Adding a __has_include_next check sounds good to me. We already do this for several other headers.
Hal: if you're interested in driving this (rebasing the patch and adding the include_next machinery), please go ahead. On Thu, Sep 25, 2014 at 1:42 PM, Ed Schouten <[email protected]> wrote: > On 22 September 2014 16:38, Hal Finkel <[email protected]> wrote: > > Understood. I suggested exactly this ( > http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20140915/115196.html). > Does that work for you? > > Yes. That's perfect. Thanks! > > On a more general note, I'd love to see if Clang would eventually > focus on providing built-ins that would allow OS authors to come up > with the standard headers easily, instead of providing the headers > themselves. > > For example, it would be pretty nice if we could write down things like: > > typedef __builtin_uint32_t uint32_t; > > #define LONG_MAX __builtin_max(long) > > etc. > > But don't let me hijack this thread. :-) > > -- > Ed Schouten <[email protected]> >
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
