dexonsmith closed this revision. dexonsmith added a comment. In https://reviews.llvm.org/D34578#801519, @hfinkel wrote:
> In https://reviews.llvm.org/D34578#801357, @dexonsmith wrote: > > > Ping! Hal, you committed r283051. Do you have a problem with this? > > > It looks like you're just renaming `__libcpp_isnan` and friends to > __libcpp_isnan_or_builtin` and similar. LGTM Thanks; committed in r307357. >> (Incidentally, I noticed that r283051 was optimizing the implementation of >> <complex>. I wonder why we have so much code in that header, instead of >> calling out to libc's already-optimized <complex.h> implementation?) > > This is definitely worth looking at (especially given that we don't support > complex of user-defined types with those functions). Would making > complex<double> use _Complex double, and so on, be an API break? Would that > matter? I think paragraph 4 of complex.numbers (from n4640) effectively enforces that `complex<T>` and `_Complex T` are ABI-compatible. https://reviews.llvm.org/D34578 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
