===================================================================
RCS file: /cvs/mandoc/term.c,v
retrieving revision 1.200
retrieving revision 1.209
diff -u -p -r1.200 -r1.209
--- mandoc/term.c	2011/09/19 22:36:16	1.200
+++ mandoc/term.c	2013/08/08 20:07:47	1.209
@@ -1,7 +1,7 @@
-/*	$Id: term.c,v 1.200 2011/09/19 22:36:16 schwarze Exp $ */
+/*	$Id: term.c,v 1.209 2013/08/08 20:07:47 schwarze Exp $ */
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
- * Copyright (c) 2010, 2011 Ingo Schwarze <schwarze@openbsd.org>
+ * Copyright (c) 2010, 2011, 2012, 2013 Ingo Schwarze <schwarze@openbsd.org>
  *
  * Permission to use, copy, modify, and distribute this software for any
  * purpose with or without fee is hereby granted, provided that the above
@@ -33,6 +33,7 @@
 #include "term.h"
 #include "main.h"
 
+static	size_t		 cond_width(const struct termp *, int, int *);
 static	void		 adjbuf(struct termp *p, int);
 static	void		 bufferc(struct termp *, char);
 static	void		 encode(struct termp *, const char *, size_t);
@@ -105,6 +106,7 @@ void
 term_flushln(struct termp *p)
 {
 	int		 i;     /* current input position in p->buf */
+	int		 ntab;	/* number of tabs to prepend */
 	size_t		 vis;   /* current visual position on output */
 	size_t		 vbl;   /* number of blanks to prepend to output */
 	size_t		 vend;	/* end of word visual position on output */
@@ -143,10 +145,12 @@ term_flushln(struct termp *p)
 		 * Handle literal tab characters: collapse all
 		 * subsequent tabs into a single huge set of spaces.
 		 */
+		ntab = 0;
 		while (i < p->col && '\t' == p->buf[i]) {
 			vend = (vis / p->tabwidth + 1) * p->tabwidth;
 			vbl += vend - vis;
 			vis = vend;
+			ntab++;
 			i++;
 		}
 
@@ -158,7 +162,7 @@ term_flushln(struct termp *p)
 		 */
 
 		for (j = i, jhy = 0; j < p->col; j++) {
-			if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j])
+			if (' ' == p->buf[j] || '\t' == p->buf[j])
 				break;
 
 			/* Back over the the last printed character. */
@@ -184,15 +188,18 @@ term_flushln(struct termp *p)
 		if (vend > bp && 0 == jhy && vis > 0) {
 			vend -= vis;
 			(*p->endline)(p);
+			p->viscol = 0;
 			if (TERMP_NOBREAK & p->flags) {
-				p->viscol = p->rmargin;
-				(*p->advance)(p, p->rmargin);
+				vbl = p->rmargin;
 				vend += p->rmargin - p->offset;
-			} else {
-				p->viscol = 0;
+			} else
 				vbl = p->offset;
-			}
 
+			/* use pending tabs on the new line */
+
+			if (0 < ntab)
+				vbl += ntab * p->tabwidth;
+
 			/* Remove the p->overstep width. */
 
 			bp += (size_t)p->overstep;
@@ -266,16 +273,11 @@ term_flushln(struct termp *p)
 		p->overstep = (int)(vis - maxvis + (*p->width)(p, ' '));
 
 		/*
-		 * Behave exactly the same way as groff:
 		 * If we have overstepped the margin, temporarily move
 		 * it to the right and flag the rest of the line to be
 		 * shorter.
-		 * If we landed right at the margin, be happy.
-		 * If we are one step before the margin, temporarily
-		 * move it one step LEFT and flag the rest of the line
-		 * to be longer.
 		 */
-		if (p->overstep < -1)
+		if (p->overstep < 0)
 			p->overstep = 0;
 		return;
 
@@ -318,7 +320,10 @@ term_vspace(struct termp *p)
 
 	term_newln(p);
 	p->viscol = 0;
-	(*p->endline)(p);
+	if (0 < p->skipvsp)
+		p->skipvsp--;
+	else
+		(*p->endline)(p);
 }
 
 void
@@ -371,7 +376,7 @@ void
 term_fontpopq(struct termp *p, const void *key)
 {
 
-	while (p->fonti >= 0 && key != &p->fontq[p->fonti])
+	while (p->fonti >= 0 && key < (void *)(p->fontq + p->fonti))
 		p->fonti--;
 	assert(p->fonti >= 0);
 }
@@ -401,14 +406,14 @@ term_word(struct termp *p, const char *word)
 
 	if ( ! (TERMP_NOSPACE & p->flags)) {
 		if ( ! (TERMP_KEEP & p->flags)) {
-			if (TERMP_PREKEEP & p->flags)
-				p->flags |= TERMP_KEEP;
 			bufferc(p, ' ');
 			if (TERMP_SENTENCE & p->flags)
 				bufferc(p, ' ');
 		} else
 			bufferc(p, ASCII_NBRSP);
 	}
+	if (TERMP_PREKEEP & p->flags)
+		p->flags |= TERMP_KEEP;
 
 	if ( ! (p->flags & TERMP_NONOSPACE))
 		p->flags &= ~TERMP_NOSPACE;
@@ -418,12 +423,17 @@ term_word(struct termp *p, const char *word)
 	p->flags &= ~(TERMP_SENTENCE | TERMP_IGNDELIM);
 
 	while ('\0' != *word) {
-		if ((ssz = strcspn(word, "\\")) > 0)
+		if ('\\' != *word) {
+			if (TERMP_SKIPCHAR & p->flags) {
+				p->flags &= ~TERMP_SKIPCHAR;
+				word++;
+				continue;
+			}
+			ssz = strcspn(word, "\\");
 			encode(p, word, ssz);
-
-		word += (int)ssz;
-		if ('\\' != *word)
+			word += (int)ssz;
 			continue;
+		}
 
 		word++;
 		esc = mandoc_escape(&word, &seq, &sz);
@@ -470,6 +480,9 @@ term_word(struct termp *p, const char *word)
 		case (ESCAPE_FONTITALIC):
 			term_fontrepl(p, TERMFONT_UNDER);
 			break;
+		case (ESCAPE_FONTBI):
+			term_fontrepl(p, TERMFONT_BI);
+			break;
 		case (ESCAPE_FONT):
 			/* FALLTHROUGH */
 		case (ESCAPE_FONTROMAN):
@@ -479,9 +492,14 @@ term_word(struct termp *p, const char *word)
 			term_fontlast(p);
 			break;
 		case (ESCAPE_NOSPACE):
-			if ('\0' == *word)
+			if (TERMP_SKIPCHAR & p->flags)
+				p->flags &= ~TERMP_SKIPCHAR;
+			else if ('\0' == *word)
 				p->flags |= TERMP_NOSPACE;
 			break;
+		case (ESCAPE_SKIPCHAR):
+			p->flags |= TERMP_SKIPCHAR;
+			break;
 		default:
 			break;
 		}
@@ -521,29 +539,40 @@ encode1(struct termp *p, int c)
 {
 	enum termfont	  f;
 
-	if (p->col + 4 >= p->maxcols)
-		adjbuf(p, p->col + 4);
+	if (TERMP_SKIPCHAR & p->flags) {
+		p->flags &= ~TERMP_SKIPCHAR;
+		return;
+	}
 
+	if (p->col + 6 >= p->maxcols)
+		adjbuf(p, p->col + 6);
+
 	f = term_fonttop(p);
 
-	if (TERMFONT_NONE == f) {
-		p->buf[p->col++] = c;
-		return;
-	} else if (TERMFONT_UNDER == f) {
+	if (TERMFONT_UNDER == f || TERMFONT_BI == f) {
 		p->buf[p->col++] = '_';
-	} else
-		p->buf[p->col++] = c;
-
-	p->buf[p->col++] = 8;
+		p->buf[p->col++] = 8;
+	}
+	if (TERMFONT_BOLD == f || TERMFONT_BI == f) {
+		if (ASCII_HYPH == c)
+			p->buf[p->col++] = '-';
+		else
+			p->buf[p->col++] = c;
+		p->buf[p->col++] = 8;
+	}
 	p->buf[p->col++] = c;
 }
 
 static void
 encode(struct termp *p, const char *word, size_t sz)
 {
-	enum termfont	  f;
 	int		  i, len;
 
+	if (TERMP_SKIPCHAR & p->flags) {
+		p->flags &= ~TERMP_SKIPCHAR;
+		return;
+	}
+
 	/* LINTED */
 	len = sz;
 
@@ -553,7 +582,7 @@ encode(struct termp *p, const char *word, size_t sz)
 	 * character by character.
 	 */
 
-	if (TERMFONT_NONE == (f = term_fonttop(p))) {
+	if (TERMFONT_NONE == term_fonttop(p)) {
 		if (p->col + len >= p->maxcols) 
 			adjbuf(p, p->col + len);
 		for (i = 0; i < len; i++)
@@ -563,25 +592,15 @@ encode(struct termp *p, const char *word, size_t sz)
 
 	/* Pre-buffer, assuming worst-case. */
 
-	if (p->col + 1 + (len * 3) >= p->maxcols)
-		adjbuf(p, p->col + 1 + (len * 3));
+	if (p->col + 1 + (len * 5) >= p->maxcols)
+		adjbuf(p, p->col + 1 + (len * 5));
 
 	for (i = 0; i < len; i++) {
-		if (ASCII_HYPH != word[i] &&
-		    ! isgraph((unsigned char)word[i])) {
-			p->buf[p->col++] = word[i];
-			continue;
-		}
-
-		if (TERMFONT_UNDER == f)
-			p->buf[p->col++] = '_';
-		else if (ASCII_HYPH == word[i])
-			p->buf[p->col++] = '-';
+		if (ASCII_HYPH == word[i] ||
+		    isgraph((unsigned char)word[i]))
+			encode1(p, word[i]);
 		else
 			p->buf[p->col++] = word[i];
-
-		p->buf[p->col++] = 8;
-		p->buf[p->col++] = word[i];
 	}
 }
 
@@ -592,12 +611,22 @@ term_len(const struct termp *p, size_t sz)
 	return((*p->width)(p, ' ') * sz);
 }
 
+static size_t
+cond_width(const struct termp *p, int c, int *skip)
+{
 
+	if (*skip) {
+		(*skip) = 0;
+		return(0);
+	} else
+		return((*p->width)(p, c));
+}
+
 size_t
 term_strlen(const struct termp *p, const char *cp)
 {
 	size_t		 sz, rsz, i;
-	int		 ssz, c;
+	int		 ssz, skip, c;
 	const char	*seq, *rhs;
 	enum mandoc_esc	 esc;
 	static const char rej[] = { '\\', ASCII_HYPH, ASCII_NBRSP, '\0' };
@@ -609,10 +638,11 @@ term_strlen(const struct termp *p, const char *cp)
 	 */
 
 	sz = 0;
+	skip = 0;
 	while ('\0' != *cp) {
 		rsz = strcspn(cp, rej);
 		for (i = 0; i < rsz; i++)
-			sz += (*p->width)(p, *cp++);
+			sz += cond_width(p, *cp++, &skip);
 
 		c = 0;
 		switch (*cp) {
@@ -629,14 +659,14 @@ term_strlen(const struct termp *p, const char *cp)
 						(seq + 1, ssz - 1);
 					if ('\0' == c)
 						break;
-					sz += (*p->width)(p, c);
+					sz += cond_width(p, c, &skip);
 					continue;
 				case (ESCAPE_SPECIAL):
 					c = mchars_spec2cp
 						(p->symtab, seq, ssz);
 					if (c <= 0)
 						break;
-					sz += (*p->width)(p, c);
+					sz += cond_width(p, c, &skip);
 					continue;
 				default:
 					break;
@@ -646,12 +676,12 @@ term_strlen(const struct termp *p, const char *cp)
 
 			switch (esc) {
 			case (ESCAPE_UNICODE):
-				sz += (*p->width)(p, '?');
+				sz += cond_width(p, '?', &skip);
 				break;
 			case (ESCAPE_NUMBERED):
 				c = mchars_num2char(seq, ssz);
 				if ('\0' != c)
-					sz += (*p->width)(p, c);
+					sz += cond_width(p, c, &skip);
 				break;
 			case (ESCAPE_SPECIAL):
 				rhs = mchars_spec2str
@@ -663,6 +693,9 @@ term_strlen(const struct termp *p, const char *cp)
 				rhs = seq;
 				rsz = ssz;
 				break;
+			case (ESCAPE_SKIPCHAR):
+				skip = 1;
+				break;
 			default:
 				break;
 			}
@@ -670,15 +703,20 @@ term_strlen(const struct termp *p, const char *cp)
 			if (NULL == rhs)
 				break;
 
+			if (skip) {
+				skip = 0;
+				break;
+			}
+
 			for (i = 0; i < rsz; i++)
 				sz += (*p->width)(p, *rhs++);
 			break;
 		case (ASCII_NBRSP):
-			sz += (*p->width)(p, ' ');
+			sz += cond_width(p, ' ', &skip);
 			cp++;
 			break;
 		case (ASCII_HYPH):
-			sz += (*p->width)(p, '-');
+			sz += cond_width(p, '-', &skip);
 			cp++;
 			break;
 		default: