I submitted your patch for deallocing a base - there was no patch to the man page nor regress.c though :-)
Thank you, Niels. On 3/26/06, Nick Mathewson <[EMAIL PROTECTED]> wrote: > This patch introduces a new function, event_base_free, that releases > memory held by a libevent base. Freeing bases is useful for people > who want to use dmalloc (or similar tools) to verify that their > programs aren't leaking memory, for program subcomponents that wneed > to be loaded and unloaded, and so on. > > There are a couple potentially controversial aspects: > 1. It requires that all events be deleted from the base before it is > called. If not, an assertion is raised. (IMO, this is the only > sensible behavior, since otherwise we have no good way to free > the events' ev_arg field.) > > 2. It doesn't do anything useful about freeing resources held in > signal.c. (Refactoring signal.c to play nicely with multiple > bases seems like a separate process.) > > I've built it and checked it to be superficially vvalid on linux and > mac os 10.4; somebody should check out the devpoll.c and win32.c > changes. > > -- > Nick Mathewson > > > _______________________________________________ > Libevent-users mailing list > Libevent-users@monkey.org > http://monkey.org/mailman/listinfo/libevent-users > > > > _______________________________________________ Libevent-users mailing list Libevent-users@monkey.org http://monkey.org/mailman/listinfo/libevent-users
