search

mesh-adaptation / goalie

Goal-oriented error estimation and mesh adaptation for finite element problems solved using Firedrake
Other
2 stars 0 forks source link

Introduce R-space time `Function` #172

Closed jwallwork23 closed 7 months ago

jwallwork23 commented 7 months ago

Apologies for the delay on this. Here is what I mean by avoiding having some special err_ind_time. I'd rather have a $R$-space time Function that is used both in the forms and in the error indication.

If you're happy with this then we should use the $R$-space time for the time integration in the other demos, too.

Note also that I proposed using set_meshes when setting adapted meshes, so that the various other properties (including the new time) get re-initialised. Ideally, we would move the per-subinterval code here to the setter method of MeshSeq so that it wouldn't make a difference anyway. (See #122.)

jwallwork23 commented 7 months ago

Note that this is proposed to be merged into the open PR #137, not main.

jwallwork23 commented 7 months ago

[Needed to rebase on top of main to fix some issues, so the same will be required of #137.]