LLVM semantics vs mmap/munmap

[jyknight]

Oh, that’s apparently already known and been discussed, but not yet resolved.

RFC: Decomposing deref(N) into deref(N) + nofree

TLDR: We should change the existing dereferenceability related attributes to imply point in time facts only, and re-infer stronger global dereferenceability facts where needed. Meta If you prefer to read proposals in a browser, you can read this email here. This proposal greatly benefited from multiple rounds of feedback from Johannes, Artur, and Nick. All remaining mistakes are my own. Johannes deserves a lot of credit for driving previous iterations on this design. In particular, I want t…