ndpi_memmem: optimized, fixed bug, added tests

[pasabanov]

Please sign (check) the below before submitting the Pull Request:

• I have signed the ntop Contributor License Agreement at https://github.com/ntop/legal/blob/main/individual-contributor-licence-agreement.md
• I have read the contributing guide lines at https://github.com/ntop/nDPI/blob/dev/CONTRIBUTING.md
• I have updated the documentation (in doc/) to reflect the changes made (if applicable)

#2378 #2424 #2494

Describe changes:

Similar changes to this PR (#2494).

1. Made the function ndpi_memmem return haystack when needle_len == 0 instead of NULL. This is more logical and aligns with the default behavior for standard implementations, such as GNU libc and FreeBSD libc.
2. Optimized performance by replacing the variable haystack_end with the variable end_of_search, which reduces the number of checks during the search.
3. Added unit tests for the function ndpi_memmem.

[IvanNardi]

LGTM; I particularly appreciated the tests!

[IvanNardi]

Thank you

[IvanNardi]

@pasabanov, @0xA50C1A1: might it be possible to implement ndpi_strnstr() using ndpi_memmem()? They are basically the same function....

[0xA50C1A1]

@pasabanov, @0xA50C1A1: might it be possible to implement ndpi_strnstr() using ndpi_memmem()? They are basically the same function....

That's a great idea. I created a PR with a refactored function. #2500