[ADT]Fix AddPointer() in FoldingSet

[kevgs]

Summary

Put a pointer value instead of some garbage from stack to hash buffer.

[bkramer]

What makes you think that this was adding garbage?

[kevgs]

&Ptr is an address of a local variable which depends on a previous call stack and doesn't depends on a pointer value, isn't it? I just do not understand why do we need a this. If it has some purpose why passing a value and ignore it? We could just write
void FoldingSetNodeID::AddPointer() {

const void *Ptr;
Bits.append(reinterpret_cast<unsigned *>(&Ptr),
            reinterpret_cast<unsigned *>(&Ptr+1));

}

[bkramer]

We take a pointer (of the argument on the stack) and it gets dereferenced inside of append. What platform are you on that this causes problems for you?

[kevgs]

Linux AMD64, but I have no problems with that code. I just thought it was incorrect and feel pretty silly now. However, patched method looks nicer and it's probably faster because compiler is not forced to put variable on the stack ;)

Anyway, thanks four your help and sorry for taking your time on such a stupid case.

lib/Support/FoldingSet.cpp

   // depend on the host. It doesn't matter, however, because hashing on
   // pointer values is inherently unstable. Nothing should depend on the
   // ordering of nodes in the folding set.
-  Bits.append(reinterpret_cast<unsigned *>(&Ptr),
-              reinterpret_cast<unsigned *>(&Ptr+1));
+  AddInteger((unsigned long)Ptr);
 }
 void FoldingSetNodeID::AddInteger(signed I) {
   Bits.push_back(I);