[320] in bug-owl

home help back first fref pref prev next nref lref last post

Re: owl : zwrite.c patch

daemon@ATHENA.MIT.EDU (James M. Kretchmar)
Mon Mar 1 22:59:30 2004

Message-Id: <200403020359.WAA15489@mission-control.mit.edu>
To: Alex Rolfe <arolfe@MIT.EDU>
cc: "James M. Kretchmar" <kretch@MIT.EDU>, bug-owl@ktools.org
Reply-To: bug-owl@ktools.org
In-Reply-To: Your message of "Mon, 23 Feb 2004 22:02:44 EST."
             <ubzad39uqln.fsf@metacircular-evaluator.mit.edu> 
Date: Mon, 01 Mar 2004 22:59:12 -0500
From: "James M. Kretchmar" <kretch@MIT.EDU>

> I guess the other option is something gross to send zsig proc the
> recipients on stdin, but I think a variable to set is a better idea.  
> 
> I hadn't really worried about this since I run owl on my own machine to
> which few other people can login.  I also suspect the problem isn't too
> bad since zsigproc presumably runs relatively quickly so it'd be hard to
> catch it running in most cases.

The only other idea I'd thought of was a stdin hack as well, but a
conversation on zephyr came up with a better (and now obvious) idea:
have a perl function from the owlconf return the zsig.  The perl
function could be passed a (possibly incomplete) message object with
all the relevant data and more, would be secure, and most zsigproc's
are written in perl anyway.  And if you hate that you can always exec
out from the perl.

It's also a good opportunity to clean up some of the zsig variable
mess in Owl, which admittedly needs some fixing.

kretch

home help back first fref pref prev next nref lref last post