Page MenuHomeSoftware Heritage

List the revisions added between two subsequent visits of an origin
Closed, MigratedEdits Locked

Description

We need this to populate the cache_revision_origin table of the content provenance information database.

Of course, a first (easy) step is doing the addition for the first visit.

Some interesting cases arise when we consider what to do between two subsequent visits:

  • A new branch appears. Do we want to register all the history as belonging to the new branch, or just the new revisions?
  • A branch disappears. Do we record that some merged revisions appeared in the destination branch?
  • Or do we keep things simple and only ever record revisions that we haven't recorded before in the history of the given origin?

The choice boils down to either:

  • computing the differences branch per branch
  • computing the differences globally