Since moving to JIRA, jcheck doesn't distinguish between master bug IDs and backport IDs. Many changesets are getting pushed with incorrect bug IDs as a result.

This has consequences for bug records while the fix moves throughout different repositories towards the master repo.

One solution is for jcheck to maintain a list of all known "non-backport" type IDs on a public webserver. The list could be autogenerated from a cronjob run every 5-10 minutes. That list can be consulted by jcheck which can then reject changesets if the ID doesn't match one on the reference list.