https://github.com/zyn0217 commented:
My two cents: While I appreciate the value of inspecting more semantic information about lambdas, I still have mixed feelings about user experience. Admittedly, presenting implicitly captured variables provides users with insight into ODR-used variables, but I'm unsure about our strategy: We would probably end up showing too many hints inside the lambda capture, which would visually elongate the lambda header and distract users from reading. Indeed, we have a configuration entry to restrict the length, but I feel that's insufficient (Although we're baking it with TypeNameLimit currently, which is also inappropriate). For example, a user might have a lambda capturing several variables named in single letters e.g. a-z. As a result, we probably eventually display `[&: a, b, c, d, e, f, g, ...](...)`, which is IMO too spammy. That said, I wonder if we can implement the capture stuff in hovers instead. >From what I can think of, we present something e.g. ```cpp // implicitly captured by reference/copy: int a; std::string s; ... ``` when user hovers over the capture defaults. (I don't have a strong opinion on the hints for default arguments. It would also be great to hear @HighCommander4 and @kadircet's input.) https://github.com/llvm/llvm-project/pull/95712 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
