With the help of some debugging code, I've confirmed that the patch committed to the branch results in a correct (and speedy) exit in the case where a SIGTERM is received before a worker thread has accepted an event for processing. Will push this to the oneiric queue shortly.
With the help of some debugging code, I've confirmed that the patch committed to the branch results in a correct (and speedy) exit in the case where a SIGTERM is received before a worker thread has accepted an event for processing. Will push this to the oneiric queue shortly.