===================================================================
RCS file: /cvs/mandoc/term.c,v
retrieving revision 1.143
retrieving revision 1.164
diff -u -p -r1.143 -r1.164
--- mandoc/term.c	2010/06/07 21:03:02	1.143
+++ mandoc/term.c	2010/07/25 22:56:47	1.164
@@ -1,6 +1,7 @@
-/*	$Id: term.c,v 1.143 2010/06/07 21:03:02 kristaps Exp $ */
+/*	$Id: term.c,v 1.164 2010/07/25 22:56:47 kristaps Exp $ */
 /*
- * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se>
+ * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
+ * Copyright (c) 2010 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
@@ -26,63 +27,23 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
-#include <time.h>
 
 #include "mandoc.h"
 #include "chars.h"
 #include "out.h"
 #include "term.h"
-#include "man.h"
-#include "mdoc.h"
 #include "main.h"
 
-#define	PS_CHAR_WIDTH	  6
-#define	PS_CHAR_HEIGHT	  12
-#define	PS_CHAR_TOPMARG	 (792 - 24)
-#define	PS_CHAR_TOP	 (PS_CHAR_TOPMARG - 36)
-#define	PS_CHAR_LEFT	  36
-#define	PS_CHAR_BOTMARG	  24
-#define	PS_CHAR_BOT	 (PS_CHAR_BOTMARG + 36)
-
-static	struct termp	 *alloc(char *, enum termenc, enum termtype);
-static	void		  term_free(struct termp *);
-static	void		  spec(struct termp *, const char *, size_t);
+static	void		  spec(struct termp *, enum roffdeco,
+				const char *, size_t);
 static	void		  res(struct termp *, const char *, size_t);
 static	void		  buffera(struct termp *, const char *, size_t);
 static	void		  bufferc(struct termp *, char);
 static	void		  adjbuf(struct termp *p, size_t);
 static	void		  encode(struct termp *, const char *, size_t);
-static	void		  advance(struct termp *, size_t);
-static	void		  endline(struct termp *);
-static	void		  letter(struct termp *, char);
-static	void		  pageopen(struct termp *);
 
 
-void *
-ascii_alloc(char *outopts)
-{
-
-	return(alloc(outopts, TERMENC_ASCII, TERMTYPE_CHAR));
-}
-
-
-void *
-ps_alloc(void)
-{
-
-	return(alloc(NULL, TERMENC_ASCII, TERMTYPE_PS));
-}
-
-
 void
-terminal_free(void *arg)
-{
-
-	term_free((struct termp *)arg);
-}
-
-
-static void
 term_free(struct termp *p)
 {
 
@@ -90,65 +51,11 @@ term_free(struct termp *p)
 		free(p->buf);
 	if (p->symtab)
 		chars_free(p->symtab);
+
 	free(p);
 }
 
 
-/*
- * Push a single letter into our output engine.
- */
-static void
-letter(struct termp *p, char c)
-{
-	
-	if (TERMTYPE_CHAR == p->type) {
-		/*
-		 * If using the terminal device, just push the letter
-		 * out into the screen.
-		 */
-		putchar(c);
-		return;
-	}
-
-	if ( ! (PS_INLINE & p->psstate)) {
-		/*
-		 * If we're not in a PostScript "word" context, then
-		 * open one now at the current cursor.
-		 */
-		printf("%zu %zu moveto\n", p->pscol, p->psrow);
-		putchar('(');
-		p->psstate |= PS_INLINE;
-	}
-
-	/*
-	 * We need to escape these characters as per the PostScript
-	 * specification.  We would also escape non-graphable characters
-	 * (like tabs), but none of them would get to this point and
-	 * it's superfluous to abort() on them.
-	 */
-
-	switch (c) {
-	case ('('):
-		/* FALLTHROUGH */
-	case (')'):
-		/* FALLTHROUGH */
-	case ('\\'):
-		putchar('\\');
-		break;
-	default:
-		break;
-	}
-
-	/* Write the character and adjust where we are on the page. */
-	putchar(c);
-	p->pscol += PS_CHAR_WIDTH;
-}
-
-
-/*
- * Begin a "terminal" context.  Since terminal encompasses PostScript,
- * the actual terminal, etc., there are a few things we can do here.
- */
 void
 term_begin(struct termp *p, term_margin head, 
 		term_margin foot, const void *arg)
@@ -157,174 +64,30 @@ term_begin(struct termp *p, term_margin head, 
 	p->headf = head;
 	p->footf = foot;
 	p->argf = arg;
-
-	if (TERMTYPE_CHAR == p->type) {
-		/* Emit the header and be done. */
-		(*p->headf)(p, p->argf);
-		return;
-	}
-	
-	/*
-	 * Emit the standard PostScript prologue, set our initial page
-	 * position, then run pageopen() on the initial page.
-	 */
-
-	printf("%s\n", "%!PS");
-	printf("%s\n", "/Courier");
-	printf("%s\n", "10 selectfont");
-
-	p->pspage = 1;
-	p->psstate = 0;
-	pageopen(p);
+	(*p->begin)(p);
 }
 
 
-/*
- * Open a page.  This is only used for -Tps at the moment.  It opens a
- * page context, printing the header and the footer.  THE OUTPUT BUFFER
- * MUST BE EMPTY.  If it is not, output will ghost on the next line and
- * we'll be all gross and out of state.
- */
-static void
-pageopen(struct termp *p)
-{
-	
-	assert(TERMTYPE_PS == p->type);
-	assert(0 == p->psstate);
-
-	p->pscol = PS_CHAR_LEFT;
-	p->psrow = PS_CHAR_TOPMARG;
-	p->psstate |= PS_MARGINS;
-
-	(*p->headf)(p, p->argf);
-	endline(p);
-
-	p->psstate &= ~PS_MARGINS;
-	assert(0 == p->psstate);
-
-	p->pscol = PS_CHAR_LEFT;
-	p->psrow = PS_CHAR_BOTMARG;
-	p->psstate |= PS_MARGINS;
-
-	(*p->footf)(p, p->argf);
-	endline(p);
-
-	p->psstate &= ~PS_MARGINS;
-	assert(0 == p->psstate);
-
-	p->pscol = PS_CHAR_LEFT;
-	p->psrow = PS_CHAR_TOP;
-
-}
-
-
 void
 term_end(struct termp *p)
 {
 
-	if (TERMTYPE_CHAR == p->type) {
-		(*p->footf)(p, p->argf);
-		return;
-	}
-
-	printf("%s\n", "%%END");
+	(*p->end)(p);
 }
 
 
-static void
-endline(struct termp *p)
+struct termp *
+term_alloc(enum termenc enc)
 {
-
-	if (TERMTYPE_CHAR == p->type) {
-		putchar('\n');
-		return;
-	}
-
-	if (PS_INLINE & p->psstate) {
-		printf(") show\n");
-		p->psstate &= ~PS_INLINE;
-	} 
-
-	if (PS_MARGINS & p->psstate)
-		return;
-
-	p->pscol = PS_CHAR_LEFT;
-	if (p->psrow >= PS_CHAR_HEIGHT + PS_CHAR_BOT) {
-		p->psrow -= PS_CHAR_HEIGHT;
-		return;
-	}
-
-	/* 
-	 * XXX: can't run pageopen() until we're certain a flushln() has
-	 * occured, else the buf will reopen in an awkward state on the
-	 * next line.
-	 */
-	printf("showpage\n");
-	p->psrow = PS_CHAR_TOP;
-}
-
-
-/*
- * Advance the output engine by a certain amount of whitespace.
- */
-static void
-advance(struct termp *p, size_t len)
-{
-	size_t	 	i;
-
-	if (TERMTYPE_CHAR == p->type) {
-		/* Just print whitespace on the terminal. */
-		for (i = 0; i < len; i++)
-			putchar(' ');
-		return;
-	}
-
-	if (PS_INLINE & p->psstate) {
-		/* Dump out any existing line scope. */
-		printf(") show\n");
-		p->psstate &= ~PS_INLINE;
-	}
-
-	p->pscol += len ? len * PS_CHAR_WIDTH : 0;
-}
-
-
-static struct termp *
-alloc(char *outopts, enum termenc enc, enum termtype type)
-{
 	struct termp	*p;
-	const char	*toks[2];
-	char		*v;
-	size_t		 width;
 
-	toks[0] = "width";
-	toks[1] = NULL;
-
 	p = calloc(1, sizeof(struct termp));
 	if (NULL == p) {
 		perror(NULL);
 		exit(EXIT_FAILURE);
 	}
 
-	p->type = type;
-	p->tabwidth = 5;
 	p->enc = enc;
-
-	width = 80;
-
-	while (outopts && *outopts)
-		switch (getsubopt(&outopts, UNCONST(toks), &v)) {
-		case (0):
-			width = (size_t)atoi(v);
-			break;
-		default:
-			break;
-		}
-
-	/* Enforce some lower boundary. */
-	if (width < 60)
-		width = 60;
-	p->defrmargin = width - 2;
 	return(p);
 }
 
@@ -372,9 +135,10 @@ term_flushln(struct termp *p)
 	size_t		 vbl;   /* number of blanks to prepend to output */
 	size_t		 vend;	/* end of word visual position on output */
 	size_t		 bp;    /* visual right border position */
-	int		 j;     /* temporary loop index */
-	int		 jhy;	/* last hyphen before line overflow */
-	size_t		 maxvis, mmax;
+	int		 j;     /* temporary loop index for p->buf */
+	int		 jhy;	/* last hyph before overflow w/r/t j */
+	size_t		 maxvis; /* output position of visible boundary */
+	size_t		 mmax; /* used in calculating bp */
 
 	/*
 	 * First, establish the maximum columns of "visible" content.
@@ -399,21 +163,17 @@ term_flushln(struct termp *p)
 	 */
 	vbl = p->flags & TERMP_NOLPAD ? 0 : p->offset;
 
-	/* 
-	 * FIXME: if bp is zero, we still output the first word before
-	 * breaking the line.
-	 */
-
 	vis = vend = i = 0;
-	while (i < (int)p->col) {
 
+	while (i < (int)p->col) {
 		/*
-		 * Handle literal tab characters.
+		 * Handle literal tab characters: collapse all
+		 * subsequent tabs into a single huge set of spaces.
 		 */
 		for (j = i; j < (int)p->col; j++) {
 			if ('\t' != p->buf[j])
 				break;
-			vend = (vis/p->tabwidth+1)*p->tabwidth;
+			vend = (vis / p->tabwidth + 1) * p->tabwidth;
 			vbl += vend - vis;
 			vis = vend;
 		}
@@ -429,13 +189,21 @@ term_flushln(struct termp *p)
 		for (jhy = 0; j < (int)p->col; j++) {
 			if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j])
 				break;
-			if (8 != p->buf[j]) {
-				if (vend > vis && vend < bp &&
-				    ASCII_HYPH == p->buf[j])
-					jhy = j;
-				vend++;
-			} else
-				vend--;
+
+			/* Back over the the last printed character. */
+			if (8 == p->buf[j]) {
+				assert(j);
+				vend -= (*p->width)(p, p->buf[j - 1]);
+				continue;
+			}
+
+			/* Regular word. */
+			/* Break at the hyphen point if we overrun. */
+			if (vend > vis && vend < bp && 
+					ASCII_HYPH == p->buf[j])
+				jhy = j;
+
+			vend += (*p->width)(p, p->buf[j]);
 		}
 
 		/*
@@ -444,10 +212,10 @@ term_flushln(struct termp *p)
 		 */
 		if (vend > bp && 0 == jhy && vis > 0) {
 			vend -= vis;
-			endline(p);
+			(*p->endline)(p);
 			if (TERMP_NOBREAK & p->flags) {
 				p->viscol = p->rmargin;
-				advance(p, p->rmargin);
+				(*p->advance)(p, p->rmargin);
 				vend += p->rmargin - p->offset;
 			} else {
 				p->viscol = 0;
@@ -474,14 +242,14 @@ term_flushln(struct termp *p)
 			if ('\t' == p->buf[i])
 				break;
 			if (' ' == p->buf[i]) {
-				while (' ' == p->buf[i]) {
-					vbl++;
+				j = i;
+				while (' ' == p->buf[i])
 					i++;
-				}
+				vbl += (i - j) * (*p->width)(p, ' ');
 				break;
 			}
 			if (ASCII_NBRSP == p->buf[i]) {
-				vbl++;
+				vbl += (*p->width)(p, ' ');
 				continue;
 			}
 
@@ -491,17 +259,18 @@ term_flushln(struct termp *p)
 			 * so write preceding white space now.
 			 */
 			if (vbl) {
-				advance(p, vbl);
+				(*p->advance)(p, vbl);
 				p->viscol += vbl;
 				vbl = 0;
 			}
 
-			if (ASCII_HYPH == p->buf[i])
-				letter(p, '-');
-			else
-				letter(p, p->buf[i]);
-
-			p->viscol += 1;
+			if (ASCII_HYPH == p->buf[i]) {
+				(*p->letter)(p, '-');
+				p->viscol += (*p->width)(p, '-');
+			} else {
+				(*p->letter)(p, p->buf[i]);
+				p->viscol += (*p->width)(p, p->buf[i]);
+			}
 		}
 		vend += vbl;
 		vis = vend;
@@ -512,14 +281,14 @@ term_flushln(struct termp *p)
 
 	if ( ! (TERMP_NOBREAK & p->flags)) {
 		p->viscol = 0;
-		endline(p);
+		(*p->endline)(p);
 		return;
 	}
 
 	if (TERMP_HANG & p->flags) {
 		/* We need one blank after the tag. */
 		p->overstep = /* LINTED */
-			vis - maxvis + 1;
+			vis - maxvis + (*p->width)(p, ' ');
 
 		/*
 		 * Behave exactly the same way as groff:
@@ -543,14 +312,15 @@ term_flushln(struct termp *p)
 
 	/* Right-pad. */
 	if (maxvis > vis + /* LINTED */
-			((TERMP_TWOSPACE & p->flags) ? 1 : 0)) {
+			((TERMP_TWOSPACE & p->flags) ? 
+			 (*p->width)(p, ' ') : 0)) {
 		p->viscol += maxvis - vis;
-		advance(p, maxvis - vis);
+		(*p->advance)(p, maxvis - vis);
 		vis += (maxvis - vis);
 	} else {	/* ...or newline break. */
-		endline(p);
+		(*p->endline)(p);
 		p->viscol = p->rmargin;
-		advance(p, p->rmargin);
+		(*p->advance)(p, p->rmargin);
 	}
 }
 
@@ -586,19 +356,21 @@ term_vspace(struct termp *p)
 
 	term_newln(p);
 	p->viscol = 0;
-	endline(p);
+	(*p->endline)(p);
 }
 
 
 static void
-spec(struct termp *p, const char *word, size_t len)
+spec(struct termp *p, enum roffdeco d, const char *word, size_t len)
 {
 	const char	*rhs;
 	size_t		 sz;
 
-	rhs = chars_a2ascii(p->symtab, word, len, &sz);
+	rhs = chars_spec2str(p->symtab, word, len, &sz);
 	if (rhs) 
 		encode(p, rhs, sz);
+	else if (DECO_SSPECIAL == d)
+		encode(p, word, len);
 }
 
 
@@ -608,7 +380,7 @@ res(struct termp *p, const char *word, size_t len)
 	const char	*rhs;
 	size_t		 sz;
 
-	rhs = chars_a2res(p->symtab, word, len, &sz);
+	rhs = chars_res2str(p->symtab, word, len, &sz);
 	if (rhs)
 		encode(p, rhs, sz);
 }
@@ -719,9 +491,14 @@ term_word(struct termp *p, const char *word)
 		}
 
 	if ( ! (TERMP_NOSPACE & p->flags)) {
-		bufferc(p, ' ');
-		if (TERMP_SENTENCE & 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 ( ! (p->flags & TERMP_NONOSPACE))
@@ -729,14 +506,13 @@ term_word(struct termp *p, const char *word)
 
 	p->flags &= ~TERMP_SENTENCE;
 
-	/* FIXME: use strcspn. */
-
 	while (*word) {
-		if ('\\' != *word) {
-			encode(p, word, 1);
-			word++;
+		if ((ssz = strcspn(word, "\\")) > 0)
+			encode(p, word, ssz);
+
+		word += ssz;
+		if ('\\' != *word)
 			continue;
-		}
 
 		seq = ++word;
 		sz = a2roffdeco(&deco, &seq, &ssz);
@@ -746,7 +522,9 @@ term_word(struct termp *p, const char *word)
 			res(p, seq, ssz);
 			break;
 		case (DECO_SPECIAL):
-			spec(p, seq, ssz);
+			/* FALLTHROUGH */
+		case (DECO_SSPECIAL):
+			spec(p, deco, seq, ssz);
 			break;
 		case (DECO_BOLD):
 			term_fontrepl(p, TERMFONT_BOLD);
@@ -773,7 +551,7 @@ term_word(struct termp *p, const char *word)
 	 * Note that we don't process the pipe: the parser sees it as
 	 * punctuation, but we don't in terms of typography.
 	 */
-	if (sv[0] && 0 == sv[1])
+	if (sv[0] && '\0' == sv[1])
 		switch (sv[0]) {
 		case('('):
 			/* FALLTHROUGH */
@@ -838,12 +616,9 @@ encode(struct termp *p, const char *word, size_t sz)
 	 * character by character.
 	 */
 
-	if (TERMTYPE_PS == p->type) {
+	if (TERMFONT_NONE == (f = term_fonttop(p))) {
 		buffera(p, word, sz);
 		return;
-	} else if (TERMFONT_NONE == (f = term_fonttop(p))) {
-		buffera(p, word, sz);
-		return;
 	}
 
 	for (i = 0; i < (int)sz; i++) {
@@ -864,8 +639,29 @@ encode(struct termp *p, const char *word, size_t sz)
 
 
 size_t
-term_vspan(const struct roffsu *su)
+term_len(const struct termp *p, size_t sz)
 {
+
+	return((*p->width)(p, ' ') * sz);
+}
+
+
+size_t
+term_strlen(const struct termp *p, const char *cp)
+{
+	size_t		 sz;
+
+	for (sz = 0; *cp; cp++)
+		sz += (*p->width)(p, *cp);
+
+	return(sz);
+}
+
+
+/* ARGSUSED */
+size_t
+term_vspan(const struct termp *p, const struct roffsu *su)
+{
 	double		 r;
 
 	switch (su->unit) {
@@ -900,41 +696,13 @@ term_vspan(const struct roffsu *su)
 
 
 size_t
-term_hspan(const struct roffsu *su)
+term_hspan(const struct termp *p, const struct roffsu *su)
 {
-	double		 r;
+	double		 v;
 
-	/* XXX: CM, IN, and PT are approximations. */
-
-	switch (su->unit) {
-	case (SCALE_CM):
-		r = 4 * su->scale;
-		break;
-	case (SCALE_IN):
-		/* XXX: this is an approximation. */
-		r = 10 * su->scale;
-		break;
-	case (SCALE_PC):
-		r = (10 * su->scale) / 6;
-		break;
-	case (SCALE_PT):
-		r = (10 * su->scale) / 72;
-		break;
-	case (SCALE_MM):
-		r = su->scale / 1000; /* FIXME: double-check. */
-		break;
-	case (SCALE_VS):
-		r = su->scale * 2 - 1; /* FIXME: double-check. */
-		break;
-	default:
-		r = su->scale;
-		break;
-	}
-
-	if (r < 0.0)
-		r = 0.0;
-	return((size_t)/* LINTED */
-			r);
+	v = ((*p->hspan)(p, su));
+	if (v < 0.0)
+		v = 0.0;
+	return((size_t) /* LINTED */
+			v);
 }
-
-