I am curious if there are plans to implement a libc (not libc++ which
already exists) under the umbrella of the llvm project.
glibc is crufty and bloated, musl is not as fast as glibc and not
supported by some software.
It would be really nice to see a fresh start, a fast yet lightweight
but still well supported libc; and the llvm project seems to be a good
candidate to tackle this.
They already had been discussed in the past. The answer is probably never. Libc are far more system dependent.
Each system has it’s own for good reasons. They need to interact closely with the kernel, and so can’t be generic. Of course there is a bunch of functions that may be shared, but there is also large part of the API that can’t.
Moreover, you will have a hard time convincing people to use yet another libc. I don’t expect linux to replace glibc, Apple need a very tight control of its libc and is not going to anything but its own.