New C++ standard

[MatthewRock]

Hello, I came across this on G+
http://wiki.hsr.ch/PeterSommerlad/files/TFZ-C++14Standard.pdf

It looks like huge changes. What do you think about these ideas?

[TwilightSpectre]

Yeah, I have seen the C++14 standard already (and I have been using some of its features in Clang quite a bit recently, mostly generic lambdas, the improved constexpr rules and std::make_unique).

If you look closely, though, there aren't actually a very large number of changes: basically its just adding in all the things that should have been in C++11 but they either didn't think of or didn't have time to standardize, for example std::make_unique is the obvious counterpart to std::make_shared, and the constexpr restrictions used to be far too large for what modern compilers can parse.

[LB]

They are following a plan of large changes followed by small fixes. C++98 was followed by C++03, C++11 is followed by C++14, and C++17 is intended to be followed by C++20 (assuming they don't get pushed back).

[BHX]

http://meetingcpp.com/index.php/br/items/cpp-status.html

I'm still not sure what to think about networking being added to C++14.

[LB]

Wow, that's a nice page to show people who don't believe that C++ has changed! It's too bad they're predicting that modules won't make it into C++17 :(

EDIT: What's with the weird grammar in that article? I had a hard time understanding a lot of the sentences.

[Cubbi]

Off the top of my head, besides the much-anticipated std::make_unique, the standard library got std::make_reverse_iterator, std::index_sequence/integer_sequence/etc, std::is_null_pointer, std::is_final, std::quoted, std::exchange, std::bit_not (finally!), std::shared_timed_mutex and std::shared_lock (not to mention additional overloads, operators, and aliases)... Not a lot.

[LB]

I'm disappointed that std::optional didn't make the cut, I was looking forward to it since it would mean I would no longer have any valid use cases for raw pointers.

EDIT: Woah, resumable functions...I haven't even started messing with the threading or atomic libraries yet, haha.

[ResidentBiscuit]

Networking in C++ has me so excited. I personally think it's awesome and will be a huge step forward for modernizing C++.

[LB]

How high will the abstraction levels go for the standard C++ networking support? If it stays relatively low level, I will probably still use external libraries instead.

[NoXzema]

I'd imagine it would be like boost::asio, which is very high level while allowing the developer to access lower-level and OS specific mechanisms if need be. I've not seen the drafts or proposals though.

[BHX]

LB wrote:

What's with the weird grammar in that article? I had a hard time understanding a lot of the sentences.

Jens Weller's native language isn't English so I'd imagine that is why his article seems odd, but it made sense to me after the second read through.