From: James S. <jbs...@fa...> - 2012-01-10 20:07:55
|
Hi - I've been attempting to use EIDORS on a Mac, and while I ultimately was able to get netgen to compile on Mac OS X 10.7 (using the 4.9.13 sources), it has been a bit difficult to work with because it's throwing quite a few malloc errors during operation of various examples and occasionally segfaults on exit. Has there been further interest or work on using Gmsh? The current code in the most recently release doesn't seem to parse the mesh output from Gmsh 2.5 successfully because it attempts to pull out columns for node numbers that don't exist in the elements section, and I've made some modifications to correct for this to get that functionality working, and would be happy to send a minimal and/or slightly more extensive patch to try and follow the file format specification a little more closely, but I'm not sure how usable EIDORS is without netgen as a mesher. It appears that a bit more work might be needed in order to have Gmsh be able to be usable as a replacement. I'm curious what other users or the maintainers thoughts might be? Is there a reason why Gmsh might not be a suitable mesher? I've become somewhat familiar with Gmsh from working with it with a Finite Volume Method solver (http://www.ctcms.nist.gov/fipy/), and it has seemed to work reasonably well, but I don't have experience with that many mesh generators. -jsnyder |