Never mind - RTFD plus some experimentation answered the question. Interesting how it ran for years before finally exposing the problem.
On Feb 25, 2012, at 10:53 AM, Ralph Castain wrote: > Hi folks > > I have a hopefully quick question. I've been chasing down memory corruption > issues, and think I'm coming to closure on the root cause. > > If I setup an event by malloc'ing an event_t, am I allowed to free that > memory in the event handler when called by libevent? Or does libevent still > need to do something with the event after the callback completes? > > In other words, can I do this: > > In code that sets up the event: > > struct event_t *ev; > > ev = (struct event_t*)malloc(sizeof(struct event_t); > … > trigger_event > > > > In event handler: > > struct event_t *ev = (struct event_t*)cbdata; > ….. > free(ev); > return > > *********************************************************************** To unsubscribe, send an e-mail to majord...@freehaven.net with unsubscribe libevent-users in the body.
