[PATCH] D137020: [clang][AST] Handle variable declaration with unknown typedef in C

[Sam McCall via Phabricator via cfe-commits]

sammccall added a comment.

Not sure this is ready for review again, ignore me if not...

I'm still not sure why this is correct in principle. Without that, if someone finds a misparse 6 months from now I don't know how we determine where to fix it.

For example, this path is called from `Parser::isKnownToBeDeclarationSpecifier()` whose contract is `Return true if we know that we are definitely looking at a decl-specifier... Return false if it's no a decl-specifier, or we're not sure.` There doesn't seem to be any room for heuristics, unless we're going to change that contract and audit all the callers. If this *isn't* a heuristic (it sure looks like one) it needs some comments on why it's correct.



================
Comment at: clang/test/Parser/recovery.c:105
+  unknown_t a; // expected-error {{unknown type name 'unknown_t'}}
+  unknown_t *b; // expected-error {{unknown type name 'unknown_t'}}
+  unknown_t const *c; // expected-error {{unknown type name 'unknown_t'}}
----------------
this diagnostic is worse than the old one (less accurate).

(I think it's OK to trade off diagnostics quality if it's better on balance, maybe leave a comment?)


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D137020/new/

https://reviews.llvm.org/D137020