Ah crap. The problem is that the mesh dimension doesn't appear to get set >> until the very end, in particular following the sideset read that trips up >> the 1D read. I'm looking into a workaround. Go ahead and remove the error >> check for now - sorry. > > I believe the attached patch should be a quick fix. The check to set the > mesh dimension was at the end, but it looks to me as though it didn't need > to be - everything needed is known after the first block read. Thus, I moved > the mesh dimension set there, which is before the sideset read, and then > test the mesh dimension there. Note that checking get_num_dim() from > exio_helper is not enough because it sets the spatial dimension, not the > mesh dimension. > The longer term workaround (that I hope to get to later today) is to just > set the boundary ids manually for 1D meshes and bypass the ExodusII read for > that part. > Sorry for the trouble. Let me know if this still breaks for you (it breaks > as expected on my ex0 test).
I think this patch should be OK, we are running our tests here and will check-in if everything passes. Note: libmesh *could* actually be forced to write exodus files with the correct dimension... right now we write everything with dim=3, hence needing to determine the actual dimension after reading the elements. -- John ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 _______________________________________________ Libmesh-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-devel
