Create a hash of the repo path and store it as its own result, allowing for a different cached result per repo path.
Use the existing result file if there is one.