New functionality: assign here document to string variable directly,

without cat and temp files. Hacked in Lëtzebuerg ☺

This was the third try. Where to put this was not palpable… same for =(…)

check.t

@@ -1,9 +1,9 @@
-# $MirOS: src/bin/mksh/check.t,v 1.397 2010/12/19 20:00:53 tg Exp $
+# $MirOS: src/bin/mksh/check.t,v 1.398 2011/01/09 21:57:22 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 $
 #-
-# Copyright © 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
+# Copyright © 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
 #	Thorsten Glaser <tg@mirbsd.org>
 #
 # Provided that these terms and disclaimer and all copyright notices
@@ -25,7 +25,7 @@
 # http://www.research.att.com/~gsf/public/ifs.sh
 
 expected-stdout:
-	@(#)MIRBSD KSH R39 2010/12/19
+	@(#)MIRBSD KSH R39 2011/01/08
 description:
 	Check version of shell.
 stdin:
@@ -2007,6 +2007,45 @@ stdin:
 expected-stdout:
 	one
 ---
+name: heredoc-10
+description:
+	Check direct here document assignment
+stdin:
+	x=u
+	va=<<EOF
+	=a $x \x40=
+	EOF
+	vb=<<'EOF'
+	=b $x \x40=
+	EOF
+	function foo {
+		vc=<<-EOF
+			=c $x \x40=
+		EOF
+	}
+	typeset -f foo
+	foo
+	# rather nonsensical, but…
+	vd=<<<"=d $x \x40="
+	ve=<<<'=e $x \x40='
+	vf=<<<$'=f $x \x40='
+	# now check
+	print -r -- "| va={$va} vb={$vb} vc={$vc} vd={$vd} ve={$ve} vf={$vf} |"
+expected-stdout:
+	function foo {
+	    vc= <<- EOF 
+	=c $x \x40=
+	EOF
+	
+	} 
+	| va={=a u \x40=
+	} vb={=b $x \x40=
+	} vc={=c u \x40=
+	} vd={=d u \x40=
+	} ve={=e $x \x40=
+	} vf={=f $x @=
+	} |
+---
 name: heredoc-quoting-unsubst
 description:
 	Check for correct handling of quoted characters in

exec.c

@@ -1,7 +1,7 @@
 /*	$OpenBSD: exec.c,v 1.49 2009/01/29 23:27:26 jaredy Exp $	*/
 
 /*-
- * Copyright (c) 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
+ * Copyright (c) 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
  *	Thorsten Glaser <tg@mirbsd.org>
  *
  * Provided that these terms and disclaimer and all copyright notices
@@ -22,18 +22,18 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/exec.c,v 1.83 2010/09/15 21:08:17 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/exec.c,v 1.84 2011/01/09 21:57:25 tg Exp $");
 
 #ifndef MKSH_DEFAULT_EXECSHELL
 #define MKSH_DEFAULT_EXECSHELL	"/bin/sh"
 #endif
 
-static int comexec(struct op *, struct tbl *volatile, const char **,
+static int comexec(struct op *, struct tbl * volatile, const char **,
     int volatile, volatile int *);
 static void scriptexec(struct op *, const char **) MKSH_A_NORETURN;
 static int call_builtin(struct tbl *, const char **);
 static int iosetup(struct ioword *, struct tbl *);
-static int herein(const char *, int);
+static int herein(const char *, int, char **);
 static const char *do_selectargs(const char **, bool);
 static Test_op dbteste_isa(Test_env *, Test_meta);
 static const char *dbteste_getopnd(Test_env *, Test_op, bool);
@@ -43,7 +43,7 @@ static void dbteste_error(Test_env *, int, const char *);
  * execute command tree
  */
 int
-execute(struct op *volatile t,
+execute(struct op * volatile t,
     volatile int flags,		/* if XEXEC don't fork */
     volatile int * volatile xerrok)
 {
@@ -52,9 +52,10 @@ execute(struct op *volatile t,
 	int pv[2];
 	const char ** volatile ap;
 	char ** volatile up;
-	const char *s, *cp;
+	const char *s, *ccp;
 	struct ioword **iowp;
 	struct tbl *tp = NULL;
+	char *cp;
 
 	if (t == NULL)
 		return (0);
@@ -71,7 +72,48 @@ execute(struct op *volatile t,
 	if (trap)
 		runtraps(0);
 
-	if (t->type == TCOM) {
+	if (t->type == TCOM /* we want to run an executable... */) {
+		/* check if this is 'var=<<EOF' */
+		if (
+		    /* we have zero arguments, i.e. no programme to run */
+		    t->args[0] == NULL &&
+		    /* we have exactly one variable assignment */
+		    t->vars[0] != NULL && t->vars[1] == NULL &&
+		    /* we have exactly one I/O redirection */
+		    t->ioact != NULL && t->ioact[0] != NULL &&
+		    t->ioact[1] == NULL &&
+		    /* of type "here document" (or "here string") */
+		    (t->ioact[0]->flag & IOTYPE) == IOHERE &&
+		    /* the variable assignment begins with a valid varname */
+		    (ccp = skip_wdvarname(t->vars[0], true)) != t->vars[0] &&
+		    /* and has no right-hand side (i.e. "varname=") */
+		    ccp[0] == CHAR && ccp[1] == '=' && ccp[2] == EOS &&
+		    /* plus we can have a here document content */
+		    herein(t->ioact[0]->heredoc, t->ioact[0]->flag & IOEVAL,
+		    &cp) == 0 && cp && *cp) {
+			char *sp = cp, *dp;
+			size_t n = ccp - t->vars[0] + 2, z;
+
+			/* drop redirection (will be garbage collected) */
+			t->ioact = NULL;
+
+			/* set variable to its expanded value */
+			z = strlen(cp) + 1;
+			if (notoktomul(z, 2) || notoktoadd(z * 2, n))
+				internal_errorf(T_oomem, (unsigned long)-1);
+			dp = alloc(z * 2 + n, ATEMP);
+			memcpy(dp, t->vars[0], n);
+			t->vars[0] = dp;
+			dp += n;
+			while (*sp) {
+				*dp++ = QCHAR;
+				*dp++ = *sp++;
+			}
+			*dp = EOS;
+			/* free the expanded value */
+			afree(cp, APERM);
+		}
+
 		/* Clear subst_exstat before argument expansion. Used by
 		 * null commands (see comexec() and c_eval()) and by c_set().
 		 */
@@ -293,12 +335,12 @@ execute(struct op *volatile t,
 			}
 		} else { /* TSELECT */
 			for (;;) {
-				if (!(cp = do_selectargs(ap, is_first))) {
+				if (!(ccp = do_selectargs(ap, is_first))) {
 					rv = 1;
 					break;
 				}
 				is_first = false;
-				setstr(global(t->str), cp, KSH_UNWIND_ERROR);
+				setstr(global(t->str), ccp, KSH_UNWIND_ERROR);
 				execute(t->left, flags & XERROK, xerrok);
 			}
 		}
@@ -337,11 +379,11 @@ execute(struct op *volatile t,
 		break;
 
 	case TCASE:
-		cp = evalstr(t->str, DOTILDE);
+		ccp = evalstr(t->str, DOTILDE);
 		for (t = t->left; t != NULL && t->type == TPAT; t = t->right)
 		    for (ap = (const char **)t->vars; *ap; ap++)
 			if ((s = evalstr(*ap, DOTILDE|DOPAT)) &&
-			    gmatchx(cp, s, false))
+			    gmatchx(ccp, s, false))
 				goto Found;
 		break;
  Found:
@@ -400,7 +442,7 @@ execute(struct op *volatile t,
  */
 
 static int
-comexec(struct op *t, struct tbl *volatile tp, const char **ap,
+comexec(struct op *t, struct tbl * volatile tp, const char **ap,
     volatile int flags, volatile int *xerrok)
 {
 	int i;
@@ -566,7 +608,7 @@ comexec(struct op *t, struct tbl *volatile tp, const char **ap,
 	case CFUNC: {			/* function call */
 		volatile unsigned char old_xflag;
 		volatile Tflag old_inuse;
-		const char *volatile old_kshname;
+		const char * volatile old_kshname;
 
 		if (!(tp->flag & ISSET)) {
 			struct tbl *ftp;
@@ -1171,7 +1213,7 @@ iosetup(struct ioword *iop, struct tbl *tp)
 	case IOHERE:
 		do_open = 0;
 		/* herein() returns -2 if error has been printed */
-		u = herein(iop->heredoc, iop->flag & IOEVAL);
+		u = herein(iop->heredoc, iop->flag & IOEVAL, NULL);
 		/* cp may have wrong name */
 		break;
 
@@ -1261,66 +1303,91 @@ iosetup(struct ioword *iop, struct tbl *tp)
 }
 
 /*
- * open here document temp file.
+ * Process here documents by providing the content, either as
- * if unquoted here, expand here temp file into second temp file.
+ * result (globally allocated) string or in a temp file; if
+ * unquoted, the string is expanded first.
  */
 static int
-herein(const char *content, int sub)
+hereinval(const char *content, int sub, char **resbuf, struct shf *shf)
 {
-	volatile int fd = -1;
+	const char *ccp;
-	struct source *s, *volatile osource;
+	struct source *s, *osource;
-	struct shf *volatile shf;
-	struct temp *h;
-	int i;
-
-	/* ksh -c 'cat << EOF' can cause this... */
-	if (content == NULL) {
-		warningf(true, "%s missing", "here document");
-		return (-2); /* special to iosetup(): don't print error */
-	}
-
-	/* Create temp file to hold content (done before newenv so temp
-	 * doesn't get removed too soon).
-	 */
-	h = maketemp(ATEMP, TT_HEREDOC_EXP, &e->temps);
-	if (!(shf = h->shf) || (fd = open(h->name, O_RDONLY, 0)) < 0) {
-		fd = errno;
-		warningf(true, "can't %s temporary file %s: %s",
-		    !shf ? "create" : "open",
-		    h->name, strerror(fd));
-		if (shf)
-			shf_close(shf);
-		return (-2 /* special to iosetup(): don't print error */);
-	}
 
 	osource = source;
 	newenv(E_ERRH);
-	i = sigsetjmp(e->jbuf, 0);
+	if (sigsetjmp(e->jbuf, 0)) {
-	if (i) {
 		source = osource;
 		quitenv(shf);
-		close(fd);
+		/* special to iosetup(): don't print error */
-		return (-2); /* special to iosetup(): don't print error */
+		return (-2);
 	}
 	if (sub) {
-		/* Do substitutions on the content of heredoc */
+		/* do substitutions on the content of heredoc */
 		s = pushs(SSTRING, ATEMP);
 		s->start = s->str = content;
 		source = s;
 		if (yylex(ONEWORD|HEREDOC) != LWORD)
 			internal_errorf("%s: %s", "herein", "yylex");
 		source = osource;
-		shf_puts(evalstr(yylval.cp, 0), shf);
+		ccp = evalstr(yylval.cp, 0);
 	} else
-		shf_puts(content, shf);
+		ccp = content;
+
+	if (resbuf == NULL)
+		shf_puts(ccp, shf);
+	else
+		strdupx(*resbuf, ccp, APERM);
 
 	quitenv(NULL);
+	return (0);
+}
+
+static int
+herein(const char *content, int sub, char **resbuf)
+{
+	int fd = -1;
+	struct shf *shf;
+	struct temp *h;
+	int i;
+
+	/* ksh -c 'cat << EOF' can cause this... */
+	if (content == NULL) {
+		warningf(true, "%s missing", "here document");
+		/* special to iosetup(): don't print error */
+		return (-2);
+	}
+
+	/* skip all the fd setup if we just want the value */
+	if (resbuf != NULL)
+		return (hereinval(content, sub, resbuf, NULL));
+
+	/*
+	 * Create temp file to hold content (done before newenv
+	 * so temp doesn't get removed too soon).
+	 */
+	h = maketemp(ATEMP, TT_HEREDOC_EXP, &e->temps);
+	if (!(shf = h->shf) || (fd = open(h->name, O_RDONLY, 0)) < 0) {
+		i = errno;
+		warningf(true, "can't %s temporary file %s: %s",
+		    !shf ? "create" : "open", h->name, strerror(i));
+		if (shf)
+			shf_close(shf);
+		/* special to iosetup(): don't print error */
+		return (-2);
+	}
+
+	if (hereinval(content, sub, NULL, shf) == -2) {
+		close(fd);
+		/* special to iosetup(): don't print error */
+		return (-2);
+	}
 
 	if (shf_close(shf) == EOF) {
 		i = errno;
 		close(fd);
 		warningf(true, "%s: %s: %s", "write", h->name, strerror(i));
-		return (-2);	/* special to iosetup(): don't print error */
+		/* special to iosetup(): don't print error */
+		return (-2);
 	}
 
 	return (fd);

funcs.c

@@ -4,7 +4,8 @@
 /*	$OpenBSD: c_ulimit.c,v 1.17 2008/03/21 12:51:19 millert Exp $	*/
 
 /*-
- * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,
+ *		 2010, 2011
  *	Thorsten Glaser <tg@mirbsd.org>
  *
  * Provided that these terms and disclaimer and all copyright notices
@@ -25,7 +26,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/funcs.c,v 1.164 2010/11/01 17:29:02 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/funcs.c,v 1.165 2011/01/09 21:57:26 tg Exp $");
 
 #if HAVE_KILLPG
 /*
@@ -1174,8 +1175,10 @@ c_typeset(const char **wp)
 							char *s = str_val(vp);
 
 							shf_putc('=', shl_stdout);
-							/* AT&T ksh can't have
+							/*
-							 * justified integers.. */
+							 * AT&T ksh can't have
+							 * justified integers...
+							 */
 							if ((vp->flag &
 							    (INTEGER|LJUST|RJUST)) ==
 							    INTEGER)
@@ -3464,8 +3467,7 @@ c_cat(const char **wp)
 
 	/* XXX uses malloc instead of lalloc (for alignment/speed) */
 	if ((buf = malloc(MKSH_CAT_BUFSIZ)) == NULL) {
-		bi_errorf("can't allocate %lu data bytes",
+		bi_errorf(T_oomem, (unsigned long)MKSH_CAT_BUFSIZ);
-		    (unsigned long)MKSH_CAT_BUFSIZ);
 		return (1);
 	}
 

lalloc.c

@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2009, 2010
+ * Copyright (c) 2009, 2010, 2011
  *	Thorsten Glaser <tg@mirbsd.org>
  *
  * Provided that these terms and disclaimer and all copyright notices
@@ -20,7 +20,7 @@
 
 #include "sh.h"
 
-__RCSID("$MirOS: src/bin/mksh/lalloc.c,v 1.13 2010/09/14 21:26:14 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/lalloc.c,v 1.14 2011/01/09 21:57:27 tg Exp $");
 
 /* build with CPPFLAGS+= -DUSE_REALLOC_MALLOC=0 on ancient systems */
 #if defined(USE_REALLOC_MALLOC) && (USE_REALLOC_MALLOC == 0)
@@ -70,7 +70,7 @@ findptr(ALLOC_ITEM **lpp, char *ptr, Area *ap)
 void *
 aresize2(void *ptr, size_t fac1, size_t fac2, Area *ap)
 {
-	if (fac1 && fac2 && (SIZE_MAX / fac1 < fac2))
+	if (notoktomul(fac1, fac2))
 		internal_errorf(T_intovfl, (unsigned long)fac1, '*',
 		    (unsigned long)fac2);
 	return (aresize(ptr, fac1 * fac2, ap));
@@ -95,8 +95,7 @@ aresize(void *ptr, size_t numb, Area *ap)
 	    || ALLOC_ISUNALIGNED(lp)
 #endif
 	    )
-		internal_errorf("can't allocate %lu data bytes",
+		internal_errorf(T_oomem, (unsigned long)numb);
-		    (unsigned long)numb);
 	/* this only works because Area is an ALLOC_ITEM */
 	lp->next = ap->next;
 	ap->next = lp;

misc.c

@@ -2,7 +2,7 @@
 /*	$OpenBSD: path.c,v 1.12 2005/03/30 17:16:37 deraadt Exp $	*/
 
 /*-
- * Copyright (c) 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
+ * Copyright (c) 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
  *	Thorsten Glaser <tg@mirbsd.org>
  *
  * Provided that these terms and disclaimer and all copyright notices
@@ -29,7 +29,7 @@
 #include <grp.h>
 #endif
 
-__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.148 2010/09/19 19:28:22 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.149 2011/01/09 21:57:27 tg Exp $");
 
 unsigned char chtypes[UCHAR_MAX + 1];	/* type bits for unsigned char */
 
@@ -921,25 +921,33 @@ void
 print_value_quoted(const char *s)
 {
 	const char *p;
-	int inquote = 0;
+	bool inquote = false;
 
-	/* Test if any quotes are needed */
+	/* first, check whether any quotes are needed */
 	for (p = s; *p; p++)
 		if (ctype(*p, C_QUOTE))
 			break;
 	if (!*p) {
+		/* nope, use the shortcut */
 		shf_puts(s, shl_stdout);
 		return;
 	}
+
+	/* quote via state machine */
 	for (p = s; *p; p++) {
 		if (*p == '\'') {
-			if (inquote)
+			/*
+			 * multiple '''s or any ' at beginning of string
+			 * look nicer this way than when simply substituting
+			 */
+			if (inquote) {
 				shf_putc('\'', shl_stdout);
+				inquote = false;
+			}
 			shf_putc('\\', shl_stdout);
-			inquote = 0;
 		} else if (!inquote) {
 			shf_putc('\'', shl_stdout);
-			inquote = 1;
+			inquote = true;
 		}
 		shf_putc(*p, shl_stdout);
 	}

mksh.1

@@ -1,7 +1,8 @@
-.\" $MirOS: src/bin/mksh/mksh.1,v 1.243 2010/12/19 20:00:55 tg Exp $
+.\" $MirOS: src/bin/mksh/mksh.1,v 1.244 2011/01/09 21:57:27 tg Exp $
 .\" $OpenBSD: ksh.1,v 1.138 2010/09/20 07:41:17 jmc Exp $
 .\"-
-.\" Copyright © 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
+.\" Copyright © 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,
+.\"		2010, 2011
 .\"	Thorsten Glaser <tg@mirbsd.org>
 .\"
 .\" Provided that these terms and disclaimer and all copyright notices
@@ -71,7 +72,7 @@
 .\" with -mandoc, it might implement .Mx itself, but we want to
 .\" use our own definition. And .Dd must come *first*, always.
 .\"
-.Dd $Mdocdate: December 19 2010 $
+.Dd $Mdocdate: January 9 2011 $
 .\"
 .\" Check which macro package we use
 .\"
@@ -6073,7 +6074,14 @@ code fails because of the parenthesis asymmetry
 .Pq RedHat BZ#496791 .
 A workaround exists; use
 .Ql x=$(cat) \*(Lt\*(Lt"EOF"
+or the newline-keeping
+.Ql x=\*(Lt\*(Lt"EOF"
 instead to merely slurp the string.
+Of course, in this case, the form
+.Ql x=$(case $foo in (bar) ...
+actually recommended by
+.St -p1003.1
+would work as well, but we use it as an example of this parser bug.
 .Bd -literal -offset indent
 x=$(case $foo in bar) echo $bar ;; *) echo $baz ;; esac)
 # above fails to parse; below is the workaround

sh.h

@@ -9,7 +9,7 @@
 /*	$OpenBSD: tty.h,v 1.5 2004/12/20 11:34:26 otto Exp $	*/
 
 /*-
- * Copyright (c) 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
+ * Copyright (c) 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
  *	Thorsten Glaser <tg@mirbsd.org>
  *
  * Provided that these terms and disclaimer and all copyright notices
@@ -154,9 +154,9 @@
 #endif
 
 #ifdef EXTERN
-__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.420 2010/12/19 20:00:56 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.421 2011/01/09 21:57:29 tg Exp $");
 #endif
-#define MKSH_VERSION "R39 2010/12/19"
+#define MKSH_VERSION "R39 2011/01/08"
 
 #ifndef MKSH_INCLUDES_ONLY
 
@@ -630,6 +630,7 @@ extern const struct shoption options[];
 EXTERN char null[] I__("");
 /* helpers for string pooling */
 EXTERN const char T_intovfl[] I__("integer overflow %lu %c %lu prevented");
+EXTERN const char T_oomem[] I__("can't allocate %lu data bytes");
 #if defined(__GNUC__)
 /* trust this to have string pooling; -Wformat bitches otherwise */
 #define T_synerr	"syntax error"
@@ -1199,7 +1200,7 @@ struct ioword {
 #define DB_BE	4	/* an inserted -BE */
 #define DB_PAT	5	/* a pattern argument */
 
-#define X_EXTRA	8	/* this many extra bytes in X string */
+#define X_EXTRA	20	/* this many extra bytes in X string */
 
 typedef struct XString {
 	char *end, *beg;	/* end, begin of string */
@@ -1397,6 +1398,8 @@ EXTERN int histsize;	/* history size */
 /* user and system time of last j_waitjed job */
 EXTERN struct timeval j_usrtime, j_systime;
 
+#define notoktomul(fac1, fac2)	((fac1) && (fac2) && \
+				    (SIZE_MAX / (fac1) < (fac2)))
 #define notoktoadd(val, cnst)	((val) > (SIZE_MAX - (cnst)))
 #define checkoktoadd(val, cnst) do {					\
 	if (notoktoadd((val), (cnst)))					\
@@ -1637,6 +1640,7 @@ int xstrcmp(const void *, const void *);
 void ksh_getopt_reset(Getopt *, int);
 int ksh_getopt(const char **, Getopt *, const char *);
 void print_value_quoted(const char *);
+char *quote_value(const char *);
 void print_columns(struct shf *, int,
     char *(*)(char *, int, int, const void *),
     const void *, int, int, bool);