Wow, did I miss a lot of devel messages; you go away from email for
one or two hours...

On Fri, 23 Jan 2009, Derek Gaston wrote:

> In the (mysterious) future we could make a
> recalculate_n_subdomains()... it would just run through all of the
> elements and get the std::set of subdomain id's then call size().

This is a good idea.

We could also keep that subdomain id set precomputed like John
suggested, but doing it safely would take more than just a
modification of add_elem() - we'd also want to modify the
subdomain_id() function that returns a writeable reference, so that
any attempt to rewrite a subdomain id got caught: either replace it
with a set_subdomain_id(), or if we suddenly "got religion" again
about API backwards compatibility, cause it to return a proxy class
with an overloaded operator=.

> But that's overkill when we really do know the number of
> subdomains... so for now I'm just going to make that public.

Sounds good.

We really need a third category beyond public and protected/private,
though; maybe start using namespaces called "Private" as shorthand for
"the compiler won't stop you from using this, but we're not going to
support the API or guarantee it will work like you expect".
---
Roy

------------------------------------------------------------------------------
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