Re: PATCH to implement `restrict' in C

[Nathan Sidwell <nathan at acm dot org>]

Jeffrey A Law wrote:
> 
>   In message <k23e8jsb68.fsf@zero.aec.at>you write:
>   > There is one problem with that: once restrict is ported to g++ it will
>   > need an own option for it. -flang-c9x doesn't make sense for a C++ compiler
>   > (and a new C++ standard including it is unlikely to appear anytime soon)
>   > So an -frestrict is needed anyways, and it would be inconsistent to have
>   > it for g++, but not for gcc. So I propose to keep -frestrict, but to have
>   > -flang-c9x imply it.
> Actually I'd prefer "-lang-c9x", not "-flang-c9x".  Again, consistency with
> existing language variant options.
Just a thought, should we (eventually) allow g++ to parse C9X header
files? If so, doesn't there need to be some other mechanism to flag
this. `-lang-c++ -lang-c9x' would be a rather stupid way to do it! I
don't think an extern "C9X" { ... } construct would be appropriate, as
that's specifying linkage not syntax. Although -lang-c9x on its own is
sufficient to select C9X flavour C, it's not approriate for a C9X aware
C++ compiler.

> For C++, I would recommend using __restrict__ until it appears in a C++
> standard of some sort.
(and have it always understood? -- no -frestrict flag?) sound's good to
me.

nathan
-- 
Dr Nathan Sidwell :: Computer Science Department :: Bristol University
      You can up the bandwidth, but you can't up the speed of light      
nathan@acm.org  http://www.cs.bris.ac.uk/~nathan/  nathan@cs.bris.ac.uk