Wolfgang Bangerth
Wolfgang Bangerth
I've reviewed a number of very large patches recently and found it extraordinarily difficult to make progress with them. It also led to long delays for the patch author because...
On 12/25/2015 05:49 AM, Martin Kronbichler wrote: > ``` > Indices are stored in MPI-local index space. DoFHandler will keep a > |Utilities::MPI::Partitioner| object to translate between the local index...
I'd like to get this done, given that step-19 is new in this release.
In principle yes, at least the documentation says so. In practice, it is a boundary point for these cells, and so round-off is a problem. It would probably be useful...
Correct. Not all mappings have support points. Support points only make sense for (a subclass of) polynomial mappings that *interpolate* the boundary.
I don't think this is going to make it for 9.4. Re-tagging for 9.5.
If I see this right, if we don't find a cell we simply return an empty iterator object, right? If so, should the function simply call itself without marked vertices...
Needs to wait for the next release.
I'll add that we have a number of open issues about bugs in the Nedelec element for which we have never been able to find the problem: * #4092 uses...
I've come to realize that at least for the RT element, the function is correct -- though probably by sheer coincidence. I'll fix the coincidence later today. I don't know...