[vimpulse] Bug: TAB doesn't autocomplete in minibuffer.

Fri Jul 2 17:57:51 CEST 2010

On Fri, Jul 02, 2010 at 04:55:04PM +0200, Vegard Øye wrote:
> On 2010-07-02 14:17, Alessandro Piras wrote:
> 
> > vimpulse-misc-keybindings:564
> > ---
> > (global-set-key [tab] (or (key-binding [tab]) (key-binding "\C-i")))
> > ---
> > I guess this doesn't take into account mode changes, i.e., it sets
> > the tab key and when the binding changes (when changing to a minor
> > or major mode that sets the tab key) the tab key still does what it
> > was set to with the global set key.
> 
> It depends. If the mode does indeed bind the Tab /key/ (i.e., [tab],
> as in (define-key map [tab] 'cmd), it will work as expected. But if
> the mode binds the Tab /character/ (i.e., C-i, also known as "\t"),
> only the C-i key is bound, because the code above breaks the bond
> between [tab] and C-i.
> 
> In the words of the Emacs manual (my emphasis):
> 
>     <TAB>, <RET>, <BS>, <LFD>, <ESC> and <DEL> started out as names
>     for certain ASCII control characters, used so often that they have
>     special keys of their own. For instance, <TAB> was another name
>     for C-i. [...] If you do not want to distinguish between (for
>     example) <TAB> and C-i, make just one binding, for the ASCII
>     character <TAB> (octal code 011). IF YOU DO WANT TO DISTINGUISH,
>     MAKE ONE BINDING FOR THIS ASCII CHARACTER, AND ANOTHER FOR THE
>     "FUNCTION KEY" TAB.

... with the the caveat that in the terminal (emacs -nw) there is no
such distinction (as the section you quote also mentions).

In any case, we should IMO most definitely NOT mess with global map at
all. Currently the only exception seems to be

vimpulse-misc-keybindings.el:580:
(global-set-key "\C-c\C-o" 'open-line)) ; some may miss this command

IIUC, it can be safely removed, as C-o is only rebound in
`viper-vi-basic-map'.

> 
> http://www.gnu.org/software/emacs/manual/html_node/emacs/Named-ASCII-Chars.html
> 
> Before the distinction is made, one can bind Tab by binding C-i.
> After the distinction is made, one can bind Tab only by binding [tab].
> 
> Minibuffer completion breaks because it binds Tab the first way.
> Commit [ee2d1f9527] fixes this.

Somewhat off-topic, I've found Viper's messing with minibuffer
intolerable, have put the following in my init file:

(remove-hook 'minibuffer-setup-hook 'viper-minibuffer-setup-sentinel)

... and live happily ever since ;-)

(I also set viper-vi-style-in-minibuffer to nil before, don't remember
if it still matters after removing the hook.)

Štěpán