git harvester should default to HEAD of master

harvester should default to the HEAD of the master branch instead of defaulting to "null" which doesn't work.

this satisfies the use case where you don't care about crawling a specific ref, but just the repo.

The issue identifies the git harvester as the entry point but names no file or test. Locate how an unspecified ref is handled, then verify that the default resolves to the HEAD of the master branch and allows crawling the repository without a specific ref.

Written by the indexing model from the issue text.