implement the roff(7) \p (break output line) escape sequence

5 files changed, 78 insertions, 13 deletions

diff --git a/usr.bin/mandoc/html.c b/usr.bin/mandoc/html.c
index f25c21da748..58576242ace 100644
--- a/usr.bin/mandoc/html.c
+++ b/usr.bin/mandoc/html.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: html.c,v 1.83 2017/06/08 12:54:40 schwarze Exp $ */
+/*	$OpenBSD: html.c,v 1.84 2017/06/14 01:31:19 schwarze Exp $ */
 /*
  * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2011-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
@@ -343,16 +343,18 @@ static int
 print_encode(struct html *h, const char *p, const char *pend, int norecurse)
 {
 	char		 numbuf[16];
-	size_t		 sz;
-	int		 c, len, nospace;
+	struct tag	*t;
 	const char	*seq;
+	size_t		 sz;
+	int		 c, len, breakline, nospace;
 	enum mandoc_esc	 esc;
-	static const char rejs[9] = { '\\', '<', '>', '&', '"',
+	static const char rejs[10] = { ' ', '\\', '<', '>', '&', '"',
 		ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' };
 
 	if (pend == NULL)
 		pend = strchr(p, '\0');
 
+	breakline = 0;
 	nospace = 0;
 
 	while (p < pend) {
@@ -363,14 +365,28 @@ print_encode(struct html *h, const char *p, const char *pend, int norecurse)
 		}
 
 		for (sz = strcspn(p, rejs); sz-- && p < pend; p++)
-			if (*p == ' ')
-				print_endword(h);
-			else
-				print_byte(h, *p);
+			print_byte(h, *p);
+
+		if (breakline &&
+		    (p >= pend || *p == ' ' || *p == ASCII_NBRSP)) {
+			t = print_otag(h, TAG_DIV, "");
+			print_text(h, "\\~");
+			print_tagq(h, t);
+			breakline = 0;
+			while (p < pend && (*p == ' ' || *p == ASCII_NBRSP))
+				p++;
+			continue;
+		}
 
 		if (p >= pend)
 			break;
 
+		if (*p == ' ') {
+			print_endword(h);
+			p++;
+			continue;
+		}
+
 		if (print_escape(h, *p++))
 			continue;
 
@@ -415,6 +431,9 @@ print_encode(struct html *h, const char *p, const char *pend, int norecurse)
 			if (c <= 0)
 				continue;
 			break;
+		case ESCAPE_BREAK:
+			breakline = 1;
+			continue;
 		case ESCAPE_NOSPACE:
 			if ('\0' == *p)
 				nospace = 1;
diff --git a/usr.bin/mandoc/mandoc.c b/usr.bin/mandoc/mandoc.c
index 8000e4ad664..3aff3c5b77f 100644
--- a/usr.bin/mandoc/mandoc.c
+++ b/usr.bin/mandoc/mandoc.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: mandoc.c,v 1.69 2017/06/11 19:36:31 schwarze Exp $ */
+/*	$OpenBSD: mandoc.c,v 1.70 2017/06/14 01:31:19 schwarze Exp $ */
 /*
  * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2011-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
@@ -94,6 +94,8 @@ mandoc_escape(const char **end, const char **start, int *sz)
 	case ',':
 	case '/':
 		return ESCAPE_IGNORE;
+	case 'p':
+		return ESCAPE_BREAK;
 
 	/*
 	 * The \z escape is supposed to output the following
diff --git a/usr.bin/mandoc/mandoc.h b/usr.bin/mandoc/mandoc.h
index 6f14e23ec31..ed4d1d7fc17 100644
--- a/usr.bin/mandoc/mandoc.h
+++ b/usr.bin/mandoc/mandoc.h
@@ -1,4 +1,4 @@
-/*	$OpenBSD: mandoc.h,v 1.171 2017/06/11 19:36:31 schwarze Exp $ */
+/*	$OpenBSD: mandoc.h,v 1.172 2017/06/14 01:31:19 schwarze Exp $ */
 /*
  * Copyright (c) 2010, 2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010-2017 Ingo Schwarze <schwarze@openbsd.org>
@@ -423,6 +423,7 @@ enum	mandoc_esc {
 	ESCAPE_FONTPREV, /* previous font mode */
 	ESCAPE_NUMBERED, /* a numbered glyph */
 	ESCAPE_UNICODE, /* a unicode codepoint */
+	ESCAPE_BREAK, /* break the output line */
 	ESCAPE_NOSPACE, /* suppress space if the last on a line */
 	ESCAPE_HORIZ, /* horizontal movement */
 	ESCAPE_HLINE, /* horizontal line drawing */
diff --git a/usr.bin/mandoc/mdoc_markdown.c b/usr.bin/mandoc/mdoc_markdown.c
index 19d564a2dbd..1a8a70ada5d 100644
--- a/usr.bin/mandoc/mdoc_markdown.c
+++ b/usr.bin/mandoc/mdoc_markdown.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: mdoc_markdown.c,v 1.22 2017/05/30 16:31:25 schwarze Exp $ */
+/*	$OpenBSD: mdoc_markdown.c,v 1.23 2017/06/14 01:31:19 schwarze Exp $ */
 /*
  * Copyright (c) 2017 Ingo Schwarze <schwarze@openbsd.org>
  *
@@ -493,7 +493,7 @@ md_word(const char *s)
 {
 	const char	*seq, *prevfont, *currfont, *nextfont;
 	char		 c;
-	int		 bs, sz, uc;
+	int		 bs, sz, uc, breakline;
 
 	/* No spacing before closing delimiters. */
 	if (s[0] != '\0' && s[1] == '\0' &&
@@ -510,6 +510,7 @@ md_word(const char *s)
 	if ((s[0] == '(' || s[0] == '[') && s[1] == '\0')
 		outflags &= ~MD_spc;
 
+	breakline = 0;
 	prevfont = currfont = "";
 	while ((c = *s++) != '\0') {
 		bs = 0;
@@ -595,6 +596,9 @@ md_word(const char *s)
 			case ESCAPE_FONTPREV:
 				nextfont = prevfont;
 				break;
+			case ESCAPE_BREAK:
+				breakline = 1;
+				break;
 			case ESCAPE_NOSPACE:
 			case ESCAPE_SKIPCHAR:
 			case ESCAPE_OVERSTRIKE:
@@ -642,6 +646,13 @@ md_word(const char *s)
 		if (bs)
 			putchar('\\');
 		md_char(c);
+		if (breakline &&
+		    (*s == '\0' || *s == ' ' || *s == ASCII_NBRSP)) {
+			printf("  \n");
+			breakline = 0;
+			while (*s == ' ' || *s == ASCII_NBRSP)
+				s++;
+		}
 	}
 	if (*currfont != '\0') {
 		outflags &= ~MD_spc;
diff --git a/usr.bin/mandoc/term.c b/usr.bin/mandoc/term.c
index 07b7f57a2e5..e6aebdc379a 100644
--- a/usr.bin/mandoc/term.c
+++ b/usr.bin/mandoc/term.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: term.c,v 1.129 2017/06/12 18:55:42 schwarze Exp $ */
+/*	$OpenBSD: term.c,v 1.130 2017/06/14 01:31:19 schwarze Exp $ */
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010-2017 Ingo Schwarze <schwarze@openbsd.org>
@@ -114,6 +114,7 @@ term_flushln(struct termp *p)
 	size_t		 jhy;	/* last hyph before overflow w/r/t j */
 	size_t		 maxvis; /* output position of visible boundary */
 	int		 ntab;	/* number of tabs to prepend */
+	int		 breakline; /* after this word */
 
 	vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ?
 	    0 : p->tcol->offset - p->viscol;
@@ -153,7 +154,13 @@ term_flushln(struct termp *p)
 		 */
 
 		jhy = 0;
+		breakline = 0;
 		for (j = p->tcol->col; j < p->tcol->lastcol; j++) {
+			if (p->tcol->buf[j] == '\n') {
+				if ((p->flags & TERMP_BRIND) == 0)
+					breakline = 1;
+				continue;
+			}
 			if (p->tcol->buf[j] == ' ' || p->tcol->buf[j] == '\t')
 				break;
 
@@ -219,6 +226,8 @@ term_flushln(struct termp *p)
 		for ( ; p->tcol->col < p->tcol->lastcol; p->tcol->col++) {
 			if (vend > bp && jhy > 0 && p->tcol->col > jhy)
 				break;
+			if (p->tcol->buf[p->tcol->col] == '\n')
+				continue;
 			if (p->tcol->buf[p->tcol->col] == '\t')
 				break;
 			if (p->tcol->buf[p->tcol->col] == ' ') {
@@ -258,6 +267,26 @@ term_flushln(struct termp *p)
 				    p->tcol->buf[p->tcol->col]);
 		}
 		vis = vend;
+
+		if (breakline == 0)
+			continue;
+
+		/* Explicitly requested output line break. */
+
+		if (p->flags & TERMP_MULTICOL)
+			return;
+
+		endline(p);
+		breakline = 0;
+		vis = vend = 0;
+
+		/* Re-establish indentation. */
+
+		vbl = p->tcol->offset;
+		maxvis = p->tcol->rmargin > vbl ?
+		    p->tcol->rmargin - vbl : 0;
+		bp = !(p->flags & TERMP_NOBREAK) ? maxvis :
+		    p->maxrmargin > vbl ?  p->maxrmargin - vbl : 0;
 	}
 
 	/*
@@ -485,6 +514,9 @@ term_word(struct termp *p, const char *word)
 		case ESCAPE_FONTPREV:
 			term_fontlast(p);
 			continue;
+		case ESCAPE_BREAK:
+			bufferc(p, '\n');
+			continue;
 		case ESCAPE_NOSPACE:
 			if (p->flags & TERMP_BACKAFTER)
 				p->flags &= ~TERMP_BACKAFTER;