Comment 2 for bug 1672434

To summarize the idea I mentioned during the hackfest, we could consider adjusting the test bib data so that biblio.record_entry.last_xact_id is used (abused?) as a stable identifier for sample bib record. Each distinct file of bib records could have last_xact_id set to a per-file prefix + a sequentially-assigned number; e.g., 'concerto-1', 'concert-2'. last_xact_id would then be used for the purpose of

- linking specific test copies and volumes to specific bibs, allowing the biblio.record_entry.id values to float
- generating copy barcodes