Re: Problem with an exported array

[Peter Stephenson <pws@xxxxxxx>]

As a consequence of a somewhat tangential discussion with Oliver, here
is a new proposal for tied arrays which should satisfy all the
requirements.

There is no new option; instead, the split/join character is given as
the third argument, which was previously an error so is entirely
backwardly compatible:

typeset -T PAGER pager ' '

Internally, instead of storing just the pointer to the array, we
allocate a new structure for both that and the join character and store
a pointer to that.  We then just have to remember to free the new
structure when we unset the parameter, so there's a new unset function.

I tried to be careful with the possibility that the join character could
be an arbitrary eight-bit character.

The second zsh hunk is now irrelevant but a bit neater than before.

I fixed the problem that setting using the scalar didn't respect -U by
brute force.

Index: Doc/Zsh/builtins.yo
===================================================================
RCS file: /cvsroot/zsh/zsh/Doc/Zsh/builtins.yo,v
retrieving revision 1.64
diff -u -r1.64 builtins.yo
--- Doc/Zsh/builtins.yo	11 Sep 2003 07:00:07 -0000	1.64
+++ Doc/Zsh/builtins.yo	23 Sep 2003 15:58:33 -0000
@@ -1160,7 +1160,7 @@
 xitem(tt(typeset) [ {tt(PLUS())|tt(-)}tt(AEFHLRUZafghilprtuxm) [var(n)]] [ \
 var(name)[tt(=)var(value)] ... ])
 item(tt(typeset) -T [ {tt(PLUS()|tt(-))}tt(LRUZrux) ] \
-  var(SCALAR)[tt(=)var(value)] var(array))(
+  var(SCALAR)[tt(=)var(value)] var(array) tt([) var(sep) tt(]))(
 Set or display attributes and values for shell parameters.
 
 A parameter is created for each var(name) that does not already refer
@@ -1189,20 +1189,24 @@
 and options.  Note that the tt(-h) flag on parameters is respected; no
 value will be shown for these parameters.
 
-If the tt(-T) option is given, exactly two (or zero) var(name)
-arguments must be present.  They represent a scalar and an array (in
-that order) that will be tied together in the manner of tt($PATH) and
-tt($path).  In other words, an array present in the latter variable
-appears as a scalar with the elements of the array joined by colons in
-the former.  Only the scalar may have an initial value.  Both the
-scalar and the array may otherwise be manipulated as normal.  If one
-is unset, the other will automatically be unset too.  There is no way
-of untying the variables without unsetting them, or converting the
-type of one of them with another tt(typeset) command; tt(+T) does not
-work, assigning an array to var(SCALAR) is an error, and assigning a
-scalar to var(array) sets it to be a single-element array.  Note that
-both `tt(typeset -xT ...)' and `tt(export -T ...)' work, but only the
-scalar will be marked for export.
+If the tt(-T) option is given, two or three arguments must be present (an
+exception is that zero arguments are allowed to show the list of parameters
+created in this fashion).  The first two are the name of a scalar and an
+array parameter (in that order) that will be tied together in the manner of
+tt($PATH) and tt($path).  The optional third argument is a single-character
+separator which will be used to join the elements of the array to form the
+scalar; if absent, a colon is used, as with tt($PATH).  Only the first
+character of the separator is significant; any remaining characters are
+ignored.  Only the scalar parameter may be assigned an initial value.  Both
+the scalar and the array may otherwise be manipulated as normal.  If one is
+unset, the other will automatically be unset too.  There is no way of
+untying the variables without unsetting them, or converting the type of one
+of them with another tt(typeset) command; tt(+T) does not work, assigning
+an array to var(SCALAR) is an error, and assigning a scalar to var(array)
+sets it to be a single-element array.  Note that both `tt(typeset -xT ...)'
+and `tt(export -T ...)' work, but only the scalar will be marked for
+export.  Setting the value using the scalar version causes a split on all
+separators (which cannot be quoted).
 
 The tt(-g) (global) flag is treated specially: it means that any
 resulting parameter will not be restricted to local scope.  Note that this
Index: Src/builtin.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/builtin.c,v
retrieving revision 1.105
diff -u -r1.105 builtin.c
--- Src/builtin.c	11 Sep 2003 07:00:07 -0000	1.105
+++ Src/builtin.c	23 Sep 2003 15:58:34 -0000
@@ -1677,7 +1677,7 @@
 static Param
 typeset_single(char *cname, char *pname, Param pm, int func,
 	       int on, int off, int roff, char *value, Param altpm,
-	       Options ops, int auxlen)
+	       Options ops, int auxlen, int joinchar)
 {
     int usepm, tc, keeplocal = 0, newspecial = NS_NONE, readonly;
     char *subscript;
@@ -1997,9 +1997,17 @@
 	 * to make sure we only ever use the colonarr functions
 	 * when u.data is correctly set.
 	 */
-	pm->sets.cfn = colonarrsetfn;
-	pm->gets.cfn = colonarrgetfn;
-	pm->u.data = &altpm->u.arr;
+	struct tieddata *tdp = (struct tieddata *)
+	    zalloc(sizeof(struct tieddata));
+	if (!tdp)
+	    return NULL;
+	tdp->joinchar = joinchar;
+	tdp->arrptr = &altpm->u.arr;
+
+	pm->sets.cfn = tiedarrsetfn;
+	pm->gets.cfn = tiedarrgetfn;
+	pm->unsetfn = tiedarrunsetfn;
+	pm->u.data = tdp;
     }
 
     if (keeplocal)
@@ -2155,6 +2163,7 @@
 	Param apm;
 	struct asgment asg0;
 	char *oldval = NULL;
+	int joinchar;
 
 	if (OPT_ISSET(ops,'m')) {
 	    zwarnnam(name, "incompatible options for -T", NULL, 0);
@@ -2162,12 +2171,25 @@
 	    return 1;
 	}
 	on &= ~off;
-	if (!argv[1] || argv[2]) {
+	if (!argv[1] || argv[3]) {
 	    zwarnnam(name, "-T requires names of scalar and array", NULL, 0);
 	    unqueue_signals();
 	    return 1;
 	}
 
+	/*
+	 * Third argument, if given, is character used to join
+	 * the elements of the array in the scalar.
+	 */
+	if (!argv[2])
+	    joinchar = ':';
+	else if (!*argv[2])
+	    joinchar = 0;
+	else if (*argv[2] == Meta)
+	    joinchar = argv[2][1] ^ 32;
+	else
+	    joinchar = *argv[2];
+
 	if (!(asg = getasg(argv[0]))) {
 	    unqueue_signals();
 	    return 1;
@@ -2212,7 +2234,8 @@
 				 (Param)paramtab->getnode(paramtab,
 							  asg->name),
 				 func, (on | PM_ARRAY) & ~PM_EXPORTED,
-				 off, roff, asg->value, NULL, ops, auxlen))) {
+				 off, roff, asg->value, NULL, ops, auxlen,
+				 0))) {
 	    unqueue_signals();
 	    return 1;
 	}
@@ -2224,7 +2247,7 @@
 				(Param)paramtab->getnode(paramtab,
 							 asg0.name),
 				func, on, off, roff, asg0.value, apm,
-				ops, auxlen))) {
+				ops, auxlen, joinchar))) {
 	    if (oldval)
 		zsfree(oldval);
 	    unsetparam_pm(apm, 1, 1);
@@ -2291,7 +2314,7 @@
 	    for (pmnode = firstnode(pmlist); pmnode; incnode(pmnode)) {
 		pm = (Param) getdata(pmnode);
 		if (!typeset_single(name, pm->nam, pm, func, on, off, roff,
-				    asg->value, NULL, ops, auxlen))
+				    asg->value, NULL, ops, auxlen, 0))
 		    returnval = 1;
 	    }
 	}
@@ -2306,7 +2329,7 @@
 				     gethashnode2(paramtab, asg->name) :
 				     paramtab->getnode(paramtab, asg->name)),
 			    func, on, off, roff, asg->value, NULL,
-			    ops, auxlen))
+			    ops, auxlen, 0))
 	    returnval = 1;
     }
     unqueue_signals();
Index: Src/params.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/params.c,v
retrieving revision 1.72
diff -u -r1.72 params.c
--- Src/params.c	30 Aug 2003 19:00:20 -0000	1.72
+++ Src/params.c	23 Sep 2003 15:58:34 -0000
@@ -2595,22 +2595,77 @@
 colonarrsetfn(Param pm, char *x)
 {
     char ***dptr = (char ***)pm->u.data;
-
     /*
-     * If this is tied to a parameter (rather than internal) array,
-     * the array itself may be NULL.  Otherwise, we have to make
-     * sure it doesn't ever get null.
+     * We have to make sure this is never NULL, since that
+     * can cause problems.
      */
     if (*dptr)
 	freearray(*dptr);
-    *dptr = x ? colonsplit(x, pm->flags & PM_UNIQUE) :
-	(pm->flags & PM_TIED) ? NULL : mkarray(NULL);
+    if (x)
+	*dptr = colonsplit(x, pm->flags & PM_UNIQUE);
+    else
+	*dptr = mkarray(NULL);
     if (pm->ename)
 	arrfixenv(pm->nam, *dptr);
     zsfree(x);
 }
 
 /**/
+char *
+tiedarrgetfn(Param pm)
+{
+    struct tieddata *dptr = (struct tieddata *)pm->u.data;
+    return *dptr->arrptr ? zjoin(*dptr->arrptr, dptr->joinchar, 1) : "";
+}
+
+/**/
+void
+tiedarrsetfn(Param pm, char *x)
+{
+    struct tieddata *dptr = (struct tieddata *)pm->u.data;
+
+    if (*dptr->arrptr)
+	freearray(*dptr->arrptr);
+    if (x) {
+	char sepbuf[3];
+	if (imeta(dptr->joinchar))
+	{
+	    sepbuf[0] = Meta;
+	    sepbuf[1] = dptr->joinchar;
+	    sepbuf[2] = '\0';
+	}
+	else
+	{
+	    sepbuf[0] = dptr->joinchar;
+	    sepbuf[1] = '\0';
+	}
+	*dptr->arrptr = sepsplit(x, sepbuf, 0, 0);
+	if (pm->flags & PM_UNIQUE)
+	    uniqarray(*dptr->arrptr);
+    } else
+	*dptr->arrptr = NULL;
+    if (pm->ename)
+	arrfixenv(pm->nam, *dptr->arrptr);
+    zsfree(x);
+}
+
+/**/
+void
+tiedarrunsetfn(Param pm, int exp)
+{
+    /*
+     * Special unset function because we allocated a struct tieddata
+     * in typeset_single to hold the special data which we now
+     * need to delete.
+     */
+    pm->sets.cfn(pm, NULL);
+    zfree(pm->u.data, sizeof(struct tieddata));
+    /* paranoia -- shouldn't need these, but in case we reuse the struct... */
+    pm->u.data = NULL;
+    pm->flags &= ~PM_TIED;
+}
+
+/**/
 void
 uniqarray(char **x)
 {
@@ -3187,6 +3242,7 @@
 arrfixenv(char *s, char **t)
 {
     Param pm;
+    int joinchar;
 
     if (t == path)
 	cmdnamtab->emptytable(cmdnamtab);
@@ -3208,8 +3264,15 @@
      * Do not "fix" parameters that were not exported
      */
 
-    if (pm->flags & PM_EXPORTED)
-	pm->env = addenv(s, t ? zjoin(t, ':', 1) : "", pm->flags);
+    if (!(pm->flags & PM_EXPORTED))
+	return;
+
+    if (pm->flags & PM_TIED)
+	joinchar = ((struct tieddata *)pm->u.data)->joinchar;
+    else
+	joinchar = ':';
+
+    pm->env = addenv(s, t ? zjoin(t, joinchar, 1) : "", pm->flags);
 }
 
 
Index: Src/utils.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/utils.c,v
retrieving revision 1.53
diff -u -r1.53 utils.c
--- Src/utils.c	1 Aug 2003 16:29:21 -0000	1.53
+++ Src/utils.c	23 Sep 2003 15:58:34 -0000
@@ -1793,14 +1793,20 @@
     char **s, *ret, *ptr;
 
     for (s = arr; *s; s++)
-	len += strlen(*s) + 1;
+	len += strlen(*s) + 1 + (imeta(delim) ? 1 : 0);
     if (!len)
 	return heap? "" : ztrdup("");
     ptr = ret = (heap ? (char *) hcalloc(len) : (char *) zcalloc(len));
     for (s = arr; *s; s++) {
 	strucpy(&ptr, *s);
-	if (delim)
-	    *ptr++ = delim;
+	if (delim) {
+	    if (imeta(delim)) {
+		*ptr++ = Meta;
+		*ptr++ = delim ^ 32;
+	    }
+	    else
+		*ptr++ = delim;
+	}
     }
     ptr[-1] = '\0';
     return ret;
@@ -1856,7 +1862,15 @@
     return i;
 }
 
-/* see findsep() below for handling of `quote' argument */
+/*
+ * haven't worked out what allownull does; it's passed down from
+ *   sepsplit but all the cases it's used are either 0 or 1 without
+ *   a comment.  it seems to be something to do with the `nulstring'
+ *   which i think is some kind of a metafication thing, so probably
+ *   allownull's value is associated with whether we are using
+ *   metafied strings.
+ * see findsep() below for handling of `quote' argument
+ */
 
 /**/
 mod_export char **
Index: Src/zsh.h
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/zsh.h,v
retrieving revision 1.49
diff -u -r1.49 zsh.h
--- Src/zsh.h	3 Sep 2003 10:15:36 -0000	1.49
+++ Src/zsh.h	23 Sep 2003 15:58:34 -0000
@@ -1156,6 +1156,12 @@
     int level;			/* if (old != NULL), level of localness  */
 };
 
+/* structure stored in struct param's u.data by tied arrays */
+struct tieddata {
+    char ***arrptr;		/* pointer to corresponding array */
+    int joinchar;		/* character used to join arrays */
+};
+
 /* flags for parameters */
 
 /* parameter types */
@@ -1193,16 +1199,16 @@
 #define PM_TIED 	(1<<16)	/* array tied to colon-path or v.v.         */
 
 /* Remaining flags do not correspond directly to command line arguments */
-#define PM_LOCAL	(1<<17) /* this parameter will be made local        */
-#define PM_SPECIAL	(1<<18) /* special builtin parameter                */
-#define PM_DONTIMPORT	(1<<19)	/* do not import this variable              */
-#define PM_RESTRICTED	(1<<20) /* cannot be changed in restricted mode     */
-#define PM_UNSET	(1<<21)	/* has null value                           */
-#define PM_REMOVABLE	(1<<22)	/* special can be removed from paramtab     */
-#define PM_AUTOLOAD	(1<<23) /* autoloaded from module                   */
-#define PM_NORESTORE	(1<<24)	/* do not restore value of local special    */
-#define PM_HASHELEM     (1<<25) /* is a hash-element */
-#define PM_NAMEDDIR     (1<<26) /* has a corresponding nameddirtab entry    */
+#define PM_LOCAL	(1<<21) /* this parameter will be made local        */
+#define PM_SPECIAL	(1<<22) /* special builtin parameter                */
+#define PM_DONTIMPORT	(1<<23)	/* do not import this variable              */
+#define PM_RESTRICTED	(1<<24) /* cannot be changed in restricted mode     */
+#define PM_UNSET	(1<<25)	/* has null value                           */
+#define PM_REMOVABLE	(1<<26)	/* special can be removed from paramtab     */
+#define PM_AUTOLOAD	(1<<27) /* autoloaded from module                   */
+#define PM_NORESTORE	(1<<28)	/* do not restore value of local special    */
+#define PM_HASHELEM     (1<<29) /* is a hash-element */
+#define PM_NAMEDDIR     (1<<30) /* has a corresponding nameddirtab entry    */
 
 /* The option string corresponds to the first of the variables above */
 #define TYPESET_OPTSTR "aiEFALRZlurtxUhHT"


-- 
Peter Stephenson <pws@xxxxxxx>                  Software Engineer
CSR Ltd., Science Park, Milton Road,
Cambridge, CB4 0WH, UK                          Tel: +44 (0)1223 692070


**********************************************************************
The information transmitted is intended only for the person or
entity to which it is addressed and may contain confidential 
and/or privileged material. 
Any review, retransmission, dissemination or other use of, or
taking of any action in reliance upon, this information by 
persons or entities other than the intended recipient is 
prohibited.  
If you received this in error, please contact the sender and 
delete the material from any computer.
**********************************************************************