PATCH: Re: ${-str}

PATCH: Re: ${-str}

[Sven Wischnowsky]

Peter Stephenson wrote:

> Sven wrote: 
> > Geoff Wing wrote:
> > > :serein% print -lr ${-str}
> > > :569XZfims
> > > % setopt norcs; echo ${-}
> > > 569XZfims
> > 
> > Yep. The same as for the other special parameters (`${$foo}').
> > Shouldn't we change this to be an error? Or do we really need this for 
> > compatibility?
> 
> I think this should be an error.  This was never documented behaviour and
> most forms of bad parameter syntax produce complaints.

Hope I haven't missed a special (and allowed) character...

This is done quite lately to be able to catch `${foo[1][2]xxx}'. It
doesn't catch `${}', but since `${:-...}' is valid, that's probably ok.

The other hunks avoid a `unused variable' warning.

Bye
 Sven

Index: Src/subst.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/subst.c,v
retrieving revision 1.4
diff -u -r1.4 subst.c
--- Src/subst.c	2000/05/08 08:58:37	1.4
+++ Src/subst.c	2000/05/15 14:56:35
@@ -1230,6 +1230,15 @@
 	if (!inbrace)
 	    break;
     }
+    if (inbrace &&
+	(c = *s) != '-' && c != '+' && c != ':' && c != '%'  && c != '/' &&
+	c != '=' && c != Equals &&
+	c != '#' && c != Pound &&
+	c != '?' && c != Quest &&
+	c != '}' && c != Outbrace) {
+	zerr("bad substitution", NULL, 0);
+	return NULL;
+    }
     if (isarr) {
 	if (nojoin)
 	    isarr = -1;
@@ -1463,9 +1472,6 @@
 		}
 	    }
 	    {
-		char t = s[-1];
-
-		singsub(&s);
 #if 0
 		/*
 		 * This allows # and % to be at the start of
@@ -1473,6 +1479,11 @@
 		 * a bit nasty, and can be done (although
 		 * less efficiently) with anchors.
 		 */
+
+		char t = s[-1];
+
+		singsub(&s);
+
 		if (t == '/' && (flags & SUB_SUBSTR)) {
 		    if ((c = *s) == '#' || c == '%') {
 			flags &= ~SUB_SUBSTR;
@@ -1483,6 +1494,8 @@
 			s++;
 		    }
 		}
+#else
+		singsub(&s);
 #endif
 	    }
 

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

PATCH: Re: ${-str}

[Peter Stephenson]

Sven wrote
> > Tanaka Akira <akr@m17n.org> typed:
> > :Z(4):akr@serein% Src/zsh -f             
> > :serein% print -lr ${-str}
> > :569XZfims
>
> Shouldn't we change this to be an error? Or do we really need this for 
> compatibility?

Is this correct?  All the parameter tests still seem to work.  The first
hunk simply removes an unneeded local variable.  The change already picked
up an extraneous closing brace in _vars_eq, which is also patched.

Index: Completion/Builtins/_vars_eq
===================================================================
RCS file: /cvsroot/zsh/zsh/Completion/Builtins/_vars_eq,v
retrieving revision 1.6
diff -u -r1.6 _vars_eq
--- Completion/Builtins/_vars_eq	2000/05/15 00:31:20	1.6
+++ Completion/Builtins/_vars_eq	2000/05/15 15:01:42
@@ -45,7 +45,7 @@
   export) use="${${use/g/}/x/}" ;;
 esac
 
-[[ -z "${words[(r)-*f*]]}" ]] || func=f
+[[ -z "${words[(r)-*f*]}" ]] || func=f
 [[ -z "${words[(r)-*[aA]*]}" ]] || func=p
    
 for ((i=1;i<=$#use;++i)); do
Index: Src/subst.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/subst.c,v
retrieving revision 1.4
diff -u -r1.4 subst.c
--- Src/subst.c	2000/05/08 08:58:37	1.4
+++ Src/subst.c	2000/05/15 15:01:42
@@ -1463,10 +1463,12 @@
 		}
 	    }
 	    {
+#if 1
+		singsub(&s);
+#else
 		char t = s[-1];
 
 		singsub(&s);
-#if 0
 		/*
 		 * This allows # and % to be at the start of
 		 * a parameter in the substitution, which is
@@ -1543,6 +1545,10 @@
 	}
 	if (!inbrace)
 	    fstr = s;
+	else if (s != fstr && *s) {
+	    zerr("garbage in parameter substitution: %s", s, 0);
+	    return NULL;
+	}
     }
     if (errflag)
 	return NULL;

-- 
Peter Stephenson <pws@cambridgesiliconradio.com>
Cambridge Silicon Radio, Unit 300, Science Park, Milton Road,
Cambridge, CB4 0XL, UK                          Tel: +44 (0)1223 392070