wimlib

I've uploaded wimlib v1.13.3-BETA1 which contains the optimization you suggested to avoid unnecessarily falling back to a recursive scan. Can you try it out?

No, this isn't currently supported, except in the special case of the --update-of option. FWIW, I've wanted to implement this, but it's a bit difficult since wimlib-imagex uses the getopt_long() function to parse its command-line arguments, and getopt_long() considers anything beginning with "-" to ...

1. I think I did it that way just because it was simpler, and at the time not many directories on a default Windows install had object IDs (or otherwise triggered using the recursive scan). It's possible that Windows has changed to start assigning object IDs to more directories and as a result the c...

No, the minimum XPRESS chunk size is 4096 bytes.

Using 2048-byte chunks would probably make it slower, not faster.

wimlib 1.13.2 has been released. This is a maintenance release. List of changes: Prevented miscompilation with gcc 10 at -O3 (https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94994). Avoided some compiler warnings with gcc 9 and later. The 'mkwinpeimg' script now also looks for syslinux modules in /usr/...

It's expected that creating a solid-compressed WIM takes a long time and uses all CPUs (or however many threads you set). It's very computationally intensive.

What wimlib does is correct for WIM files. There could be other variants of LZX though.

With --compress=XPRESS:20, compression shouldn't be a bottleneck at all. You might have better results with a higher compression setting, e.g. --compress=XPRESS, --compress=LZX:20, or even the default of --compress=LZX. You could also try changing the number of threads (--num-threads). But if that d...

If you're using the --source-list option to capture all these drives at once, then I don't think there's any way to do that currently, as the excluded paths are always matched relative to each top-level directory being captured. But if you're using a separate wimlib-imagex command to capture each dr...

wimlib hardly ever uses temporary files, so choosing a different temporary directory wouldn't actually matter.