C++ Standards Committee Meeting in Rapperswil, June 2014

http://botondballo.wordpress.com/2014/07/17/trip-report-c-standards-committee-meeting-in-rapperswil-june-2014/

by dochtman • 12 years ago

72 32 12 years ago

botondballo.wordpress.com

r-s2• 12 years ago

Great Write Up. There is a ton of information packed in there. C++17 looks like it will be a major improvement like C++11 was. Very much looking forward to Modules.

The (elem : range) proposal seen here: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2014/n385...

I'm really looking forward to as well. I wish it could have been in C++14. STL (the guy) is right when he says (auto elem : range) is very tempting. Its so easy for a novice user to not understand they are making a ton of potentially (and likely) unneeded copies. Also in code review its quite common, to see (const auto &elem : range) when it should be (const auto &&elem : range).

After getting bogged down years ago in a crappy C++03 codebase I eased off C++ for a few years, but i must say, the language is really getting better. 5 years ago if you would have told me I would be looking forward to writing something major in C++ I would have laughed.

cageface2• 12 years ago

I've had the same experience. Hadn't really touched C++ since the late 90's and had traumatic memories from those times but working in my own C++11 codebase recently has been both pleasant and surprisingly productive. The language is headed in the right direction but it's a big ship to steer.

pjmlp5• 12 years ago

I love the language, but mostly on personal projects.

For the typical enterprise developer, at least in many of our projects, I am happy that we manly use other languages.

I feel like crying every time I see C style coding in C++, with its usual set of security exploits.

StephanTLavavej4• 12 years ago

It took me some time to recognize the problem with range-for (after testing its implementation in VC 2012), and more time to work up the courage to write a Core Language proposal. I also wish next-gen range-for had gotten into C++14, but that's mostly a formality - compilers can (and often do) ship features as soon as they're voted into the Working Paper.

Note that nothing's wrong with "const auto& elem", except that it prohibits modification. On the other hand, "const auto&& elem" will typically not compile (it'll insist on binding to rvalues, which only proxy iterators return).

r-s2• 12 years ago

First off, its Great to see you on HN! For the few of you who don't know, Stephan maintains Visual Studio's C++ Standard Library implementation and he is extremely well regarded in C++ circles.

Thats of course right about const auto &&.. It won't work, I did mean (auto &&elem : range).

jzwinck• 12 years ago

As I was reading through your "for (elem : range)" proposal I became quite anxious to know how adding const would be handled. I was hoping that "for (const elem : range)" would be added to mean "for (const auto& elem : range)" because I very often find myself within a non-const method having a non-const container but wanting to iterate with no chance of modifying it.

I was at first relieved to see this topic addressed at the top of Q&A, then disappointed to see "Just do it the old way if you want const." I feel that we don't use const enough in C++, and making it easier (yet still clear) to use would be of significant benefit.

If you agree, perhaps there is still time to support "for (const elem : range)" - it seems to me natural and no more disruptive than the rest of the proposal.

cdi• 12 years ago

Making Cairo C API a C++ standard, through automatic conversion, looks all kinds of wrong to me. The result is not properly generified like Boost GIL, and it carries vaguely defined implementation details like for example font hinting style from freetype (not even cairo library itself). I hope this and the ABI proposal will never make it as the C++ TS or standard.

jokoon• 12 years ago

(cheering)

modules ! modules ! modules !

synthos• 12 years ago

Glad to see named parameters and introspection on the table

ape4• 12 years ago

The extern "abi" for an Application Binary Interface would be so useful.

en4bz• 12 years ago

Shameless plug for my implementation of the Executors & Schedulers TS: https://github.com/en4bz/executors

frozenport• 12 years ago

What is the motivation behind including Cairo?

zanny2• 12 years ago

Herb wants new users to be able to draw graphics without using third party toolkits. It is not meant to be all encompassing, but if you need custom 2d it is a nice feature, albeit very out of scope of a standard library implementation.

pjmlp5• 12 years ago

It is pretty much in scope. We old timers remember the days when compilers came with graphics libraries.

So why forcing everyone to jump through hurdles just to do some simple uses of 2D programming?

cheepin• 12 years ago

The information seems pretty exciting, but I still don't feel like I have fully grasped the changes in C++11 yet. What are some good resources for getting up to speed with the latest and greatest C++?

packetslave• 12 years ago

Scott Meyers has an Introduction: http://www.artima.com/shop/overview_of_the_new_cpp

He's also writing a new book that's available in Rough Cuts from O'Reilly that covers C++11/14: http://scottmeyers.blogspot.com/2014/06/effective-modern-c-s... and has given a talk covering some of the same ground: http://vimeo.com/97318797

Herb Sutter has "Elements of Modern C++ Style": http://herbsutter.com/elements-of-modern-c-style/ and has given several talks, such as http://channel9.msdn.com/Events/Build/2014/2-661

Finally, you can't really beat Stroustrup's book: http://www.amazon.com/The-Programming-Language-4th-Edition/d...

SupremumLimit• 12 years ago

If you already know C++ and just want to get up to speed with C++11/14 features, I wrote several books aimed at specific feature sets supported by different compilers (GCC, Clang, MSVC). Check them out at http://cpprocks.com.

_kst_• 12 years ago

"Modules is, in my opinion, one of the most sorely needed features in C++. They have the potential of increasing compile times by an order of magnitude or more, thus bringing compile-time performance more in line with more modern languages, ..."

Presumably that should be "decreasing", not "increasing" (or "compilation speed" rather than "compile times").

C++ Standards Committee Meeting in Rapperswil, June 2014

Related Stories