Re: PATCH: Corrected style test in prediction

Re: PATCH: Corrected style test in prediction

[Sven Wischnowsky]

Bart Schaefer wrote:

> } The number of colons is still correct, but the original context name
> } was correct, too. The `predict' was stored in the `function' field of
> } the context name, indicating that completion was called from predict.
> 
> But in insert-and-predict we have:
> 
> 	  local crs curcontext="${curcontext}"
> 
>           [[ -z "$curcontext" ]] && curcontext=:::
>           curcontext="${curcontext#*:}predict:"
> 
> which makes the context be :completion:::predict::, with `predict' in
> the `command' field.  Surely the two should be the same?  If not, why
> not?

Ouch. Thanks for the hint.

Bye
 Sven

diff -ru ../z.old/Functions/Zle/predict-on Functions/Zle/predict-on
--- ../z.old/Functions/Zle/predict-on	Wed Feb 23 14:44:05 2000
+++ Functions/Zle/predict-on	Wed Feb 23 16:47:41 2000
@@ -56,7 +56,7 @@
 	  local crs curcontext="${curcontext}"
 
           [[ -z "$curcontext" ]] && curcontext=:::
-          curcontext="${curcontext#*:}predict:"
+          curcontext="predict:${curcontext#*:}"
 
 	  comppostfuncs=( predict-limit-list )
 	  zle complete-word

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

Re: PATCH: Corrected style test in prediction

[Sven Wischnowsky]

I wrote:

> -          curcontext="${curcontext#*:}predict:"
> +          curcontext="predict:${curcontext#*:}"

Getting a bit slow in the brain...

Bye
 Sven

diff -ru ../z.old/Functions/Zle/incremental-complete-word Functions/Zle/incremental-complete-word
--- ../z.old/Functions/Zle/incremental-complete-word	Wed Feb 23 14:44:05 2000
+++ Functions/Zle/incremental-complete-word	Wed Feb 23 16:52:31 2000
@@ -21,7 +21,7 @@
   local curcontext="${curcontext}" stop brk
 
   [[ -z "$curcontext" ]] && curcontext=:::
-  curcontext="${curcontext#*:}incremental:"
+  curcontext="incremental:${curcontext#*:}"
 
   zstyle -s ":completion:${curcontext}" prompt pmpt ||
     pmpt='incremental (%c): %u%s  %l'

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

Re: PATCH: Corrected style test in prediction

[Bart Schaefer]

On Feb 23, 11:46am, Sven Wischnowsky wrote:
} Subject: Re: PATCH: Corrected style test in prediction
}
} Bart Schaefer wrote:
} 
} > Sven, please confirm ... I'm never sure about the number of trailing colons.
} > 
} > Index: Functions/Zle/predict-on
} > ===================================================================
} > -  elif zstyle -t ":completion:predict::::" list always
} > +  elif zstyle -t ":completion:::predict::" list always
} 
} The number of colons is still correct, but the original context name
} was correct, too. The `predict' was stored in the `function' field of
} the context name, indicating that completion was called from predict.

But in insert-and-predict we have:

	  local crs curcontext="${curcontext}"

          [[ -z "$curcontext" ]] && curcontext=:::
          curcontext="${curcontext#*:}predict:"

which makes the context be :completion:::predict::, with `predict' in
the `command' field.  Surely the two should be the same?  If not, why
not?

} However, the other call to zstyle had a missing colon, the one before
} the (not-given) tag.

Well, that part at least is less confusing now.

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

Re: PATCH: Corrected style test in prediction

[Sven Wischnowsky]

Bart Schaefer wrote:

> Sven, please confirm ... I'm never sure about the number of trailing colons.
> 
> Index: Functions/Zle/predict-on
> ===================================================================
> @@ -122,7 +122,7 @@
>    then
>      compstate[list]=''
>      compstate[force_list]=yes
> -  elif zstyle -t ":completion:predict::::" list always
> +  elif zstyle -t ":completion:::predict::" list always
>    then
>      compstate[force_list]=yes
>    fi

The number of colons is still correct, but the original context name
was correct, too. The `predict' was stored in the `function' field of
the context name, indicating that completion was called from predict.

However, the other call to zstyle had a missing colon, the one before
the (not-given) tag.


The patch also fixes a type in _gunzip -- as shown in Peter's guide.

Bye
 Sven

diff -ru ../z.old/Completion/User/_gunzip Completion/User/_gunzip
--- ../z.old/Completion/User/_gunzip	Wed Feb 23 10:37:24 2000
+++ Completion/User/_gunzip	Wed Feb 23 11:20:56 2000
@@ -3,4 +3,4 @@
 local expl
 
 _description files expl 'compressed file'
-_files "$expl[@]" -g '*.([tT]|)[gG][z]'
+_files "$expl[@]" -g '*.([tT]|)[gG][zZ]'
diff -ru ../z.old/Functions/Zle/predict-on Functions/Zle/predict-on
--- ../z.old/Functions/Zle/predict-on	Wed Feb 23 09:03:49 2000
+++ Functions/Zle/predict-on	Wed Feb 23 11:45:21 2000
@@ -64,7 +64,7 @@
 	  # get out of that `case'.
 	  repeat 1
 	  do
-	    zstyle -s ":completion:${curcontext}" cursor crs
+	    zstyle -s ":completion:${curcontext}:" cursor crs
 	    case $crs in
 	    (complete)
 	      # At the place where the completion left it, if it is after

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

PATCH: Corrected style test in prediction

[Bart Schaefer]

Sven, please confirm ... I'm never sure about the number of trailing colons.

Index: Functions/Zle/predict-on
===================================================================
@@ -122,7 +122,7 @@
   then
     compstate[list]=''
     compstate[force_list]=yes
-  elif zstyle -t ":completion:predict::::" list always
+  elif zstyle -t ":completion:::predict::" list always
   then
     compstate[force_list]=yes
   fi

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