Andrew Ho

You can use `configure_file(src dst COPY_ONLY)` to have CMake automatically track edits for you.

I tried out the current state of this branch (23edd7525), and I noticed a few things: 1. If you don't build any of the executables which depend on `copy_data` (right...

> Hi @helloworld922, > > Do you mind trying https://github.com/farscape-project/mfem/tree/refactor-cmake-data-directory? I'd push here, but I regret doing that earlier, and I'd rather @cjvogl say he's okay with that beforehand. @cjvogl...

> > I tried out the current state of this branch ([23edd75](https://github.com/mfem/mfem/commit/23edd75253db3f21e39907b3ed419a167f41d1d9)), and I noticed a few things: > > ``` > > 1. If you don't build any of...

I ran a test using ex1p, modifying `par_ref_levels=4` to get larger outputs (17,057,025 DOFs, first orde). mpirun -np 16 ./ex1p -m ../data/beam-hex.mesh No compression: `ParMesh` writeout took 1.97s, `ParGridFunction` writeout...

Summary of the current state of things which need to be resolved: - What should `GetVectorDim` return for fespaces which include traces. One option is for this function to raise...

Added a change which fixes #4732

I split off the fix for #4732 and the macro namespace related changes into separate PR's.

Hmm, the target being `UNKNOWN_LIBRARY` is a known issue with CMake (https://gitlab.kitware.com/cmake/cmake/-/issues/23516), however it shouldn't show up if `MFEM_USE_CUDA` isn't enable. I can dig into how to fix this.

Hmm, I think I figured out why. The logic for [finding targets](https://github.com/mfem/mfem/blob/7c80866b4930d3572035a96ea14dceb74d98694e/CMakeLists.txt#L604) only excludes a CMake-style target if the target doesn't exist, not if it was requested to be used....