On Thu, 7 Jun 2001, Greg Stein wrote:
> On Thu, Jun 07, 2001 at 01:04:35PM -0400, Jeff Trawick wrote:
> >...
> > > @@ -479,12 +477,9 @@
> > >
> > > /* we want to ignore HUPs and WINCH while we're busy processing one */
> > > sigaddset(&sa.sa_mask, SIGHUP);
> > > - sigaddset(&sa.sa_mask, SIGWINCH);
> > > sa.sa_handler = restart;
> > > ap_log_error(APLOG_MARK, APLOG_WARNING, errno, ap_server_conf,
>"sigaction(SIGHUP)");
> > > - if (sigaction(SIGWINCH, &sa, NULL) < 0)
> > > - ap_log_error(APLOG_MARK, APLOG_WARNING, errno, ap_server_conf,
> > "sigaction(SIGWINCH)");
> > ditto
> > > #else
> > > if (!one_process) {
> > > apr_signal(SIGSEGV, sig_coredump);
> > > @@ -513,7 +508,7 @@
> > > apr_signal(SIGHUP, restart);
> > > #endif /* SIGHUP */
> > > #ifdef SIGWINCH
> > > - apr_signal(SIGWINCH, restart);
> > > + apr_signal(SIGWINCH, SIG_IGN);
> >
> > I'm extremely confused. Don't we still use SIGWINCH vs. SIGHUP to
> > indicate the type of restart?
> >
> > Nothing is happening now in the parent for SIGWINCH (apachectl
> > graceful). These are the suspect changes.
>
> When I run httpd using:
>
> ./httpd -DONE_PROCESS -DNO_DETACH
>
> I find that I can no longer use ctrl-c to stop the thing. I've got to ctrl-z
> to suspend it, then kill it from the command line.
>
> Something is chewing up/ignoring the SIGQUIT. And pretty recently.
This has been happening for a while. For a long time, I haven't been able
to ctrl-c the server while it is busy serving a request.
Ryan
_______________________________________________________________________________
Ryan Bloom [EMAIL PROTECTED]
406 29th St.
San Francisco, CA 94131
-------------------------------------------------------------------------------
