Re: 3 readline bugs: source positions, readline-options, endless loop on error

[Martin Grabmueller]

> Date: Thu, 25 Jan 2001 11:43:23 +0100 (MET)
> From: Dirk Herrmann <address@hidden>
> 
> > Two possible fixes came to my mind:
> > 
> > - Check the argument in `set-readline-input-port!' and only accept
> >   valid open input ports.
> > - When the argument check in `%readline' fails, reset the current
> >   input port to the standard input file port.
> > 
> > I think the first is preferable.
> 
> That's what I think, too.  Could you provide a patch?

I modified the `set-readline-*put-port!' procedures (see below), but
got two questions:

- Is scm-error the right way to report the error?  The output is
    guile> (set-readline-output-port! #t)
    ERROR: In procedure set-readline-output-port!:
    ERROR: Not an open output port: #t
    ABORT: (wrong-type-arg)
  , which seems okay to me.

- How can I test whether a given port is a file port?  The problem is
  that my patch works when you pass a stupid value like #t to the
  procedures, but not for string ports, for example. They lead to the
  same endless loop like before the patch.

(define-public (set-readline-input-port! p)
  (if (and (input-port? p)
           (not (port-closed? p)))
      (set! input-port p)
      (scm-error 'wrong-type-arg "set-readline-input-port!"
                 "Not an open input port: ~S"
                 (list p)
                 #f)))

(define-public (set-readline-output-port! p)
  (if (and (output-port? p)
           (not (port-closed? p)))
      (set! output-port p)
      (scm-error 'wrong-type-arg "set-readline-output-port!"
                 "Not an open output port: ~S"
                 (list p)
                 #f)))
Regards,
  'Martin
-- 
Martin Grabmueller              address@hidden
http://www.pintus.de/mgrabmue/  address@hidden on EFnet