As discussed in several issues and MRs, our tooling (particularly pmbootstrap) has a lot of really awesome ideas, but doesn't always live up to its goals.

this milestone should be used to coordinate quality of life improvements to our tooling. Some examples:

• improved cross compile support (e.g. native meson cross)
• further reduce the amount of heavy stuff run in qemu (e.g. find a way to run mkinitfs natively)
• support multiple pmbootstrap work dirs
• improve chroot state tracking to speed up repeated actions
• support incremental building from local source checkout

An evolving idea is to separate the package and image building parts of pmbootstrap, and investigate reducing chroot spin-up time with Copy-On-Write, etc.