Semi-related, but maybe less ambitious: wouldn't it be nice if each jira ticket has or could have a corresponding unit test that is expected to fail, with known (bad) output, and (un)expected good output?
That could maybe enable a lot of goodies. From the top of my head, with CI, that could be used to automatically determine affected versions, and detect if some change interacts with some bug, either by unexpectedly producing the good output (yay!) or by producing output different from both the known bad output or the good output, possibly giving a hint to where the bug lives. For compiler crashes with stack traces, that could enable a search tool/site.