NoQ added a comment. > These functions are not called at all by your tests!
Of course they aren't. Because they're dead code. You just introduced them and haven't called them yet. But that code is taken and re-used from the checker. And the code in the checker has the problem. And after you tweaked the code a bit they still have that problem. And i'm pointing out the problem. And you //don't// want me to stop pointing out problems. >> zero test coverage to demonstrate correctness of your solution. > > Zero test coverage? Then what are the unit tests? Unittests do not demonstrate correctness of the overall solution; that's what integration tests do. Unittests only demonstrate that the specific API behaves as expected. I'm questioning the expected behavior. > I am really sorry to tell that, but now I began adding support for raw > pointers as iterators (I will upload them in a separate patch when fully > ready) and then tried your test. It passes, the error is found using this > particular patch. This particular patch is NFC. It could not have helped. Sigh. CHANGES SINCE LAST ACTION https://reviews.llvm.org/D81718/new/ https://reviews.llvm.org/D81718 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
