Peter Hill
Peter Hill
This now has #2471 merged in, as I was testing the performance with it. I think I've now eliminated the performance hit this PR was introducing -- it seems to...
Fixed conflicts and an actual bug that clang-tidy caught. This is now ready to go in
I'm merging this now so that we can get some actual experience with it in next before the imminent 5.0 release
As you said, nobody's complained in 5 years, so let's get rid of it.
I am absolutely for this. What's the best type for the `*_boundary_condition` options? Could users in principle define their own boundary conditions to be applied? If so, `string` makes sense....
Is this still an issue?
Ah I see! I was thinking that might be suspicious just because I was tracing through where the a/b/c from `Laplacian::tridagMatrix` are used. That deals with `INVERT_AC_GRAD` for both PCR...
Is this a "fundamental" problem that PCF is incompatible with `INVERT_BNDRY_ONE`? Or it would work, if the number of interior points were different such that we'd still use a power...
The plan is now to make `solve` throw if `INVERT_AC_GRAD` is set. This means we won't use PCR as the default solver.
It looks like it's possible to get them to auto-build, but we'd need to have the Dockerfile in this repo -- currently, there's a separate repo (https://github.com/boutproject/docker-bout) with all the...