Conversation
Failure seems unrelated, outdated actions? |
xref #32 |
|
CI is back online. |
| @@ -0,0 +1,79 @@ | |||
| function fix_node_order!(grid::Grid{2}) | |||
There was a problem hiding this comment.
Can we add a meaningful error error for 3D and a bypass for 1D?
We also had this for grids from abaqus input files (ref), perhaps should be added to Ferrite then? |
|
Should we add such geometry processing stuff into Ferrite directly or a package like FerriteGeometry/FerriteMeshTools/...? |
|
I think I’m Ferrite directly makes more sense as long as not too exotic to avoid a jungle of packages and if it doesn’t add more deps |
As this happens from time to time for users loading from Gmsh, maybe this functionality that I did for an internal project we have could be relevant here.
Would be a solution for e.g. #15 without having to use Gmsh commands, which is useful if Gmsh is run from another software (e.g. neper).
Of course, this functionality could be in Ferrite too since it doesn't really interact with Gmsh, but on the other hand, this problem only seems to arise for Gmsh.
The few tests I've conducted indicate that it's fast enough to be used as a safe (perhaps even default) option when loading a Gmsh grid, but in that case, I guess the order of the vertex- and facetsets should be preserved?
Let me know if this functionality is desired here.
As reference, here is also code I used to visually validate it in addition to the tests,