Support nested `{, }' subexpressions in eqn. Document in code.

4 files changed, 77 insertions, 24 deletions

diff --git a/eqn.c b/eqn.c
index 8383a878..9aa0af78 100644
--- a/eqn.c
+++ b/eqn.c
@@ -1,4 +1,4 @@
-/*	$Id: eqn.c,v 1.14 2011/07/21 11:57:56 kristaps Exp $ */
+/*	$Id: eqn.c,v 1.15 2011/07/21 12:30:44 kristaps Exp $ */
 /*
  * Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -54,7 +54,8 @@ static	const char	*eqn_nexttok(struct eqn_node *, size_t *);
 static	const char	*eqn_nextrawtok(struct eqn_node *, size_t *);
 static	const char	*eqn_next(struct eqn_node *, 
 				char, size_t *, int);
-static	int		 eqn_box(struct eqn_node *, struct eqn_box *);
+static	int		 eqn_box(struct eqn_node *, 
+				struct eqn_box *, struct eqn_box **);
 
 static	const struct eqnpart eqnparts[EQN__MAX] = {
 	{ "define", 6, eqn_do_define }, /* EQN_DEFINE */
@@ -119,7 +120,8 @@ eqn_alloc(int pos, int line, struct mparse *parse)
 enum rofferr
 eqn_end(struct eqn_node *ep)
 {
-	struct eqn_box	*root;
+	struct eqn_box	*root, *last;
+	int		 c;
 
 	ep->eqn.root = root = 
 		mandoc_calloc(1, sizeof(struct eqn_box));
@@ -129,21 +131,38 @@ eqn_end(struct eqn_node *ep)
 		return(ROFF_IGN);
 
 	/*
-	 * Validate the expression.
-	 * Use the grammar found in the literature.
+	 * Run the parser.
+	 * If we return before reaching the end of our input, our scope
+	 * is still open somewhere.
+	 * If we return alright but don't have a symmetric scoping, then
+	 * something's not right either.
+	 * Otherwise, return the equation.
 	 */
 
-	return(eqn_box(ep, root) < 0 ? ROFF_IGN : ROFF_EQN);
+	if ((c = eqn_box(ep, root, &last)) > 0) {
+		EQN_MSG(MANDOCERR_EQNNSCOPE, ep);
+		c = 0;
+	} else if (0 == c && last != root)
+		EQN_MSG(MANDOCERR_EQNSCOPE, ep);
+
+	return(1 == c ? ROFF_EQN : ROFF_IGN);
 }
 
 static int
-eqn_box(struct eqn_node *ep, struct eqn_box *last)
+eqn_box(struct eqn_node *ep, struct eqn_box *last, struct eqn_box **sv)
 {
 	size_t		 sz;
 	const char	*start;
-	int		 i, nextc;
+	int		 c, i, nextc;
 	struct eqn_box	*bp;
 
+	/* 
+	 * Mark our last level of subexpression. 
+	 * Also mark whether that the next node should be a
+	 * subexpression node.
+	 */
+
+	*sv = last;
 	nextc = 1;
 again:
 	if (NULL == (start = eqn_nexttok(ep, &sz)))
@@ -160,19 +179,39 @@ again:
 		goto again;
 	} 
 
-	bp = mandoc_calloc(1, sizeof(struct eqn_box));
-	bp->type = EQN_TEXT;
+	/* Exit this [hopefully] subexpression. */
 
+	if (sz == 1 && 0 == strncmp("}", start, 1)) 
+		return(1);
+
+	bp = mandoc_calloc(1, sizeof(struct eqn_box));
 	if (nextc)
 		last->child = bp;
 	else
 		last->next = bp;
 
+	last = bp;
+
+	/* 
+	 * See if we're to open a new subexpression.
+	 * If so, mark our node as such and descend.
+	 */
+
+	if (sz == 1 && 0 == strncmp("{", start, 1)) {
+		bp->type = EQN_SUBEXPR;
+		c = eqn_box(ep, bp, sv);
+
+		nextc = 0;
+		goto again;
+	}
+
+	/* A regular text node. */
+
+	bp->type = EQN_TEXT;
 	bp->text = mandoc_malloc(sz + 1);
 	*bp->text = '\0';
 	strlcat(bp->text, start, sz + 1);
 
-	last = bp;
 	nextc = 0;
 	goto again;
 }
diff --git a/mandoc.h b/mandoc.h
index 7c46c763..b585ace8 100644
--- a/mandoc.h
+++ b/mandoc.h
@@ -1,4 +1,4 @@
-/*	$Id: mandoc.h,v 1.83 2011/07/21 11:34:53 kristaps Exp $ */
+/*	$Id: mandoc.h,v 1.84 2011/07/21 12:30:44 kristaps Exp $ */
 /*
  * Copyright (c) 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -112,6 +112,8 @@ enum	mandocerr {
 	/* related to equations */
 	MANDOCERR_EQNARGS, /* bad equation macro arguments */
 	MANDOCERR_EQNNEST, /* too many nested equation defines */
+	MANDOCERR_EQNNSCOPE, /* unexpected equation scope closure*/
+	MANDOCERR_EQNSCOPE, /* equation scope open on exit */
 
 	/* related to tables */
 	MANDOCERR_TBL, /* bad table syntax */
@@ -278,19 +280,24 @@ struct	tbl_span {
 };
 
 enum	eqn_boxt {
-	EQN_ROOT,
-	EQN_TEXT
+	EQN_ROOT, /* root of parse tree */
+	EQN_TEXT, /* text (number, variable, whatever) */
+	EQN_SUBEXPR /* nested subexpression */
 };
 
+/*
+ * A "box" is a parsed mathematical expression as defined by the eqn.7
+ * grammar.
+ */
 struct	eqn_box {
-	enum eqn_boxt	  type;
-	struct eqn_box	 *child;
-	struct eqn_box	 *next;
-	char		 *text;
+	enum eqn_boxt	  type; /* type of node */
+	struct eqn_box	 *child; /* child node */
+	struct eqn_box	 *next; /* next in tree */
+	char		 *text; /* text (or NULL) */
 };
 
 struct	eqn {
-	struct eqn_box	 *root;
+	struct eqn_box	 *root; /* root mathematical expression */
 	int		  ln; /* invocation line */
 	int		  pos; /* invocation position */
 };
diff --git a/read.c b/read.c
index 8d59b23f..c79f080f 100644
--- a/read.c
+++ b/read.c
@@ -1,4 +1,4 @@
-/*	$Id: read.c,v 1.19 2011/07/18 14:30:51 kristaps Exp $ */
+/*	$Id: read.c,v 1.20 2011/07/21 12:30:44 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
@@ -154,6 +154,8 @@ static	const char * const	mandocerrs[MANDOCERR_MAX] = {
 	/* related to equations */
 	"bad equation macro syntax",
 	"too many nested equation defines",
+	"unexpected equation scope closure",
+	"equation scope open on exit",
 
 	/* related to tables */
 	"bad table syntax",
diff --git a/tree.c b/tree.c
index a446fb14..0b05b5ea 100644
--- a/tree.c
+++ b/tree.c
@@ -1,4 +1,4 @@
-/*	$Id: tree.c,v 1.38 2011/07/21 11:34:53 kristaps Exp $ */
+/*	$Id: tree.c,v 1.39 2011/07/21 12:30:44 kristaps Exp $ */
 /*
  * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -272,14 +272,19 @@ print_box(const struct eqn_box *ep, int indent)
 	case (EQN_ROOT):
 		puts("eqn-root");
 		print_box(ep->child, indent + 1);
-		return;
+		break;
+	case (EQN_SUBEXPR):
+		puts("eqn-subxpr");
+		print_box(ep->child, indent + 1);
+		break;
 	case (EQN_TEXT):
 		printf("eqn-text: [%s]\n", ep->text);
-		print_box(ep->next, indent);
-		return;
+		break;
 	default:
 		break;
 	}
+
+	print_box(ep->next, indent);
 }
 
 static void