Description
Hi, thanks for maintaining this wonderful project :)
If I understood correctly, one of this project's goals is to isolate as good as possible for builds on macOS, which is why a fixed version of LLVM is pulled in during build time.
I wanted to try making specific builds for the various Apple M processors by adding e.g. -mcpu=apple-m3
to the target_cflags
and target_ldflags
in the aarch64-apple-darwin
target in targets.yml
. (Btw, is that a cursed thing to try?)
For -mcpu=apple-m2
everything works nicely, but for -mcpu=apple-m3
(or m4) the build fails. I couldn't find anything informative in the logs that would point to the issue, so maybe someone can reproduce this?
But that kind of fits a suspicion I have, namely that the LLVM from XCode is notably involved in the build process (which it shouldn't?) and it does not know about M3 or M4 processors, which is why the build fails. When I run clang --version
I get Apple clang version 16.0.0 (clang-1600.0.26.3)
and I'm using macOS 15.1.1 (24B91)
on a M3 MacBook so the LLVM version indeed does not know about M3 and M4. See https://releases.llvm.org/ and https://en.wikipedia.org/wiki/Apple_M3 for the according dates.
Don't know how relevant this is to you but maybe this points to some unintended non-isolation.