Implement the roff \z escape sequence, intended to output the next

character without advancing the cursor position; implement it to
simply skip the next character, as it will usually be overwritten.

With this change, the pod2man(1) preamble user-defined string \*:,
intended to render as a diaeresis or umlaut diacritic above the
preceding character, is rendered in a slightly less ugly way,
though still not correctly.  It was rendered as "z.." and is now
rendered as ".".

Given that the definition of \*: uses elaborate manual \h positioning,
there is little chance for mandoc(1) to ever render it correctly,
but at least we can refrain from printing out a spurious "z", and
we can make the \z do something semi-reasonable for easier cases.

"just commit" kristaps@

6 files changed, 125 insertions, 39 deletions

diff --git a/html.c b/html.c
index fe8a3ba8..abb868b5 100644
--- a/html.c
+++ b/html.c
@@ -1,7 +1,7 @@
 /*	$Id$ */
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
- * Copyright (c) 2011 Ingo Schwarze <schwarze@openbsd.org>
+ * Copyright (c) 2011, 2012 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
@@ -262,8 +262,8 @@ print_metaf(struct html *h, enum mandoc_esc deco)
 int
 html_strlen(const char *cp)
 {
-	int		 ssz, sz;
-	const char	*seq, *p;
+	size_t		 rsz;
+	int		 skip, sz;
 
 	/*
 	 * Account for escaped sequences within string length
@@ -274,10 +274,21 @@ html_strlen(const char *cp)
 	 */
 
 	sz = 0;
-	while (NULL != (p = strchr(cp, '\\'))) {
-		sz += (int)(p - cp);
-		++cp;
-		switch (mandoc_escape(&cp, &seq, &ssz)) {
+	skip = 0;
+	while (1) {
+		rsz = strcspn(cp, "\\");
+		if (rsz) {
+			cp += rsz;
+			if (skip) {
+				skip = 0;
+				rsz--;
+			}
+			sz += rsz;
+		}
+		if ('\0' == *cp)
+			break;
+		cp++;
+		switch (mandoc_escape(&cp, NULL, NULL)) {
 		case (ESCAPE_ERROR):
 			return(sz);
 		case (ESCAPE_UNICODE):
@@ -285,15 +296,19 @@ html_strlen(const char *cp)
 		case (ESCAPE_NUMBERED):
 			/* FALLTHROUGH */
 		case (ESCAPE_SPECIAL):
-			sz++;
+			if (skip)
+				skip = 0;
+			else
+				sz++;
+			break;
+		case (ESCAPE_SKIPCHAR):
+			skip = 1;
 			break;
 		default:
 			break;
 		}
 	}
-
-	assert(sz >= 0);
-	return(sz + strlen(cp));
+	return(sz);
 }
 
 static int
@@ -308,6 +323,12 @@ print_encode(struct html *h, const char *p, int norecurse)
 	nospace = 0;
 
 	while ('\0' != *p) {
+		if (HTML_SKIPCHAR & h->flags && '\\' != *p) {
+			h->flags &= ~HTML_SKIPCHAR;
+			p++;
+			continue;
+		}
+
 		sz = strcspn(p, rejs);
 
 		fwrite(p, 1, sz, stdout);
@@ -338,6 +359,31 @@ print_encode(struct html *h, const char *p, int norecurse)
 			break;
 
 		switch (esc) {
+		case (ESCAPE_FONT):
+			/* FALLTHROUGH */
+		case (ESCAPE_FONTPREV):
+			/* FALLTHROUGH */
+		case (ESCAPE_FONTBOLD):
+			/* FALLTHROUGH */
+		case (ESCAPE_FONTITALIC):
+			/* FALLTHROUGH */
+		case (ESCAPE_FONTROMAN):
+			if (0 == norecurse)
+				print_metaf(h, esc);
+			continue;
+		case (ESCAPE_SKIPCHAR):
+			h->flags |= HTML_SKIPCHAR;
+			continue;
+		default:
+			break;
+		}
+
+		if (h->flags & HTML_SKIPCHAR) {
+			h->flags &= ~HTML_SKIPCHAR;
+			continue;
+		}
+
+		switch (esc) {
 		case (ESCAPE_UNICODE):
 			/* Skip passed "u" header. */
 			c = mchars_num2uc(seq + 1, len - 1);
@@ -356,19 +402,6 @@ print_encode(struct html *h, const char *p, int norecurse)
 			else if (-1 == c && 1 == len)
 				putchar((int)*seq);
 			break;
-		case (ESCAPE_FONT):
-			/* FALLTHROUGH */
-		case (ESCAPE_FONTPREV):
-			/* FALLTHROUGH */
-		case (ESCAPE_FONTBOLD):
-			/* FALLTHROUGH */
-		case (ESCAPE_FONTITALIC):
-			/* FALLTHROUGH */
-		case (ESCAPE_FONTROMAN):
-			if (norecurse)
-				break;
-			print_metaf(h, esc);
-			break;
 		case (ESCAPE_NOSPACE):
 			if ('\0' == *p)
 				nospace = 1;
diff --git a/html.h b/html.h
index 7c95e2dc..2615060e 100644
--- a/html.h
+++ b/html.h
@@ -117,6 +117,7 @@ struct	html {
 #define	HTML_PREKEEP	 (1 << 3)
 #define	HTML_NONOSPACE	 (1 << 4) /* never add spaces */
 #define	HTML_LITERAL	 (1 << 5) /* literal (e.g., <PRE>) context */
+#define	HTML_SKIPCHAR	 (1 << 6) /* skip the next character */
 	struct tagq	  tags; /* stack of open tags */
 	struct rofftbl	  tbl; /* current table */
 	struct tag	 *tblt; /* current open table scope */
diff --git a/mandoc.c b/mandoc.c
index a7d8a92d..e363765b 100644
--- a/mandoc.c
+++ b/mandoc.c
@@ -1,7 +1,7 @@
 /*	$Id$ */
 /*
  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
- * Copyright (c) 2011 Ingo Schwarze <schwarze@openbsd.org>
+ * Copyright (c) 2011, 2012 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
@@ -143,6 +143,16 @@ mandoc_escape(const char **end, const char **start, int *sz)
 		break;
 
 	/*
+	 * The \z escape is supposed to output the following
+	 * character without advancing the cursor position.  
+	 * Since we are mostly dealing with terminal mode,
+	 * let us just skip the next character.
+	 */
+	case ('z'):
+		(*end)++;
+		return(ESCAPE_SKIPCHAR);
+
+	/*
 	 * Handle all triggers matching \X(xy, \Xx, and \X[xxxx], where
 	 * 'X' is the trigger.  These have opaque sub-strings.
 	 */
diff --git a/mandoc.h b/mandoc.h
index bd7dfdc1..5aaac54c 100644
--- a/mandoc.h
+++ b/mandoc.h
@@ -379,7 +379,8 @@ enum	mandoc_esc {
 	ESCAPE_FONTPREV, /* previous font mode */
 	ESCAPE_NUMBERED, /* a numbered glyph */
 	ESCAPE_UNICODE, /* a unicode codepoint */
-	ESCAPE_NOSPACE /* suppress space if the last on a line */
+	ESCAPE_NOSPACE, /* suppress space if the last on a line */
+	ESCAPE_SKIPCHAR /* skip the next character */
 };
 
 typedef	void	(*mandocmsg)(enum mandocerr, enum mandoclevel,
diff --git a/term.c b/term.c
index 889f72af..c9166b54 100644
--- a/term.c
+++ b/term.c
@@ -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);
@@ -419,12 +420,17 @@ 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);
@@ -480,9 +486,14 @@ 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;
 		}
@@ -522,6 +533,11 @@ encode1(struct termp *p, int c)
 {
 	enum termfont	  f;
 
+	if (TERMP_SKIPCHAR & p->flags) {
+		p->flags &= ~TERMP_SKIPCHAR;
+		return;
+	}
+
 	if (p->col + 4 >= p->maxcols)
 		adjbuf(p, p->col + 4);
 
@@ -545,6 +561,11 @@ 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;
 
@@ -593,12 +614,22 @@ 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' };
@@ -610,10 +641,11 @@ 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) {
@@ -630,14 +662,14 @@ 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;
@@ -647,12 +679,12 @@ 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
@@ -664,6 +696,9 @@ term_strlen(const struct termp *p, const char *cp)
 				rhs = seq;
 				rsz = ssz;
 				break;
+			case (ESCAPE_SKIPCHAR):
+				skip = 1;
+				break;
 			default:
 				break;
 			}
@@ -671,15 +706,20 @@ 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:
diff --git a/term.h b/term.h
index ce1b4ed5..ccac729e 100644
--- a/term.h
+++ b/term.h
@@ -78,6 +78,7 @@ struct	termp {
 #define	TERMP_ANPREC	 (1 << 13)	/* See termp_an_pre(). */
 #define	TERMP_KEEP	 (1 << 14)	/* Keep words together. */
 #define	TERMP_PREKEEP	 (1 << 15)	/* ...starting with the next one. */
+#define	TERMP_SKIPCHAR	 (1 << 16)	/* Skip the next character. */
 	int		 *buf;		/* Output buffer. */
 	enum termenc	  enc;		/* Type of encoding. */
 	struct mchars	 *symtab;	/* Encoded-symbol table. */