https://github.com/hokein commented:
Thanks, the diagnostics in the test look much better. There is an ambiguous case that I'm a little concerned about: if the alias template parameter and the underlying template parameter have the same name, the synthesized deduction guide may end up with two template parameters with the same name. For instance, see [this example](https://godbolt.org/z/q1n7Edzcz). In such cases, we cannot distinguish between the `T1` parameters from the function signature `auto (T1, T1) -> X<T1, double>`. However, the diagnostic improvement looks promising, and this seems like the right tradeoff. https://github.com/llvm/llvm-project/pull/99840 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
