We use the PreLoad.cmake to enforce a generator for MinGW compatibility, and avoid having the user provide an argument manually to simplify the build process. As you note, the downside is that it prevents an experienced user from overriding such an argument from the command line as they usually would.

For your own purposes, it should be safe to remove the file or adjust it to your needs. I only ask that when you submit a PR that you do not commit that particular change.

It should also be noted that we have not tested the build system or library using any makefile generators other than Unix Makefiles, and while I don't expect issues, I can't be 100% sure of the behavior you'll see. If Ninja works much better for you, either in performance or some other way, please let us know, either here or on our discourse forum.

For your own purposes, it should be safe to remove the file or adjust it to your needs. I only ask that when you submit a PR that you do not commit that particular change.

Makes sense, that's what I've been doing.

If Ninja works much better for you, either in performance or some other way, please let us know

I was previously told Ninja would typically be faster than GNU Make.
I did some quick and dirty benchmarking to confirm this, and for openfhe-development's default CMake options make takes on average ~85s while Ninja takes ~69s on a Ryzen 5900X (24 threads).