AaronBallman wrote: > I saw this got updates over break. Did LWG make its decision already? Are you > expecting re-review? I didn't see anything go across my emails about LWG, but > I could definitely have missed it.
The LWG issue is still open: https://cplusplus.github.io/LWG/issue4149 but doesn't seem to be under active discussion any longer. My reading of the reflector thread is that my suggested approach is not quite correct. Consider a case like: ``` #define maybe_unused 12 #include <algorithm> ``` where the namespace isn't going to be declared (aka header not included) at the time the `#define` happens. We could aim for a perfect analysis here, but it gets pretty complicated, so I am leaning towards always issuing a `-Wreserved-attribute-identifier` warning when we see the `#define` of a standard attribute name. It does mean false positives are possible, but the amount of C++ code that does not include a standard library header anywhere in the TU *and* defines standard attribute identifiers as macro names is hopefully quite small. WDYT @erichkeane? https://github.com/llvm/llvm-project/pull/106036 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
