[flang] Fix regression from generic fix

[klausler]

Summary

A recent change that corrected the name resolution of a generic interface
when the same name was visible in scope incorrectly prevented a local
generic from shadowing an outer name that is not a generic, subprogram,
or derived type -- e.g., a simple variable -- leading to an inappropriate
error message.

flang/lib/Semantics/resolve-names.cpp

if (const auto *existingGeneric{ultimate.detailsIf<GenericDetails>()}) {
           AddGenericUse(genericDetails, existing->name(), use->symbol());
           EraseSymbol(*existing);
         }
         existing = &MakeSymbol(symbolName, Attrs{}, std::move(genericDetails));
       }
       info.Resolve(existing);
       return;
     }
+    if (&existing->owner() == &currScope()) {
       if (ultimate.has<SubprogramDetails>() ||
           ultimate.has<SubprogramNameDetails>()) {
         genericDetails.set_specific(ultimate);
       } else if (ultimate.has<DerivedTypeDetails>()) {
         genericDetails.set_derivedType(ultimate);
       } else {
         SayAlreadyDeclared(symbolName, *existing);
         return;
       }
       EraseSymbol(*existing);
     }
+  }
   info.Resolve(&MakeSymbol(symbolName, Attrs{}, std::move(genericDetails)));
 }
 
 void ResolveNamesVisitor::FinishSpecificationPart(
     const std::list<parser::DeclarationConstruct> &decls) {
   badStmtFuncFound_ = false;
   funcResultStack().CompleteFunctionResultType();
   CheckImports();