merge the nameref code, using mksh standard scoping as discussed

check.t

@@ -1,4 +1,4 @@
-# $MirOS: src/bin/mksh/check.t,v 1.302 2009/08/30 21:01:59 tg Exp $
+# $MirOS: src/bin/mksh/check.t,v 1.303 2009/09/06 17:42:11 tg Exp $
 # $OpenBSD: bksl-nl.t,v 1.2 2001/01/28 23:04:56 niklas Exp $
 # $OpenBSD: history.t,v 1.5 2001/01/28 23:04:56 niklas Exp $
 # $OpenBSD: read.t,v 1.3 2003/03/10 03:48:16 david Exp $
@@ -4197,7 +4197,7 @@ name: xxx-param-subst-qmark-1
 description:
 	Check suppresion of error message with null string.  According to
 	POSIX, it shouldn't print the error as 'word' isn't ommitted.
-	ksh88, Solaris /bin/sh and /usr/xpg4/bin/sh all print the error,
+	ksh88/93, Solaris /bin/sh and /usr/xpg4/bin/sh all print the error,
 	that's why the condition is reversed.
 stdin:
 	unset foo
@@ -4731,6 +4731,7 @@ expected-stdout:
 	integer='typeset -i'
 	local=typeset
 	login='exec login'
+	nameref='typeset -n'
 	nohup='nohup '
 	r='fc -e -'
 	source='PATH=$PATH:. command .'
@@ -4753,6 +4754,7 @@ expected-stdout:
 	integer='typeset -i'
 	local=typeset
 	login='exec login'
+	nameref='typeset -n'
 	nohup='nohup '
 	r='fc -e -'
 	source='PATH=$PATH:. command .'
@@ -4799,6 +4801,7 @@ expected-stdout:
 	integer='typeset -i'
 	local=typeset
 	login='exec login'
+	nameref='typeset -n'
 	nohup='nohup '
 	r='fc -e -'
 	source='PATH=$PATH:. command .'
@@ -4823,6 +4826,7 @@ expected-stdout:
 	integer='typeset -i'
 	local=typeset
 	login='exec login'
+	nameref='typeset -n'
 	nohup='nohup '
 	r='fc -e -'
 	source='PATH=$PATH:. command .'
@@ -4846,6 +4850,7 @@ expected-stdout:
 	integer='typeset -i'
 	local=typeset
 	login='exec login'
+	nameref='typeset -n'
 	nohup='nohup '
 	r='fc -e -'
 	source='PATH=$PATH:. command .'
@@ -4868,6 +4873,7 @@ expected-stdout:
 	integer='typeset -i'
 	local=typeset
 	login='exec login'
+	nameref='typeset -n'
 	nohup='nohup '
 	r='fc -e -'
 	source='PATH=$PATH:. command .'
@@ -5006,6 +5012,33 @@ expected-stdout:
 	g 0<0>1<1>2<4> = g
 	h 0<0>1<1>2<4> = h
 ---
+name: arrays-7
+description:
+	Check if we can get the array keys (indices) for indexed arrays,
+	Korn shell style, in some corner cases
+stdin:
+	print !arz: ${!arz}
+	print !arz[0]: ${!arz[0]}
+	print !arz[1]: ${!arz[1]}
+	arz=foo
+	print !arz: ${!arz}
+	print !arz[0]: ${!arz[0]}
+	print !arz[1]: ${!arz[1]}
+	unset arz
+	print !arz: ${!arz}
+	print !arz[0]: ${!arz[0]}
+	print !arz[1]: ${!arz[1]}
+expected-stdout:
+	!arz: 0
+	!arz[0]:
+	!arz[1]:
+	!arz: arz
+	!arz[0]: 0
+	!arz[1]:
+	!arz: 0
+	!arz[0]:
+	!arz[1]:
+---
 name: varexpand-substr-1
 description:
 	Check if bash-style substring expansion works
@@ -6100,3 +6133,81 @@ expected-stdout:
 	mypid=nz
 	=15
 ---
+name: nameref-1
+description:
+	Testsuite for nameref (bound variables)
+stdin:
+	bar=global
+	typeset -n ir2=bar
+	typeset -n ind=ir2
+	print !ind: ${!ind}
+	print ind: $ind
+	print !ir2: ${!ir2}
+	print ir2: $ir2
+	typeset +n ind
+	print !ind: ${!ind}
+	print ind: $ind
+	typeset -n ir2=ind
+	print !ir2: ${!ir2}
+	print ir2: $ir2
+	set|grep ^ir2|sed 's/^/s1: /'
+	typeset|grep ' ir2'|sed -e 's/^/s2: /' -e 's/nameref/typeset -n/'
+	blub=(e1 e2 e3)
+	typeset -n ind=blub
+	typeset -n ir2=blub[2]
+	print !ind[1]: ${!ind[1]}
+	print !ir2: $!ir2
+	print ind[1]: ${ind[1]}
+	print ir2: $ir2
+expected-stdout:
+	!ind: bar
+	ind: global
+	!ir2: bar
+	ir2: global
+	!ind: ind
+	ind: ir2
+	!ir2: ind
+	ir2: ir2
+	s1: ir2=ind
+	s2: typeset -n ir2
+	!ind[1]: 1
+	!ir2: ir2
+	ind[1]: e2
+	ir2: e3
+---
+name: nameref-2da
+description:
+	Testsuite for nameref (bound variables)
+	Functions, argument given directly, after local
+stdin:
+	function foo {
+		typeset bar=lokal baz=auch
+		typeset -n v=bar
+		print entering
+		print !v: ${!v}
+		print !bar: ${!bar}
+		print !baz: ${!baz}
+		print bar: $bar
+		print v: $v
+		v=123
+		print bar: $bar
+		print v: $v
+		print exiting
+	}
+	bar=global
+	print bar: $bar
+	foo bar
+	print bar: $bar
+expected-stdout:
+	bar: global
+	entering
+	!v: bar
+	!bar: bar
+	!baz: baz
+	bar: lokal
+	v: lokal
+	bar: 123
+	v: 123
+	exiting
+	bar: global
+---

eval.c

@@ -22,7 +22,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/eval.c,v 1.68 2009/08/28 22:39:09 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/eval.c,v 1.69 2009/09/06 17:42:12 tg Exp $");
 
 /*
  * string expansion
@@ -1021,8 +1021,7 @@ varsub(Expand *xp, const char *sp, const char *word,
 				if (!(vp->flag&ISSET))
 					continue;
 				XPput(wv, c == '!' ? shf_smprintf("%lu",
-				    vp->flag & AINDEX ?
+				    arrayindex(vp)) :
-				    (unsigned long)vp->ua.index : 0) :
 				    str_val(vp));
 			}
 			if (XPsize(wv) == 0) {
@@ -1037,14 +1036,27 @@ varsub(Expand *xp, const char *sp, const char *word,
 				state = XARG;
 			}
 		} else {
-			if (*sp == '!' && sp[1])
-				return (-1);
 			/* Can't assign things like $! or $1 */
 			if ((stype & 0x7f) == '=' &&
 			    ctype(*sp, C_VAR1 | C_DIGIT))
 				return (-1);
+			if (*sp == '!' && sp[1]) {
+				++sp;
+				xp->var = global(sp);
+				if (cstrchr(sp, '[')) {
+					if (xp->var->flag & ISSET)
+						xp->str = shf_smprintf("%lu",
+						    arrayindex(xp->var));
+					else
+						xp->str = null;
+				} else if (xp->var->flag & ISSET)
+					xp->str = xp->var->name;
+				else
+					xp->str = "0";	/* ksh93 compat */
+			} else {
 				xp->var = global(sp);
 				xp->str = str_val(xp->var);
+			}
 			state = XSUB;
 		}
 	}

funcs.c

@@ -25,7 +25,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/funcs.c,v 1.128 2009/08/30 21:02:00 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/funcs.c,v 1.129 2009/09/06 17:42:12 tg Exp $");
 
 #if HAVE_KILLPG
 /*
@@ -47,6 +47,8 @@ __RCSID("$MirOS: src/bin/mksh/funcs.c,v 1.128 2009/08/30 21:02:00 tg Exp $");
 #define c_ulimit c_label
 #endif
 
+extern uint8_t set_refflag;
+
 /* A leading = means assignments before command are kept;
  * a leading * means a POSIX special builtin;
  * a leading + means a POSIX regular builtin
@@ -896,7 +898,7 @@ c_typeset(const char **wp)
 	}
 
 	/* see comment below regarding possible opions */
-	opts = istset ? "L#R#UZ#afi#lprtux" : "p";
+	opts = istset ? "L#R#UZ#afi#lnprtux" : "p";
 
 	fieldstr = basestr = NULL;
 	builtin_opt.flags |= GF_PLUSOPT;
@@ -947,6 +949,9 @@ c_typeset(const char **wp)
 		case 'l':
 			flag = LCASEV;
 			break;
+		case 'n':
+			set_refflag = (builtin_opt.info & GI_PLUS) ? 2 : 1;
+			break;
 		case 'p':
 			/* export, readonly: POSIX -p flag */
 			/* typeset: show values as well */
@@ -995,13 +1000,14 @@ c_typeset(const char **wp)
 		builtin_opt.optind++;
 	}
 
-	if (func && ((fset|fclr) & ~(TRACE|UCASEV_AL|EXPORT))) {
+	if (func && (((fset|fclr) & ~(TRACE|UCASEV_AL|EXPORT)) || set_refflag)) {
 		bi_errorf("only -t, -u and -x options may be used with -f");
+		set_refflag = 0;
 		return (1);
 	}
 	if (wp[builtin_opt.optind]) {
 		/* Take care of exclusions.
-		 * At this point, flags in fset are cleared in fclr and vise
+		 * At this point, flags in fset are cleared in fclr and vice
 		 * versa. This property should be preserved.
 		 */
 		if (fset & LCASEV)	/* LCASEV has priority over UCASEV_AL */
@@ -1015,8 +1021,8 @@ c_typeset(const char **wp)
 		/* Setting these attributes clears the others, unless they
 		 * are also set in this command
 		 */
-		if (fset & (LJUST | RJUST | ZEROFIL | UCASEV_AL | LCASEV |
+		if ((fset & (LJUST | RJUST | ZEROFIL | UCASEV_AL | LCASEV |
-		    INTEGER | INT_U | INT_L))
+		    INTEGER | INT_U | INT_L)) || set_refflag)
 			fclr |= ~fset & (LJUST | RJUST | ZEROFIL | UCASEV_AL |
 			    LCASEV | INTEGER | INT_U | INT_L);
 	}
@@ -1047,9 +1053,11 @@ c_typeset(const char **wp)
 					    "%s() %T\n", wp[i], f->val.t);
 			} else if (!typeset(wp[i], fset, fclr, field, base)) {
 				bi_errorf("%s: not identifier", wp[i]);
+				set_refflag = 0;
 				return (1);
 			}
 		}
+		set_refflag = 0;
 		return (rv);
 	}
 
@@ -1112,6 +1120,8 @@ c_typeset(const char **wp)
 						 * be suitable for re-entry...
 						 */
 						shf_puts("typeset ", shl_stdout);
+						if (((vp->flag&(ARRAY|ASSOC))==ASSOC))
+							shf_puts("-n ", shl_stdout);
 						if ((vp->flag&INTEGER))
 							shf_puts("-i ", shl_stdout);
 						if ((vp->flag&EXPORT))
@@ -1160,9 +1170,7 @@ c_typeset(const char **wp)
 						if ((vp->flag&ARRAY) && any_set)
 							shprintf("%s[%lu]",
 							    vp->name,
-							    vp->flag & AINDEX ?
+							    arrayindex(vp));
-							    (unsigned long)vp->ua.index :
-							    0);
 						else
 							shf_puts(vp->name, shl_stdout);
 						if (thing == '-' && (vp->flag&ISSET)) {

main.c

@@ -33,7 +33,7 @@
 #include <locale.h>
 #endif
 
-__RCSID("$MirOS: src/bin/mksh/main.c,v 1.143 2009/08/29 11:26:44 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/main.c,v 1.144 2009/09/06 17:42:13 tg Exp $");
 
 extern char **environ;
 
@@ -67,6 +67,7 @@ static const char *initcoms[] = {
 	"autoload=typeset -fu",
 	"functions=typeset -f",
 	"history=fc -l",
+	"nameref=typeset -n",
 	"nohup=nohup ",
 	r_fc_e_,
 	"source=PATH=$PATH:. command .",

mksh.1

@@ -1,4 +1,4 @@
-.\" $MirOS: src/bin/mksh/mksh.1,v 1.180 2009/09/05 17:12:49 tg Exp $
+.\" $MirOS: src/bin/mksh/mksh.1,v 1.181 2009/09/06 17:42:13 tg Exp $
 .\" $OpenBSD: ksh.1,v 1.129 2009/05/28 06:09:06 jmc Exp $
 .\"-
 .\" Copyright © 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009
@@ -48,7 +48,7 @@
 .el .xD \\$1 \\$2 \\$3 \\$4 \\$5 \\$6 \\$7 \\$8
 ..
 .\"-
-.Dd $Mdocdate: September 5 2009 $
+.Dd $Mdocdate: September 6 2009 $
 .Dt MKSH 1
 .Os MirBSD
 .Sh NAME
@@ -883,6 +883,7 @@ history=\*(aqfc \-l\*(aq
 integer=\*(aqtypeset \-i\*(aq
 local=\*(aqtypeset\*(aq
 login=\*(aqexec login\*(aq
+nameref=\*(aqtypeset -n\*(aq
 nohup=\*(aqnohup \*(aq
 r=\*(aqfc \-e \-\*(aq
 stop=\*(aqkill \-STOP\*(aq
@@ -1261,6 +1262,18 @@ of the string value of parameter
 The number of elements in the array
 .Ar name .
 .Pp
+.It Pf ${! Ns Ar name Ns }
+The name of the variable referred to by
+.Ar name .
+This will be
+.Ar name
+except when
+.Ar name
+is a name reference (bound variable), created by the
+.Ic nameref
+command (which is an alias for
+.Ic typeset Fl n ) .
+.Pp
 .It Pf ${! Ns Ar name Ns [*]}
 .It Pf ${! Ns Ar name Ns [@]}
 The names of indices (keys) in the array
@@ -4110,7 +4123,7 @@ A command that exits with a zero value.
 .Pp
 .It Xo
 .Ic typeset
-.Oo Op Ic +\-alprtUux
+.Oo Op Ic +\-alpnrtUux
 .Op Fl L Ns Op Ar n
 .Op Fl R Ns Op Ar n
 .Op Fl Z Ns Op Ar n
@@ -4191,6 +4204,18 @@ All upper case characters in values are converted to lower case.
 when used with the
 .Fl i
 option.)
+.It Fl n
+Create a bound variable (name reference): any access to the variable
+.Ar name
+will access the variable
+.Ar value
+in the current scope (this is different from
+.At
+.Nm ksh93 ! )
+instead.
+This can be used by functions to access variables whose names are
+passed as parametres, instead of using
+.Ic eval .
 .It Fl p
 Print complete
 .Ic typeset

sh.h

@@ -134,7 +134,7 @@
 #endif
 
 #ifdef EXTERN
-__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.337 2009/08/30 21:02:01 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.338 2009/09/06 17:42:14 tg Exp $");
 #endif
 #define MKSH_VERSION "R39 2009/08/30"
 
@@ -906,6 +906,7 @@ struct tbl {			/* table item */
 #define EXPRINEVAL	BIT(23)	/* contents currently being evaluated */
 #define EXPRLVALUE	BIT(24)	/* useable as lvalue (temp flag) */
 #define AINDEX		BIT(25) /* array index >0 = ua.index filled in */
+#define ASSOC		BIT(26) /* ARRAY ? associative : reference */
 /* flag bits used for taliases/builtins/aliases/keywords/functions */
 #define KEEPASN		BIT(8)	/* keep command assignments (eg, var=x cmd) */
 #define FINUSE		BIT(9)	/* function being executed */
@@ -919,6 +920,9 @@ struct tbl {			/* table item */
 #define USERATTRIB	(EXPORT|INTEGER|RDONLY|LJUST|RJUST|ZEROFIL|\
 			    LCASEV|UCASEV_AL|INT_U|INT_L)
 
+#define arrayindex(vp)	((unsigned long)((vp)->flag & AINDEX ? \
+			    (vp)->ua.index : 0))
+
 /* command types */
 #define CNONE		0	/* undefined */
 #define CSHELL		1	/* built-in */

var.c

@@ -22,7 +22,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/var.c,v 1.86 2009/08/28 22:44:47 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/var.c,v 1.87 2009/09/06 17:42:15 tg Exp $");
 
 /*
  * Variables
@@ -47,6 +47,8 @@ static const char *array_index_calc(const char *, bool *, uint32_t *);
 static int rnd_get(void);
 static void rnd_set(unsigned long);
 
+uint8_t set_refflag = 0;
+
 /*
  * create a new block for function calls and simple commands
  * assume caller has allocated and set up e->loc
@@ -135,7 +137,7 @@ initvar(void)
 }
 
 /* Used to calculate an array index for global()/local(). Sets *arrayp to
- * non-zero if this is an array, sets *valp to the array index, returns
+ * true if this is an array, sets *valp to the array index, returns
  * the basename of the array.
  */
 static const char *
@@ -143,9 +145,39 @@ array_index_calc(const char *n, bool *arrayp, uint32_t *valp)
 {
 	const char *p;
 	int len;
+	char *ap = NULL;
 
 	*arrayp = false;
+ redo_from_ref:
 	p = skip_varname(n, false);
+	if (!set_refflag && (p != n) && ksh_isalphx(n[0])) {
+		struct block *l = e->loc;
+		struct tbl *vp;
+		char *vn;
+		uint32_t h;
+
+		strndupx(vn, n, p - n, ATEMP);
+		h = hash(vn);
+		/* check if this is a reference */
+		for (l = e->loc; ; l = l->next) {
+			if ((vp = ktsearch(&l->vars, vn, h)) != NULL)
+				break;
+			if (l->next == NULL)
+				break;
+		}
+		afree(vn, ATEMP);
+		if (vp && (vp->flag & (DEFINED|ASSOC|ARRAY)) ==
+		    (DEFINED|ASSOC)) {
+			char *cp;
+
+			/* gotcha! */
+			cp = shf_smprintf("%s%s", str_val(vp), p);
+			afree(ap, ATEMP);
+			n = ap = cp;
+			goto redo_from_ref;
+		}
+	}
+
 	if (p != n && *p == '[' && (len = array_ref_len(p))) {
 		char *sub, *tmp;
 		mksh_ari_t rval;
@@ -646,6 +678,9 @@ typeset(const char *var, Tflag set, Tflag clr, int field, int base)
 	if (*val == '[') {
 		int len;
 
+		if (set_refflag)
+			errorf("%s: reference variable cannot be an array",
+			    var);
 		len = array_ref_len(val);
 		if (len == 0)
 			return (NULL);
@@ -680,6 +715,26 @@ typeset(const char *var, Tflag set, Tflag clr, int field, int base)
 
 	vp = (set&LOCAL) ? local(tvar, (set & LOCAL_COPY) ? true : false) :
 	    global(tvar);
+	if (set_refflag == 2 && (vp->flag & (ARRAY|ASSOC)) == ASSOC)
+		vp->flag &= ~ASSOC;
+	else if (set_refflag == 1) {
+		if (vp->flag & ARRAY) {
+			struct tbl *a, *tmp;
+
+			/* Free up entire array */
+			for (a = vp->u.array; a; ) {
+				tmp = a;
+				a = a->u.array;
+				if (tmp->flag & ALLOC)
+					afree(tmp->val.s, tmp->areap);
+				afree(tmp, tmp->areap);
+			}
+			vp->u.array = NULL;
+			vp->flag &= ~ARRAY;
+		}
+		vp->flag |= ASSOC;
+	}
+
 	set &= ~(LOCAL|LOCAL_COPY);
 
 	vpbase = (vp->flag & ARRAY) ? global(arrayname(var)) : vp;
@@ -803,8 +858,8 @@ unset(struct tbl *vp, int array_ref)
 }
 
 /* return a pointer to the first char past a legal variable name (returns the
- * argument if there is no legal name, returns * a pointer to the terminating
+ * argument if there is no legal name, returns a pointer to the terminating
- * null if whole string is legal).
+ * NUL if whole string is legal).
  */
 const char *
 skip_varname(const char *s, int aok)
@@ -1240,7 +1295,7 @@ arraysearch(struct tbl *vp, uint32_t val)
 	struct tbl *prev, *curr, *new;
 	size_t len;
 
-	vp->flag |= ARRAY|DEFINED;
+	vp->flag = (vp->flag | (ARRAY|DEFINED)) & ~ASSOC;
 	/* The table entry is always [0] */
 	if (val == 0)
 		return (vp);