On some aspects of Fisher's exact test and its extensions