Ignore .gzi index files and clean test output files

Created by: jmarshall

On a different topic from BioPerl — I'm glad that after five years they're getting on with splitting it up, but yeah living through the transition is not going to be fun 😄 — here is a small tidy up.

Most times when I run ./Build test I wind up with

Changes not staged for commit: modified: t/data/Saccharomyces_cerevisiae.R64-1-1.dna.chromosome.I.fa.gz.gzi

The .gzi index is generated alongside the .fai index when you faidx-index a compressed .fa file. (And it keeps changing in length for weird HTSlib reasons that are too boring to describe.) So it's appropriate not to track it in the repository; instead ignore it the same as .fai.

Also clean up the test output files (based on the list in .gitignore) when ./Build clean is used.