On Wed, 2 Mar 2011, Roman Vetter wrote:

> Indeed, this did it. I followed the
> MeshTools::Generation::build_cube() procedure, assigning the new nodes
> to DofObject::invalid_processor_id, and then call
> prepare_for_use(true).

Glad to hear it worked!

>> The catch is that I'm not sure our System::project_vector code is up
>> to the task.  It will try to project a solution on the changed mesh
>> from the corresponding solution on the old mesh, which won't exist.
>> Even if that's the case it ought to be simple to fix, though, just by
>> adding a test for the old_dof_object and skipping projections on any
>> elements where that doesn't exist.
>
> Thank you for the detailed description, I have been able to implement
> it as you suggested. Both System::ProjectVector::operator() and
> System::BuildProjectionList::operator() needed an additional continue
> in the case that the old_dof_object doesn't exist for a given element
> or it has no dofs for the current system (this is relevant for
> equation systems with more than one system).
>
> Attached please find my diff. I suppose it should be fully backwards
> compatible and might come in handy for others who also want to modify
> their mesh.

I don't suppose you want to write up an example that does mesh
extension/projection?  Feel free to say no, as this is mostly a greedy
request on my part (the more neat things people can see examples of,
the less they have to ask the mailing lists about), but it might also
be useful to you to have that feature running in our regression
tests.
---
Roy

------------------------------------------------------------------------------
Free Software Download: Index, Search & Analyze Logs and other IT data in 
Real-Time with Splunk. Collect, index and harness all the fast moving IT data 
generated by your applications, servers and devices whether physical, virtual
or in the cloud. Deliver compliance at lower cost and gain new business 
insights. http://p.sf.net/sfu/splunk-dev2dev 
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to