Prior to C++11, for every user-defined type we wrote in C++03, all standards-conforming C++ compilers gave us:
- a “free” copy constructor
- a “free” copy assignment operator
- a “free” default constructor
- a “free” destructor
The caveat is that we only got them for free if we didn’t manually override the compiler and write them ourselves. And unless we defined reference or pointer members inside of our type, we didn’t have to manually write them.
Starting from C++11 on, we not only get those operations for free for our user-defined types, we also get these turbo-boosters:
- a “free” move constructor
- a “free” move assignment operator
In addition, all of the C++ standard library containers have been “move enabled“.
When I first learned how move semantics worked and why this new core language feature dramatically improved program performance over copying, I started wondering about user-defined types that wrapped move-enabled, standard library types. For example, check out this simple user-defined Msg structure that encapsulates a move-enabled std::vector.
Logic would dictate that since I get “move” operations from the compiler for free with the Msg type as written, if I manually “moved” a Msg object in some application code, the compiler would “move” the vDoubs member under the covers along with it – for free.
Until now, I didn’t test out that deduction because I heard my bruh Herb Sutter say in a video talk that deep moves came for free with user-defined types as long as each class member in the hierarchical composition is also move-enabled. However, in a more recent video, I saw an excellent C++ teacher explicitly write a move constructor for a class similar to the Msg struct above:
D’oh! So now I was confused – and determined to figure out was was going on. Here is the program that I wrote to not only verify that manually written “move” operations are not required for the Msg struct, but to also measure the performance difference between moving and copying:
First, the program built cleanly as expected because the compiler provided the free “move” operations for the Msg struct. Second, the following, 5-run, output results proved that the compiler did indeed perform the deep, under the covers, “move” that my man Herb promised it would do. If the deep move wasn’t executed, there would have been no noticeable difference in performance between the move and copy operations.
From the eye-popping performance difference shown in the results, we should conclude that it’s time to start replacing copy operations in our code with “move” operations wherever it makes sense. The only thing to watch out for when moving objects from one place to another is that in the scope of the code that performs the move, the internal state of the moved-from object is not needed or used by the code following the move. The following code snippet, which prints out 0, highlights this behavior.
Check out this snippet from one of Bjarne Stroustrup’s CppCon keynote slides:
Now take a glance at one of Herb Sutter’s CppCon talk slides:
Bjarne’s talk was titled “Make Simple Tasks Simple” and Herb’s talk was tiled “Back To The Basics!: Modern C++ Style“. Since I abhor unessential complexity, I absolutely love the fact that these two dedicated gentlemen are spearheading the effort to evolve C++ in two directions simultaneously: increasing both expert-friendliness AND novice-friendliness.
By counterbalancing the introduction of advanced features like variadic templates and forwarding references with simpler features like range-for loops, nullptr, and brace-initialization, I think Bjarne and Herb (and perhaps other community members I don’t know about) are marvelously succeeding at the monumental task of herding cats. To understand what I mean, take a look at another one of Herb’s slides:
Do you see that teeny tiny dot at the end of the big arrow down on the lower right edge of the circle? Well, I don’t come anywhere close to qualifying for membership with the cats inside that dot… and I’d speculate that most advanced feature proposals and idiom ideas, whether they are understandable/teachable to mere mortals or not, originate from the really smart cats within that dot.
By gently but doggedly communicating the need for lowering the barriers to entry for potentially new C++ users while still navigating the language forward into unchartered waters, I’m grateful to Herb and Bjarne. Because of these men, the ISO C++ committee actually works – and it is indeed amazing for any committee to “work“.
The addition, err, redefinition of the auto keyword in C++11 was a great move to reduce code verbosity during the definition of local variables:
In addition to this convenient usage, employing auto in conjunction with the new (and initially weird) function-trailing-return-type syntax is useful for defining function templates that manipulate multiple parameterized types (see the third entry in the list of function definitions below).
In the upcoming C++14 standard, auto will also become useful for defining normal, run-of-the-mill, non-template functions. As the fourth entry below illustrates, we’ll be able to use auto in function definitions without having to use the funky function-trailing-return-type syntax (see the useless, but valid, second entry in the list).
For a more in depth treatment of C++11’s automatic type deduction capability, check out Herb Sutter’s masterful post on the new AAA (Almost Always Auto) idiom.
Since I have the privilege of using C++11/14 on my current project, I’ve been using the new language idioms as fast as I can discover and learn them. For example, instead of writing risky, exception-unsafe, naked “new“, code like this:
I’ve been writing code like this instead:
By using std::unique_ptr instead of a naked pointer, I don’t have to veer away from the local code I’m writing to write matching delete statements in destructors or in catch() exception clauses to prevent inadvertent memory leaks.
I could’ve used a std::shared_ptr (which can be copied instead of “moved“) in place of the std::unique_ptr, but std::shared_ptr is required to maintain a fatter internal state in the form of strong and weak owner counters. Unless I really need shared ownership of a dynamically allocated object, which I haven’t so far, I stick to the slimmer and more performant std::unique_ptr.
When I first wrote the std::unique_ptr code above, I was concerned that using the std::move() function to transfer encapsulated memory ownership into the safeTgtList vector would add some runtime overhead to the code (relative to the C++98/03 style of simply copying the naked pointer into the scaryTgtList vector). It is, after all, a function, so I thought it must insert some code into my own code.
However, after digging a little deeper into my concern, I discovered (via Stroustrup, Sutter, and Meyers) that std::move() adds zero runtime overhead to the code. Its use is equivalent to performing a static_cast on its argument – which is evaluated at compile time.
As Scott Meyers stated at GoingNative13, std::move() doesn’t really move anything. It simply prepares for a subsequent real move by casting its argument from an lvalue to an rvalue – which is required for movement of an object’s innards. In the previous code, the move is actually performed within the std::vector::emplace_back() function.
Quoting Scott Meyers: “think of std::move() as an rvalue_cast“. I’m not sure why the ISO C++ committee didn’t define a new rvalue_cast keyword instead of std::move() to drive home the point that no runtime overhead is imposed, but I’d speculate that the issue was debated. Perhaps they thought rvalue_cast was too technical a term for most users?
As I said early in the post, the code example is “like” the code I’ve been writing. The real code that triggered this post is as shown here:
Since each “entry” CfarDetState object must be uniquely intialized form it’s associated CfarCrossing object, I can’t simply insert estd::make_unique<CfarDetState>() into the emplace_back() function. All of the members of each “entry” must be initialized first. Regardless of whether I use emplace_back() or push_back(), std::move(entry) must be used as the argument of the chosen function.
Herb Sutter is by far the biggest cheerleader for the C++ programming language – even more so than the language’s soft spoken creator, Bjarne Stroustrup. Herb speaks with such enthusiasm and optimism that it’s infectious.
In his talk at the recently concluded GoingNative2013 C++ conference, Herb presented this slide to convey the structure of the ISO C++ Working Group 21 (WG21):
On the left, we have the language core and language evolution working groups. On the right, we have the standard library working group.
But wait! That was the organizational structure as of 18 months ago. As of now, we have this decomposition:
As you can see, there’s a lot of volunteer effort being applied to the evolution of the language – especially in the domain of libraries. In fact, most of the core language features on the left side exist to support the development of the upcoming libraries on the right.
In addition to the forthcoming minor 2014 release of C++, which adds a handful of new features and fixes some bugs/omissions from C++11, the next major release is slated for 2017. Of course, we won’t get all of the features and libraries enumerated on the slide, but the future looks bright for C++.
The biggest challenge for Herb et al will be to ensure the conceptual integrity of the language as a whole remains intact in spite of the ambitious growth plan. The faster the growth, the higher the chance of the wheels falling off the bus.
“The entire system also must have conceptual integrity, and that requires a system architect to design it all, from the top down.” – Fred Brooks
“Who advocates … for the product itself—its conceptual integrity, its efficiency, its economy, its robustness? Often, no one.” – Fred Brooks
I’m not a fan of committees in general, but in this specific case I’m confident that Herb, Bjarne, and their fellow WG21 friends can pull it off. I think they did a great job on C++11 and I think they’ll perform just as admirably in hatching future C++ releases.
Relatively recently, I participated in a debate with a peer regarding the sacredness of the Single Entry / Single Exit (SE/SE) “rule” of programming. I wish I had this eloquent Herb Sutter treatise on hand when it occurred:
Woot! Now that I’ve stashed the case against SE/SE nazi “enforcement” on this blawg, I’m armed and ready to confront the next brainwashed purist on the matter.
On my current project, we’re joyfully using C++11 to write our computationally-dense target processing software. We’ve found that std::shared_ptr and std::unique_ptr are extremely useful classes for avoiding dreaded memory leaks. However, I find it mildly irritating that there is no std::make_unique to complement std::make_shared. It’s great that std::make_unique will be included in the C++14 standard, but whenever we use a std::unique_ptr we gotta include a fugly “new” in our C++11 code until then:
But wait! I stumbled across this helpful Herb Sutter slide:
A variadic function template that uses perfect forwarding. It’s outta my league, but….. Whoo Hoo! I’m gonna add this sucker to our platform library and start using it ASAP.