[clang] [BoundsSafety][Sema] Allow counted_by and counted_by_or_null on pointers where the pointee type is incomplete but potentially completable (PR #106321)

[Dan Liew via cfe-commits]

================
@@ -2438,6 +2438,22 @@ bool Type::isIncompleteType(NamedDecl **Def) const {
   }
 }
 
+bool Type::isIncompletableIncompleteType() const {
+  if (!isIncompleteType())
+    return false;
+
+  // Forward declarations of structs, classes, enums, and unions could be later
+  // completed in a compilation unit by providing a type definition.
+  if (isStructureOrClassType() || isEnumeralType() || isUnionType())
+    return false;
+
+  // Other types are incompletable.
+  //
+  // E.g. `char[]` and `void`. The type is incomplete and no future
+  // type declarations can make the type complete.
+  return true;
+}
+
----------------
delcypher wrote:

This implementation is more concise (and is probably slightly faster) but it is much less readable. I could potentially fix that by using the conditions if your patch but not writing them like this.

```
if (!isCompleteType())
  return false;

// Forward declarations of structs, classes, enums, and unions could be later
// completed in a compilation unit by providing a type definition.
TagDecl *TD = getAsTagDecl();
if (TD)
  return false;

// Other types are incompletable.
//
// E.g. `char[]` and `void`. The type is incomplete and no future
// type declarations can make the type complete.
return true;
```

Why is the `!TD->isCompleteDefinition()` necessary? If `isIncompleteType()` returned true then that would imply that `!TD->isCompleteDefinition()` is true, right? Or am I missing something?

https://github.com/llvm/llvm-project/pull/106321