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

9 files changed, 117 insertions, 16 deletions

diff --git a/html.c b/html.c
index 4bb3ca56..d92ec16a 100644
--- a/html.c
+++ b/html.c
@@ -1,4 +1,4 @@
-/*	$Id: html.c,v 1.213 2017/06/08 12:54:58 schwarze Exp $ */
+/*	$Id: html.c,v 1.214 2017/06/14 01:31:26 schwarze Exp $ */
 /*
  * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2011-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
@@ -345,16 +345,18 @@ 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) {
@@ -365,14 +367,28 @@ 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;
 
@@ -417,6 +433,9 @@ 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/mandoc.c b/mandoc.c
index 003bc066..5b541fa4 100644
--- a/mandoc.c
+++ b/mandoc.c
@@ -1,4 +1,4 @@
-/*	$Id: mandoc.c,v 1.101 2017/06/11 19:37:01 schwarze Exp $ */
+/*	$Id: mandoc.c,v 1.102 2017/06/14 01:31:26 schwarze Exp $ */
 /*
  * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2011-2015, 2017 Ingo Schwarze <schwarze@openbsd.org>
@@ -96,6 +96,8 @@ 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/mandoc.h b/mandoc.h
index 7499a4d3..3acebc36 100644
--- a/mandoc.h
+++ b/mandoc.h
@@ -1,4 +1,4 @@
-/*	$Id: mandoc.h,v 1.229 2017/06/11 19:37:01 schwarze Exp $ */
+/*	$Id: mandoc.h,v 1.230 2017/06/14 01:31:26 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/mdoc_markdown.c b/mdoc_markdown.c
index 7f5368fb..0b0f1848 100644
--- a/mdoc_markdown.c
+++ b/mdoc_markdown.c
@@ -1,4 +1,4 @@
-/*	$Id: mdoc_markdown.c,v 1.22 2017/05/30 16:31:29 schwarze Exp $ */
+/*	$Id: mdoc_markdown.c,v 1.23 2017/06/14 01:31:26 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/regress/roff/esc/Makefile b/regress/roff/esc/Makefile
index f0860368..4c8e492b 100644
--- a/regress/roff/esc/Makefile
+++ b/regress/roff/esc/Makefile
@@ -1,6 +1,6 @@
 # $OpenBSD: Makefile,v 1.11 2015/04/29 18:32:57 schwarze Exp $
 
-REGRESS_TARGETS = one two multi B c c_man e f h o w z ignore
+REGRESS_TARGETS = one two multi B c c_man e f h o p w z ignore
 LINT_TARGETS	= B h w ignore
 
 .include <bsd.regress.mk>
diff --git a/regress/roff/esc/p.in b/regress/roff/esc/p.in
new file mode 100644
index 00000000..b94ab13a
--- /dev/null
+++ b/regress/roff/esc/p.in
@@ -0,0 +1,15 @@
+.Dd June 14, 2017
+.Dt ESC-P 1
+.Os OpenBSD
+.Sh NAME
+.Nm esc-p
+.Nd line break escape sequence
+.Sh DESCRIPTION
+no blank: line one\pline two
+.Pp
+blank after esc: line one\p line two
+.Pp
+blank before esc: line one \pline two
+.Pp
+at eol: line one\p
+line two
diff --git a/regress/roff/esc/p.out_ascii b/regress/roff/esc/p.out_ascii
new file mode 100644
index 00000000..488e2c61
--- /dev/null
+++ b/regress/roff/esc/p.out_ascii
@@ -0,0 +1,19 @@
+ESC-P(1)                    General Commands Manual                   ESC-P(1)
+
+NNAAMMEE
+     eesscc--pp - line break escape sequence
+
+DDEESSCCRRIIPPTTIIOONN
+     no blank: line oneline
+     two
+
+     blank after esc: line one
+     line two
+
+     blank before esc: line one line
+     two
+
+     at eol: line one
+     line two
+
+OpenBSD                          June 14, 2017                         OpenBSD
diff --git a/roff.7 b/roff.7
index e329f161..c2200a2a 100644
--- a/roff.7
+++ b/roff.7
@@ -1,4 +1,4 @@
-.\"	$Id: roff.7,v 1.87 2017/06/10 16:32:27 schwarze Exp $
+.\"	$Id: roff.7,v 1.88 2017/06/14 01:31:26 schwarze Exp $
 .\"
 .\" Copyright (c) 2010, 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv>
 .\" Copyright (c) 2010,2011,2013-2015,2017 Ingo Schwarze <schwarze@openbsd.org>
@@ -15,7 +15,7 @@
 .\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 .\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 .\"
-.Dd $Mdocdate: June 10 2017 $
+.Dd $Mdocdate: June 14 2017 $
 .Dt ROFF 7
 .Os
 .Sh NAME
@@ -1991,6 +1991,8 @@ Overstrike, writing all the characters contained in the
 to the same output position.
 In terminal and HTML output modes,
 only the last one of the characters is visible.
+.Ss \ep
+Break the output line at the end of the current word.
 .Ss \eR\(aq Ns Ar name Oo +|- Oc Ns Ar number Ns \(aq
 Set number register; ignored by
 .Xr mandoc 1 .
diff --git a/term.c b/term.c
index b97b9be3..4a227988 100644
--- a/term.c
+++ b/term.c
@@ -1,4 +1,4 @@
-/*	$Id: term.c,v 1.269 2017/06/12 19:05:47 schwarze Exp $ */
+/*	$Id: term.c,v 1.270 2017/06/14 01:31:27 schwarze Exp $ */
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  * Copyright (c) 2010-2017 Ingo Schwarze <schwarze@openbsd.org>
@@ -116,6 +116,7 @@ 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;
@@ -155,7 +156,13 @@ 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;
 
@@ -221,6 +228,8 @@ 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] == ' ') {
@@ -260,6 +269,26 @@ 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;
 	}
 
 	/*
@@ -487,6 +516,9 @@ 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;