PATCH: Re: 3.1.6-bart-8: menu select + oldlist oddities

PATCH: Re: 3.1.6-bart-8: menu select + oldlist oddities

[Sven Wischnowsky]

Andrej Borsenkow wrote:

> 3.1.6-bart-8 just compiled and installed.
> 
> ...
> 
> Case 1 (O.K.)
         ^^^^^^

Uff. Good. ;-)

> bor@itsrm2:~%> l /t/s/z/f/_<TAB>
> bor@itsrm2:~%> l /tools/share/zsh/functions<CURSOR>/_
> functions/      functions.old/
> Now press cursror-right,TAB
> bor@itsrm2:~%> l /tools/share/zsh/functions/_
> zsh: do you wish to see all 194 possibilities? 

Remember this line...

> so far so good
> 
> Case 2 (strange)
> bor@itsrm2:~%> l /t/s/z/f/_<TAB><TAB>
> bor@itsrm2:~%> l /tools/share/zsh/functions<CURSOR>/_
> functions/      functions.old/
> ^^ highlighted
> Now press "/" (accept-and-infer-next-history)
> bor@itsrm2:~%> l /tools/share/zsh/functions/_a2ps
> 
> Oops! No more completion there inside /tools/share/zsh/functions

Because (as it said above), there are too many matches to display
them. I was aware of this when I wrote that code but didn't know what
to do - -so I just made it give up. But what do you think, it should
do? It *can't* display the menu in this case... Should we disallow
a-a-i-n-h in this case? I think this would be even weirder.

> Case 3 (weird)
> bor@itsrm2:~%> l /t/s/z/f/_<TAB><TAB>
> bor@itsrm2:~%> l /tools/share/zsh/functions<CURSOR>/_
> functions/      functions.old/
> ^^ highlighted
> Now press RETURN, TAB
> bor@itsrm2:~%> l /tools/share/zsh/functions.old/_
> functions/      functions.old/
> TAB once more
> bor@itsrm2:~%> l /tools/share/zsh/functions/_    
> functions/      functions.old/

You don't like it that it continues menu-completion of the same
matches (because of `oldlist_list=_complete:...'). Yes, I agree, after 
a menu-selection we almost certainly don't want to re-use such a old
list.


Bye
 Sven

diff -u oldsrc/Zle/complist.c Src/Zle/complist.c
--- oldsrc/Zle/complist.c	Wed Nov  3 11:47:59 1999
+++ Src/Zle/complist.c	Wed Nov  3 12:46:55 1999
@@ -856,8 +856,7 @@
 	mselect = -1;
 	inselect = 0;
 	if (acc) {
-	    menucmp = 0;
-	    lastambig = 0;
+	    menucmp = lastambig = hasoldlist = 0;
 	    do_single(*(minfo.cur));
 	}
 	if (!noselect) {

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

Re: PATCH: Re: 3.1.6-bart-8: menu select + oldlist oddities

[Bart Schaefer]

On Nov 3, 12:53pm, Sven Wischnowsky wrote:
} Subject: PATCH: Re: 3.1.6-bart-8: menu select + oldlist oddities
}
} > Oops! No more completion there inside /tools/share/zsh/functions
} 
} Because (as it said above), there are too many matches to display
} them. I was aware of this when I wrote that code but didn't know what
} to do - -so I just made it give up. But what do you think, it should
} do? It *can't* display the menu in this case... Should we disallow
} a-a-i-n-h in this case? I think this would be even weirder.

How about this:  Drop back into regular menu completion and produce the
"do you wish to see ...?" prompt.

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

RE: PATCH: Re: 3.1.6-bart-8: menu select + oldlist oddities

[Andrej Borsenkow]

> On Nov 3, 12:53pm, Sven Wischnowsky wrote:
> } Subject: PATCH: Re: 3.1.6-bart-8: menu select + oldlist oddities
> }
> } > Oops! No more completion there inside /tools/share/zsh/functions
> }
> } Because (as it said above), there are too many matches to display
> } them. I was aware of this when I wrote that code but didn't know what
> } to do - -so I just made it give up. But what do you think, it should
> } do? It *can't* display the menu in this case... Should we disallow
> } a-a-i-n-h in this case? I think this would be even weirder.
>
> How about this:  Drop back into regular menu completion and produce the
> "do you wish to see ...?" prompt.
>

Exactly. Actually, I again fail to see, how these three cases differ. In all
cases I first complete the first ambiguous prefix and then want to complete the
rest. The way I leave this first completion (using cursor keys, accept or
a-a-i-n-h) should not matter. IMHO :-)

/andrej

Re: PATCH: Re: 3.1.6-bart-8: menu select + oldlist oddities

[Sven Wischnowsky]

Bart Schaefer wrote:

> On Nov 3, 12:53pm, Sven Wischnowsky wrote:
> } Subject: PATCH: Re: 3.1.6-bart-8: menu select + oldlist oddities
> }
> } > Oops! No more completion there inside /tools/share/zsh/functions
> } 
> } Because (as it said above), there are too many matches to display
> } them. I was aware of this when I wrote that code but didn't know what
> } to do - -so I just made it give up. But what do you think, it should
> } do? It *can't* display the menu in this case... Should we disallow
> } a-a-i-n-h in this case? I think this would be even weirder.
> 
> How about this:  Drop back into regular menu completion and produce the
> "do you wish to see ...?" prompt.

Blink. Of course...

Bye
 Sven

diff -u oldsrc/Zle/compcore.c Src/Zle/compcore.c
--- oldsrc/Zle/compcore.c	Wed Nov  3 12:53:50 1999
+++ Src/Zle/compcore.c	Thu Nov  4 11:23:30 1999
@@ -428,7 +428,7 @@
     /* If we are doing a menu-completion... */
 
     if (menucmp && *lst != COMP_LIST_EXPAND && 
-	(!compwidget || compwidget == lastcompwidget)) {
+	(menucmp != 1 || !compwidget || compwidget == lastcompwidget)) {
 	do_menucmp(*lst);
 	return 1;
     }
diff -u oldsrc/Zle/complist.c Src/Zle/complist.c
--- oldsrc/Zle/complist.c	Wed Nov  3 12:53:50 1999
+++ Src/Zle/complist.c	Thu Nov  4 11:30:04 1999
@@ -545,7 +545,7 @@
     Cmgroup *pg;
     Thingy cmd;
     Menustack u = NULL;
-    int i = 0, acc = 0, wishcol = 0, setwish = 0, oe = onlyexpl;
+    int i = 0, acc = 0, wishcol = 0, setwish = 0, oe = onlyexpl, wasnext = 0;
     char *s;
 
     HEAPALLOC {
@@ -591,7 +591,7 @@
 		while (mcol < mcols - 1 && p[1] == minfo.cur)
 		    mcol++, p++, pg++;
 	    }
-	    setwish = 0;
+	    setwish = wasnext = 0;
 
 	getk:
 
@@ -634,7 +634,7 @@
 		}
 		clearlist = listshown = 1;
 		mselect = (*(minfo.cur))->gnum;
-		setwish = 1;
+		setwish = wasnext = 1;
 		continue;
 	    } else if (cmd == Th(z_acceptandhold) ||
 		       cmd == Th(z_acceptandmenucomplete)) {
@@ -858,6 +858,11 @@
 	if (acc) {
 	    menucmp = lastambig = hasoldlist = 0;
 	    do_single(*(minfo.cur));
+	}
+	if (wasnext) {
+	    menucmp = 2;
+	    showinglist = -2;
+	    minfo.asked = 0;
 	}
 	if (!noselect) {
 	    showinglist = -2;
diff -u oldsrc/Zle/zle_tricky.c Src/Zle/zle_tricky.c
--- oldsrc/Zle/zle_tricky.c	Wed Nov  3 12:53:51 1999
+++ Src/Zle/zle_tricky.c	Thu Nov  4 11:24:49 1999
@@ -77,7 +77,8 @@
 /**/
 int usemenu, useglob;
 
-/* != 0 if we are in the middle of a menu completion. */
+/* != 0 if we are in the middle of a menu completion. May be == 2 to force *
+ * menu completion even if using different widgets.                        */
 
 /**/
 int menucmp;

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