Fix: Huge delay posting myriad error messages on run abort (e.g. overdeterminations) [5534]
Bug Number: 5534
Release notes (y/n): Yes
For Release Nums: 6.7

Gnats 5534: Program acted as in endless loop, forced shutdown. [Gary Palhegyi, EDMUD]

Normally, when a run is aborted, only a single rwError is posted.  But in this case of many overdetermination errors, an rwError is posted for every series slot for every timestep having an overdetermination -- about 32,000 errors in the case of the bug model.  Each posting of an rwError releases all queued diagnostic messages for display, which is somewhat expensive.  This mechanism hadn't anticipated that there would be many rwErrors posted at once.

This has been modified such that the existing QTimer used in that mechanicm to throttle low-priority messages (warnings) is now also used to defer _subsequent_ rwErrors.  The first one -- generally received when the QTimer isn't current scheduled -- releases all messages right away.  But the QTimer is then also scheduled, and when subsequent rwErrors are received, they do not release queued messages (because the timer has been scheduled).

See also notes in the Gnats 5534 record.