Sender:	nate AT cartsys DOT com
Message-ID:	<36B79E0F.E9023D53@cartsys.com>
Date:	Tue, 02 Feb 1999 16:53:35 -0800
From:	Nate Eldredge <nate AT cartsys DOT com>
X-Mailer:	Mozilla 4.08 [en] (X11; I; Linux 2.0.36 i586)
MIME-Version:	1.0
To:	djgpp AT delorie DOT com
Subject:	Re: Clarification re: signals
References:	<3 DOT 0 DOT 6 DOT 32 DOT 19990202002923 DOT 008bd370 AT pop DOT netaddress DOT com>
	<3 DOT 0 DOT 6 DOT 32 DOT 19990202002923 DOT 008bd370 AT pop DOT netaddress DOT com> <3 DOT 0 DOT 6 DOT 32 DOT 19990202112714 DOT 008bf210 AT pop DOT netaddress DOT com>
Reply-To:	djgpp AT delorie DOT com

Paul Derbyshire wrote:
> 
> At 10:46 AM 2/2/99 -0500, you wrote:
> 
> >No, SIGKILL is the only signal a process can't catch, so it's the
> >"kill of last resort" to try to end a process.  SIGKILL can only be
> >sent, not trapped.  The kill() function can send *any* signal to a
> >process.
> 
> A parent can use WIFSIGNALED and WTERMSIG to spot a SIGKILL kill of a child
> though right?
> 
> >> SIGQUIT    -- ???
> >
> >Not sure about this one, but I think it can sometimes be generated
> >from the keyboard under Unix.
> 
> Other sources have informed me that this is true, and that it's sometimes
> used in interpreters where you want SIGINT to be a user interrupt of the
> interpreted code, and SIGQUIT can be used for a kind of "instant
> breakpoint" to drop into the debugger. Sounds nice. I remember debugging
> tons of QBASIC and having to add keypress checks for some key everywhere to
> issue "STOP", which in QBASIC drops into the debugger for inspecting or
> modifying things, resuming, or quitting.

Also, under normal cirucmstances, it typically dumps core, so if your
program is misbehaving, you can hit C-\ and then debug it postmortem. 
On DJGPP it dumps a traceback, IIRC.  SIGINT (Ctrl-C) is intended for
when you just get sick of running it.
-- 

Nate Eldredge
nate AT cartsys DOT com

- Raw text -