Re: lynx-dev libreadline

[Vlad Harchev]

On Sat, 11 Dec 1999, Klaus Weide wrote:

> On Sat, 11 Dec 1999, brian j pardy wrote:
> 
> > Actually, I just downloaded and compiled readline-4.0.  It has an
> > option to side-scroll as opposed to wrap (set horizontal-scroll-mode
> > in .inputrc, assume can be done in the source).
> > 
> > Also -- I'm not proposing completely switching over to readline, just
> > the possibility of a compile-time choice.  From a quick skim of the
> > source, I think it might be possible to #ifdef out a bit of the edit
> > handling code, if someone wants to use readline.  The only area where
> > I think it could possibly be used is when there's a simple line
> > prompt, such as for save file name, or when composing mail.  I'm not
> > sure it could be manipulated into handling form fields or anything,
> > nor if that would even be desirable.
> > 
> > As for the benefit, that would be wholly subjective.  I like using vi
> > commandline editing, with a bunch of additional keybindings for the
> > emacs control keys.  Using readline would allow users to set bindings
> > only in .inputrc, and have all readline-using apps follow that.
> 
> So in some situations input would act one way (and be controlled in one
> way), and in other situations input would act in another way (and be
> controlled in another way - .inputrc).  I just think this isn't very
> useful, to most people.  That shouldn't prevent you from playing around
> with it, of course.  (Any chance you'll do something wrt cookie saving
> before 2.8.3 ? :))

  May be libreadline should be hacked to allow specifying different config
file to read (like adding readline_init(char* fname) - this could be useful
for other projects too - I mean include this hack in the standard
libreadline distro)?
  
 Also, I guess it's impossible to have multiplie history buffers in
libreadline. This could be a limitation. 
 Also ut8 input could cause problems (but they could be solved by improving 
libreadline itself).

 Best regards,
  -Vlad