Re: [Libevent-users] coverity and event_new/bufferevent_socket_new

[Nick Mathewson <nickm@xxxxxxxxxxxxx>]

On Fri, Nov 22, 2013 at 5:52 PM, Vincent Bernat <bernat@xxxxxxxx> wrote:
> Hi!
>
> Using coverity with libevent triggers a resource leaks for socket
> descriptors registered to event_new or bufferevent_socket_new. To my
> understanding, coverity can be taught globally for such cases.
>
> Since I have seen some coverity-related bugfixes in the changelog, I
> wonder if something has already been tried on this front?

Can you say more about the warnings in question and why they occur?  I
maintain at least one application that uses Libevent and that's
scanned by Coverity, and I don't believe we've seen that kind of
warning.

I know that ntp also uses Libevent and Coverity Scan together, and so
far as I know event_new and/or bufferevent_socket_new hasn't been an
issue there. (I hope Harlan will let me know otherwise if I'm wrong.)

(I'm not saying you're mistaken about the warnings -- only that AFAICT
they don't seem to be an inevitable consequence of using event_new(),
and that I need more info to know what the trouble could be.)


yrs,
-- 
Nick
***********************************************************************
To unsubscribe, send an e-mail to majordomo@xxxxxxxxxxxxx with
unsubscribe libevent-users    in the body.