"The New C++"
started 4/18/2002; 6:40:49 PM - last post 4/19/2002; 3:37:55 PM
|
|
Keith Devens - "The New C++"
4/18/2002; 6:40:49 PM (reads: 456, responses: 3)
|
|
Not as dramatic as it sounds... it's not D (oh wait, there already is a D), but a list of extentions to C++ that are under consideration for the next level standard. The article, from the C/C++ User's Journal, is by Herb Sutter.
On the list are "hash-based containers", new kinds of smart pointers, some stuff from the C99 standard such as specified sized integers (like int32_t or int_least32_t), threading, and "type traits", which let you query the type of a variable, like is_class<T>, or is_member_pointer<T>.
You may also be interested in this interview with him at DevX.
http://www.keithdevens.com/weblog/?id1847
|
|
Chris Rathman - Re:
4/18/2002; 7:31:43 PM (reads: 454, responses: 0)
|
|
Guess I'll have to moan & groan that I think all the new primitive C/C++ types over the last 10 years seem to have ugly names?
|
|
Jo Totland - Re:
4/19/2002; 1:06:56 AM (reads: 465, responses: 1)
|
|
Ehh, what would the new primitive C/C++-types be? The only one I can think up is C9X complex, otherwise new C++-types have usually been implemented in the library (and C9X complex is not in C++). Oh, there is one more of course: bool, but I hardly find that name ugly, quite the opposite in fact!
My bitching and moaning about C++ would mostly be that it has somewhat unpredictable semantics of all the constructs it supports. If you think this construct ought to do that, without having consulted the manual or standard for exactly that case, you are more often wrong than not (at least I am).
The additions mentioned in the article is pretty much straightforward library-additions that many people do themselves anyway. No big surprises there...
|
|
Ehud Lamm - Re:
4/19/2002; 3:37:55 PM (reads: 416, responses: 0)
|
|
Adding to the standard library is in general a good idea. Much much easier than changing the language itself. C++ is in a pretty good place, having the STL. Other languages have some catching up to do.
Some of us are trying to do something about that, and extend the Ada standard library.
|
|
|
|