===================================================================
RCS file: /cvs/mandoc/out.c,v
retrieving revision 1.16
retrieving revision 1.33
diff -u -p -r1.16 -r1.33
--- mandoc/out.c	2010/06/19 20:46:28	1.16
+++ mandoc/out.c	2011/01/10 14:40:30	1.33
@@ -1,6 +1,6 @@
-/*	$Id: out.c,v 1.16 2010/06/19 20:46:28 kristaps Exp $ */
+/*	$Id: out.c,v 1.33 2011/01/10 14:40:30 kristaps Exp $ */
 /*
- * Copyright (c) 2009 Kristaps Dzonsons <kristaps@bsd.lv>
+ * Copyright (c) 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
  *
  * Permission to use, copy, modify, and distribute this software for any
  * purpose with or without fee is hereby granted, provided that the above
@@ -27,28 +27,16 @@
 #include <string.h>
 #include <time.h>
 
+#include "mandoc.h"
 #include "out.h"
 
-/* See a2roffdeco(). */
-#define	C2LIM(c, l) do { \
-	(l) = 1; \
-	if ('[' == (c) || '\'' == (c)) \
-		(l) = 0; \
-	else if ('(' == (c)) \
-		(l) = 2; } \
-	while (/* CONSTCOND */ 0)
+static	void	tblcalc_data(struct rofftbl *, struct roffcol *,
+			const struct tbl *, const struct tbl_dat *);
+static	void	tblcalc_literal(struct rofftbl *, struct roffcol *,
+			const struct tbl_dat *);
+static	void	tblcalc_number(struct rofftbl *, struct roffcol *,
+			const struct tbl *, const struct tbl_dat *);
 
-/* See a2roffdeco(). */
-#define	C2TERM(c, t) do { \
-	(t) = 0; \
-	if ('\'' == (c)) \
-		(t) = 1; \
-	else if ('[' == (c)) \
-		(t) = 2; \
-	else if ('(' == (c)) \
-		(t) = 3; } \
-	while (/* CONSTCOND */ 0)
-
 /* 
  * Convert a `scaling unit' to a consistent form, or fail.  Scaling
  * units are documented in groff.7, mdoc.7, man.7.
@@ -136,11 +124,10 @@ a2roffsu(const char *src, struct roffsu *dst, enum rof
 		return(0);
 	}
 
+	/* FIXME: do this in the caller. */
 	if ((dst->scale = atof(buf)) < 0)
 		dst->scale = 0;
 	dst->unit = unit;
-	dst->pt = hasd;
-
 	return(1);
 }
 
@@ -187,213 +174,390 @@ time2a(time_t t, char *dst, size_t sz)
 }
 
 
-/* 
- * Returns length of parsed string (the leading "\" should NOT be
- * included).  This can be zero if the current character is the nil
- * terminator.  "d" is set to the type of parsed decorator, which may
- * have an adjoining "word" of size "sz" (e.g., "(ab" -> "ab", 2).
- */
 int
-a2roffdeco(enum roffdeco *d,
-		const char **word, size_t *sz)
+a2roffdeco(enum roffdeco *d, const char **word, size_t *sz)
 {
-	int		 j, term, lim;
-	char		 set;
-	const char	*wp, *sp;
+	int		 i, j, lim;
+	char		 term, c;
+	const char	*wp;
+	enum roffdeco	 dd;
 
 	*d = DECO_NONE;
+	lim = i = 0;
+	term = '\0';
 	wp = *word;
 
-	switch ((set = *wp)) {
-	case ('\0'):
-		return(0);
-
+	switch ((c = wp[i++])) {
 	case ('('):
-		if ('\0' == *(++wp))
-			return(1);
-		if ('\0' == *(wp + 1))
-			return(2);
-
 		*d = DECO_SPECIAL;
-		*sz = 2;
-		*word = wp;
-		return(3);
-
+		lim = 2;
+		break;
 	case ('F'):
 		/* FALLTHROUGH */
 	case ('f'):
-		/*
-		 * FIXME: this needs work and consolidation (it should
-		 * follow the sequence that special characters do, for
-		 * one), but isn't a priority at the moment.  Note, for
-		 * one, that in reality \fB != \FB, although here we let
-		 * these slip by.
-		 */
-		switch (*(++wp)) {
-		case ('\0'):
-			return(1);
+		*d = 'F' == c ? DECO_FFONT : DECO_FONT;
+
+		switch (wp[i++]) {
+		case ('('):
+			lim = 2;
+			break;
+		case ('['):
+			term = ']';
+			break;
 		case ('3'):
 			/* FALLTHROUGH */
 		case ('B'):
 			*d = DECO_BOLD;
-			return(2);
+			return(i);
 		case ('2'):
 			/* FALLTHROUGH */
 		case ('I'):
 			*d = DECO_ITALIC;
-			return(2);
+			return(i);
 		case ('P'):
 			*d = DECO_PREVIOUS;
-			return(2);
+			return(i);
 		case ('1'):
 			/* FALLTHROUGH */
 		case ('R'):
 			*d = DECO_ROMAN;
-			return(2);
-		case ('('):
-			if ('\0' == *(++wp))
-				return(2);
-			if ('\0' == *(wp + 1))
-				return(3);
+			return(i);
+		default:
+			i--;
+			lim = 1;
+			break;
+		}
+		break;
+	case ('k'):
+		/* FALLTHROUGH */
+	case ('M'):
+		/* FALLTHROUGH */
+	case ('m'):
+		/* FALLTHROUGH */
+	case ('*'):
+		if ('*' == c)
+			*d = DECO_RESERVED;
 
-			*d = 'F' == set ? DECO_FFONT : DECO_FONT;
-			*sz = 2;
-			*word = wp;
-			return(4);
+		switch (wp[i++]) {
+		case ('('):
+			lim = 2;
+			break;
 		case ('['):
-			*word = ++wp;
-			for (j = 0; *wp && ']' != *wp; wp++, j++)
-				/* Loop... */ ;
+			term = ']';
+			break;
+		default:
+			i--;
+			lim = 1;
+			break;
+		}
+		break;
+	case ('h'):
+		/* FALLTHROUGH */
+	case ('v'):
+		/* FALLTHROUGH */
+	case ('s'):
+		j = 0;
+		if ('+' == wp[i] || '-' == wp[i]) {
+			i++;
+			j = 1;
+		}
 
-			if ('\0' == *wp)
-				return(j + 2);
-
-			*d = 'F' == set ? DECO_FFONT : DECO_FONT;
-			*sz = (size_t)j;
-			return(j + 3);
+		switch (wp[i++]) {
+		case ('('):
+			lim = 2;
+			break;
+		case ('['):
+			term = ']';
+			break;
+		case ('\''):
+			term = '\'';
+			break;
+		case ('0'):
+			j = 1;
+			/* FALLTHROUGH */
 		default:
+			i--;
+			lim = 1;
 			break;
 		}
 
-		*d = 'F' == set ? DECO_FFONT : DECO_FONT;
-		*sz = 1;
-		*word = wp;
-		return(2);
+		if ('+' == wp[i] || '-' == wp[i]) {
+			if (j)
+				return(i);
+			i++;
+		} 
 
-	case ('*'):
-		switch (*(++wp)) {
-		case ('\0'):
-			return(1);
+		/* Handle embedded numerical subexp or escape. */
 
-		case ('('):
-			if ('\0' == *(++wp))
-				return(2);
-			if ('\0' == *(wp + 1))
-				return(3);
+		if ('(' == wp[i]) {
+			while (wp[i] && ')' != wp[i])
+				if ('\\' == wp[i++]) {
+					/* Handle embedded escape. */
+					*word = &wp[i];
+					i += a2roffdeco(&dd, word, sz);
+				}
 
-			*d = DECO_RESERVED;
-			*sz = 2;
-			*word = wp;
-			return(4);
+			if (')' == wp[i++])
+				break;
 
-		case ('['):
-			*word = ++wp;
-			for (j = 0; *wp && ']' != *wp; wp++, j++)
-				/* Loop... */ ;
+			*d = DECO_NONE;
+			return(i - 1);
+		} else if ('\\' == wp[i]) {
+			*word = &wp[++i];
+			i += a2roffdeco(&dd, word, sz);
+		}
 
-			if ('\0' == *wp)
-				return(j + 2);
+		break;
+	case ('['):
+		*d = DECO_SPECIAL;
+		term = ']';
+		break;
+	case ('c'):
+		*d = DECO_NOSPACE;
+		return(i);
+	case ('z'):
+		*d = DECO_NONE;
+		if ('\\' == wp[i]) {
+			*word = &wp[++i];
+			return(i + a2roffdeco(&dd, word, sz));
+		} else
+			lim = 1;
+		break;
+	case ('o'):
+		/* FALLTHROUGH */
+	case ('w'):
+		if ('\'' == wp[i++]) {
+			term = '\'';
+			break;
+		} 
+		/* FALLTHROUGH */
+	default:
+		*d = DECO_SSPECIAL;
+		i--;
+		lim = 1;
+		break;
+	}
 
-			*d = DECO_RESERVED;
-			*sz = (size_t)j;
-			return(j + 3);
+	assert(term || lim);
+	*word = &wp[i];
 
-		default:
-			break;
+	if (term) {
+		j = i;
+		while (wp[i] && wp[i] != term)
+			i++;
+		if ('\0' == wp[i]) {
+			*d = DECO_NONE;
+			return(i);
 		}
 
-		*d = DECO_RESERVED;
-		*sz = 1;
-		*word = wp;
-		return(2);
+		assert(i >= j);
+		*sz = (size_t)(i - j);
 
-	case ('s'):
-		sp = wp;
-		if ('\0' == *(++wp))
-			return(1);
+		return(i + 1);
+	}
 
-		C2LIM(*wp, lim);
-		C2TERM(*wp, term);
+	assert(lim > 0);
+	*sz = (size_t)lim;
 
-		if (term) 
-			wp++;
+	for (j = 0; wp[i] && j < lim; j++)
+		i++;
+	if (j < lim)
+		*d = DECO_NONE;
 
-		*word = wp;
+	return(i);
+}
 
-		if (*wp == '+' || *wp == '-')
-			++wp;
+/*
+ * Calculate the abstract widths and decimal positions of columns in a
+ * table.  This routine allocates the columns structures then runs over
+ * all rows and cells in the table.  The function pointers in "tbl" are
+ * used for the actual width calculations.
+ */
+void
+tblcalc(struct rofftbl *tbl, const struct tbl_span *sp)
+{
+	const struct tbl_dat	*dp;
+	const struct tbl_head	*hp;
+	struct roffcol		*col;
 
-		switch (*wp) {
-		case ('\''):
-			/* FALLTHROUGH */
-		case ('['):
-			/* FALLTHROUGH */
-		case ('('):
-			if (term) 
-				return((int)(wp - sp));
+	/*
+	 * Allocate the master column specifiers.  These will hold the
+	 * widths and decimal positions for all cells in the column.  It
+	 * must be freed and nullified by the caller.
+	 */
 
-			C2LIM(*wp, lim);
-			C2TERM(*wp, term);
-			wp++;
-			break;
-		default:
-			break;
-		}
+	assert(NULL == tbl->cols);
+	tbl->cols = calloc(sp->tbl->cols, sizeof(struct roffcol));
 
-		if ( ! isdigit((u_char)*wp))
-			return((int)(wp - sp));
+	hp = sp->head;
 
-		for (j = 0; isdigit((u_char)*wp); j++) {
-			if (lim && j >= lim)
-				break;
-			++wp;
+	for ( ; sp; sp = sp->next) {
+		if (TBL_SPAN_DATA != sp->pos)
+			continue;
+		/*
+		 * Account for the data cells in the layout, matching it
+		 * to data cells in the data section.
+		 */
+		for (dp = sp->first; dp; dp = dp->next) {
+			assert(dp->layout);
+			col = &tbl->cols[dp->layout->head->ident];
+			tblcalc_data(tbl, col, sp->tbl, dp);
 		}
+	}
 
-		if (term && term < 3) {
-			if (1 == term && *wp != '\'')
-				return((int)(wp - sp));
-			if (2 == term && *wp != ']')
-				return((int)(wp - sp));
-			++wp;
+	/* 
+	 * Calculate width of the spanners.  These get one space for a
+	 * vertical line, two for a double-vertical line. 
+	 */
+
+	for ( ; hp; hp = hp->next) {
+		col = &tbl->cols[hp->ident];
+		switch (hp->pos) {
+		case (TBL_HEAD_VERT):
+			col->width = (*tbl->len)(1, tbl->arg);
+			break;
+		case (TBL_HEAD_DVERT):
+			col->width = (*tbl->len)(2, tbl->arg);
+			break;
+		default:
+			break;
 		}
+	}
+}
 
-		*d = DECO_SIZE;
-		return((int)(wp - sp));
+static void
+tblcalc_data(struct rofftbl *tbl, struct roffcol *col,
+		const struct tbl *tp, const struct tbl_dat *dp)
+{
+	size_t		 sz;
 
-	case ('['):
-		*word = ++wp;
+	/* Branch down into data sub-types. */
 
-		for (j = 0; *wp && ']' != *wp; wp++, j++)
-			/* Loop... */ ;
+	switch (dp->layout->pos) {
+	case (TBL_CELL_HORIZ):
+		/* FALLTHROUGH */
+	case (TBL_CELL_DHORIZ):
+		sz = (*tbl->len)(1, tbl->arg);
+		if (col->width < sz)
+			col->width = sz;
+		break;
+	case (TBL_CELL_LONG):
+		/* FALLTHROUGH */
+	case (TBL_CELL_CENTRE):
+		/* FALLTHROUGH */
+	case (TBL_CELL_LEFT):
+		/* FALLTHROUGH */
+	case (TBL_CELL_RIGHT):
+		tblcalc_literal(tbl, col, dp);
+		break;
+	case (TBL_CELL_NUMBER):
+		tblcalc_number(tbl, col, tp, dp);
+		break;
+	default:
+		abort();
+		/* NOTREACHED */
+	}
+}
 
-		if ('\0' == *wp)
-			return(j + 1);
+static void
+tblcalc_literal(struct rofftbl *tbl, struct roffcol *col,
+		const struct tbl_dat *dp)
+{
+	size_t		 sz, bufsz, spsz;
 
-		*d = DECO_SPECIAL;
-		*sz = (size_t)j;
-		return(j + 2);
+	/* 
+	 * Calculate our width and use the spacing, with a minimum
+	 * spacing dictated by position (centre, e.g,. gets a space on
+	 * either side, while right/left get a single adjacent space).
+	 */
 
-	case ('c'):
-		*d = DECO_NOSPACE;
-		*sz = 1;
-		return(1);
+	sz = bufsz = spsz = 0;
+	if (dp->string)
+		sz = (*tbl->slen)(dp->string, tbl->arg);
 
+	assert(dp->layout);
+	switch (dp->layout->pos) {
+	case (TBL_CELL_LONG):
+		/* FALLTHROUGH */
+	case (TBL_CELL_CENTRE):
+		bufsz = (*tbl->len)(2, tbl->arg);
+		break;
 	default:
+		bufsz = (*tbl->len)(1, tbl->arg);
 		break;
 	}
 
-	*d = DECO_SPECIAL;
-	*word = wp;
-	*sz = 1;
-	return(1);
+	if (dp->layout->spacing) {
+		spsz = (*tbl->len)(dp->layout->spacing, tbl->arg);
+		bufsz = bufsz > spsz ? bufsz : spsz;
+	}
+
+	sz += bufsz;
+	if (col->width < sz)
+		col->width = sz;
 }
+
+static void
+tblcalc_number(struct rofftbl *tbl, struct roffcol *col,
+		const struct tbl *tp, const struct tbl_dat *dp)
+{
+	int 		 i;
+	size_t		 sz, psz, ssz, d, max;
+	char		*cp;
+	const char	*str;
+	char		 buf[2];
+
+	/*
+	 * First calculate number width and decimal place (last + 1 for
+	 * no-decimal numbers).  If the stored decimal is subsequent
+	 * ours, make our size longer by that difference
+	 * (right-"shifting"); similarly, if ours is subsequent the
+	 * stored, then extend the stored size by the difference.
+	 * Finally, re-assign the stored values.
+	 */
+
+	str = dp && dp->string ? dp->string : "";
+	max = dp && dp->layout ? dp->layout->spacing : 0;
+
+	sz = (*tbl->slen)(str, tbl->arg);
+
+	buf[0] = tp->decimal;
+	buf[1] = '\0';
+
+	psz = (*tbl->slen)(buf, tbl->arg);
+
+	if (NULL != (cp = strrchr(str, tp->decimal))) {
+		buf[1] = '\0';
+		for (ssz = 0, i = 0; cp != &str[i]; i++) {
+			buf[0] = str[i];
+			ssz += (*tbl->slen)(buf, tbl->arg);
+		}
+		d = ssz + psz;
+	} else
+		d = sz + psz;
+
+	/* Padding. */
+
+	sz += (*tbl->len)(2, tbl->arg);
+	d += (*tbl->len)(1, tbl->arg);
+
+	/* Adjust the settings for this column. */
+
+	if (col->decimal > d) {
+		sz += col->decimal - d;
+		d = col->decimal;
+	} else
+		col->width += d - col->decimal;
+
+	if (sz > col->width)
+		col->width = sz;
+	if (d > col->decimal)
+		col->decimal = d;
+
+	/* Adjust for stipulated width. */
+
+	if (col->width < max)
+		col->width = max;
+}
+
+