https://github.com/Endilll commented:
It appears that CWG400 has two parts to it: 1. An arguably primary part of the issue is concerned with how many declarations using-declaration can name (one vs many). This is what the recorded discussion is about, and what Richard wrote tests for back in 2014, which we based our conformance claims on without C++98 caveat. 2. Requirement that nested-name-specifier names a base class. This aspect wasn't discussed anywhere, and we didn't test for it, but I believe it was intentional, because it is still in the wording, and is implemented by everyone else. ------------------- As for actionable feedback, we need to update CWG400 tests with cases from [namespace.udecl]/3 tests that were affected. Also, I would like us to mention in release notes that we're conforming with CWG400 now. https://github.com/llvm/llvm-project/pull/143492 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
