[MOAB-dev] Problems with remove_adjacencies

[Iulian Grindeanu]

----- Original Message -----
| Hi again,
| Last email hopefully. I thought that some of the stuff I was saying
| was a bit confusing so thought I would email a short version of my
| code. Basically this is a shortened version of what I am trying to do
| but if you run the code with the attached mesh (of 2 tets) you can see
| what the problem is. The adjacent faces of the tet is not changing
| even though the connectivity is changing and I am removing adjacencies
| of the face from the tet.
| Regards
| Graeme
| --------------------
| Graeme A. Edwards
| PhD Student - Materials & Mechanics Research Group
| School of Engineering
| University of Glasgow
| Rankine Building
| Oakfield Avenue
| G12 8LT
Hi Graeme, Thank you for the testing code. Your code seems to negate/contradict all the suggestions I made before :( . So I made an error in what I said before. Indeed, with this example, the adjacency of the tet did not seem to change, when you created a new node on face 4. If you modify your code like this: diff testing.cpp testing_Graeme.cpp 37c37 < MBEntityHandle old_node=1;//one node on the shared face, but a corner node!!!! --- > MBEntityHandle old_node=7;//one node on the shared face 55c55 < //result = moab->remove_adjacencies(T,&face,1);RR//remove old face --- > result = moab->remove_adjacencies(T,&face,1);RR//remove old face you would get expected result, new faces created, adjacent to the modified tet. testing Tri 1 - 7 Tri 4 Tri 1 - 4 Tri 1 Tri 8 - 10 So the problem is that the adjacency routine seems to overlook the mid-nodes. I can look into it deeper, but I do not think we are going to change the code there. Or maybe there is a flag, to consider midnodes too. I suggest, as a workaround, to always duplicate corner nodes (at least one), to get the desired effect. You do not have to remove adjacency between face and tet, it has no effect, I commented it out. It is not stored, by default, and I suggest to not store it. I hope this helps. Iulian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/moab-dev/attachments/20120509/393baca8/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: testing.cpp
Type: text/x-c++src
Size: 2051 bytes
Desc: testing.cpp
URL: <http://lists.mcs.anl.gov/pipermail/moab-dev/attachments/20120509/393baca8/attachment.cpp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 2_tets.vtk
Type: application/octet-stream
Size: 557 bytes
Desc: 2_tets.vtk
URL: <http://lists.mcs.anl.gov/pipermail/moab-dev/attachments/20120509/393baca8/attachment.obj>