Szelethus added a comment. In https://reviews.llvm.org/D53274#1265625, @george.karpenkov wrote:
> I'm not sure why you could get away with removing those llvm_unreachable > cases? Because I got a warning for using `default` when every enum value was handled in the switch. Since whether the flag is set or not can be retrieved via `llvm::Optional::hasValue()`, I removed the `.*NotSet` flags. The future maintainer who adds a new value will get a warning for not handling it anyways. In https://reviews.llvm.org/D53274#1265618, @george.karpenkov wrote: > > The main motivation behind here is to emit warnings if an invalid > > I'm totally with you here, but IIRC (@NoQ might want to correct me here), > the design decision was made specifically to ignore incorrect options, so > that e.g. old versions of Xcode used with old projects would still work. I intend to emit a warning, but go on with the analysis. That's fair I believe? https://reviews.llvm.org/D53274 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits