Re: Finally, some info about the zombie child

[Jim Jagielski <ji...@jaguNET.com>]

Good point.

Roy T. Fielding wrote:
> 
> >Here's a "clue"... the child process _does_ seem to be stuck in the select()
> >call in ap_bhalfduplex(). Sooooo either the select is blocking or
> >what's I'm really seeing is the looping inside.
> >
> >Ideas?
> 
> Well, for one thing, ap_bhalfduplex isn't checking to see if the
> socket is still valid.  It should be doing
> 
>     if (fb == NULL || fb->fd_in < 0 || fb->incnt > 0 || fb->outcnt == 0) {
>         return;
>     }
> 
> I think (not having compiled this snippet).
> 
> ....Roy
> 


-- 
===========================================================================
   Jim Jagielski   |||   jim@jaguNET.com   |||   http://www.jaguNET.com/
            "That's no ordinary rabbit... that's the most foul,
            cruel and bad-tempered rodent you ever laid eyes on"