better error reporting, bug fixes, Markov shuffle, reduced RAM needs
Do a better job of reporting chase errors back to the user (in rally).
Fix some validation bugs, etc.
Add basic Markov database shuffle.
Memory usage now does not depend on the number of spectra being searched.
Likewise, filehandle usage does not depend on the number of spectrum files
being searched. (Validation memory is still linear in the number of spectra,
but this seems not to be a practical problem at the moment.)
Add --mass-error-histogram to validate.
Add greylag-flatten-fasta utility.