Re: #compdef tag and menu-select

[Sven Wischnowsky <wischnow@informatik.hu-berlin.de>]

Bart Schaefer wrote:

> On May 30, 10:16am, Sven Wischnowsky wrote:
> } 
> } Bart Schaefer wrote:
> } 
> } > At other places in the completion system we automatically load complist when
> } > it seems wanted.  Perhaps compdef should too?
> } 
> } Nowadays I'm a bit reluctant to change that because I've too often
> } thought about removing menu-select -- after all, it's just
> } menu-complete with enforcing selection, which can be achieved by other 
> } means.
> 
> Really?  What other means?  Specifically, I have a widget that is entirely
> independent of the completion system (except that it's created with zle -C
> and uses compadd; but it has nothing to do with _main_complete) and *for
> that widget only* I want menu-select behavior *all the time*.
> 
> It's convenient to put `#compdef -k menu-select ^X:' at the top of the
> file that defines this widget and let compinstall bind it for me, even
> though it's otherwise not part of the completion system.

Yes.

> What can I put *in the definition file* for this widget that will enable
> menu selection "by other means"?  I tried everything I could think of,
> poking into various values of compstate and assigning to MENUSELECT and
> whatnot, and I couldn't force menu-selection to start if the widget was
> e.g. bound with `#compdef -k menu-complete ^X:'.

  foo() {
    compadd ...
    MENUSELECT=0
    compstate[insert]=menu
  }
  zle -C foo complete-word foo
  bindkey '...' foo

Works for me.

> } And there is the question if menu-select should be re-bound to
> } _main_complete when compdef finds a widget using menu-select.
> 
> In the situation described above I definitely do not want that.

Err, I meant the automatic re-binding compinit already does for all
the completion widgets (including menu-select if complist is already
loaded), at line 452. So if compdef can make complist be loaded
automatically, should it do the same for menu-select if that isn't
re-bound already? Probably not very important...


Should I apply the patch? Shouldn't be able to do any harm, should it?

Bye
 Sven

Index: Completion/Core/compinit
===================================================================
RCS file: /cvsroot/zsh/zsh/Completion/Core/compinit,v
retrieving revision 1.3
diff -u -r1.3 compinit
--- Completion/Core/compinit	2000/05/25 14:48:38	1.3
+++ Completion/Core/compinit	2000/05/31 06:49:03
@@ -232,6 +232,7 @@
 	fi
 	[[ $1 = _* ]] || 1="_$1"
 	[[ $2 = .* ]] || 2=".$2"
+        [[ $2 = .menu-select ]] && zmodload -i zsh/complist
 	zle -C "$1" "$2" "$func"
 	if [[ -n $new ]]; then
 	  bindkey "$3" | read -A opt
@@ -250,8 +251,10 @@
 
       # Define the widget.
       if [[ $1 = .* ]]; then
+        [[ $1 = .menu-select ]] && zmodload -i zsh/complist
 	zle -C "$func" "$1" "$func"
       else
+        [[ $1 = menu-select ]] && zmodload -i zsh/complist
 	zle -C "$func" ".$1" "$func"
       fi
       shift

--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de