Fwiw I've never been really satisfied with how the dumped sequences look like, it is often hard to tell the actual sequence vs the expected sequence from looking at the diff-like output dump_sequence prints.
I think maybe the code should not be so smart and instead print the full expected / actual sequences (or maybe only the actual sequence if it's too long) and let us decide how to match it. This is just a thought, I'm not completely sure about it either.