Skip duplicate pages based on digest

Currently caching is done by timestamp only. A later version of a page is always downloaded, however if it has the same digest we should just be able to skip it instead.

Start by tracing where timestamp-based caching decides whether a later page is downloaded, then inspect how page digests are obtained or compared. The change is complete when a later page with the same digest is skipped while a changed digest is still downloaded; no file or test location is specified in the issue.

Written by the indexing model from the issue text.