mbid added a comment. Hi Nathan, thanks for looking into this. I did indeed not know about the existing check in cppcoreguidelines -- maybe I should've been more suspicious about assuming something so obvious supposedly doesn't exist yet :)
But perhaps the replacements could still be integrated into the existing check. Note that replacements in my patch are generated only if you have a defaulted copying version. So e.g. the defaulted move constructor is emitted only if the class has a default copy constructor. This can still introduce bugs, but I think the probability for that is fairly low. For context, I'm trying to run this on the chromium codebase, where copy constructors are required to be defined (even if = default) in the .cpp file for all non-trivial classes, which is why the implicit move constructors are often missing. I want to understand what the performance impact of that is. > the replacements that this check generates would often not improve performance Why would you expect there to be no improvement? If there's a default copy constructor and hence no (implicit) move constructor, than the more expensive copying version will be invoked when a move would be sufficient. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D138031/new/ https://reviews.llvm.org/D138031 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
