Off topic, but I would love to eliminate the "c++ness" notion of the driver. I think it is confusing, and most users don't really understand whether or not it is important (and it usually isn't).
It only effects two things: 1. The behavior of the driver when used to link. 2. Some input files are "auto-detected" as C++. The latter behavior is gross, produces a warning, and I believe has been deprecated in gcc for a while. Maybe we should just eliminate it? If we did that, then it would be entirely sensible to provide another (more sensibly named) option to request linking with the C++ standard library (and other associated behavior). Then clang -flink-c++ (or whatever we called it) and clang++ would be equivalent, and we could also kill -ccc-cxx. I doubt we could eliminate the actual "clang++" file even if we did that, but it would at least be nice to tell people that it is just a shortcut for providing the -flink-c++ flag. - Daniel On Mon, Oct 29, 2012 at 5:10 PM, Jordan Rose <[email protected]> wrote: > > On Oct 29, 2012, at 17:04 , Eli Friedman <[email protected]> wrote: > > > On Mon, Oct 29, 2012 at 4:59 PM, Rafael Espíndola > > <[email protected]> wrote: > >> In the past "production" clang builds would not be used for c++, and > >> we had the -ccc-clang-cxx and -ccc-no-clang-cxx options to force them > >> on or off for testing. C++ is now production quality and I think these > >> options are dead. Is it ok to remove them? > > > > Yes, please do. > > Is there any other way to force clang to behave like clang++ though? I > have aliases for built clangs, but clang -x c++ is not the same as clang++. > > > _______________________________________________ > cfe-commits mailing list > [email protected] > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits >
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
