C ++ Lambda functions may be less efficient than imagined

In C ++ Lambda functions, where are the parameters stored when the function parameters are captured by means of replication?

I thought it was wrong in the stack.

With this question, write a recursive function and name the address of the parameter.

ReferenceCodeAs follows:

# Include <stdio. h> # Include <Functional> Void Fun2 (STD: function < Void ()> Callback) {(callback )();}  Void Fun1 ( Int  N ){  If (N <= 0 ) Return ; Printf (  "  Stack address = % P,  " ,& N); fun2 ([N] () {printf (  "  Capture address = % P \ n  " ,& N); fun1 (n - 1  );});}  Int  Main () {fun1 (  200  ); Return   0  ;} 

(Build Environment: mingw64, version x64-4.8.0-release-posix-seh-rev2)

The results were surprising. The output result is as follows. Note that the address of the capture parameter is notProgramStack region,

And the address is not always continuous.

Obviously, the replication capture parameters expressed by Lambda are dynamically allocated and stored.

Be careful when paying special attention to efficiency or dynamic allocation. Reality is very skinny.

Stack address = 000000000022f1e0, capture address = Destination Address = 000000000022f0c0, capture address = Destination Address = 000000000022efa0, capture address = Destination Address = 000000000022ee80, capture address = Destination Address = 000000000022ed60, capture address = Destination Address = 000000000022ec40, capture address = Destination Address = 000000000022eb20, capture address = Destination Address = 000000000022ea00, capture address = Destination Address = destination, capture address = Destination Address = 000000000022e7c0, capture address = 00000000007a7840
 

(Reprint please mark: http://www.cnblogs.com/xhawk18)