Good point. I added

/* `dst` has room for strlen(src) - 1 bytes */

in the C setup, and allocated dst like so:

const size_t mem_size = strlen(input) - 1;
char *user_string = memset(malloc(mem_size), '@', mem_size);

I did not fix NASM because it uses a completely different test style than C, the output buffers have a fixed length and are embedded inside a struct.

i removed the JavaScript code snippet from the description

the C code template/boilerplate is not match with the function signature in the description, it makes people confused.

not a kata suggestion

OP solved the kata, not enough code shared to conclude. most likely, dst was reassigned to another memory location somewhere in the function body

Yes, you're right, then i rewrite my solution

OP solved it, closing!

Yes, you're right, then i rewrite my solution. Thanks for help

No, it is not "in an absolutely random time". It is always. For "abcb", your solution always return "cb" instead of "bc". Your solution has a bug, and handles some inputs incorrectly. You can also check "abc abc".

But in some moments, in an absolutly random time it may be s='cb'

You need to check exactly, because for s="abcb", your solution does not return "bc" (as it should), but something else.

What will your solution return for string s="abcb"?

It is very difficult to tell why your code is wrong without seeing your code, or even without knowing what language you are trying to solve.

Hello everyone, I have a problem, i don't know why, but with true solution, each time, when i try to answer, randow number of tests be wrong.
Like 25 true answer, 21 wrong OR 40 true, 6 wrong.
And symbols in string replaced in random place, first and last places removed, but in some tests, like "khgjhbkjgjhb" answer is "hghhbkjgjj" but true answer is "hgjhbkjgjh".
What's a problem?