rjmccall accepted this revision. rjmccall added a comment. This revision is now accepted and ready to land.
Good catch! Is there a reasonable way to just make FileCheck itself enforce this, or does that cause too many false positives with tests that are using e.g. `x86_64` as the entire prefix? If we were starting over again, I think ideally the check-prefix would be unconditionally treated as flagging something for FileCheck, so that any sort of unrecognized command after it would produce an error. I doubt there's any real use-case for writing e.g. `CHECK` in a test case. But that's no longer possible because of the widespread idiom of using things like `CHECK-NATIVE` as conditionally-enabled prefixes, which is of course ambiguous with typoing a command like `CHECK-NEXT` or thinking that `CHECK-CONT` should exist. But it'd be nice to get as close to that as we can without rewriting ten thousand `FileCheck` tests. CHANGES SINCE LAST ACTION https://reviews.llvm.org/D58061/new/ https://reviews.llvm.org/D58061 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
