hfinkel added a comment.

In https://reviews.llvm.org/D42366#1014546, @rjmccall wrote:

> In https://reviews.llvm.org/D42366#1014157, @kosarev wrote:
> > I think zero would serve better as the unknown-size value. People who are 
> > not aware of TBAA internals would guess that since zero-sized accesses make 
> > no sense, they are likely to have some special meaning. Similarly, for code 
> > that is supposed to process the size fields of access descriptors zero 
> > would be an obvious "illegal size value". In contrast, UINT64_MAX is just a 
> > very large number that doesn't hint anything on its special purpose.
> My thoughts exactly.
> John.

SGTM. Let's do that.

  rC Clang


cfe-commits mailing list

Reply via email to