Best practices for returning objects within the c ++ 11 function using move Semantics

Best practices for returning objects within the c ++ 11 function using move Semantics

In one sentence, return directly without any changes.

After the c ++ 11 option is enabled, the Code returned through the function has not changed, but the move semantics has been used without the previous NRVO Compiler optimization technology.

Note: The right-side reference rvalue reference is a temporary variable that no longer exists after expression computation. The left-side value is a variable after expression computation. If you can use & to find the address, it is the left value.

 

The following is a discussion post on stackoverflow, which is more valuable:

 

246 down voteaccepted

First example std::vector return_vector(void){ std::vector tmp {1,2,3,4,5}; return tmp;}std::vector &&rval_ref = return_vector(); The first example returns a temporary which is caughtrval_ref. That temporary will have its life extended beyondrval_refDefinition and you can use it as if you had caught it by value. This is very similar to the following: const std::vector & rval_ref = return_vector(); Could t that in my rewrite you obviusly can't use rval_ref in a non-const manner. Second example std::vector && return_vector(void){ std::vector tmp {1,2,3,4,5}; return std::move(tmp);}std::vector &&rval_ref = return_vector(); In the second example you have created a run time error. rval_ref now holds a reference to the destructed tmp inside the function. With any luck, this code wowould immediately crash. Third example std::vector return_vector(void){ std::vector tmp {1,2,3,4,5}; return std::move(tmp);}std::vector &&rval_ref = return_vector(); Your third example is roughly equivalent to your first. The std: move on tmp is unnecessary and can actually be a performance pessimization as it will inhibit return value optimization. The best way to code what you're doing is: Best practice std::vector return_vector(void){ std::vector tmp {1,2,3,4,5}; return tmp;}std::vector rval_ref = return_vector(); I. e. just as you wocould in C ++ 03. tmp is implicitly treated as an rvalue in the return statement. it will either be returned via return-value-optimization (no copy, no move), or if the compiler decides it can not perform RVO, then it will use vector's move constructor to do the return. only if RVO is not saved med, and if the returned type did not have a move constructor wocould the copy constructor be used for the return. Shareeditflag Answered Feb 13 '11 at 20: 52 Howard Hinnant 74.1k12145133

6       Great thanks for the elaborated answer! -Tarantula Feb 14' 11 10       Compilers will RVO when you return a local object by value, and the type of the local and the return of the function are the same, and neither is cv-qualified (don't return const types ). stay away from returning with the condition (:?) Statement as it can inhibit RVO. Don't wrap the local in some other function that returns a reference to the local. Justreturn my_local;. Multiple return statements are OK and will not inhibit RVO.-Howard Hinnant Feb 25 '13 3       There is a caveat: when returningMemberOf a local object, the move must be explicit.-boycy Feb 26 '13 4       @ NoSenseEtAl: There is no temporary created on the return line.moveDoesn't create a temporary. it casts an lvalue to an xvalue, making no copies, creating nothing, destroying nothing. that example is the exact same situation as if you returned by lvalue-reference and removedmoveFrom the return line: Either way you 've got a dangling reference to a local variable inside the function and which has been destructed.-Howard Hinnant Feb 27 '13 at 11 1 Upvote   Flag Just a nit: Since youNamedThe variable (tmp) In the "Best practice" section, it is the NRVO that kicks in, not the RVO. These are two different optimizations. Other than that, great answer! -Daniel Frey Feb

 

When this is the case, I returned a vector in a program a few days ago. , Encountered core dump, replaced with vector The & output parameter method is solved. We also need to further track whether it is a clang ++ 3.5 bug.