Aliasees no longer have to be GlobalObjects

[mkuper]

Summary

r210062 reverted the need for aliasees of GlobalAliases to be GlobalObjects, but left a cast<> in CloneModule() that was added in r209007 in place.

Removing the cast.

[• rafael]

LGTM as is since this can be read as a code cleanup.

But it would be awesome to have a testcase :-)

[mkuper]

Thanks Rafael.

So, the problem is that I'm not sure how to construct a test-case for this.
I've ran into it when debugging a crash in InstCombine and having bugpoint itself crash on the module while trying to reduce it. And it looks like bugpoint is also the only thing actually using this interface.

I'll commit it as is, if you have a good idea for how to test this, let me know.

[mkuper]

Closed by rL224756.

lib/Transforms/Utils/CloneModule.cpp

for (Module::const_iterator I = M->begin(), E = M->end(); I != E; ++I) {
     }
   }
 
   // And aliases
   for (Module::const_alias_iterator I = M->alias_begin(), E = M->alias_end();
        I != E; ++I) {
     GlobalAlias *GA = cast<GlobalAlias>(VMap[I]);
     if (const Constant *C = I->getAliasee())
-      GA->setAliasee(cast<GlobalObject>(MapValue(C, VMap)));
+      GA->setAliasee(MapValue(C, VMap));
   }
 
   // And named metadata....
   for (Module::const_named_metadata_iterator I = M->named_metadata_begin(),
          E = M->named_metadata_end(); I != E; ++I) {
     const NamedMDNode &NMD = *I;
     NamedMDNode *NewNMD = New->getOrInsertNamedMetadata(NMD.getName());
     for (unsigned i = 0, e = NMD.getNumOperands(); i != e; ++i)