Refactor detection of upstream rebuilds and build starts.

We're connecting to jenkins API in two different places in pift, both in pift_ci_job_ping_page() as well as drush_pift_cron_fetch_results() and gathering a list of jobs from the API, and attempting to detect whether a queued job has either started running, or if jenkins has failed somehow and kicked off a new job that corresponds to the same pift_ci_job id on drupal.org.

So, theres a few scenarios that would be good to make sure are handled properly here:

1. there can be multiple different drupal.org urls: drupal.org/staging.drupal.org/dev etc. This is embedded in the Drupal_JobID that gets sent to jenkins.
2. There can be multiple different dispatcher urls. Occasionally we'll rename dispatcher, or maybe set up a staging dispatcher, or have a legacy one.

We should refactor this check so that its only in one place, and handles these scenarios cleanly, as right now theres a sort of bug where it doesnt take into account that dispatcher urls can change.