===================================================================
RCS file: /cvs/mandoc/mdoc.c,v
retrieving revision 1.112
retrieving revision 1.125
diff -u -p -r1.112 -r1.125
--- mandoc/mdoc.c	2009/10/27 08:26:12	1.112
+++ mandoc/mdoc.c	2010/05/08 07:30:19	1.125
@@ -1,4 +1,4 @@
-/*	$Id: mdoc.c,v 1.112 2009/10/27 08:26:12 kristaps Exp $ */
+/*	$Id: mdoc.c,v 1.125 2010/05/08 07:30:19 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -14,6 +14,10 @@
  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  */
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+
 #include <sys/types.h>
 
 #include <assert.h>
@@ -22,14 +26,15 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
+#include <time.h>
 
 #include "libmdoc.h"
+#include "libmandoc.h"
 
 const	char *const __mdoc_merrnames[MERRMAX] = {		 
 	"trailing whitespace", /* ETAILWS */
 	"unexpected quoted parameter", /* EQUOTPARM */
 	"unterminated quoted parameter", /* EQUOTTERM */
-	"system: malloc error", /* EMALLOC */
 	"argument parameter suggested", /* EARGVAL */
 	"macro disallowed in prologue", /* EBODYPROL */
 	"macro disallowed in body", /* EPROLBODY */
@@ -79,6 +84,7 @@ const	char *const __mdoc_merrnames[MERRMAX] = {		 
 	"no description found for library", /* ELIB */
 	"bad child for parent context", /* EBADCHILD */
 	"list arguments preceding type", /* ENOTYPE */
+	"deprecated comment style", /* EBADCOMMENT */
 };
 
 const	char *const __mdoc_macronames[MDOC_MAX] = {		 
@@ -93,11 +99,11 @@ const	char *const __mdoc_macronames[MDOC_MAX] = {		 
 	"Nm",		"Op",		"Ot",		"Pa",
 	"Rv",		"St",		"Va",		"Vt",
 	/* LINTED */
-	"Xr",		"\%A",		"\%B",		"\%D",
+	"Xr",		"%A",		"%B",		"%D",
 	/* LINTED */
-	"\%I",		"\%J",		"\%N",		"\%O",
+	"%I",		"%J",		"%N",		"%O",
 	/* LINTED */
-	"\%P",		"\%R",		"\%T",		"\%V",
+	"%P",		"%R",		"%T",		"%V",
 	"Ac",		"Ao",		"Aq",		"At",
 	"Bc",		"Bf",		"Bo",		"Bq",
 	"Bsx",		"Bx",		"Db",		"Dc",
@@ -114,11 +120,11 @@ const	char *const __mdoc_macronames[MDOC_MAX] = {		 
 	"Fr",		"Ud",		"Lb",		"Lp",
 	"Lk",		"Mt",		"Brq",		"Bro",
 	/* LINTED */
-	"Brc",		"\%C",		"Es",		"En",
+	"Brc",		"%C",		"Es",		"En",
 	/* LINTED */
-	"Dx",		"\%Q",		"br",		"sp",
+	"Dx",		"%Q",		"br",		"sp",
 	/* LINTED */
-	"\%U"
+	"%U"
 	};
 
 const	char *const __mdoc_argnames[MDOC_ARG_MAX] = {		 
@@ -136,23 +142,20 @@ const	char *const __mdoc_argnames[MDOC_ARG_MAX] = {		 
 const	char * const *mdoc_macronames = __mdoc_macronames;
 const	char * const *mdoc_argnames = __mdoc_argnames;
 
+static	void		  mdoc_node_free(struct mdoc_node *);
+static	void		  mdoc_node_unlink(struct mdoc *, 
+				struct mdoc_node *);
 static	void		  mdoc_free1(struct mdoc *);
-static	int		  mdoc_alloc1(struct mdoc *);
+static	void		  mdoc_alloc1(struct mdoc *);
 static	struct mdoc_node *node_alloc(struct mdoc *, int, int, 
-				int, enum mdoc_type);
+				enum mdoct, enum mdoc_type);
 static	int		  node_append(struct mdoc *, 
 				struct mdoc_node *);
-static	int		  parsetext(struct mdoc *, int, char *);
-static	int		  parsemacro(struct mdoc *, int, char *);
+static	int		  mdoc_ptext(struct mdoc *, int, char *);
+static	int		  mdoc_pmacro(struct mdoc *, int, char *);
 static	int		  macrowarn(struct mdoc *, int, const char *);
-static	int		  pstring(struct mdoc *, int, int, 
-				const char *, size_t);
 
-#ifdef __linux__
-extern	size_t	  	  strlcpy(char *, const char *, size_t);
-#endif
 
-
 const struct mdoc_node *
 mdoc_node(const struct mdoc *m)
 {
@@ -177,7 +180,7 @@ mdoc_free1(struct mdoc *mdoc)
 {
 
 	if (mdoc->first)
-		mdoc_node_freelist(mdoc->first);
+		mdoc_node_delete(mdoc, mdoc->first);
 	if (mdoc->meta.title)
 		free(mdoc->meta.title);
 	if (mdoc->meta.os)
@@ -194,21 +197,17 @@ mdoc_free1(struct mdoc *mdoc)
 /*
  * Allocate all volatile resources (parse tree, meta-data, fields).
  */
-static int
+static void
 mdoc_alloc1(struct mdoc *mdoc)
 {
 
 	memset(&mdoc->meta, 0, sizeof(struct mdoc_meta));
 	mdoc->flags = 0;
 	mdoc->lastnamed = mdoc->lastsec = SEC_NONE;
-	mdoc->last = calloc(1, sizeof(struct mdoc_node));
-	if (NULL == mdoc->last)
-		return(0);
-
+	mdoc->last = mandoc_calloc(1, sizeof(struct mdoc_node));
 	mdoc->first = mdoc->last;
 	mdoc->last->type = MDOC_ROOT;
 	mdoc->next = MDOC_NEXT_CHILD;
-	return(1);
 }
 
 
@@ -218,12 +217,12 @@ mdoc_alloc1(struct mdoc *mdoc)
  * and the parser is ready for re-invocation on a new tree; however,
  * cross-parse non-volatile data is kept intact.
  */
-int
+void
 mdoc_reset(struct mdoc *mdoc)
 {
 
 	mdoc_free1(mdoc);
-	return(mdoc_alloc1(mdoc));
+	mdoc_alloc1(mdoc);
 }
 
 
@@ -248,21 +247,17 @@ mdoc_alloc(void *data, int pflags, const struct mdoc_c
 {
 	struct mdoc	*p;
 
-	if (NULL == (p = calloc(1, sizeof(struct mdoc))))
-		return(NULL);
+	p = mandoc_calloc(1, sizeof(struct mdoc));
+
 	if (cb)
-		(void)memcpy(&p->cb, cb, sizeof(struct mdoc_cb));
+		memcpy(&p->cb, cb, sizeof(struct mdoc_cb));
 
-	mdoc_hash_init();
-
 	p->data = data;
 	p->pflags = pflags;
 
-	if (mdoc_alloc1(p))
-		return(p);
-
-	free(p);
-	return(NULL);
+	mdoc_hash_init();
+	mdoc_alloc1(p);
+	return(p);
 }
 
 
@@ -285,7 +280,7 @@ mdoc_endparse(struct mdoc *m)
 
 /*
  * Main parse routine.  Parses a single line -- really just hands off to
- * the macro (parsemacro()) or text parser (parsetext()).
+ * the macro (mdoc_pmacro()) or text parser (mdoc_ptext()).
  */
 int
 mdoc_parseln(struct mdoc *m, int ln, char *buf)
@@ -294,8 +289,8 @@ mdoc_parseln(struct mdoc *m, int ln, char *buf)
 	if (MDOC_HALT & m->flags)
 		return(0);
 
-	return('.' == *buf ? parsemacro(m, ln, buf) :
-			parsetext(m, ln, buf));
+	return('.' == *buf ? mdoc_pmacro(m, ln, buf) :
+			mdoc_ptext(m, ln, buf));
 }
 
 
@@ -350,19 +345,33 @@ mdoc_err(struct mdoc *m, int line, int pos, int iserr,
 
 
 int
-mdoc_macro(struct mdoc *m, int tok, 
+mdoc_macro(struct mdoc *m, enum mdoct tok, 
 		int ln, int pp, int *pos, char *buf)
 {
-	/*
-	 * If we're in the prologue, deny "body" macros.  Similarly, if
-	 * we're in the body, deny prologue calls.
-	 */
+	assert(tok < MDOC_MAX);
+
+	/* If we're in the body, deny prologue calls. */
+
 	if (MDOC_PROLOGUE & mdoc_macros[tok].flags && 
 			MDOC_PBODY & m->flags)
 		return(mdoc_perr(m, ln, pp, EPROLBODY));
+
+	/* If we're in the prologue, deny "body" macros.  */
+
 	if ( ! (MDOC_PROLOGUE & mdoc_macros[tok].flags) && 
-			! (MDOC_PBODY & m->flags))
-		return(mdoc_perr(m, ln, pp, EBODYPROL));
+			! (MDOC_PBODY & m->flags)) {
+		if ( ! mdoc_pwarn(m, ln, pp, EBODYPROL))
+			return(0);
+		if (NULL == m->meta.title)
+			m->meta.title = mandoc_strdup("unknown");
+		if (NULL == m->meta.vol)
+			m->meta.vol = mandoc_strdup("local");
+		if (NULL == m->meta.os)
+			m->meta.os = mandoc_strdup("local");
+		if (0 == m->meta.date)
+			m->meta.date = time(NULL);
+		m->flags |= MDOC_PBODY;
+	}
 
 	return((*mdoc_macros[tok].fp)(m, tok, ln, pp, pos, buf));
 }
@@ -433,35 +442,28 @@ node_append(struct mdoc *mdoc, struct mdoc_node *p)
 
 
 static struct mdoc_node *
-node_alloc(struct mdoc *m, int line, 
-		int pos, int tok, enum mdoc_type type)
+node_alloc(struct mdoc *m, int line, int pos, 
+		enum mdoct tok, enum mdoc_type type)
 {
 	struct mdoc_node *p;
 
-	if (NULL == (p = calloc(1, sizeof(struct mdoc_node)))) {
-		(void)mdoc_nerr(m, m->last, EMALLOC);
-		return(NULL);
-	}
-
+	p = mandoc_calloc(1, sizeof(struct mdoc_node));
 	p->sec = m->lastsec;
 	p->line = line;
 	p->pos = pos;
 	p->tok = tok;
-	if (MDOC_TEXT != (p->type = type))
-		assert(p->tok >= 0);
+	p->type = type;
 
 	return(p);
 }
 
 
 int
-mdoc_tail_alloc(struct mdoc *m, int line, int pos, int tok)
+mdoc_tail_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 {
 	struct mdoc_node *p;
 
 	p = node_alloc(m, line, pos, tok, MDOC_TAIL);
-	if (NULL == p)
-		return(0);
 	if ( ! node_append(m, p))
 		return(0);
 	m->next = MDOC_NEXT_CHILD;
@@ -470,7 +472,7 @@ mdoc_tail_alloc(struct mdoc *m, int line, int pos, int
 
 
 int
-mdoc_head_alloc(struct mdoc *m, int line, int pos, int tok)
+mdoc_head_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 {
 	struct mdoc_node *p;
 
@@ -478,8 +480,6 @@ mdoc_head_alloc(struct mdoc *m, int line, int pos, int
 	assert(m->last);
 
 	p = node_alloc(m, line, pos, tok, MDOC_HEAD);
-	if (NULL == p)
-		return(0);
 	if ( ! node_append(m, p))
 		return(0);
 	m->next = MDOC_NEXT_CHILD;
@@ -488,13 +488,11 @@ mdoc_head_alloc(struct mdoc *m, int line, int pos, int
 
 
 int
-mdoc_body_alloc(struct mdoc *m, int line, int pos, int tok)
+mdoc_body_alloc(struct mdoc *m, int line, int pos, enum mdoct tok)
 {
 	struct mdoc_node *p;
 
 	p = node_alloc(m, line, pos, tok, MDOC_BODY);
-	if (NULL == p)
-		return(0);
 	if ( ! node_append(m, p))
 		return(0);
 	m->next = MDOC_NEXT_CHILD;
@@ -504,13 +502,11 @@ mdoc_body_alloc(struct mdoc *m, int line, int pos, int
 
 int
 mdoc_block_alloc(struct mdoc *m, int line, int pos, 
-		int tok, struct mdoc_arg *args)
+		enum mdoct tok, struct mdoc_arg *args)
 {
 	struct mdoc_node *p;
 
 	p = node_alloc(m, line, pos, tok, MDOC_BLOCK);
-	if (NULL == p)
-		return(0);
 	p->args = args;
 	if (p->args)
 		(args->refcnt)++;
@@ -523,13 +519,11 @@ mdoc_block_alloc(struct mdoc *m, int line, int pos, 
 
 int
 mdoc_elem_alloc(struct mdoc *m, int line, int pos, 
-		int tok, struct mdoc_arg *args)
+		enum mdoct tok, struct mdoc_arg *args)
 {
 	struct mdoc_node *p;
 
 	p = node_alloc(m, line, pos, tok, MDOC_ELEM);
-	if (NULL == p)
-		return(0);
 	p->args = args;
 	if (p->args)
 		(args->refcnt)++;
@@ -540,22 +534,16 @@ mdoc_elem_alloc(struct mdoc *m, int line, int pos, 
 }
 
 
-static int
-pstring(struct mdoc *m, int line, int pos, const char *p, size_t len)
+int
+mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
 {
 	struct mdoc_node *n;
-	size_t		  sv;
+	size_t		  sv, len;
 
-	n = node_alloc(m, line, pos, -1, MDOC_TEXT);
-	if (NULL == n)
-		return(mdoc_nerr(m, m->last, EMALLOC));
+	len = strlen(p);
 
-	n->string = malloc(len + 1);
-	if (NULL == n->string) {
-		free(n);
-		return(mdoc_nerr(m, m->last, EMALLOC));
-	}
-
+	n = node_alloc(m, line, pos, MDOC_MAX, MDOC_TEXT);
+	n->string = mandoc_malloc(len + 1);
 	sv = strlcpy(n->string, p, len + 1);
 
 	/* Prohibit truncation. */
@@ -563,25 +551,16 @@ pstring(struct mdoc *m, int line, int pos, const char 
 
 	if ( ! node_append(m, n))
 		return(0);
+
 	m->next = MDOC_NEXT_SIBLING;
 	return(1);
 }
 
 
-int
-mdoc_word_alloc(struct mdoc *m, int line, int pos, const char *p)
-{
-
-	return(pstring(m, line, pos, p, strlen(p)));
-}
-
-
 void
 mdoc_node_free(struct mdoc_node *p)
 {
 
-	if (p->parent)
-		p->parent->nchild--;
 	if (p->string)
 		free(p->string);
 	if (p->args)
@@ -590,16 +569,53 @@ mdoc_node_free(struct mdoc_node *p)
 }
 
 
-void
-mdoc_node_freelist(struct mdoc_node *p)
+static void
+mdoc_node_unlink(struct mdoc *m, struct mdoc_node *n)
 {
 
-	if (p->child)
-		mdoc_node_freelist(p->child);
-	if (p->next)
-		mdoc_node_freelist(p->next);
+	/* Adjust siblings. */
 
+	if (n->prev)
+		n->prev->next = n->next;
+	if (n->next)
+		n->next->prev = n->prev;
+
+	/* Adjust parent. */
+
+	if (n->parent) {
+		n->parent->nchild--;
+		if (n->parent->child == n)
+			n->parent->child = n->prev ? n->prev : n->next;
+	}
+
+	/* Adjust parse point, if applicable. */
+
+	if (m && m->last == n) {
+		if (n->prev) {
+			m->last = n->prev;
+			m->next = MDOC_NEXT_SIBLING;
+		} else {
+			m->last = n->parent;
+			m->next = MDOC_NEXT_CHILD;
+		}
+	}
+
+	if (m && m->first == n)
+		m->first = NULL;
+}
+
+
+void
+mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
+{
+
+	while (p->child) {
+		assert(p->nchild);
+		mdoc_node_delete(m, p->child);
+	}
 	assert(0 == p->nchild);
+
+	mdoc_node_unlink(m, p);
 	mdoc_node_free(p);
 }
 
@@ -609,68 +625,67 @@ mdoc_node_freelist(struct mdoc_node *p)
  * control character.
  */
 static int
-parsetext(struct mdoc *m, int line, char *buf)
+mdoc_ptext(struct mdoc *m, int line, char *buf)
 {
-	int		 i, j;
+	int		 i;
 
+	/* Ignore bogus comments. */
+
+	if ('\\' == buf[0] && '.' == buf[1] && '\"' == buf[2])
+		return(mdoc_pwarn(m, line, 0, EBADCOMMENT));
+
+	/* No text before an initial macro. */
+
 	if (SEC_NONE == m->lastnamed)
 		return(mdoc_perr(m, line, 0, ETEXTPROL));
-	
-	/*
-	 * If in literal mode, then pass the buffer directly to the
-	 * back-end, as it should be preserved as a single term.
-	 */
 
+	/* Literal just gets pulled in as-is. */
+	
 	if (MDOC_LITERAL & m->flags)
 		return(mdoc_word_alloc(m, line, 0, buf));
 
-	/* Disallow blank/white-space lines in non-literal mode. */
+	/* Check for a blank line, which may also consist of spaces. */
 
 	for (i = 0; ' ' == buf[i]; i++)
-		/* Skip leading whitespace. */ ;
-	if (0 == buf[i])
-		return(mdoc_perr(m, line, 0, ENOBLANK));
+		/* Skip to first non-space. */ ;
 
-	/*
-	 * Break apart a free-form line into tokens.  Spaces are
-	 * stripped out of the input.
-	 */
+	if ('\0' == buf[i]) {
+		if ( ! mdoc_pwarn(m, line, 0, ENOBLANK))
+			return(0);
 
-	for (j = i; buf[i]; i++) {
-		if (' ' != buf[i])
-			continue;
+		/*
+		 * Insert a `Pp' in the case of a blank line.  Technically,
+		 * blank lines aren't allowed, but enough manuals assume this
+		 * behaviour that we want to work around it.
+		 */
+		if ( ! mdoc_elem_alloc(m, line, 0, MDOC_Pp, NULL))
+			return(0);
 
-		/* Escaped whitespace. */
-		if (i && ' ' == buf[i] && '\\' == buf[i - 1])
-			continue;
+		m->next = MDOC_NEXT_SIBLING;
+		return(1);
+	}
 
-		buf[i++] = 0;
-		if ( ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
-			return(0);
+	/* Warn if the last un-escaped character is whitespace. */
 
-		for ( ; ' ' == buf[i]; i++)
-			/* Skip trailing whitespace. */ ;
+	i = (int)strlen(buf);
+	assert(i);
 
-		j = i;
-		if (0 == buf[i])
-			break;
-	}
+	if (' ' == buf[i - 1] || '\t' == buf[i - 1])
+		if (1 == i || ('\\' != buf[i - 2]))
+			if ( ! mdoc_pwarn(m, line, i - 1, ETAILWS))
+				return(0);
 
-	if (j != i && ! pstring(m, line, j, &buf[j], (size_t)(i - j)))
-		return(0);
+	/* Allocate the whole word. */
 
-	m->next = MDOC_NEXT_SIBLING;
-	return(1);
+	return(mdoc_word_alloc(m, line, 0, buf));
 }
 
 
-
 static int
 macrowarn(struct mdoc *m, int ln, const char *buf)
 {
 	if ( ! (MDOC_IGN_MACRO & m->pflags))
-		return(mdoc_verr(m, ln, 0, 
-				"unknown macro: %s%s", 
+		return(mdoc_verr(m, ln, 0, "unknown macro: %s%s", 
 				buf, strlen(buf) > 3 ? "..." : ""));
 	return(mdoc_vwarn(m, ln, 0, "unknown macro: %s%s",
 				buf, strlen(buf) > 3 ? "..." : ""));
@@ -682,14 +697,15 @@ macrowarn(struct mdoc *m, int ln, const char *buf)
  * character.
  */
 int
-parsemacro(struct mdoc *m, int ln, char *buf)
+mdoc_pmacro(struct mdoc *m, int ln, char *buf)
 {
-	int		  i, j, c;
-	char		  mac[5];
+	enum mdoct	tok;
+	int		i, j;
+	char		mac[5];
 
 	/* Empty lines are ignored. */
 
-	if (0 == buf[1])
+	if ('\0' == buf[1])
 		return(1);
 
 	i = 1;
@@ -700,14 +716,14 @@ parsemacro(struct mdoc *m, int ln, char *buf)
 		i++;
 		while (buf[i] && ' ' == buf[i])
 			i++;
-		if (0 == buf[i])
+		if ('\0' == buf[i])
 			return(1);
 	}
 
 	/* Copy the first word into a nil-terminated buffer. */
 
 	for (j = 0; j < 4; j++, i++) {
-		if (0 == (mac[j] = buf[i]))
+		if ('\0' == (mac[j] = buf[i]))
 			break;
 		else if (' ' == buf[i])
 			break;
@@ -727,7 +743,7 @@ parsemacro(struct mdoc *m, int ln, char *buf)
 		return(1);
 	} 
 	
-	if (MDOC_MAX == (c = mdoc_hash_find(mac))) {
+	if (MDOC_MAX == (tok = mdoc_hash_find(mac))) {
 		if ( ! macrowarn(m, ln, mac))
 			goto err;
 		return(1);
@@ -739,10 +755,19 @@ parsemacro(struct mdoc *m, int ln, char *buf)
 		i++;
 
 	/* 
+	 * Trailing whitespace.  Note that tabs are allowed to be passed
+	 * into the parser as "text", so we only warn about spaces here.
+	 */
+
+	if ('\0' == buf[i] && ' ' == buf[i - 1])
+		if ( ! mdoc_pwarn(m, ln, i - 1, ETAILWS))
+			goto err;
+
+	/* 
 	 * Begin recursive parse sequence.  Since we're at the start of
 	 * the line, we don't need to do callable/parseable checks.
 	 */
-	if ( ! mdoc_macro(m, c, ln, 1, &i, buf)) 
+	if ( ! mdoc_macro(m, tok, ln, 1, &i, buf)) 
 		goto err;
 
 	return(1);