 @@ -1,14 +1,14 @@
-/*	$NetBSD: parser.c,v 1.110 2016/03/27 14:36:29 christos Exp $	*/
+/*	$NetBSD: parser.c,v 1.111 2016/03/27 14:39:33 christos Exp $	*/
 /*-
  * Copyright (c) 1991, 1993
  *	The Regents of the University of California.  All rights reserved.
+ *
  * This code is derived from software contributed to Berkeley by
  * Kenneth Almquist.
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
  * 1. Redistributions of source code must retain the above copyright
  *    notice, this list of conditions and the following disclaimer.
 @@ -27,27 +27,27 @@
  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
  */
 #include <sys/cdefs.h>
 #ifndef lint
 #if 0
 static char sccsid[] = "@(#)parser.c	8.7 (Berkeley) 5/16/95";
 #else
-__RCSID("$NetBSD: parser.c,v 1.110 2016/03/27 14:36:29 christos Exp $");
+__RCSID("$NetBSD: parser.c,v 1.111 2016/03/27 14:39:33 christos Exp $");
 #endif
 #endif /* not lint */
 #include <stdio.h>
 #include <stdlib.h>
 #include <limits.h>
 #include "shell.h"
 #include "parser.h"
 #include "nodes.h"
 #include "expand.h"	/* defines rmescapes() */
 #include "eval.h"	/* defines commandname */
 #include "redir.h"	/* defines copyfd() */
 @@ -59,28 +59,26 @@ __RCSID("$NetBSD: parser.c,v 1.110 2016/
 #include "error.h"
 #include "memalloc.h"
 #include "mystring.h"
 #include "alias.h"
 #include "show.h"
 #ifndef SMALL
 #include "myhistedit.h"
 #endif
 /*
  * Shell command parser.
  */
 #define EOFMARKLEN 79
 /* values returned by readtoken */
 #include "token.h"
 #define OPENBRACE '{'
 #define CLOSEBRACE '}'
 struct heredoc {
 	struct heredoc *next;	/* next here document in list */
 	union node *here;		/* redirection node */
 	char *eofmark;		/* string indicating end of input */
 	int striptabs;		/* if set, strip leading tabs */
 };
 @@ -101,31 +99,32 @@ union node *redirnode;
 struct heredoc *heredoc;
 int quoteflag;			/* set if (part of) last token was quoted */
 int startlinno;			/* line # where last token started */
 int funclinno;			/* line # where the current function started */
 STATIC union node *list(int, int);
 STATIC union node *andor(void);
 STATIC union node *pipeline(void);
 STATIC union node *command(void);
 STATIC union node *simplecmd(union node **, union node *);
 STATIC union node *makename(void);
 STATIC void parsefname(void);
-STATIC void parseheredoc(void);
+STATIC void slurp_heredoc(char *const, int, int);
 STATIC void readheredocs(void);
 STATIC int peektoken(void);
 STATIC int readtoken(void);
 STATIC int xxreadtoken(void);
-STATIC int readtoken1(int, char const *, char *, int);
+STATIC int readtoken1(int, char const *, int);
 STATIC int noexpand(char *);
 STATIC void synexpect(int, const char *) __dead;
 STATIC void synerror(const char *) __dead;
 STATIC void setprompt(int);
 static const char EOFhere[] = "EOF reading here (<<) document";
 /*
  * Read and parse a command.  Returns NEOF on end of file.  (NULL is a
  * valid parse tree indicating a blank line.)
  */
 @@ -186,39 +185,39 @@ list(int nlflag, int erflag)
 			n3 = stalloc(sizeof(struct nbinary));
 			n3->type = NSEMI;
 			n3->nbinary.ch1 = n1;
 			n3->nbinary.ch2 = n2;
 			n1 = n3;
+		}
 		switch (tok) {
 		case TBACKGND:
 		case TSEMI:
 			tok = readtoken();
 			/* FALLTHROUGH */
 		case TNL:
 			if (tok == TNL) {
-				parseheredoc();
+				readheredocs();
 				if (nlflag)
 					return n1;
 			} else {
 				tokpushback++;
+			}
 			checkkwd = 2;
 			if (tokendlist[peektoken()])
 				return n1;
 			break;
 		case TEOF:
 			if (heredoclist)
-				parseheredoc();
+				readheredocs();
 			else
 				pungetc();	/* push back EOF on input */
 			return n1;
 		default:
 			if (nlflag || erflag)
 				synexpect(-1, 0);
 			tokpushback++;
 			return n1;
+		}
+	}
+}
 STATIC union node *
 @@ -661,89 +660,230 @@ fixredir(union node *n, const char *text
+}
 STATIC void
 parsefname(void)
+{
 	union node *n = redirnode;
 	if (readtoken() != TWORD)
 		synexpect(-1, 0);
 	if (n->type == NHERE) {
 		struct heredoc *here = heredoc;
 		struct heredoc *p;
 		int i;
 		if (quoteflag == 0)
 			n->type = NXHERE;
 		TRACE(("Here document %d\n", n->type));
 		if (here->striptabs) {
 			while (*wordtext == '\t')
 				wordtext++;
+		}
 		if (! noexpand(wordtext) || (i = strlen(wordtext)) == 0 || i > EOFMARKLEN)
 		/*
 		 * this test is not really necessary, we are not
 		 * required to expand wordtext, but there's no reason
 		 * it cannot be $$ or something like that - that would
 		 * not mean the pid, but literally two '$' characters.
 		 * There is no need for limits on what the word can be.
 		 * However, it needs to stay literal as entered, not
 		 * have $ converted to CTLVAR or something, which as
 		 * the parser is, at the minute, is impossible to prevent.
 		 * So, leave it like this until the rest of the parser is fixed.
 		 */
 		if (! noexpand(wordtext))
 			synerror("Illegal eof marker for << redirection");
 		rmescapes(wordtext);
 		here->eofmark = wordtext;
 		here->next = NULL;
 		if (heredoclist == NULL)
 			heredoclist = here;
 		else {
 			for (p = heredoclist ; p->next ; p = p->next)
 				continue;
 			p->next = here;
+		}
 	} else if (n->type == NTOFD || n->type == NFROMFD) {
 		fixredir(n, wordtext, 0);
 	} else {
 		n->nfile.fname = makename();
+	}
+}
 /*
  * Check to see whether we are at the end of the here document.  When this
  * is called, c is set to the first character of the next input line.  If
  * we are at the end of the here document, this routine sets the c to PEOF.
  * The new value of c is returned.
  */
 static int
 checkend(int c, char * const eofmark, const int striptabs)
+{
 	if (striptabs) {
 		while (c == '\t')
 			c = pgetc();
+	}
 	if (c == PEOF) {
 		if (*eofmark == '\0')
 			return (c);
 		synerror(EOFhere);
+	}
 	if (c == *eofmark) {
 		int c2;
 		char *q;
 		for (q = eofmark + 1; c2 = pgetc(), *q != '\0' && c2 == *q; q++)
+			;
 		if ((c2 == PEOF || c2 == '\n') && *q == '\0') {
 			c = PEOF;
 			if (c2 == '\n') {
 				plinno++;
 				needprompt = doprompt;
+			}
 		} else {
 			pungetc();
 			pushstring(eofmark + 1, q - (eofmark + 1), NULL);
+		}
 	} else if (c == '\n' && *eofmark == '\0') {
 		c = PEOF;
 		plinno++;
 		needprompt = doprompt;
+	}
 	return (c);
+}
 /*
  * Input any here documents.
  */
 STATIC void
-parseheredoc(void)
+slurp_heredoc(char *const eofmark, int striptabs, int sq)
+{
 	int c;
 	char *out;
 	c = pgetc();
 	/*
 	 * If we hit EOF on the input, and the eofmark is a null string ('')
 	 * we consider this empty line to be the eofmark, and exit without err.
 	 */
 	if (c == PEOF && *eofmark != '\0')
 		synerror(EOFhere);
 	STARTSTACKSTR(out);
 	while ((c = checkend(c, eofmark, striptabs)) != PEOF) {
 		do {
 			if (sq) {
 				/*
 				 * in single quoted mode (eofmark quoted)
 				 * all we look for is \n so we can check
 				 * for the epfmark - everything saved literally.
 				 */
 				STPUTC(c, out);
 				if (c == '\n')
 					break;
 				continue;
+			}
 			/*
 			 * In double quoted (non-quoted eofmark)
 			 * we must handle \ followed by \n here
 			 * otherwise we can mismatch the end mark.
 			 * All other uses of \ will be handled later
 			 * when the here doc is expanded.
+			 *
 			 * This also makes sure \\ followed by \n does
 			 * not suppress the newline (the \ quotes itself)
 			 */
 			if (c == '\\') {		/* A backslash */
 				c = pgetc();		/* followed by */
 				if (c == '\n')		/* a newline?  */
 					continue;	/* y:drop both */
 				STPUTC('\\', out);	/* else keep \ */
+			}
 			STPUTC(c, out);			/* keep the char */
 			if (c == '\n')			/* at end of line */
 				break;			/* look for eofmark */
 		} while ((c = pgetc()) != PEOF);
 		/*
 		 * If we have read a line, and reached EOF, without
 		 * finding the eofmark, whether the EOF comes before
 		 * or immediately after the \n, that is an error.
 		 */
 		if (c == PEOF || (c = pgetc()) == PEOF)
 			synerror(EOFhere);
+	}
 	STPUTC('\0', out);
 	c = out - stackblock();
 	out = stackblock();
 	grabstackblock(c);
 	wordtext = out;
 	TRACE(("Slurped a heredoc (to '%s')%s: len %d, \"%.16s\"...\n",
 		eofmark, striptabs ? " tab stripped" : "", c, wordtext));
+}
 STATIC void
 readheredocs(void)
+{
 	struct heredoc *here;
 	union node *n;
 	while (heredoclist) {
 		int c;
 		here = heredoclist;
 		heredoclist = here->next;
 		if (needprompt) {
 			setprompt(2);
 			needprompt = 0;
+		}
 		if ((c = pgetc()) == PEOF) {
-			synerror(EOFhere);
+		slurp_heredoc(here->eofmark, here->striptabs,
-			/* NOTREACHED */
+		    here->here->nhere.type == NHERE);
 		readtoken1(c, here->here->type == NHERE? SQSYNTAX : DQSYNTAX,
 		    here->eofmark, here->striptabs);
 		n = stalloc(sizeof(struct narg));
 		n->narg.type = NARG;
 		n->narg.next = NULL;
 		n->narg.text = wordtext;
 		n->narg.backquote = backquotelist;
 		here->here->nhere.doc = n;
+	}
+}
 void
 parse_heredoc(union node *n)
+{
 	if (n->narg.type != NARG)
 		abort();
 	if (n->narg.text[0] == '\0')		/* nothing to do */
 		return;
 	setinputstring(n->narg.text, 1);
 	readtoken1(pgetc(), DQSYNTAX, 1);
 	n->narg.text = wordtext;
 	n->narg.backquote = backquotelist;
 	popfile();
+}
 STATIC int
 peektoken(void)
+{
 	int t;
 	t = readtoken();
 	tokpushback++;
 	return (t);
+}
 STATIC int
 readtoken(void)
+{
 @@ -754,27 +894,27 @@ readtoken(void)
 #endif
 	struct alias *ap;
 	top:
 	t = xxreadtoken();
 	if (checkkwd) {
 		/*
 		 * eat newlines
 		 */
 		if (checkkwd == 2) {
 			checkkwd = 0;
 			while (t == TNL) {
-				parseheredoc();
+				readheredocs();
 				t = xxreadtoken();
+			}
 		} else
 			checkkwd = 0;
 		/*
 		 * check for keywords and aliases
 		 */
 		if (t == TWORD && !quoteflag) {
 			const char *const *pp;
 			for (pp = parsekwd; *pp; pp++) {
 				if (**pp == *wordtext && equal(*pp, wordtext)) {
 					lasttoken = t = pp -
 @@ -877,27 +1017,27 @@ xxreadtoken(void)
 				if (doprompt)
 					setprompt(2);
 				else
 					setprompt(0);
 				continue;
 			case PEOF:
 				RETURN(TEOF);
 			default:
 				pungetc();
 				break;
+			}
 			/* FALLTHROUGH */
 		default:
-			return readtoken1(c, BASESYNTAX, NULL, 0);
+			return readtoken1(c, BASESYNTAX, 0);
+		}
+	}
 #undef RETURN
+}
 /*
  * If eofmark is NULL, read a word or a redirection symbol.  If eofmark
  * is not NULL, read a here document.  In the latter case, eofmark is the
  * word which marks the end of the document and striptabs is true if
  * leading tabs should be stripped from the document.  The argument firstc
  * is the first character of the input token or document.
 @@ -1029,27 +1169,26 @@ drop_state_level(VSS *stack)
 	--stack->cur;
 	return stack;
+}
 static inline void
 cleanup_state_stack(VSS *stack)
+{
 	while (stack->prev != NULL) {
 		stack->cur = 0;
 		stack = drop_state_level(stack);
+	}
+}
 #define	CHECKEND()	{goto checkend; checkend_return:;}
 #define	PARSEREDIR()	{goto parseredir; parseredir_return:;}
 #define	PARSESUB()	{goto parsesub; parsesub_return:;}
 #define	PARSEARITH()	{goto parsearith; parsearith_return:;}
 /*
  * The following macros all assume the existance of a local var "stack"
  * which contains a pointer to the current struct stackstate
  */
 /*
  * These are macros rather than inline funcs to avoid code churn as much
  * as possible - they replace macros of the same name used previously.
  */
 @@ -1222,149 +1361,145 @@ done:
 		INTON;
+	}
 	parsebackquote = savepbq;
 	handler = savehandler;
 	if (arinest || ISDBLQUOTE())
 		USTPUTC(CTLBACKQ | CTLQUOTE, out);
 	else
 		USTPUTC(CTLBACKQ, out);
 	return out;
+}
 STATIC int
-readtoken1(int firstc, char const *syn, char *eofmark, int striptabs)
+readtoken1(int firstc, char const *syn, int magicq)
+{
 	int c = firstc;
 	char * out;
 	int len;
 	char line[EOFMARKLEN + 1];
 	struct nodelist *bqlist;
 	int quotef;
 	VSS static_stack;
 	VSS *stack = &static_stack;
 	stack->prev = NULL;
 	stack->cur = 0;
 	syntax = syn;
 	startlinno = plinno;
 	varnest = 0;
 	quoted = 0;
 	if (syntax == DQSYNTAX)
 		SETDBLQUOTE();
 	quotef = 0;
 	bqlist = NULL;
 	arinest = 0;
 	parenlevel = 0;
 	STARTSTACKSTR(out);
 	loop: {	/* for each line, until end of word */
 		CHECKEND();	/* set c to PEOF if at end of here document */
 		for (;;) {	/* until end of line or end of word */
 			CHECKSTRSPACE(4, out);	/* permit 4 calls to USTPUTC */
 			switch(syntax[c]) {
 			case CNL:	/* '\n' */
 				if (syntax == BASESYNTAX)
 					goto endword;	/* exit outer loop */
 				USTPUTC(c, out);
 				plinno++;
 				if (doprompt)
 					setprompt(2);
 				else
 					setprompt(0);
 				c = pgetc();
 				goto loop;		/* continue outer loop */
 			case CWORD:
 				USTPUTC(c, out);
 				break;
 			case CCTL:
-				if (eofmark == NULL || ISDBLQUOTE())
+				if (!magicq || ISDBLQUOTE())
 					USTPUTC(CTLESC, out);
 				USTPUTC(c, out);
 				break;
 			case CBACK:	/* backslash */
 				c = pgetc();
 				if (c == PEOF) {
 					USTPUTC('\\', out);
 					pungetc();
 					break;
+				}
 				if (c == '\n') {
 					plinno++;
 					if (doprompt)
 						setprompt(2);
 					else
 						setprompt(0);
 					break;
+				}
 				quotef = 1;
 				if (ISDBLQUOTE() && c != '\\' &&
 				    c != '`' && c != '$' &&
-				    (c != '"' || eofmark != NULL))
+				    (c != '"' || magicq))
 					USTPUTC('\\', out);
 				if (SQSYNTAX[c] == CCTL)
 					USTPUTC(CTLESC, out);
-				else if (eofmark == NULL) {
+				else if (!magicq) {
 					USTPUTC(CTLQUOTEMARK, out);
 					USTPUTC(c, out);
 					if (varnest != 0)
 						USTPUTC(CTLQUOTEEND, out);
 					break;
+				}
 				USTPUTC(c, out);
 				break;
 			case CSQUOTE:
 				if (syntax != SQSYNTAX) {
-					if (eofmark == NULL)
+					if (!magicq)
 						USTPUTC(CTLQUOTEMARK, out);
 					quotef = 1;
 					TS_PUSH();
 					syntax = SQSYNTAX;
 					quoted = SQ;
 					break;
+				}
-				if (eofmark != NULL && arinest == 0 &&
+				if (magicq && arinest == 0 && varnest == 0) {
 				    varnest == 0) {
 					/* Ignore inside quoted here document */
 					USTPUTC(c, out);
 					break;
+				}
 				/* End of single quotes... */
 				TS_POP();
 				if (syntax == BASESYNTAX && varnest != 0)
 					USTPUTC(CTLQUOTEEND, out);
 				break;
 			case CDQUOTE:
-				if (eofmark != NULL && arinest == 0 &&
+				if (magicq && arinest == 0 && varnest == 0) {
 				    varnest == 0) {
 					/* Ignore inside here document */
 					USTPUTC(c, out);
 					break;
+				}
 				quotef = 1;
 				if (arinest) {
 					if (ISDBLQUOTE()) {
 						TS_POP();
 					} else {
 						TS_PUSH();
 						syntax = DQSYNTAX;
 						SETDBLQUOTE();
 						USTPUTC(CTLQUOTEMARK, out);
+					}
 					break;
+				}
-				if (eofmark != NULL)
+				if (magicq)
 					break;
 				if (ISDBLQUOTE()) {
 					TS_POP();
 					if (varnest != 0)
 						USTPUTC(CTLQUOTEEND, out);
 				} else {
 					TS_PUSH();
 					syntax = DQSYNTAX;
 					SETDBLQUOTE();
 					USTPUTC(CTLQUOTEMARK, out);
+				}
 				break;
 			case CVAR:	/* '$' */
 @@ -1411,96 +1546,59 @@ readtoken1(int firstc, char const *syn,
 			default:
 				if (varnest == 0 && !ISDBLQUOTE())
 					goto endword;	/* exit outer loop */
 				USTPUTC(c, out);
+			}
 			c = pgetc_macro();
+		}
+	}
 endword:
 	if (syntax == ARISYNTAX) {
 		cleanup_state_stack(stack);
 		synerror("Missing '))'");
+	}
-	if (syntax != BASESYNTAX && /* ! parsebackquote && */ eofmark == NULL) {
+	if (syntax != BASESYNTAX && /* ! parsebackquote && */ !magicq) {
 		cleanup_state_stack(stack);
 		synerror("Unterminated quoted string");
+	}
 	if (varnest != 0) {
 		cleanup_state_stack(stack);
 		startlinno = plinno;
 		/* { */
 		synerror("Missing '}'");
+	}
 	USTPUTC('\0', out);
 	len = out - stackblock();
 	out = stackblock();
-	if (eofmark == NULL) {
+	if (!magicq) {
-		if ((c == '>' || c == '<')
+		if ((c == '<' || c == '>')
 		 && quotef == 0
 		 && (*out == '\0' || is_number(out))) {
 			PARSEREDIR();
 			cleanup_state_stack(stack);
 			return lasttoken = TREDIR;
 		} else {
 			pungetc();
+		}
+	}
 	quoteflag = quotef;
 	backquotelist = bqlist;
 	grabstackblock(len);
 	wordtext = out;
 	cleanup_state_stack(stack);
 	return lasttoken = TWORD;
 /* end of readtoken routine */
 /*
  * Check to see whether we are at the end of the here document.  When this
  * is called, c is set to the first character of the next input line.  If
  * we are at the end of the here document, this routine sets the c to PEOF.
  */
 checkend: {
 	if (eofmark) {
 		if (c == PEOF)
 			synerror(EOFhere);
 		if (striptabs) {
 			while (c == '\t')
 				c = pgetc();
 		if (c == *eofmark) {
 			if (pfgets(line, sizeof line) != NULL) {
 				char *p, *q;
 				p = line;
 				for (q = eofmark + 1 ; *q && *p == *q ; p++, q++)
 					continue;
 				if ((*p == '\0' || *p == '\n') && *q == '\0') {
 					c = PEOF;
 					plinno++;
 					needprompt = doprompt;
 				} else {
 					pushstring(line, strlen(line), NULL);
 			} else
 				synerror(EOFhere);
 	goto checkend_return;
 /*
  * Parse a redirection operator.  The variable "out" points to a string
  * specifying the fd to be redirected.  The variable "c" contains the
  * first character of the redirection operator.
  */
 parseredir: {
 	char fd[64];
 	union node *np;
 	strlcpy(fd, out, sizeof(fd));
 	np = stalloc(sizeof(struct nfile));
 	if (c == '>') {

 @@ -1,14 +1,14 @@
-/*	$NetBSD: parser.h,v 1.19 2016/02/22 20:02:00 christos Exp $	*/
+/*	$NetBSD: parser.h,v 1.20 2016/03/27 14:39:33 christos Exp $	*/
 /*-
  * Copyright (c) 1991, 1993
  *	The Regents of the University of California.  All rights reserved.
+ *
  * This code is derived from software contributed to Berkeley by
  * Kenneth Almquist.
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
  * 1. Redistributions of source code must retain the above copyright
  *    notice, this list of conditions and the following disclaimer.
 @@ -71,15 +71,16 @@
 /*
  * NEOF is returned by parsecmd when it encounters an end of file.  It
  * must be distinct from NULL, so we use the address of a variable that
  * happens to be handy.
  */
 extern int tokpushback;
 #define NEOF ((union node *)&tokpushback)
 extern int whichprompt;		/* 1 == PS1, 2 == PS2 */
 union node *parsecmd(int);
 void fixredir(union node *, const char *, int);
 void parse_heredoc(union node *);
 int goodname(char *);
 const char *getprompt(void *);