Make `sync` incremental by storing last scanned git hash

[shonfeder]

Followup to #17

For large repos, it will take time to process all files (since a tag could be in pretty much any source file).

But we can use the git history to optimize:

• Store the git hash at the time a repo is added.
• Whenever we sync, check the whether the repo hash has changed (after fetching the latest version of the default branch)
  • If not, we have nothing to do
  • If so, just get a list of the files that have changed between commits, and only scan those for changes (perhaps further optimizing with #46).