[PATCH] D43112: [WebAssembly] Use Symbol class heirarchy. NFC.

[Nicholas Wilson via Phabricator via llvm-commits]

ncw added a comment.

In https://reviews.llvm.org/D43112#1007692, @sbc100 wrote:

> Also, doesn't the existing assert do exactly what you want already:
>
>   assert(static_cast<Symbol *>(static_cast<T *>(nullptr)) == nullptr &&          
>            "Not a Symbol");   
>
>
> ?


D'oh! I hadn't spotted that one in amongst the rest, it covers exactly one of the cases I was worried about.

OK, I think the existing asserts cover it all then, plus your helpful comment about placement new clears that up.

Whew, thanks for adding the other assertions. I agree now it's all safe.  (Although, a standards-conforming compiler could quite legally fail the assertion, no compilers will in practice and you'd just get a build failure. So that's OK.)


Repository:
  rLLD LLVM Linker

https://reviews.llvm.org/D43112