#1428 Remove test for entity-subentity consistency from checkindexset.cc
Metadata
| Property | Value |
|---|---|
| Reported by | Oliver Sander (oliver.sander@tu-dresden.de) |
| Reported at | Feb 4, 2014 11:31 |
| Type | Bug Report |
| Version | Git (pre2.4) [autotools] |
| Operating System | Unspecified / All |
| Last edited by | Oliver Sander (oliver.sander@tu-dresden.de) |
| Last edited at | Feb 5, 2014 15:23 |
| Closed by | Oliver Sander (oliver.sander@tu-dresden.de) |
| Closed at | Feb 5, 2014 15:23 |
| Closed in version | Unknown |
| Resolution | Implemented |
| Comment | In master: 3ee3612cc15fa485b971c33fa2039ce33493a8cd |
Description
At the 2013 dev meeting in Aachen we decided that entities of all codimension should exist (if they exist) conceptually separate from any other entity. In particular, this implies that if I can access an entity as the subentity of two different elements, then the subentity will look the same no matter what element I use to obtain it.
At least I remember deciding this. For some strange reason it does not show up in the meeting minutes.
Anyway, there is one small test in checkindexset.cc which makes the opposite assumption: that subfaces have different vertex orderings depending on the element where I call the subEntity method to get to the subface. The reason this doesn't cause problems in dune-grid appears to be that a number of grids (OneDGrid, UGGrid, YaspGrid) don't implement subfaces at all, and for the others it comes out right "by coincidence" (twist-free-ness).
However, the FoamGrid grid manager implements edges in a 2d grid as separate objects, and the grid check fails for that reason. In view of the Aachen decision I request to remove the offending test. The attached patch does just that. Are there objections?