An experiment of tail recursion

I was just. NET test, for the tail recursion, in debug mode, will not be optimized to non-recursive structure, in release mode, will be optimized to non-recursive structure, there is no stack overflow problem

Stst?

This is my simulation of the file structure class

Stst

?

This is a non-tailed recursive version

Stst

This is the tail-recursive version.

Stst

This is the test code.

Stst

This is the definition of createdir.

Stst

Results:

1, in debug mode, whether it is an ill-ending recursion, will overflow

2, in release mode, the non-tail recursive version will overflow, and the tail recursive version will not overflow

Stst

Modern compilers I think it's all there.

Stst

Because tail recursion is optimized for non-recursive versions, it's a very old technology.

Stst

You can test it under Java, I don't have a Java environment

Stst

Instead of changing your method parameter to the file array type, you can turn recursion into a real tail recursion.

Stst

There's no basis for this.

0xb000:do_ tail recursion (Files)

......

......

Stack:

0xa000:-> Parameters: Files

0xa004:-> return Address: 0xa000

0xa008:? Parameters: Files

0xa012:? --Return Address: 0xa004

0xa016:??? Parameters: Files

0xa020:??? --Return Address: 0xa012

............

The above is the recursive stack structure, the tail recursion has such a fact:

Is that each time recursion, its previous call has been calculated, no need to store intermediate results, so this call can fully use the previous call stack, so the stack structure does not extend downward

An experiment of tail recursion