===================================================================
RCS file: /cvs/mandoc/out.c,v
retrieving revision 1.7
retrieving revision 1.14
diff -u -p -r1.7 -r1.14
--- mandoc/out.c	2009/10/22 18:59:00	1.7
+++ mandoc/out.c	2010/04/07 11:25:38	1.14
@@ -1,4 +1,4 @@
-/*	$Id: out.c,v 1.7 2009/10/22 18:59:00 kristaps Exp $ */
+/*	$Id: out.c,v 1.14 2010/04/07 11:25:38 kristaps Exp $ */
 /*
  * Copyright (c) 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>
@@ -25,10 +29,26 @@
 
 #include "out.h"
 
-#ifdef __linux__
-extern	size_t	  strlcat(char *, const char *, size_t);
-#endif
+/* See a2roffdeco(). */
+#define	C2LIM(c, l) do { \
+	(l) = 1; \
+	if ('[' == (c) || '\'' == (c)) \
+		(l) = 0; \
+	else if ('(' == (c)) \
+		(l) = 2; } \
+	while (/* CONSTCOND */ 0)
 
+/* See a2roffdeco(). */
+#define	C2TERM(c, t) do { \
+	(t) = 0; \
+	if ('\'' == (c)) \
+		(t) = 1; \
+	else if ('[' == (c)) \
+		(t) = 2; \
+	else if ('(' == (c)) \
+		(t) = 3; } \
+	while (/* CONSTCOND */ 0)
+
 /* 
  * Convert a `scaling unit' to a consistent form, or fail.  Scaling
  * units are documented in groff.7, mdoc.7, man.7.
@@ -166,3 +186,215 @@ time2a(time_t t, char *dst, size_t sz)
 	(void)strftime(p, sz, "%Y", &tm);
 }
 
+
+/* 
+ * Returns length of parsed string (the leading "\" should NOT be
+ * included).  This can be zero if the current character is the nil
+ * terminator.  "d" is set to the type of parsed decorator, which may
+ * have an adjoining "word" of size "sz" (e.g., "(ab" -> "ab", 2).
+ */
+int
+a2roffdeco(enum roffdeco *d,
+		const char **word, size_t *sz)
+{
+	int		 j, offs, term, lim;
+	char		 set;
+	const char	*wp, *sp;
+
+	*d = DECO_NONE;
+	wp = *word;
+	offs = 0;
+
+	switch ((set = *wp)) {
+	case ('\0'):
+		return(0);
+
+	case ('('):
+		if ('\0' == *(++wp))
+			return(1);
+		if ('\0' == *(wp + 1))
+			return(2);
+
+		*d = DECO_SPECIAL;
+		*sz = 2;
+		*word = wp;
+		return(3);
+
+	case ('F'):
+		/* FALLTHROUGH */
+	case ('f'):
+		/*
+		 * FIXME: this needs work and consolidation (it should
+		 * follow the sequence that special characters do, for
+		 * one), but isn't a priority at the moment.  Note, for
+		 * one, that in reality \fB != \FB, although here we let
+		 * these slip by.
+		 */
+		switch (*(++wp)) {
+		case ('\0'):
+			return(1);
+		case ('3'):
+			/* FALLTHROUGH */
+		case ('B'):
+			*d = DECO_BOLD;
+			return(2);
+		case ('2'):
+			/* FALLTHROUGH */
+		case ('I'):
+			*d = DECO_ITALIC;
+			return(2);
+		case ('P'):
+			*d = DECO_PREVIOUS;
+			return(2);
+		case ('1'):
+			/* FALLTHROUGH */
+		case ('R'):
+			*d = DECO_ROMAN;
+			return(2);
+		case ('('):
+			if ('\0' == *(++wp))
+				return(2);
+			if ('\0' == *(wp + 1))
+				return(3);
+
+			*d = 'F' == set ? DECO_FFONT : DECO_FONT;
+			*sz = 2;
+			*word = wp;
+			return(4);
+		case ('['):
+			*word = ++wp;
+			for (j = 0; *wp && ']' != *wp; wp++, j++)
+				/* Loop... */ ;
+
+			if ('\0' == *wp)
+				return(j + 2);
+
+			*d = 'F' == set ? DECO_FFONT : DECO_FONT;
+			*sz = (size_t)j;
+			return(j + 3);
+		default:
+			break;
+		}
+
+		*d = 'F' == set ? DECO_FFONT : DECO_FONT;
+		*sz = 1;
+		*word = wp;
+		return(2);
+
+	case ('*'):
+		switch (*(++wp)) {
+		case ('\0'):
+			return(1);
+
+		case ('('):
+			if ('\0' == *(++wp))
+				return(2);
+			if ('\0' == *(wp + 1))
+				return(3);
+
+			*d = DECO_RESERVED;
+			*sz = 2;
+			*word = wp;
+			return(4);
+
+		case ('['):
+			*word = ++wp;
+			for (j = 0; *wp && ']' != *wp; wp++, j++)
+				/* Loop... */ ;
+
+			if ('\0' == *wp)
+				return(j + 2);
+
+			*d = DECO_RESERVED;
+			*sz = (size_t)j;
+			return(j + 3);
+
+		default:
+			break;
+		}
+
+		*d = DECO_RESERVED;
+		*sz = 1;
+		*word = wp;
+		return(2);
+
+	case ('s'):
+		sp = wp;
+		if ('\0' == *(++wp))
+			return(1);
+
+		C2LIM(*wp, lim);
+		C2TERM(*wp, term);
+
+		if (term) 
+			wp++;
+
+		*word = wp;
+
+		if (*wp == '+' || *wp == '-')
+			++wp;
+
+		switch (*wp) {
+		case ('\''):
+			/* FALLTHROUGH */
+		case ('['):
+			/* FALLTHROUGH */
+		case ('('):
+			if (term) 
+				return((int)(wp - sp));
+
+			C2LIM(*wp, lim);
+			C2TERM(*wp, term);
+			wp++;
+			break;
+		default:
+			break;
+		}
+
+		if ( ! isdigit((u_char)*wp))
+			return((int)(wp - sp));
+
+		for (j = 0; isdigit((u_char)*wp); j++) {
+			if (lim && j >= lim)
+				break;
+			++wp;
+		}
+
+		if (term && term < 3) {
+			if (1 == term && *wp != '\'')
+				return((int)(wp - sp));
+			if (2 == term && *wp != ']')
+				return((int)(wp - sp));
+			++wp;
+		}
+
+		*d = DECO_SIZE;
+		return((int)(wp - sp));
+
+	case ('['):
+		*word = ++wp;
+
+		for (j = 0; *wp && ']' != *wp; wp++, j++)
+			/* Loop... */ ;
+
+		if ('\0' == *wp)
+			return(j + 1);
+
+		*d = DECO_SPECIAL;
+		*sz = (size_t)j;
+		return(j + 2);
+
+	case ('c'):
+		*d = DECO_NOSPACE;
+		*sz = 1;
+		return(1);
+
+	default:
+		break;
+	}
+
+	*d = DECO_SPECIAL;
+	*word = wp;
+	*sz = 1;
+	return(1);
+}