[Date Prev][Date Next][Thread Prev][][Date Index][Thread Index]

Re: Feature inquiries

>>>>> E Oltmanns writes:

 > > Note that with the latest code, available at:
 > >
 > >     https://alioth.debian.org/project/showfiles.php?group_id=30594
 > >
 > > this would be (w3m-type-ahead-mode 1), and this automatically
 > > sets up a minor-mode map with / bound to w3m-type-ahead and M-/
 > > bound to w3m-type-ahead-regexp.  To create a new session, use a
 > > prefix argument of C-u 2 or C-u C-u.  (Information at
 > > www.emacswiki.org updated.)

 > Nice to hear from you straight away. That gives me the opportunity to
 > ask you something regarding your package.
 > Would it be possible to implement w3m-type-ahead-internal in a way
 > such that the cursor doesn't go to the minibuffer but stays in the w3m
 > window and indicates the current match - just like the original
 > I-Search does? Is there a particular reason that this behaviour hasn't
 > been adopted from I-Search? I'm just asking because I'm visually
 > impaired and therefore using a braille terminal. Technically it is
 > much easier to bind the teerminal to the system cursor than a
 > particular soft cursor indicated by yet another colour. Since the
 > terminal displays only half a line of the screen, I'd rather like it
 > to show me the current match than the minibuffer.
 > However, I recognise, of course, that I'm certainly in a minority and
 > if there are good reasons for the current behaviour, I will certainly
 > accept them. In this case I would be very grateful if someone could
 > give me a hint how I can patch my local version to adjust it to my
 > needs - I'm not really expreienced in lisp programming and therefore
 > don't even know where to start.

The way the implementation works at the moment is that the search
occurs in a different buffer, and the minibuffer window is the
selected window; hence the problem with the cursor.  It may be
possible to implement it in a way consistent with what you want.  I
will have a think about it.

 > By the way, wouldn't it be a good idea to add something like
 > (eval-when-compile (require 'cl))

 > since ordinary (debian) user like me don't generally have cl loaded
 > once they start w3m?

Thanks for pointing this out.