[Top][All Lists]

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

Re: dabbrev-hover.el v. 0.1

From: D. Goel
Subject: Re: dabbrev-hover.el v. 0.1
Date: Mon, 24 May 2004 16:05:47 -0400
User-agent: Gnus/5.1002 (Gnus v5.10.2) Emacs/21.2 (gnu/linux)

Vagn Johansen <address@hidden> writes:

>     (defun vj-tooltip-show (str)
>         (interactive)
>         (setq header-line-format (replace-regexp-in-string "\n" ", " str))) 

> Maybe you could detect if tooltip is possible and the fallback to my
> function.  

Thanks, I will do that. 

> It would be nice with some cleanup so that header-line-format is set
> to "" when there are no dynamic expansions possible.
> Is it suppossed to show only one suggestion? I would like to use
> dabbrev-expand with a prefix argument. This would make the mode even
> better.
> I also think i would prefer if it did not fiddle with the RET keybinding
> (customizable?).

yes, it was customizable (using the way Kevin Rodgers suggested), but
I ran into the above RET problem you mentioned myself, and made TAB
the default in 0.2.

> Being in effect while in the minibuffer with the RET keybinding is a
> bad recipe.

> Is it needed at all in the minibuffer?

Hm, on my emacs 21.2 and 21.3, it never seems to get activated in the
minibuffer in the first place for some reason, but i will go ahead and
provide something like an "expand-in-minibuffer-p" option, and disable
it by default.

> I wonder if dh-complete *must* be used. dabbrev-expand seems work to just
> as well and I already have it bound to a convenient key.

Ah, I don't understand dabbrev.el too well, so I was not sure if
dabbrev-expand will expand to the same completion as shown by
dabbrev-hover.el (which, of course, dh-complete does, by
construction.)  dh-complete also makes the code general enough to
allow use with other mechanisms like hippie-expand, as suggested by
Trey Jackson privately.

> And finally, correct instances of '(keymap) with (make-sparse-keymap).
> Kevin Rodgers

thanks, done in 0.2dev already, with help from Uwe Brauer.
> If you don't need the tooltip feature in the minibuffer, then why the
> need to redefine tooltip-show? The default implementation should use
> the minibuffer if tooltips are not supported.

It does :)

Vagn Johansen <address@hidden> writes:

> You are right. A "(require 'tooltip)" was missing in dabbrev-hover.el. It
> is odd that there was no missing function error (it failed silently).

Will add.

> Now it works in the sense that instead of showing a "real" tooltip it
> prints "Error while displaying tooltip: (void-function x-show-tip)" in
> the minibuffer for half a second and then shows the possible expansion
> (also in the minibuffer).

One thing I don't understand is why fancy dh doesn't work when in -nw
mode (emacs -nw).  Everything in the code/edebug suggests that we are
in the right mode (dh-fancy-doing-mode), yet it doesn't work..  even
when I do the following for debugging:

 (define-key dh-fancy-doing-mode-map (kbd "t") 'dh-complete)
 (setq dh-fancy-doing-mode-string " DEBUG")

and then try "t" for completion.  
I would be in the dh-fancy-doing-mode, yet t will
 self-insert... strange..
If anyone knows, please do share..

(manual M-x dh-complete still works)


reply via email to

[Prev in Thread] Current Thread [Next in Thread]