Hello all,

Here is a proposal for small change in the "read" method that I had to
do for my work. Others might find it useful.
http://cmaf.ptmat.fc.ul.pt/~pvieira/patch

Currently we have a bool flag named "skip_renumber_nodes_and_elements"
declared in the method and initialized in this way:

        bool skip_renumber_nodes_and_elements = (name.rfind(".gmv") < 
name.size());

That is, true if we are reading a mesh in GMV format, false otherwise.
I think it would be better if this flag was passed as an argument of
the method, so that the user can have control over whether or not the
renumbering should be done (regardless of the input format). The
default value will be false, so nothing will change in the case the
third argument is not used. The exceptions are the GMV and XDR legacy
formats, in which case the flag will always be assigned true (even if
the method was called with false).

This is useful if the user wants to compare the results obtained with
libMesh with results from other finite element codes. In my case, I'm
interested in comparing the solution I get with libMesh and the
solution from FreeFem++ (using the same mesh as input, obviously). I
can save the solution with a precision of only 4 ou 5 decimal places
and diff the two files to check if they're equal. If the mesh is
renumbered in libMesh, I cannot do this.

Paulo

------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
Libmesh-devel mailing list
Libmesh-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-devel

Reply via email to