AaronBallman wrote: > > I kind of think "no enforcement for invalid input" is reasonable here. That > > would be consistent with things like the alias attribute where we accept a > > string that has to conform to ABI mangling rules (which could be a matter > > of poor QoI rather than precedent). > > If I'm not mistaken, `alias` does _not_ accept invalid input, because the > compiler ensures the given string matches the mangled name of another entity > in the program. Now, if Clang creates an invalid mangling for that entity > (due to, for example, an unchecked `abi_tag`), `alias` will also accept that > invalid mangling.
Fair point, I was thinking in terms of compiler construction. We don't do that semantically, we do that during codegen: https://godbolt.org/z/PMaK431nv But for `alias`, we can do that from codegen relatively easily. We can't do it for `abi_tag` though because we don't have an exposed interface for checking identifier validity outside of the lexer. https://github.com/llvm/llvm-project/pull/84272 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
