Hi!
I don't have a solution for you, but I did see a similar difference when I once experimented with building WebAssembly targets along with the main rust toolchain. Something in the result seems to be including the host information, and that should probably be fixed, but I'm not sure where that is coming from.
Regarding mingw-rust, would you really want an entirely separate toolchain? Or should we see about adding mingw targets as a subpackage of the main rust.srpm? We would still need to solve this hash difference, but I'm curious about your approach.
Thanks,
Josh