Re: [Libmesh-devel] Mesh::n_elem() and Mesh::n_nodes()

[Kirk, B. \(JSC-EG\) <ben...@na...>]

Why not wrap some safety code inside

#ifdef NDEBUG
  (compute O(N) slow values)
  (assert slow way =3D vector.size())
#endif=20

To catch when they may be wrong? =20

Are you hitting a case where you need them but they are wrong??



-Ben




-----Original Message-----
From: lib...@li...
[mailto:lib...@li...] On Behalf Of John
Peterson
Sent: Thursday, February 15, 2007 10:16 AM
To: lib...@li...
Subject: [Libmesh-devel] Mesh::n_elem() and Mesh::n_nodes()

Hi,

These functions are O(1) "fast" (they just return the sizes of the
underlying std::vectors of Nodes and Elements) but they are not always
100% correct.

For example, this is not correct during coarsening when elements are
being deleted and their vector entries are being replaced by NULLs.
It's also not correct if I build a mesh and randomly delete an element
from it, which is allowed by the current interface as well.

An alternative is to replace these implementations with something
similar to the n_active_elem() functions (defined in mesh_base.C) which
use the Mesh iterators but are O(N) complexity.

Another possible try would be to keep a "running total" of the number of
nodes/elements that would be incremented/decremented by the
Mesh::add_elem(), Mesh::delete_elem(), Mesh::add_point(), and
Mesh::delete_node() functions.  However, I see this as potentially
error-prone as well.

Thoughts?
-John

------------------------------------------------------------------------
-
Take Surveys. Earn Cash. Influence the Future of IT Join
SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3D=
DEVDE
V
_______________________________________________
Libmesh-devel mailing list
Lib...@li...
https://lists.sourceforge.net/lists/listinfo/libmesh-devel