If malformatted fastq, uLTRA silently aligns only a subset of reads

[ksahlin]

Originally reported as arising from issue #2. If the length of the quality value fields differs from the length of the sequence, uLTRA will align only a subset of the reads without reporting an error of a malformatted fastq-file.

It is, unfortunately, non-trivial to fix this (or even to output a warning or halt the program) with the current fq-parser, which runs much faster than other parsers. We could consider reverting to a slower parser that implements such checks.