===================================================================
RCS file: /cvs/mandoc/mandoc.c,v
retrieving revision 1.12
retrieving revision 1.78
diff -u -p -r1.12 -r1.78
--- mandoc/mandoc.c	2010/05/12 17:08:03	1.12
+++ mandoc/mandoc.c	2014/04/08 01:37:27	1.78
@@ -1,14 +1,15 @@
-/*	$Id: mandoc.c,v 1.12 2010/05/12 17:08:03 kristaps Exp $ */
+/*	$Id: mandoc.c,v 1.78 2014/04/08 01:37:27 schwarze Exp $ */
 /*
- * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
+ * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
+ * Copyright (c) 2011, 2012, 2013, 2014 Ingo Schwarze <schwarze@openbsd.org>
  *
  * Permission to use, copy, modify, and distribute this software for any
  * purpose with or without fee is hereby granted, provided that the above
  * copyright notice and this permission notice appear in all copies.
  *
- * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
- * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
@@ -22,232 +23,429 @@
 
 #include <assert.h>
 #include <ctype.h>
+#include <errno.h>
+#include <limits.h>
 #include <stdlib.h>
 #include <stdio.h>
 #include <string.h>
 #include <time.h>
 
+#include "mandoc.h"
+#include "mandoc_aux.h"
 #include "libmandoc.h"
 
-static int	 a2time(time_t *, const char *, const char *);
+#define DATESIZE 32
 
+static	int	 a2time(time_t *, const char *, const char *);
+static	char	*time2a(time_t);
 
-int
-mandoc_special(const char *p)
+
+enum mandoc_esc
+mandoc_escape(const char **end, const char **start, int *sz)
 {
-	int		 terminator;	/* Terminator for \s. */
-	int		 lim;		/* Limit for N in \s. */
-	int		 c, i;
-	
-	if ('\\' != *p++)
-		return(0);
+	const char	*local_start;
+	int		 local_sz;
+	char		 term;
+	enum mandoc_esc	 gly; 
 
-	switch (*p) {
-	case ('\''):
+	/*
+	 * When the caller doesn't provide return storage,
+	 * use local storage.
+	 */
+
+	if (NULL == start)
+		start = &local_start;
+	if (NULL == sz)
+		sz = &local_sz;
+
+	/*
+	 * Beyond the backslash, at least one input character
+	 * is part of the escape sequence.  With one exception
+	 * (see below), that character won't be returned.
+	 */
+
+	gly = ESCAPE_ERROR;
+	*start = ++*end;
+	*sz = 0;
+	term = '\0';
+
+	switch ((*start)[-1]) {
+	/*
+	 * First the glyphs.  There are several different forms of
+	 * these, but each eventually returns a substring of the glyph
+	 * name.
+	 */
+	case ('('):
+		gly = ESCAPE_SPECIAL;
+		*sz = 2;
+		break;
+	case ('['):
+		gly = ESCAPE_SPECIAL;
+		/*
+		 * Unicode escapes are defined in groff as \[uXXXX] to
+		 * \[u10FFFF], where the contained value must be a valid
+		 * Unicode codepoint.  Here, however, only check whether
+		 * it's not a zero-width escape.
+		 */
+		if ('u' == (*start)[0] && ']' != (*start)[1])
+			gly = ESCAPE_UNICODE;
+		term = ']';
+		break;
+	case ('C'):
+		if ('\'' != **start)
+			return(ESCAPE_ERROR);
+		*start = ++*end;
+		if ('u' == (*start)[0] && '\'' != (*start)[1])
+			gly = ESCAPE_UNICODE;
+		else
+			gly = ESCAPE_SPECIAL;
+		term = '\'';
+		break;
+
+	/*
+	 * Escapes taking no arguments at all.
+	 */
+	case ('d'):
 		/* FALLTHROUGH */
-	case ('`'):
+	case ('u'):
+		return(ESCAPE_IGNORE);
+
+	/*
+	 * The \z escape is supposed to output the following
+	 * character without advancing the cursor position.  
+	 * Since we are mostly dealing with terminal mode,
+	 * let us just skip the next character.
+	 */
+	case ('z'):
+		return(ESCAPE_SKIPCHAR);
+
+	/*
+	 * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where
+	 * 'X' is the trigger.  These have opaque sub-strings.
+	 */
+	case ('F'):
 		/* FALLTHROUGH */
-	case ('q'):
+	case ('g'):
 		/* FALLTHROUGH */
-	case ('-'):
+	case ('k'):
 		/* FALLTHROUGH */
-	case ('~'):
+	case ('M'):
 		/* FALLTHROUGH */
-	case ('^'):
+	case ('m'):
 		/* FALLTHROUGH */
-	case ('%'):
+	case ('n'):
 		/* FALLTHROUGH */
-	case ('0'):
+	case ('V'):
 		/* FALLTHROUGH */
-	case (' '):
+	case ('Y'):
+		gly = ESCAPE_IGNORE;
 		/* FALLTHROUGH */
-	case ('|'):
+	case ('f'):
+		if (ESCAPE_ERROR == gly)
+			gly = ESCAPE_FONT;
+		switch (**start) {
+		case ('('):
+			*start = ++*end;
+			*sz = 2;
+			break;
+		case ('['):
+			*start = ++*end;
+			term = ']';
+			break;
+		default:
+			*sz = 1;
+			break;
+		}
+		break;
+
+	/*
+	 * These escapes are of the form \X'Y', where 'X' is the trigger
+	 * and 'Y' is any string.  These have opaque sub-strings.
+	 * The \B and \w escapes are handled in roff.c, roff_res().
+	 */
+	case ('A'):
 		/* FALLTHROUGH */
-	case ('&'):
+	case ('b'):
 		/* FALLTHROUGH */
-	case ('.'):
+	case ('D'):
 		/* FALLTHROUGH */
-	case (':'):
+	case ('o'):
 		/* FALLTHROUGH */
-	case ('c'):
-		return(2);
-	case ('e'):
-		return(2);
-	case ('s'):
-		if ('\0' == *++p)
-			return(2);
+	case ('R'):
+		/* FALLTHROUGH */
+	case ('X'):
+		/* FALLTHROUGH */
+	case ('Z'):
+		if ('\0' == **start)
+			return(ESCAPE_ERROR);
+		gly = ESCAPE_IGNORE;
+		term = **start;
+		*start = ++*end;
+		break;
 
-		c = 2;
-		terminator = 0;
-		lim = 1;
+	/*
+	 * These escapes are of the form \X'N', where 'X' is the trigger
+	 * and 'N' resolves to a numerical expression.
+	 */
+	case ('h'):
+		/* FALLTHROUGH */
+	case ('H'):
+		/* FALLTHROUGH */
+	case ('L'):
+		/* FALLTHROUGH */
+	case ('l'):
+		/* FALLTHROUGH */
+	case ('S'):
+		/* FALLTHROUGH */
+	case ('v'):
+		/* FALLTHROUGH */
+	case ('x'):
+		if (strchr("\0 %&()*+-./0123456789:<=>", **start))
+			return(ESCAPE_ERROR);
+		gly = ESCAPE_IGNORE;
+		term = **start;
+		*start = ++*end;
+		break;
 
-		if (*p == '\'') {
-			lim = 0;
-			terminator = 1;
-			++p;
-			++c;
-		} else if (*p == '[') {
-			lim = 0;
-			terminator = 2;
-			++p;
-			++c;
-		} else if (*p == '(') {
-			lim = 2;
-			terminator = 3;
-			++p;
-			++c;
+	/*
+	 * Special handling for the numbered character escape.
+	 * XXX Do any other escapes need similar handling?
+	 */
+	case ('N'):
+		if ('\0' == **start)
+			return(ESCAPE_ERROR);
+		(*end)++;
+		if (isdigit((unsigned char)**start)) {
+			*sz = 1;
+			return(ESCAPE_IGNORE);
 		}
+		(*start)++;
+		while (isdigit((unsigned char)**end))
+			(*end)++;
+		*sz = *end - *start;
+		if ('\0' != **end)
+			(*end)++;
+		return(ESCAPE_NUMBERED);
 
-		if (*p == '+' || *p == '-') {
-			++p;
-			++c;
-		}
+	/* 
+	 * Sizes get a special category of their own.
+	 */
+	case ('s'):
+		gly = ESCAPE_IGNORE;
 
-		if (*p == '\'') {
-			if (terminator)
-				return(0);
-			lim = 0;
-			terminator = 1;
-			++p;
-			++c;
-		} else if (*p == '[') {
-			if (terminator)
-				return(0);
-			lim = 0;
-			terminator = 2;
-			++p;
-			++c;
-		} else if (*p == '(') {
-			if (terminator)
-				return(0);
-			lim = 2;
-			terminator = 3;
-			++p;
-			++c;
-		}
+		/* See +/- counts as a sign. */
+		if ('+' == **end || '-' == **end || ASCII_HYPH == **end)
+			(*end)++;
 
-		/* TODO: needs to handle floating point. */
-
-		if ( ! isdigit((u_char)*p))
-			return(0);
-
-		for (i = 0; isdigit((u_char)*p); i++) {
-			if (lim && i >= lim)
-				break;
-			++p;
-			++c;
-		}
-
-		if (terminator && terminator < 3) {
-			if (1 == terminator && *p != '\'')
-				return(0);
-			if (2 == terminator && *p != ']')
-				return(0);
-			++p;
-			++c;
-		}
-
-		return(c);
-	case ('f'):
-		/* FALLTHROUGH */
-	case ('F'):
-		/* FALLTHROUGH */
-	case ('*'):
-		if (0 == *++p || ! isgraph((u_char)*p))
-			return(0);
-		switch (*p) {
+		switch (**end) {
 		case ('('):
-			if (0 == *++p || ! isgraph((u_char)*p))
-				return(0);
-			return(4);
+			*start = ++*end;
+			*sz = 2;
+			break;
 		case ('['):
-			for (c = 3, p++; *p && ']' != *p; p++, c++)
-				if ( ! isgraph((u_char)*p))
-					break;
-			return(*p == ']' ? c : 0);
+			*start = ++*end;
+			term = ']';
+			break;
+		case ('\''):
+			*start = ++*end;
+			term = '\'';
+			break;
 		default:
+			*sz = 1;
 			break;
 		}
-		return(3);
-	case ('('):
-		if (0 == *++p || ! isgraph((u_char)*p))
-			return(0);
-		if (0 == *++p || ! isgraph((u_char)*p))
-			return(0);
-		return(4);
-	case ('['):
+
 		break;
+
+	/*
+	 * Anything else is assumed to be a glyph.
+	 * In this case, pass back the character after the backslash.
+	 */
 	default:
-		return(0);
+		gly = ESCAPE_SPECIAL;
+		*start = --*end;
+		*sz = 1;
+		break;
 	}
 
-	for (c = 3, p++; *p && ']' != *p; p++, c++)
-		if ( ! isgraph((u_char)*p))
-			break;
+	assert(ESCAPE_ERROR != gly);
 
-	return(*p == ']' ? c : 0);
-}
+	/*
+	 * Read up to the terminating character,
+	 * paying attention to nested escapes.
+	 */
 
-
-void *
-mandoc_calloc(size_t num, size_t size)
-{
-	void		*ptr;
-
-	ptr = calloc(num, size);
-	if (NULL == ptr) {
-		perror(NULL);
-		exit(EXIT_FAILURE);
+	if ('\0' != term) {
+		while (**end != term) {
+			switch (**end) {
+			case ('\0'):
+				return(ESCAPE_ERROR);
+			case ('\\'):
+				(*end)++;
+				if (ESCAPE_ERROR ==
+				    mandoc_escape(end, NULL, NULL))
+					return(ESCAPE_ERROR);
+				break;
+			default:
+				(*end)++;
+				break;
+			}
+		}
+		*sz = (*end)++ - *start;
+	} else {
+		assert(*sz > 0);
+		if ((size_t)*sz > strlen(*start))
+			return(ESCAPE_ERROR);
+		*end += *sz;
 	}
 
-	return(ptr);
-}
+	/* Run post-processors. */
 
+	switch (gly) {
+	case (ESCAPE_FONT):
+		if (2 == *sz) {
+			if ('C' == **start) {
+				/*
+				 * Treat constant-width font modes
+				 * just like regular font modes.
+				 */
+				(*start)++;
+				(*sz)--;
+			} else {
+				if ('B' == (*start)[0] && 'I' == (*start)[1])
+					gly = ESCAPE_FONTBI;
+				break;
+			}
+		} else if (1 != *sz)
+			break;
 
-void *
-mandoc_malloc(size_t size)
-{
-	void		*ptr;
-
-	ptr = malloc(size);
-	if (NULL == ptr) {
-		perror(NULL);
-		exit(EXIT_FAILURE);
+		switch (**start) {
+		case ('3'):
+			/* FALLTHROUGH */
+		case ('B'):
+			gly = ESCAPE_FONTBOLD;
+			break;
+		case ('2'):
+			/* FALLTHROUGH */
+		case ('I'):
+			gly = ESCAPE_FONTITALIC;
+			break;
+		case ('P'):
+			gly = ESCAPE_FONTPREV;
+			break;
+		case ('1'):
+			/* FALLTHROUGH */
+		case ('R'):
+			gly = ESCAPE_FONTROMAN;
+			break;
+		}
+		break;
+	case (ESCAPE_SPECIAL):
+		if (1 == *sz && 'c' == **start)
+			gly = ESCAPE_NOSPACE;
+		break;
+	default:
+		break;
 	}
 
-	return(ptr);
+	return(gly);
 }
 
-
-void *
-mandoc_realloc(void *ptr, size_t size)
+/*
+ * Parse a quoted or unquoted roff-style request or macro argument.
+ * Return a pointer to the parsed argument, which is either the original
+ * pointer or advanced by one byte in case the argument is quoted.
+ * NUL-terminate the argument in place.
+ * Collapse pairs of quotes inside quoted arguments.
+ * Advance the argument pointer to the next argument,
+ * or to the NUL byte terminating the argument line.
+ */
+char *
+mandoc_getarg(struct mparse *parse, char **cpp, int ln, int *pos)
 {
+	char	 *start, *cp;
+	int	  quoted, pairs, white;
 
-	ptr = realloc(ptr, size);
-	if (NULL == ptr) {
-		perror(NULL);
-		exit(EXIT_FAILURE);
-	}
+	/* Quoting can only start with a new word. */
+	start = *cpp;
+	quoted = 0;
+	if ('"' == *start) {
+		quoted = 1;
+		start++;
+	} 
 
-	return(ptr);
-}
+	pairs = 0;
+	white = 0;
+	for (cp = start; '\0' != *cp; cp++) {
 
+		/*
+		 * Move the following text left
+		 * after quoted quotes and after "\\" and "\t".
+		 */
+		if (pairs)
+			cp[-pairs] = cp[0];
 
-char *
-mandoc_strdup(const char *ptr)
-{
-	char		*p;
+		if ('\\' == cp[0]) {
+			/*
+			 * In copy mode, translate double to single
+			 * backslashes and backslash-t to literal tabs.
+			 */
+			switch (cp[1]) {
+			case ('t'):
+				cp[0] = '\t';
+				/* FALLTHROUGH */
+			case ('\\'):
+				pairs++;
+				cp++;
+				break;
+			case (' '):
+				/* Skip escaped blanks. */
+				if (0 == quoted)
+					cp++;
+				break;
+			default:
+				break;
+			}
+		} else if (0 == quoted) {
+			if (' ' == cp[0]) {
+				/* Unescaped blanks end unquoted args. */
+				white = 1;
+				break;
+			}
+		} else if ('"' == cp[0]) {
+			if ('"' == cp[1]) {
+				/* Quoted quotes collapse. */
+				pairs++;
+				cp++;
+			} else {
+				/* Unquoted quotes end quoted args. */
+				quoted = 2;
+				break;
+			}
+		}
+	}
 
-	p = strdup(ptr);
-	if (NULL == p) {
-		perror(NULL);
-		exit(EXIT_FAILURE);
+	/* Quoted argument without a closing quote. */
+	if (1 == quoted)
+		mandoc_msg(MANDOCERR_BADQUOTE, parse, ln, *pos, NULL);
+
+	/* NUL-terminate this argument and move to the next one. */
+	if (pairs)
+		cp[-pairs] = '\0';
+	if ('\0' != *cp) {
+		*cp++ = '\0';
+		while (' ' == *cp)
+			cp++;
 	}
+	*pos += (int)(cp - start) + (quoted ? 1 : 0);
+	*cpp = cp;
 
-	return(p);
+	if ('\0' == *cp && (white || ' ' == cp[-1]))
+		mandoc_msg(MANDOCERR_EOLNSPACE, parse, ln, *pos, NULL);
+
+	return(start);
 }
 
-
 static int
 a2time(time_t *t, const char *fmt, const char *p)
 {
@@ -256,7 +454,10 @@ a2time(time_t *t, const char *fmt, const char *p)
 
 	memset(&tm, 0, sizeof(struct tm));
 
+	pp = NULL;
+#ifdef	HAVE_STRPTIME
 	pp = strptime(p, fmt, &tm);
+#endif
 	if (NULL != pp && '\0' == *pp) {
 		*t = mktime(&tm);
 		return(1);
@@ -265,61 +466,133 @@ a2time(time_t *t, const char *fmt, const char *p)
 	return(0);
 }
 
+static char *
+time2a(time_t t)
+{
+	struct tm	*tm;
+	char		*buf, *p;
+	size_t		 ssz;
+	int		 isz;
 
-/*
- * Convert from a manual date string (see mdoc(7) and man(7)) into a
- * date according to the stipulated date type.
- */
-time_t
-mandoc_a2time(int flags, const char *p)
+	tm = localtime(&t);
+
+	/*
+	 * Reserve space:
+	 * up to 9 characters for the month (September) + blank
+	 * up to 2 characters for the day + comma + blank
+	 * 4 characters for the year and a terminating '\0'
+	 */
+	p = buf = mandoc_malloc(10 + 4 + 4 + 1);
+
+	if (0 == (ssz = strftime(p, 10 + 1, "%B ", tm)))
+		goto fail;
+	p += (int)ssz;
+
+	if (-1 == (isz = snprintf(p, 4 + 1, "%d, ", tm->tm_mday)))
+		goto fail;
+	p += isz;
+
+	if (0 == strftime(p, 4 + 1, "%Y", tm))
+		goto fail;
+	return(buf);
+
+fail:
+	free(buf);
+	return(NULL);
+}
+
+char *
+mandoc_normdate(struct mparse *parse, char *in, int ln, int pos)
 {
+	char		*out;
 	time_t		 t;
 
-	if (MTIME_MDOCDATE & flags) {
-		if (0 == strcmp(p, "$" "Mdocdate$"))
-			return(time(NULL));
-		if (a2time(&t, "$" "Mdocdate: %b %d %Y $", p))
-			return(t);
+	if (NULL == in || '\0' == *in ||
+	    0 == strcmp(in, "$" "Mdocdate$")) {
+		mandoc_msg(MANDOCERR_NODATE, parse, ln, pos, NULL);
+		time(&t);
 	}
+	else if (a2time(&t, "%Y-%m-%d", in))
+		t = 0;
+	else if (!a2time(&t, "$" "Mdocdate: %b %d %Y $", in) &&
+	    !a2time(&t, "%b %d, %Y", in)) {
+		mandoc_msg(MANDOCERR_BADDATE, parse, ln, pos, NULL);
+		t = 0;
+	}
+	out = t ? time2a(t) : NULL;
+	return(out ? out : mandoc_strdup(in));
+}
 
-	if (MTIME_CANONICAL & flags || MTIME_REDUCED & flags) 
-		if (a2time(&t, "%b %d, %Y", p))
-			return(t);
+int
+mandoc_eos(const char *p, size_t sz)
+{
+	const char	*q;
+	int		 enclosed, found;
 
-	if (MTIME_ISO_8601 & flags) 
-		if (a2time(&t, "%Y-%m-%d", p))
-			return(t);
+	if (0 == sz)
+		return(0);
 
-	if (MTIME_REDUCED & flags) {
-		if (a2time(&t, "%d, %Y", p))
-			return(t);
-		if (a2time(&t, "%Y", p))
-			return(t);
+	/*
+	 * End-of-sentence recognition must include situations where
+	 * some symbols, such as `)', allow prior EOS punctuation to
+	 * propagate outward.
+	 */
+
+	enclosed = found = 0;
+	for (q = p + (int)sz - 1; q >= p; q--) {
+		switch (*q) {
+		case ('\"'):
+			/* FALLTHROUGH */
+		case ('\''):
+			/* FALLTHROUGH */
+		case (']'):
+			/* FALLTHROUGH */
+		case (')'):
+			if (0 == found)
+				enclosed = 1;
+			break;
+		case ('.'):
+			/* FALLTHROUGH */
+		case ('!'):
+			/* FALLTHROUGH */
+		case ('?'):
+			found = 1;
+			break;
+		default:
+			return(found && (!enclosed || isalnum((unsigned char)*q)));
+		}
 	}
 
-	return(0);
+	return(found && !enclosed);
 }
 
-
+/*
+ * Convert a string to a long that may not be <0.
+ * If the string is invalid, or is less than 0, return -1.
+ */
 int
-mandoc_eos(const char *p, size_t sz)
+mandoc_strntoi(const char *p, size_t sz, int base)
 {
+	char		 buf[32];
+	char		*ep;
+	long		 v;
 
-	assert(sz);
+	if (sz > 31)
+		return(-1);
 
-	switch (p[(int)sz - 1]) {
-	case ('.'):
-		/* Escaped periods. */
-		if (sz > 1 && '\\' == p[(int)sz - 2])
-			return(0);
-		/* FALLTHROUGH */
-	case ('!'):
-		/* FALLTHROUGH */
-	case ('?'):
-		break;
-	default:
-		return(0);
-	}
+	memcpy(buf, p, sz);
+	buf[(int)sz] = '\0';
 
-	return(1);
+	errno = 0;
+	v = strtol(buf, &ep, base);
+
+	if (buf[0] == '\0' || *ep != '\0')
+		return(-1);
+
+	if (v > INT_MAX)
+		v = INT_MAX;
+	if (v < INT_MIN)
+		v = INT_MIN;
+
+	return((int)v);
 }