Implement the first steps of equation parsing from within libmdoc.

This consists of a shim around the text parser that calls out to libroff
if equation components exist on the line.  Right now this will do
nothing, as the equation delimiter always returns nil.

7 files changed, 119 insertions, 26 deletions

diff --git a/Makefile b/Makefile
index 47cacdf0..351dcc08 100644
--- a/Makefile
+++ b/Makefile
@@ -332,6 +332,11 @@ clean:
 	rm -f config.h config.log $(COMPAT_OBJS) $(COMPAT_LNS)
 	rm -f mdocml.tar.gz mdocml-win32.zip
 	rm -f index.html $(INDEX_OBJS)
+	rm -rf test-strlcpy.DSYM
+	rm -rf test-strlcat.DSYM 
+	rm -rf test-strptime.DSYM 
+	rm -rf test-mmap.DSYM 
+	rm -rf test-getsubopt.DSYM
 
 install: all
 	mkdir -p $(DESTDIR)$(BINDIR)
diff --git a/eqn.c b/eqn.c
index 9c3b9783..37f01bcb 100644
--- a/eqn.c
+++ b/eqn.c
@@ -1,4 +1,4 @@
-/*	$Id: eqn.c,v 1.37 2011/07/23 18:44:42 kristaps Exp $ */
+/*	$Id: eqn.c,v 1.38 2011/07/25 15:37:00 kristaps Exp $ */
 /*
  * Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -294,8 +294,7 @@ eqn_read(struct eqn_node **epp, int ln,
 	 */
 
 	if (0 == strncmp(p, ".EN", 3)) {
-		er = eqn_end(ep);
-		*epp = NULL;
+		er = eqn_end(epp);
 		p += 3;
 		while (' ' == *p || '\t' == *p)
 			p++;
@@ -333,7 +332,7 @@ eqn_alloc(const char *name, int pos, int line, struct mparse *parse)
 
 	p = mandoc_calloc(1, sizeof(struct eqn_node));
 
-	if ('\0' != *name) {
+	if (name && '\0' != *name) {
 		sz = strlen(name);
 		assert(sz);
 		do {
@@ -352,11 +351,15 @@ eqn_alloc(const char *name, int pos, int line, struct mparse *parse)
 }
 
 enum rofferr
-eqn_end(struct eqn_node *ep)
+eqn_end(struct eqn_node **epp)
 {
+	struct eqn_node	*ep;
 	struct eqn_box	*root;
 	enum eqn_rest	 c;
 
+	ep = *epp;
+	*epp = NULL;
+
 	ep->eqn.root = mandoc_calloc(1, sizeof(struct eqn_box));
 
 	root = ep->eqn.root;
diff --git a/libmandoc.h b/libmandoc.h
index ad3ed254..1515633a 100644
--- a/libmandoc.h
+++ b/libmandoc.h
@@ -1,4 +1,4 @@
-/*	$Id: libmandoc.h,v 1.24 2011/07/21 15:21:13 kristaps Exp $ */
+/*	$Id: libmandoc.h,v 1.25 2011/07/25 15:37:00 kristaps Exp $ */
 /*
  * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -76,6 +76,10 @@ void		 roff_endparse(struct roff *);
 int		 roff_regisset(const struct roff *, enum regs);
 unsigned int	 roff_regget(const struct roff *, enum regs);
 void		 roff_regunset(struct roff *, enum regs);
+char		 roff_eqndelim(const struct roff *);
+void		 roff_openeqn(struct roff *, const char *, 
+			int, int, const char *);
+int		 roff_closeeqn(struct roff *);
 
 const struct tbl_span	*roff_span(const struct roff *);
 const struct eqn	*roff_eqn(const struct roff *);
diff --git a/libroff.h b/libroff.h
index da4c39ec..0bdd5a36 100644
--- a/libroff.h
+++ b/libroff.h
@@ -1,4 +1,4 @@
-/*	$Id: libroff.h,v 1.26 2011/07/23 18:41:18 kristaps Exp $ */
+/*	$Id: libroff.h,v 1.27 2011/07/25 15:37:00 kristaps Exp $ */
 /*
  * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -72,9 +72,9 @@ int		 tbl_layout(struct tbl_node *, int, const char *);
 int		 tbl_data(struct tbl_node *, int, const char *);
 int		 tbl_cdata(struct tbl_node *, int, const char *);
 const struct tbl_span	*tbl_span(struct tbl_node *);
-void		 tbl_end(struct tbl_node *);
+void		 tbl_end(struct tbl_node **);
 struct eqn_node	*eqn_alloc(const char *, int, int, struct mparse *);
-enum rofferr	 eqn_end(struct eqn_node *);
+enum rofferr	 eqn_end(struct eqn_node **);
 void		 eqn_free(struct eqn_node *);
 enum rofferr 	 eqn_read(struct eqn_node **, int, 
 			const char *, int, int *);
diff --git a/mdoc.c b/mdoc.c
index 29c5ec55..786ef67e 100644
--- a/mdoc.c
+++ b/mdoc.c
@@ -1,4 +1,4 @@
-/*	$Id: mdoc.c,v 1.190 2011/07/21 10:24:35 kristaps Exp $ */
+/*	$Id: mdoc.c,v 1.191 2011/07/25 15:37:00 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org>
@@ -97,6 +97,7 @@ static	struct mdoc_node *node_alloc(struct mdoc *, int, int,
 				enum mdoct, enum mdoc_type);
 static	int		  node_append(struct mdoc *, 
 				struct mdoc_node *);
+static	int		  mdoc_preptext(struct mdoc *, int, char *, int);
 static	int		  mdoc_ptext(struct mdoc *, int, char *, int);
 static	int		  mdoc_pmacro(struct mdoc *, int, char *, int);
 
@@ -299,7 +300,7 @@ mdoc_parseln(struct mdoc *m, int ln, char *buf, int offs)
 
 	return(mandoc_getcontrol(buf, &offs) ?
 			mdoc_pmacro(m, ln, buf, offs) :
-			mdoc_ptext(m, ln, buf, offs));
+			mdoc_preptext(m, ln, buf, offs));
 }
 
 int
@@ -650,6 +651,57 @@ mdoc_node_delete(struct mdoc *m, struct mdoc_node *p)
 	mdoc_node_free(p);
 }
 
+/*
+ * Pre-treat a text line.
+ * Text lines can consist of equations, which must be handled apart from
+ * the regular text.
+ * Thus, use this function to step through a line checking if it has any
+ * equations embedded in it.
+ * This must handle multiple equations AND equations that do not end at
+ * the end-of-line, i.e., will re-enter in the next roff parse.
+ */
+static int
+mdoc_preptext(struct mdoc *m, int line, char *buf, int offs)
+{
+	char		*start, *end;
+	char		 delim;
+
+	while ('\0' != buf[offs]) {
+		/* Mark starting position if eqn is set. */
+		start = NULL;
+		if ('\0' != (delim = roff_eqndelim(m->roff)))
+			if (NULL != (start = strchr(buf + offs, delim)))
+				*start++ = '\0';
+
+		/* Parse text as normal. */
+		if ( ! mdoc_ptext(m, line, buf, offs))
+			return(0);
+
+		/* Continue only if an equation exists. */
+		if (NULL == start)
+			break;
+
+		/* Read past the end of the equation. */
+		offs += start - (buf + offs);
+		assert(start == &buf[offs]);
+		if (NULL != (end = strchr(buf + offs, delim))) {
+			*end++ = '\0';
+			while (' ' == *end)
+				end++;
+		}
+
+		/* Parse the equation itself. */
+		roff_openeqn(m->roff, NULL, line, offs, buf);
+
+		/* Process a finished equation? */
+		if (roff_closeeqn(m->roff))
+			if ( ! mdoc_addeqn(m, roff_eqn(m->roff)))
+				return(0);
+		offs += (end - (buf + offs));
+	} 
+
+	return(1);
+}
 
 /*
  * Parse free-form text, that is, a line that does not begin with the
diff --git a/roff.c b/roff.c
index 80b5cec8..b43211fb 100644
--- a/roff.c
+++ b/roff.c
@@ -1,4 +1,4 @@
-/*	$Id: roff.c,v 1.150 2011/07/23 18:41:18 kristaps Exp $ */
+/*	$Id: roff.c,v 1.151 2011/07/25 15:37:00 kristaps Exp $ */
 /*
  * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
@@ -584,15 +584,13 @@ roff_endparse(struct roff *r)
 	if (r->eqn) {
 		mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse, 
 				r->eqn->eqn.ln, r->eqn->eqn.pos, NULL);
-		eqn_end(r->eqn);
-		r->eqn = NULL;
+		eqn_end(&r->eqn);
 	}
 
 	if (r->tbl) {
 		mandoc_msg(MANDOCERR_SCOPEEXIT, r->parse, 
 				r->tbl->line, r->tbl->pos, NULL);
-		tbl_end(r->tbl);
-		r->tbl = NULL;
+		tbl_end(&r->tbl);
 	}
 }
 
@@ -1182,9 +1180,8 @@ roff_TE(ROFF_ARGS)
 	if (NULL == r->tbl)
 		mandoc_msg(MANDOCERR_NOSCOPE, r->parse, ln, ppos, NULL);
 	else
-		tbl_end(r->tbl);
+		tbl_end(&r->tbl);
 
-	r->tbl = NULL;
 	return(ROFF_IGN);
 }
 
@@ -1201,14 +1198,22 @@ roff_T_(ROFF_ARGS)
 	return(ROFF_IGN);
 }
 
-/* ARGSUSED */
-static enum rofferr
-roff_EQ(ROFF_ARGS)
+int
+roff_closeeqn(struct roff *r)
 {
-	struct eqn_node	*e;
+
+	return(r->eqn && ROFF_EQN == eqn_end(&r->eqn) ? 1 : 0);
+}
+
+void
+roff_openeqn(struct roff *r, const char *name, int line, 
+		int offs, const char *buf)
+{
+	struct eqn_node *e;
+	int		 poff;
 
 	assert(NULL == r->eqn);
-	e = eqn_alloc(*bufp + pos, ppos, ln, r->parse);
+	e = eqn_alloc(name, offs, line, r->parse);
 
 	if (r->last_eqn)
 		r->last_eqn->next = e;
@@ -1216,6 +1221,19 @@ roff_EQ(ROFF_ARGS)
 		r->first_eqn = r->last_eqn = e;
 
 	r->eqn = r->last_eqn = e;
+
+	if (buf) {
+		poff = 0;
+		eqn_read(&r->eqn, line, buf, offs, &poff);
+	}
+}
+
+/* ARGSUSED */
+static enum rofferr
+roff_EQ(ROFF_ARGS)
+{
+
+	roff_openeqn(r, *bufp + pos, ln, ppos, NULL);
 	return(ROFF_IGN);
 }
 
@@ -1236,7 +1254,7 @@ roff_TS(ROFF_ARGS)
 
 	if (r->tbl) {
 		mandoc_msg(MANDOCERR_SCOPEBROKEN, r->parse, ln, ppos, NULL);
-		tbl_end(r->tbl);
+		tbl_end(&r->tbl);
 	}
 
 	t = tbl_alloc(ppos, ln, r->parse);
@@ -1472,3 +1490,10 @@ roff_eqn(const struct roff *r)
 	
 	return(r->last_eqn ? &r->last_eqn->eqn : NULL);
 }
+
+char
+roff_eqndelim(const struct roff *r)
+{
+
+	return('\0');
+}
diff --git a/tbl.c b/tbl.c
index 6ef2f735..b3d651be 100644
--- a/tbl.c
+++ b/tbl.c
@@ -1,4 +1,4 @@
-/*	$Id: tbl.c,v 1.25 2011/04/04 23:04:38 kristaps Exp $ */
+/*	$Id: tbl.c,v 1.26 2011/07/25 15:37:00 kristaps Exp $ */
 /*
  * Copyright (c) 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2011 Ingo Schwarze <schwarze@openbsd.org>
@@ -154,8 +154,12 @@ tbl_span(struct tbl_node *tbl)
 }
 
 void
-tbl_end(struct tbl_node *tbl)
+tbl_end(struct tbl_node **tblp)
 {
+	struct tbl_node	*tbl;
+
+	tbl = *tblp;
+	*tblp = NULL;
 
 	if (NULL == tbl->first_span || NULL == tbl->first_span->first)
 		mandoc_msg(MANDOCERR_TBLNODATA, tbl->parse,