Rust’s lifetime system doesn’t really help when writing a GHC. On the other hand, the portability Rust does offer in its standard library is much more sanely managed without the abomination that is libc, so we ought to be able to remove a fair amount of CPP.
What I really want to do is right the RTS in Haskell — Go’s is in Go, and some ancient Lisps’ were in Lisp, so this is less insane than it sounds.
Perhaps, getting more to the point than my own suggestion… For me, a convincing argument against using Rust for the runtime would be that, as it is, there is already plenty of activity/interest/enthusiasm for working on the runtime. Can anyone speak to this? Is this the case? (Granted, of course, that if there is not a sense of enthusiasm for working on the project in C, then it does not necessarily follow that converting to Rust is the solution).