A matter of escape

Nick Burrett nick at dsvr.net
Tue Jan 15 02:06:41 PST 2002


"Schaefer" <sf at bonn.iz-soz.de> writes:

> > -----Original Message-----
> > From: Peter Naulls [mailto:peter at chocky.org]
> > Sent: Saturday, January 12, 2002 8:08 PM
> > To: gcc at hard-mofo.dsvr.net
> > Subject: A matter of escape
> > 
> > 
> > 
> > As you'll be well aware, pressing escape has long been a way of
> > terminating CLI programs, unless they take measures which otherwise
> > prevent it.  This includes programs built with Unixlib.
> > 
> > However, recent events have made this a problem.  Here's the deal:
> 
> [SNIP]
> 
> > The question is - what do to?  Unix uses Ctrl-C for the same 
> > meaning of
> > escape in RISC OS.  Unixlib might potentially do the same.  
> > Or it might
> > be chosen on a per-app basis.   Or we might be able to arrange for
> > canonical input to avoid playing with escape.
> > 
> > It's also worth pointing out that although we can tell an 
> > app's running
> > in a taskwindow, there's no sensible way to tell it's running under
> > Nettle.  Even if we choose to do nothing about this in UL - i.e, let
> > apps changing the escape key themselves, it's still very much an issue
> > worth being aware of.
> 
> I would opt for UnixLib interpreting Ctrl-C like Unix does and ignoring
> ESC. I oftenly try Ctrl-C first until I remember being on RISC OS and
> try ESC. I think it would be the most usable option for people being
> used to Unix like systems.

I'm against this, since we are writing a C Library that helps to port
Unix applications to the RISC OS environment and should therefore make
an attempt to turn the programs into something that RISC OS users will
be familiar with.

I think that Nettle should trap the escape, then later send an Escape
signal to whatever application it is running.


Nick.



More information about the gcc mailing list