Hey,
I am currently trying to figure out how to build my own rust toolchain (because right now the latest nightly build is way behind rustc master for some reason I don't know) but the issue is that I want to offload the building process to a cloud server because my of my crippled macbook performance (So it's basically cross-compiling to macOS from linux).
So far, I've got a working osxcross toolchain on my linux server and it basically works. I can compile some crates with cargo/rustc and it uses the target x86_64-apple-darwin with the appropriate linker configured in the config file for cargo.
So I've tried to following with rustc:
git clone https://github.com/rust-lang/rust.git && cd rust
# configured the config.toml with my target:
# target = ["x86_64-apple-darwin"]
# ...
# [target.x86_64-apple-darwin]
# cc = "..."
# cxx = "..."
# linker = "..."
./x.py build
It seems however that it doesn't actually build the rust compiler with the actual target (Judging from the produces files in the build directory. ./build/x86_64-apple-darwin only contains a folder called native and the other one ./build/x86_64-apple-darwin contains only elf files).
Seems that I am doing something clearly wrong. Is this actually possible?