===================================================================
RCS file: /cvs/mandoc/man.c,v
retrieving revision 1.28
retrieving revision 1.65
diff -u -p -r1.28 -r1.65
--- mandoc/man.c	2009/07/24 20:22:24	1.28
+++ mandoc/man.c	2010/05/12 16:46:28	1.65
@@ -1,4 +1,4 @@
-/*	$Id: man.c,v 1.28 2009/07/24 20:22:24 kristaps Exp $ */
+/*	$Id: man.c,v 1.65 2010/05/12 16:46:28 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
  *
@@ -14,6 +14,12 @@
  * 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>
 #include <ctype.h>
 #include <stdarg.h>
@@ -22,19 +28,32 @@
 #include <string.h>
 
 #include "libman.h"
+#include "libmandoc.h"
 
 const	char *const __man_merrnames[WERRMAX] = {		 
 	"invalid character", /* WNPRINT */
-	"system: malloc error", /* WNMEM */
 	"invalid manual section", /* WMSEC */
 	"invalid date format", /* WDATE */
 	"scope of prior line violated", /* WLNSCOPE */
+	"over-zealous prior line scope violation", /* WLNSCOPE2 */
 	"trailing whitespace", /* WTSPACE */
 	"unterminated quoted parameter", /* WTQUOTE */
 	"document has no body", /* WNODATA */
 	"document has no title/section", /* WNOTITLE */
 	"invalid escape sequence", /* WESCAPE */
 	"invalid number format", /* WNUMFMT */
+	"expected block head arguments", /* WHEADARGS */
+	"expected block body arguments", /* WBODYARGS */
+	"expected empty block head", /* WNHEADARGS */
+	"ill-formed macro", /* WMACROFORM */
+	"scope open on exit", /* WEXITSCOPE */
+	"no scope context", /* WNOSCOPE */
+	"literal context already open", /* WOLITERAL */
+	"no literal context open", /* WNLITERAL */
+	"invalid nesting of roff declarations", /* WROFFNEST */
+	"scope in roff instructions broken", /* WROFFSCOPE */
+	"document title should be uppercase", /* WTITLECASE */
+	"deprecated comment style", /* WBADCOMMENT */
 };
 
 const	char *const __man_macronames[MAN_MAX] = {		 
@@ -43,19 +62,28 @@ const	char *const __man_macronames[MAN_MAX] = {		 
 	"IP",		"HP",		"SM",		"SB",
 	"BI",		"IB",		"BR",		"RB",
 	"R",		"B",		"I",		"IR",
-	"RI",		"na",		"i",		"sp"
+	"RI",		"na",		"i",		"sp",
+	"nf",		"fi",		"r",		"RE",
+	"RS",		"DT",		"UC",		"PD",
+	"Sp",		"Vb",		"Ve",		"de",
+	"dei",		"am",		"ami",		"ig",
+	".",
 	};
 
 const	char * const *man_macronames = __man_macronames;
 
 static	struct man_node	*man_node_alloc(int, int, 
-				enum man_type, int);
+				enum man_type, enum mant);
 static	int		 man_node_append(struct man *, 
 				struct man_node *);
+static	void		 man_node_free(struct man_node *);
+static	void		 man_node_unlink(struct man *, 
+				struct man_node *);
 static	int		 man_ptext(struct man *, int, char *);
 static	int		 man_pmacro(struct man *, int, char *);
 static	void		 man_free1(struct man *);
-static	int		 man_alloc1(struct man *);
+static	void		 man_alloc1(struct man *);
+static	int		 macrowarn(struct man *, int, const char *);
 
 
 const struct man_node *
@@ -74,12 +102,12 @@ man_meta(const struct man *m)
 }
 
 
-int
+void
 man_reset(struct man *man)
 {
 
 	man_free1(man);
-	return(man_alloc1(man));
+	man_alloc1(man);
 }
 
 
@@ -88,9 +116,6 @@ man_free(struct man *man)
 {
 
 	man_free1(man);
-
-	if (man->htab)
-		man_hash_free(man->htab);
 	free(man);
 }
 
@@ -100,22 +125,16 @@ man_alloc(void *data, int pflags, const struct man_cb 
 {
 	struct man	*p;
 
-	if (NULL == (p = calloc(1, sizeof(struct man))))
-		return(NULL);
+	p = mandoc_calloc(1, sizeof(struct man));
 
-	if ( ! man_alloc1(p)) {
-		free(p);
-		return(NULL);
-	}
+	if (cb)
+		memcpy(&p->cb, cb, sizeof(struct man_cb));
 
+	man_hash_init();
 	p->data = data;
 	p->pflags = pflags;
-	(void)memcpy(&p->cb, cb, sizeof(struct man_cb));
 
-	if (NULL == (p->htab = man_hash_alloc())) {
-		free(p);
-		return(NULL);
-	}
+	man_alloc1(p);
 	return(p);
 }
 
@@ -137,7 +156,7 @@ int
 man_parseln(struct man *m, int ln, char *buf)
 {
 
-	return('.' == *buf ? 
+	return('.' == *buf || '\'' == *buf ? 
 			man_pmacro(m, ln, buf) : 
 			man_ptext(m, ln, buf));
 }
@@ -148,7 +167,7 @@ man_free1(struct man *man)
 {
 
 	if (man->first)
-		man_node_freelist(man->first);
+		man_node_delete(man, man->first);
 	if (man->meta.title)
 		free(man->meta.title);
 	if (man->meta.source)
@@ -158,19 +177,17 @@ man_free1(struct man *man)
 }
 
 
-static int
+static void
 man_alloc1(struct man *m)
 {
 
-	bzero(&m->meta, sizeof(struct man_meta));
+	memset(&m->meta, 0, sizeof(struct man_meta));
 	m->flags = 0;
-	m->last = calloc(1, sizeof(struct man_node));
-	if (NULL == m->last)
-		return(0);
+	m->last = mandoc_calloc(1, sizeof(struct man_node));
 	m->first = m->last;
 	m->last->type = MAN_ROOT;
+	m->last->tok = MAN_MAX;
 	m->next = MAN_NEXT_CHILD;
-	return(1);
 }
 
 
@@ -197,8 +214,25 @@ man_node_append(struct man *man, struct man_node *p)
 		/* NOTREACHED */
 	}
 	
+	assert(p->parent);
 	p->parent->nchild++;
 
+	if ( ! man_valid_pre(man, p))
+		return(0);
+
+	switch (p->type) {
+	case (MAN_HEAD):
+		assert(MAN_BLOCK == p->parent->type);
+		p->parent->head = p;
+		break;
+	case (MAN_BODY):
+		assert(MAN_BLOCK == p->parent->type);
+		p->parent->body = p;
+		break;
+	default:
+		break;
+	}
+
 	man->last = p;
 
 	switch (p->type) {
@@ -217,14 +251,11 @@ man_node_append(struct man *man, struct man_node *p)
 
 
 static struct man_node *
-man_node_alloc(int line, int pos, enum man_type type, int tok)
+man_node_alloc(int line, int pos, enum man_type type, enum mant tok)
 {
 	struct man_node *p;
 
-	p = calloc(1, sizeof(struct man_node));
-	if (NULL == p)
-		return(NULL);
-
+	p = mandoc_calloc(1, sizeof(struct man_node));
 	p->line = line;
 	p->pos = pos;
 	p->type = type;
@@ -234,54 +265,102 @@ man_node_alloc(int line, int pos, enum man_type type, 
 
 
 int
-man_elem_alloc(struct man *man, int line, int pos, int tok)
+man_elem_alloc(struct man *m, int line, int pos, enum mant tok)
 {
 	struct man_node *p;
 
 	p = man_node_alloc(line, pos, MAN_ELEM, tok);
-	if (NULL == p)
+	if ( ! man_node_append(m, p))
 		return(0);
-	return(man_node_append(man, p));
+	m->next = MAN_NEXT_CHILD;
+	return(1);
 }
 
 
 int
-man_word_alloc(struct man *man, 
-		int line, int pos, const char *word)
+man_head_alloc(struct man *m, int line, int pos, enum mant tok)
 {
-	struct man_node	*p;
+	struct man_node *p;
 
-	p = man_node_alloc(line, pos, MAN_TEXT, -1);
-	if (NULL == p)
+	p = man_node_alloc(line, pos, MAN_HEAD, tok);
+	if ( ! man_node_append(m, p))
 		return(0);
-	if (NULL == (p->string = strdup(word)))
+	m->next = MAN_NEXT_CHILD;
+	return(1);
+}
+
+
+int
+man_body_alloc(struct man *m, int line, int pos, enum mant tok)
+{
+	struct man_node *p;
+
+	p = man_node_alloc(line, pos, MAN_BODY, tok);
+	if ( ! man_node_append(m, p))
 		return(0);
-	return(man_node_append(man, p));
+	m->next = MAN_NEXT_CHILD;
+	return(1);
 }
 
 
-void
+int
+man_block_alloc(struct man *m, int line, int pos, enum mant tok)
+{
+	struct man_node *p;
+
+	p = man_node_alloc(line, pos, MAN_BLOCK, tok);
+	if ( ! man_node_append(m, p))
+		return(0);
+	m->next = MAN_NEXT_CHILD;
+	return(1);
+}
+
+
+int
+man_word_alloc(struct man *m, int line, int pos, const char *word)
+{
+	struct man_node	*n;
+	size_t		 sv, len;
+
+	len = strlen(word);
+
+	n = man_node_alloc(line, pos, MAN_TEXT, MAN_MAX);
+	n->string = mandoc_malloc(len + 1);
+	sv = strlcpy(n->string, word, len + 1);
+
+	/* Prohibit truncation. */
+	assert(sv < len + 1);
+
+	if ( ! man_node_append(m, n))
+		return(0);
+
+	m->next = MAN_NEXT_SIBLING;
+	return(1);
+}
+
+
+/*
+ * Free all of the resources held by a node.  This does NOT unlink a
+ * node from its context; for that, see man_node_unlink().
+ */
+static void
 man_node_free(struct man_node *p)
 {
 
 	if (p->string)
 		free(p->string);
-	if (p->parent)
-		p->parent->nchild--;
 	free(p);
 }
 
 
 void
-man_node_freelist(struct man_node *p)
+man_node_delete(struct man *m, struct man_node *p)
 {
 
-	if (p->child)
-		man_node_freelist(p->child);
-	if (p->next)
-		man_node_freelist(p->next);
+	while (p->child)
+		man_node_delete(m, p->child);
 
-	assert(0 == p->nchild);
+	man_node_unlink(m, p);
 	man_node_free(p);
 }
 
@@ -289,55 +368,140 @@ man_node_freelist(struct man_node *p)
 static int
 man_ptext(struct man *m, int line, char *buf)
 {
+	int		 i;
 
+	/* Ignore bogus comments. */
+
+	if ('\\' == buf[0] && '.' == buf[1] && '\"' == buf[2])
+		return(man_pwarn(m, line, 0, WBADCOMMENT));
+
+	/* Literal free-form text whitespace is preserved. */
+
+	if (MAN_LITERAL & m->flags) {
+		if ( ! man_word_alloc(m, line, 0, buf))
+			return(0);
+		goto descope;
+	}
+
+	/* Pump blank lines directly into the backend. */
+
+	for (i = 0; ' ' == buf[i]; i++)
+		/* Skip leading whitespace. */ ;
+
+	if ('\0' == buf[i]) {
+		/* Allocate a blank entry. */
+		if ( ! man_word_alloc(m, line, 0, ""))
+			return(0);
+		goto descope;
+	}
+
+	/* 
+	 * Warn if the last un-escaped character is whitespace. Then
+	 * strip away the remaining spaces (tabs stay!).   
+	 */
+
+	i = (int)strlen(buf);
+	assert(i);
+
+	if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
+		if (i > 1 && '\\' != buf[i - 2])
+			if ( ! man_pwarn(m, line, i - 1, WTSPACE))
+				return(0);
+
+		for (--i; i && ' ' == buf[i]; i--)
+			/* Spin back to non-space. */ ;
+
+		/* Jump ahead of escaped whitespace. */
+		i += '\\' == buf[i] ? 2 : 1;
+
+		buf[i] = '\0';
+	}
+
 	if ( ! man_word_alloc(m, line, 0, buf))
 		return(0);
-	m->next = MAN_NEXT_SIBLING;
 
 	/*
-	 * If this is one of the zany NLINE macros that consumes the
-	 * next line of input as being influenced, then close out the
-	 * existing macro "scope" and continue processing.
+	 * End-of-sentence check.  If the last character is an unescaped
+	 * EOS character, then flag the node as being the end of a
+	 * sentence.  The front-end will know how to interpret this.
 	 */
 
-	if ( ! (MAN_NLINE & m->flags))
-		return(1);
+	assert(i);
 
-	m->flags &= ~MAN_NLINE;
-	m->last = m->last->parent;
+	switch (buf[i - 1]) {
+	case ('.'):
+		if (i > 1 && '\\' == buf[i - 2])
+			break;
+		/* FALLTHROUGH */
+	case ('!'):
+		/* FALLTHROUGH */
+	case ('?'):
+		m->last->flags |= MAN_EOS;
+		break;
+	default:
+		break;
 
-	assert(MAN_ROOT != m->last->type);
-	if ( ! man_valid_post(m))
+	}
+
+descope:
+	/*
+	 * Co-ordinate what happens with having a next-line scope open:
+	 * first close out the element scope (if applicable), then close
+	 * out the block scope (also if applicable).
+	 */
+
+	if (MAN_ELINE & m->flags) {
+		m->flags &= ~MAN_ELINE;
+		if ( ! man_unscope(m, m->last->parent, WERRMAX))
+			return(0);
+	}
+
+	if ( ! (MAN_BLINE & m->flags))
+		return(1);
+	m->flags &= ~MAN_BLINE;
+
+	if ( ! man_unscope(m, m->last->parent, WERRMAX))
 		return(0);
-	if ( ! man_action_post(m))
-		return(0);
+	return(man_body_alloc(m, line, 0, m->last->tok));
+}
 
-	return(1);
+
+static int
+macrowarn(struct man *m, int ln, const char *buf)
+{
+	if ( ! (MAN_IGN_MACRO & m->pflags))
+		return(man_verr(m, ln, 0, "unknown macro: %s%s", 
+				buf, strlen(buf) > 3 ? "..." : ""));
+	return(man_vwarn(m, ln, 0, "unknown macro: %s%s",
+				buf, strlen(buf) > 3 ? "..." : ""));
 }
 
 
 int
 man_pmacro(struct man *m, int ln, char *buf)
 {
-	int		  i, j, c, ppos, fl;
-	char		  mac[5];
-	struct man_node	 *n;
+	int		 i, j, ppos;
+	enum mant	 tok;
+	char		 mac[5];
+	struct man_node	*n;
 
 	/* Comments and empties are quickly ignored. */
 
-	n = m->last;
-	fl = MAN_NLINE & m->flags;
+	if ('\0' == buf[1])
+		return(1);
 
-	if (0 == buf[1])
-		goto out;
-
 	i = 1;
 
-	if (' ' == buf[i]) {
+	/*
+	 * Skip whitespace between the control character and initial
+	 * text.  "Whitespace" is both spaces and tabs.
+	 */
+
+	if (' ' == buf[i] || '\t' == buf[i]) {
 		i++;
-		while (buf[i] && ' ' == buf[i])
+		while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
 			i++;
-		if (0 == buf[i])
+		if ('\0' == buf[i])
 			goto out;
 	}
 
@@ -346,35 +510,33 @@ man_pmacro(struct man *m, int ln, char *buf)
 	/* 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;
+
+		/* Check for invalid characters. */
+
+		if (isgraph((u_char)buf[i]))
+			continue;
+		return(man_perr(m, ln, i, WNPRINT));
 	}
 
-	mac[j] = 0;
+	mac[j] = '\0';
 
 	if (j == 4 || j < 1) {
 		if ( ! (MAN_IGN_MACRO & m->pflags)) {
-			(void)man_verr(m, ln, ppos, 
-				"ill-formed macro: %s", mac);
+			(void)man_perr(m, ln, ppos, WMACROFORM);
 			goto err;
 		} 
-		if ( ! man_vwarn(m, ln, ppos, 
-				"ill-formed macro: %s", mac))
+		if ( ! man_pwarn(m, ln, ppos, WMACROFORM))
 			goto err;
 		return(1);
 	}
 	
-	if (MAN_MAX == (c = man_hash_find(m->htab, mac))) {
-		if ( ! (MAN_IGN_MACRO & m->pflags)) {
-			(void)man_verr(m, ln, ppos, 
-				"unknown macro: %s", mac);
+	if (MAN_MAX == (tok = man_hash_find(mac))) {
+		if ( ! macrowarn(m, ln, mac))
 			goto err;
-		} 
-		if ( ! man_vwarn(m, ln, ppos, 
-				"unknown macro: %s", mac))
-			goto err;
 		return(1);
 	}
 
@@ -383,35 +545,107 @@ man_pmacro(struct man *m, int ln, char *buf)
 	while (buf[i] && ' ' == buf[i])
 		i++;
 
-	/* Begin recursive parse sequence. */
+	/* 
+	 * Trailing whitespace.  Note that tabs are allowed to be passed
+	 * into the parser as "text", so we only warn about spaces here.
+	 */
 
-	if ( ! man_macro(m, c, ln, ppos, &i, buf))
-		goto err;
+	if ('\0' == buf[i] && ' ' == buf[i - 1])
+		if ( ! man_pwarn(m, ln, i - 1, WTSPACE))
+			goto err;
 
-out:
-	if (fl) {
+	/* 
+	 * Remove prior ELINE macro, as it's being clobbering by a new
+	 * macro.  Note that NSCOPED macros do not close out ELINE
+	 * macros---they don't print text---so we let those slip by.
+	 * NOTE: we don't allow roff blocks (NOCLOSE) to be embedded
+	 * here because that would stipulate blocks as children of
+	 * elements!
+	 */
+
+	if ( ! (MAN_NSCOPED & man_macros[tok].flags) &&
+			m->flags & MAN_ELINE) {
+		assert(MAN_TEXT != m->last->type);
+
 		/*
-		 * A NLINE macro has been immediately followed with
-		 * another.  Close out the preceding macro's scope, and
-		 * continue.
+		 * This occurs in the following construction:
+		 *   .B
+		 *   .br
+		 *   .B
+		 *   .br
+		 *   I hate man macros.
+		 * Flat-out disallow this madness.
 		 */
-		assert(MAN_ROOT != m->last->type);
-		assert(m->last->parent);
-		assert(MAN_ROOT != m->last->parent->type);
+		if (MAN_NSCOPED & man_macros[m->last->tok].flags)
+			return(man_perr(m, ln, ppos, WLNSCOPE));
 
-		if (n != m->last)
-			m->last = m->last->parent;
+		n = m->last;
 
-		if ( ! man_valid_post(m))
+		assert(n);
+		assert(NULL == n->child);
+		assert(0 == n->nchild);
+
+		if ( ! man_nwarn(m, n, WLNSCOPE))
 			return(0);
-		if ( ! man_action_post(m))
-			return(0);
-		m->next = MAN_NEXT_SIBLING;
-		m->flags &= ~MAN_NLINE;
-	} 
 
-	return(1);
+		man_node_delete(m, n);
+		m->flags &= ~MAN_ELINE;
+	}
 
+	/*
+	 * Save the fact that we're in the next-line for a block.  In
+	 * this way, embedded roff instructions can "remember" state
+	 * when they exit.
+	 */
+
+	if (MAN_BLINE & m->flags)
+		m->flags |= MAN_BPLINE;
+
+	/* Call to handler... */
+
+	assert(man_macros[tok].fp);
+	if ( ! (*man_macros[tok].fp)(m, tok, ln, ppos, &i, buf))
+		goto err;
+
+out:
+	/* 
+	 * We weren't in a block-line scope when entering the
+	 * above-parsed macro, so return.
+	 */
+
+	if ( ! (MAN_BPLINE & m->flags)) {
+		m->flags &= ~MAN_ILINE; 
+		return(1);
+	}
+	m->flags &= ~MAN_BPLINE;
+
+	/*
+	 * If we're in a block scope, then allow this macro to slip by
+	 * without closing scope around it.
+	 */
+
+	if (MAN_ILINE & m->flags) {
+		m->flags &= ~MAN_ILINE;
+		return(1);
+	}
+
+	/* 
+	 * If we've opened a new next-line element scope, then return
+	 * now, as the next line will close out the block scope.
+	 */
+
+	if (MAN_ELINE & m->flags)
+		return(1);
+
+	/* Close out the block scope opened in the prior line.  */
+
+	assert(MAN_BLINE & m->flags);
+	m->flags &= ~MAN_BLINE;
+
+	if ( ! man_unscope(m, m->last->parent, WERRMAX))
+		return(0);
+	return(man_body_alloc(m, ln, 0, m->last->tok));
+
 err:	/* Error out. */
 
 	m->flags |= MAN_HALT;
@@ -463,4 +697,46 @@ man_err(struct man *m, int line, int pos, int iserr, e
 		return(man_verr(m, line, pos, p));
 
 	return(man_vwarn(m, line, pos, p));
+}
+
+
+/*
+ * Unlink a node from its context.  If "m" is provided, the last parse
+ * point will also be adjusted accordingly.
+ */
+static void
+man_node_unlink(struct man *m, struct man_node *n)
+{
+
+	/* 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) {
+		/*XXX: this can occur when bailing from validation. */
+		/*assert(NULL == n->next);*/
+		if (n->prev) {
+			m->last = n->prev;
+			m->next = MAN_NEXT_SIBLING;
+		} else {
+			m->last = n->parent;
+			m->next = MAN_NEXT_CHILD;
+		}
+	}
+
+	if (m && m->first == n)
+		m->first = NULL;
 }