This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: [v3] Remove uses of <cstddef> as an implementation detail
- From: Gennaro Prota <gennaro dot prota at yahoo dot com>
- To: Paolo Carlini <paolo dot carlini at oracle dot com>
- Cc: Gennaro Prota <gennaro dot prota at yahoo dot com>, libstdc++ <libstdc++ at gcc dot gnu dot org>
- Date: Mon, 07 Jun 2010 22:17:00 +0200
- Subject: Re: [v3] Remove uses of <cstddef> as an implementation detail
- References: <4C080053.9010907@oracle.com> <4C0D243D.9070106@yahoo.com> <4C0D2B3F.8010806@oracle.com>
On 07/06/2010 19.24, Paolo Carlini wrote:
On 06/07/2010 06:54 PM, Gennaro Prota wrote:
On 03/06/2010 21.19, Paolo Carlini wrote:
Hi,
this, largely boring patch essentially does what the Subject says, thus
following the spirit of LWG 1250, avoiding uses of C-style NULLs,
trimming the size of the C++ headers. I took also the occasion to add
std::nullptr_t, in C++0x mode.
Hi all,
I'm just turning up in these parts so my apologies if this has
been discussed and decided upon earlier, but why would one want
such a modification? Just to save on the amount of parsed code?
See the quick summary above. In general we have been untangling and
cleaning up the dependencies in the headers for a while.
I saw the summary but was perplexed, especially about removing
NULL; why giving up using it rather than just defining it in a
"sub-header", as done for std::size_t and std::ptrdiff_t?
Or do you mean that, then, you'd have a dependency on the
sub-header, whereas using 0 directly you haven't any?
(I'm just worrying a bit about code clarity; for this case but
also for the future, if the guideline is that anything not
indispensable is better removed)
--
Gennaro Prota | name.surname yahoo.com
Breeze C++ (preview): <https://sourceforge.net/projects/breeze/>
Do you need expertise in C++? I'm available.
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com