Date: Mon, 8 Dec 1997 16:59:54 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: djgpp-workers AT delorie DOT com
cc: DJ Delorie <dj AT delorie DOT com>, Charles Sandmann <sandmann AT clio DOT rice DOT edu>
Subject: Enhanced signal support: questions
Message-ID: <Pine.SUN.3.91.971208165845.4200t-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk

There are several aspects of the SIGQUIT signal and SIGINT/SIGQUIT key
redefinition that I'd like to discuss.  Please send your comments.

     1. Is there any magic in the ``fake'' exception numbers that
        exceptn.S invents for Crl-C and SIGFPE?  I know that 0x75 is
        the math coprocessor exception, but is anything wrong in
        reusing these numbers?  For example, 0x7a is the low-level
        Novell API interrupt; could anything bad happen by using 0x7a
        to fake the SIGQUIT exception?

     2. If a program redefines its INTR key and then invokes a child
        program, Ctrl-C will generate SIGINT in the child but the
        parent won't get SIGINT.  Is this OK (I suppose it's different
        on Unix)?

     3. I don't know enough about the PC98 variant to rewrite its
        code.  Can somebody tell me what to do to make PC98 support
        this feature as well?

     4. Would the default behavior to have SIGQUIT generated by Ctrl-\
        surprise people?  If so, maybe turning it off by default is
        better?