Yes, in the sense that almost all the C++ has been rewritten in Rust (at a raw line count about 10% more, which is not a particularly insightful number). The fish_test_helper standalone binary remains in C++, because it's basically entirely libc calls, uses no other fish code and never gets run by end users.
No, in the sense that none of this code is shipped to users. We've still got work to do to get this into a tarball that someone can (say) `brew install` - at this stage I think it looks like CMake will be hanging around, much simplified, to handle some of the configure/install targets, while cargo is used for the actual build.
Even if it we could shipped to today, there's not really any point. Performance is a bit better in some areas, a bit worse in others, but most importantly the technical rationale for the whole rewrite (see thread safety in https://github.com/fish-shell/fish-shell/blob/27c8845075078041a3376b33bea5898f2369ebe3/doc_internal/fish-riir-plan.md) has not been achieved yet. And there are significant downsides for platform support, at least in the short term: it looks like Cygwin (and I think MSys2) is not going to be supported for a while, and building our own packages on old versions of Linux distributions is a headache.
