Fwd: C++ PATCH: PR 20599 (1/3)

Gennaro Prota gennaro_prota@yahoo.com
Tue Sep 19 18:20:00 GMT 2006

On Tue, 19 Sep 2006 10:38:34 -0700, Joe Buck <Joe.Buck@synopsys.COM>

>On Tue, Sep 19, 2006 at 07:13:29PM +0200, Gabriel Dos Reis wrote:
>>    http://www.open-std.org/JTC1/SC22/WG21/docs/papers/2006/n2011.htm
>>    http://www.open-std.org/JTC1/SC22/WG21/docs/papers/2006/n2051.htm
>Thanks for the pointers.
>The new feature I most want to use is "auto", which I'm happy to see
>has made it into the working paper (I'd heard it had, but thanks for
>the confirmation).

Oh, and modules, inheriting constructors, macro scopes... :-) I'm also
eagerly waiting for strongly typed enums, which would finally make all
the static const embarrassing and core issue 454:


a recollection of the past (there's no point in "fixing" static consts
if one can have real, typed, compile-time constants without rvalue

I don't like 'long long' for the simple reason that it doesn't scale
well (what will be next? long long long?), nor I like some C99
compatibility compromises.

And finally (stopping with one foot and half in the off-topic hole) I
wonder why 'nullptr' instead of 'null_pointer' or 'null_ptr' and why
'static_assert', with a verb, instead of 'static_assertion' (with
noun). The latter has Boosthistorical reasons, of course, but once it
is a declaration I find the verbal form disturbing.

Oh, and I'd like a syntax for explicitly telling you want the
implicitly generated copy functions. Currently you can just have a
comment, which isn't very much checked by the compiler :-)

  X ( [const] X & ) = auto;
  X & operator =( [const] X & ) = auto;

(details to be defined as to whether you have to specify the correct
form, i.e. with or without the const qualifier, or not)


More information about the Libstdc++ mailing list