CDR Tickets

We submit the requests to prune the nodes we just published in batches, so we don't overwhelm poor Drupal with too many at once. Inside that loop we have a nested loop which tries again a certain number of times in the face of failures on the Drupal end so we can recover and keep going. Unfortunately, the offset is moved forward by the batch size inside both loops. So unless a publishing job were pretty small (10 summaries or fewer) we could be skipping the pruning for half the summaries published. Or more, in the uncommon case of failures on the Drupal end. A fairly benign bug, as these things go (though way up there on the silly meter), since in most cases a node which got skipped for the pruning in one job could catch up in the next job. The worst-case scenario for the bug is some nodes have incredibly back luck and never get pruned, and take up more space in the database than we want them to.