Ahh, ok, so Httpd_Suspend may be the culprit - it may not be
the right thing to do before the fcopy. Originally I think we
just turned off the fileevents, but I switched to using Httpd_Suspend
after that was added. Thanks for looking at this. It may take me
a few days to dig in and work out the best solution.
One more very important question (perhaps you've said this) -
what version of Tcl are you running with? I'm still suprised
that you can trick fcopy into not making the callback.
>>>Jeff McWhirter said:
>
>
> Mike Hoegeman wrote:
>
> >this will happen if the server never tries to read/write the socket
>
> > after the client closes it
> > are you sure an fcopy is is in progress ?
> >
> > >
>
> Yes - I trace the suspend/fcopy call. I also am tracing the copydone call. I
have a
> test scriptthat opens a socket, sends a request and closes the socket. I see
the suspend
> and fcopy get called
> but the copydone never gets called. (The suspend removes the after timer eve
nt so the
> socket
> never gets cleaned up).
>
> More on the linux test tomorrow.
>
> -Jeff
>
>
>
-- Brent Welch <[EMAIL PROTECTED]>
http://www.ajubasolutions.com
Scriptics changes to Ajuba Solutions
scriptics.com => ajubasolutions.com
