Re: [boost] Re: Standardization of Boost libraries

From: Jeremy Siek (jsiek_at_[hidden])
Date: 2002-03-07 15:53:48

On Thu, 7 Mar 2002, Toon Knapen wrote:
toon.k> >
toon.k> > Alas, until we can solve some of the low-level issues with
toon.k> > initialization and aliasing, there will still be efficiency arguments
toon.k> > for Fortran.
toon.k> If the next standard also adopts the C99 restrict, I guess we solved the
toon.k> aliasing issue. Or am I missing soth.

I'm afraid C99 restrict is not really a solution for C++. Here's the main
reason:

 restrict is only applicable to pointers

So you can do things like say the pointer parameter to strcpy is
restricted, and will not alias.

However, how useful is this in C++ where we almost never directly use
pointers? Not very.

For a fairly in depth discussion between myself and the author of C99
restrict, see the following thread on the OON-list.

http://www.oonumerics.org/MailArchives/oon-list/msg00391.php

Cheers,
Jeremy

----------------------------------------------------------------------
 Jeremy Siek                          http://php.indiana.edu/~jsiek/
 Ph.D. Student, Indiana Univ. B'ton   email: jsiek_at_[hidden]
 C++ Booster (http://www.boost.org)   office phone: [phone]
----------------------------------------------------------------------