PATCH: Redo _zle using _arguments

PATCH: Redo _zle using _arguments

[Bart Schaefer]

I started out to tweak one thing so that "zle -C" would complete properly,
and just got carried away.

This patch also fixes a couple of small documentation errors, related only
in that I saw them while re-reading the docs for zle and _arguments.

I also suspect there may be a bug in comparguments -- the '(*)-x' form of
optspec is supposed to mean that the '*:msg:act' form is not completed
when -x is on the command line, but that doesn't seem to happen (or at
least, not always).  I imagine a fix is going to have to wait a few weeks
for Sven to return, though.

Random other question: Why isn't list-expand a completion widget for the
purposes of "zle -C"?

Index: Completion/Builtins/_zle
===================================================================
@@ -1,10 +1,77 @@
 #compdef zle
 
-local expl
+local expl ret=1 st
+local -a opts compwids state
 
-if [[ "$words[2]" = -N && CURRENT -eq 3 ]]; then
-  _wanted -C -N functions expl 'widget shell function' \
-      compadd "$@" -k functions
-else
-  _wanted widgets expl widget compadd -k widgets
-fi
+compwids=(accept-and-menu-complete
+	  complete-word
+	  delete-char-or-list
+	  expand-or-complete
+	  expand-or-complete-prefix
+	  list-choices
+	  menu-complete
+	  menu-expand-or-complete
+	  reverse-menu-complete)
+
+opts=(-A -C -D -L -M -N -R -U -a -c -l \* :)
+
+_arguments -s \
+    "($opts)-A[define widget alias]:old widget:->widget :new widget:->widget" \
+    "($opts)-C[define completion widget]:new widget name:->comp-widget :completion widget:->builtin-comp-widget :widget shell function:->function" \
+    "($opts)-D[delete widget]:*:widget:->widget" \
+    "(${opts#-l})-L+[with -l, list as commands]:*:-:->listing" \
+    "(${opts#-[La]})-l+[list user-defined widgets]:*:-:->listing" \
+    "(${opts#-l})-a+[with -l, list all widgets]:*:-:->listing" \
+    "($opts)-M[display message]:message: " \
+    "($opts)-N[define new widget]:widget name:->widget-or-function ::widget shell function:->function" \
+    "(${opts#-c})-R+[redisplay]:*:-:->redisplay" \
+    "(${opts#-R})-c+[with -R, clear listing]:*:-:->redisplay" \
+    "($opts)-U[unget to input stack]:string: " \
+    '(-)::widget name:->call'
+
+[[ $state == listing ]] &&
+  _arguments -s \
+    "-l[list user-defined widgets]" \
+    "(-a)-L[list as commands]" \
+    "(-L)-a[list all widgets]" \
+    '(-)*:widget name:->widget'
+
+for st in $state; do
+  case $st in
+    (call)
+      if ((CURRENT > 2)); then
+	_arguments \
+          '(-N)-n[numeric prefix]:number: ' \
+	  '(-n)-N[reset numeric prefix]' \
+	  ':widget::' '(-)*:widget arguments: ' && ret=0
+	  # :widget:: is a placeholder so we needn't shift words
+	continue
+      fi
+      ;&
+    (widget*)
+      _wanted widgets expl widget compadd -k widgets && ret=0
+      [[ $st != *function ]] && continue
+      ;&
+    (function)
+      _wanted functions expl 'widget shell function' \
+	compadd -k functions && ret=0
+      ;;
+    (comp-widget)
+      _wanted widgets expl 'completion widget' \
+	compadd -k "widgets[(R)(*:|)(.|)(${(j(|))compwids})(|:*)]" && ret=0
+      ;&
+    (builtin-comp-widget)
+      _wanted widgets expl 'builtin completion widget' \
+	compadd -k "widgets[(I)(.|)(${(j(|))compwids})]" && ret=0
+      ;;
+    (redisplay)
+      _arguments -s \
+        "-R[redisplay]" \
+	"(*)-c[clear listing]" \
+	"(-)::status line: " "*:strings to list: " && ret=0
+      ;;
+    (*) ret=$?;;
+  esac
+done
+
+return ret
Index: Doc/Zsh/compwid.yo
===================================================================
@@ -640,7 +640,8 @@
 match, the var(n)'th element of the var(array) is removed.  Elements
 for which the corresponding var(word) is matched are retained.
 )
-item(tt(-), tt(--))(
+xitem(tt(-))
+item(tt(-)tt(-))(
 This flag ends the list of flags and options. All arguments after it
 will be taken as the words to use as matches even if they begin with
 hyphens.
Index: Doc/Zsh/mod_zle.yo
===================================================================
@@ -190,7 +190,7 @@
 cindex(calling widgets)
 cindex(widgets, defining)
 cindex(defining widgets)
-xitem(tt(zle) tt(-l) [ tt(-L) ] [ tt(-a) ] [ var(string) ... ])
+xitem(tt(zle) tt(-l) [ tt(-L) | tt(-a) ] [ var(string) ... ])
 xitem(tt(zle) tt(-D) var(widget) ...)
 xitem(tt(zle) tt(-A) var(old-widget) var(new-widget))
 xitem(tt(zle) tt(-N) var(widget) [ var(function) ])
@@ -204,7 +204,7 @@
 ZLE.  Which operation it performs depends on its options:
 
 startitem()
-item(tt(-l) [ tt(-L) ])(
+item(tt(-l) [ tt(-L) | tt(-a) ])(
 List all existing user-defined widgets.  If the tt(-L)
 option is used, list in the form of tt(zle)
 commands to create the widgets.
@@ -238,7 +238,7 @@
 ifnzman(noderef(Zsh Line Editor))\
 .
 )
-citem(completion widgets, creating)
+cindex(completion widgets, creating)
 item(tt(-C) var(widget) var(completion-widget) var(function))(
 Create a user-defined completion widget named var(widget). The 
 completion widget will behave like the built-in completion-widget

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net   

Re: PATCH: Redo _zle using _arguments

[Sven Wischnowsky]

Bart Schaefer wrote:

> On Oct 4,  2:47pm, Sven Wischnowsky wrote:
> } 
> } Bart Schaefer wrote:
> } 
> } > I also suspect there may be a bug in comparguments -- the '(*)-x' form of
> } > optspec is supposed to mean that the '*:msg:act' form is not completed
> } > when -x is on the command line, but that doesn't seem to happen (or at
> } > least, not always).
> } 
> } Hm, I had a look but couldn't find a bug. Could you give me an example?
> 
> Sure.  Completion after "zle -R -c":
> 
> zagzig[501] zle -R -c <TAB>
> Completing status line
> Completing strings to list
> 
> The call to _arguments was:
> 
>       _arguments -s \
>         "-R[redisplay]" \
> 	"(*)-c[clear listing]" \
> 	"(-)::status line: " "*:strings to list: "

Aha. In combination with an optional argument. Right, it was used the
rest-spec even though that was inactivated.

Bye
 Sven

Index: Src/Zle/computil.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/Zle/computil.c,v
retrieving revision 1.42
diff -u -r1.42 computil.c
--- Src/Zle/computil.c	2000/08/14 08:16:45	1.42
+++ Src/Zle/computil.c	2000/10/05 08:52:07
@@ -1685,7 +1685,8 @@
 
 	if (!opt) {
 	    if (arg->num >= 0 && !arg->next && miss)
-		arg = ca_laststate.d->rest;
+		arg = (ca_laststate.d->rest && ca_laststate.d->rest->active ?
+		       ca_laststate.d->rest : NULL);
 	    else {
 		onum = arg->num;
 		rest = (onum != arg->min && onum == ca_laststate.nth);
@@ -1693,7 +1694,8 @@
 		    if (arg->num != onum + 1)
 			miss = 1;
 		} else if (rest || (oopt > 0 && !opt)) {
-		    arg = ca_laststate.d->rest;
+		    arg = (ca_laststate.d->rest && ca_laststate.d->rest->active ?
+			   ca_laststate.d->rest : NULL);
 		    oopt = -1;
 		}
 	    }
@@ -1711,7 +1713,8 @@
     }
     if (!opt && oopt > 0) {
 	oopt = -1;
-	arg = ca_laststate.d->rest;
+	arg = (ca_laststate.d->rest && ca_laststate.d->rest->active ?
+	       ca_laststate.d->rest : NULL);
 
 	goto rec;
     }

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

Re: PATCH: Redo _zle using _arguments

[Bart Schaefer]

On Oct 4,  2:47pm, Sven Wischnowsky wrote:
} 
} Bart Schaefer wrote:
} 
} > I also suspect there may be a bug in comparguments -- the '(*)-x' form of
} > optspec is supposed to mean that the '*:msg:act' form is not completed
} > when -x is on the command line, but that doesn't seem to happen (or at
} > least, not always).
} 
} Hm, I had a look but couldn't find a bug. Could you give me an example?

Sure.  Completion after "zle -R -c":

zagzig[501] zle -R -c <TAB>
Completing status line
Completing strings to list

The call to _arguments was:

      _arguments -s \
        "-R[redisplay]" \
	"(*)-c[clear listing]" \
	"(-)::status line: " "*:strings to list: "

Since "(*)-c" is there, and -c is on the line, the "*:strings to list: "
clause should never be activated.  Yet there's the description string for
it, in the listing.  If I go on to add something for the status line:

zagzig[501] zle -R -c something <TAB>
Completing no more arguments

What I expected the first time was either just "Completing status line"
or that plus "no more arguments", but not "strings to list".

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net   

Re: PATCH: Redo _zle using _arguments

[Sven Wischnowsky]

Bart Schaefer wrote:

> ...
> 
> I also suspect there may be a bug in comparguments -- the '(*)-x' form of
> optspec is supposed to mean that the '*:msg:act' form is not completed
> when -x is on the command line, but that doesn't seem to happen (or at
> least, not always).  I imagine a fix is going to have to wait a few weeks
> for Sven to return, though.

Hm, I had a look but couldn't find a bug. Could you give me an example?

Bye
 Sven


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

Re: PATCH: Redo _zle using _arguments

[Sven Wischnowsky]

Bart Schaefer wrote:

> ...
> 
> I also suspect there may be a bug in comparguments -- the '(*)-x' form of
> optspec is supposed to mean that the '*:msg:act' form is not completed
> when -x is on the command line, but that doesn't seem to happen (or at
> least, not always).  I imagine a fix is going to have to wait a few weeks
> for Sven to return, though.

I'll have a look.

> Random other question: Why isn't list-expand a completion widget for the
> purposes of "zle -C"?

Because it never does completion, only the hard-wired form of
expansion which can't be programmed.

Bye
 Sven


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