This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

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


On 9/19/06, Paolo Carlini <pcarlini@suse.de> wrote:
Joe Buck wrote:

>In any case, we shouldn't accept any extension that isn't rigorously specified:
>if it isn't in a released standard, it should have a rigorous description
>with the same kind of detail that one finds in an official standard.
>
As far as I can see, Doug's variadic templates certainly qualify, from
this point of view.

For all the reasons explained again by Doug and Joe, I would like to see
some leeway for that subset of C++0x features which cannot possibly
result in incompatibilities. At the same time, we must help the users
with appropriate warnings, and do more disruptive work only in a new branch.

I agree that we need to have proper warning machinery in place. I already suggested a -Wc++0x-compat warning which would warn about c++0x keywords used in existing code. And I suggest to either warn about or disable c++0x features if -std=c++0x is not specified (barring a way to have tr1 use c++0x features even if not in c++0x mode).

It would be unfortunate if people cannot start to explore the new
language features.

Richard.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]