[BACK]Return to html.c CVS log [TXT][DIR] Up to [cvsweb.bsd.lv] / mandoc

File: [cvsweb.bsd.lv] / mandoc / html.c (download)

Revision 1.201, Wed Jan 25 02:14:43 2017 UTC (7 years, 2 months ago) by schwarze
Branch: MAIN
Changes since 1.200: +4 -1 lines

Improve HTML formatting of .Bl -tag.

In particular, when using the style sheet, put the body on the same
line as the head for short heads, or on the next line for long
heads, in a way that preserves both correct indentation and correct
vertical spacing with and without -compact, and with one or more
heads per body (hi, Zaphod) - eight use cases so far - and with and
without -tag, and with and without -offset, 32 use cases grand total.

Using many ideas from zhuk@, from <David dot Dahlberg at fkie dot
fraunhofer dot de>, and from Benny Lofgren <bl dash lists at lofgren
dot biz>, and a few of my own.

This is an excellent demonstration that CSS is an extremely hostile
language, much more trapful and much harder to use than, say, C.
When matthew@ reported this in July 2014 (!), it was already a known
issue, and i no longer remember for how long.  My first serious
attempt at fixing it (in November 2015) failed miserably.  I'd love
to see simplifications of both the generated HTML code and of the
style sheet, but without breaking any of the 32 use cases, please.

/*	$Id: html.c,v 1.201 2017/01/25 02:14:43 schwarze Exp $ */
/*
 * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
 * Copyright (c) 2011-2015, 2017 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
 * copyright notice and this permission notice appear in all copies.
 *
 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 */
#include "config.h"

#include <sys/types.h>

#include <assert.h>
#include <ctype.h>
#include <stdarg.h>
#include <stdio.h>
#include <stdint.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>

#include "mandoc.h"
#include "mandoc_aux.h"
#include "out.h"
#include "html.h"
#include "manconf.h"
#include "main.h"

struct	htmldata {
	const char	 *name;
	int		  flags;
#define	HTML_NOSTACK	 (1 << 0)
#define	HTML_AUTOCLOSE	 (1 << 1)
#define	HTML_NLBEFORE	 (1 << 2)
#define	HTML_NLBEGIN	 (1 << 3)
#define	HTML_NLEND	 (1 << 4)
#define	HTML_NLAFTER	 (1 << 5)
#define	HTML_NLAROUND	 (HTML_NLBEFORE | HTML_NLAFTER)
#define	HTML_NLINSIDE	 (HTML_NLBEGIN | HTML_NLEND)
#define	HTML_NLALL	 (HTML_NLAROUND | HTML_NLINSIDE)
#define	HTML_INDENT	 (1 << 6)
#define	HTML_NOINDENT	 (1 << 7)
};

static	const struct htmldata htmltags[TAG_MAX] = {
	{"html",	HTML_NLALL},
	{"head",	HTML_NLALL | HTML_INDENT},
	{"body",	HTML_NLALL},
	{"meta",	HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL},
	{"title",	HTML_NLAROUND},
	{"div",		HTML_NLAROUND},
	{"h1",		HTML_NLAROUND},
	{"h2",		HTML_NLAROUND},
	{"span",	0},
	{"link",	HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL},
	{"br",		HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL},
	{"a",		0},
	{"table",	HTML_NLALL | HTML_INDENT},
	{"tbody",	HTML_NLALL | HTML_INDENT},
	{"col",		HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL},
	{"tr",		HTML_NLALL | HTML_INDENT},
	{"td",		HTML_NLAROUND},
	{"li",		HTML_NLAROUND | HTML_INDENT},
	{"ul",		HTML_NLALL | HTML_INDENT},
	{"ol",		HTML_NLALL | HTML_INDENT},
	{"dl",		HTML_NLALL | HTML_INDENT},
	{"dt",		HTML_NLAROUND},
	{"dd",		HTML_NLAROUND | HTML_INDENT},
	{"pre",		HTML_NLALL | HTML_NOINDENT},
	{"b",		0},
	{"i",		0},
	{"code",	0},
	{"small",	0},
	{"style",	HTML_NLALL | HTML_INDENT},
	{"math",	HTML_NLALL | HTML_INDENT},
	{"mrow",	0},
	{"mi",		0},
	{"mo",		0},
	{"msup",	0},
	{"msub",	0},
	{"msubsup",	0},
	{"mfrac",	0},
	{"msqrt",	0},
	{"mfenced",	0},
	{"mtable",	0},
	{"mtr",		0},
	{"mtd",		0},
	{"munderover",	0},
	{"munder",	0},
	{"mover",	0},
};

static	const char	*const roffscales[SCALE_MAX] = {
	"cm", /* SCALE_CM */
	"in", /* SCALE_IN */
	"pc", /* SCALE_PC */
	"pt", /* SCALE_PT */
	"em", /* SCALE_EM */
	"em", /* SCALE_MM */
	"ex", /* SCALE_EN */
	"ex", /* SCALE_BU */
	"em", /* SCALE_VS */
	"ex", /* SCALE_FS */
};

static	void	 a2width(const char *, struct roffsu *);
static	void	 print_byte(struct html *, char);
static	void	 print_endline(struct html *);
static	void	 print_endword(struct html *);
static	void	 print_indent(struct html *);
static	void	 print_word(struct html *, const char *);

static	void	 print_ctag(struct html *, struct tag *);
static	int	 print_escape(struct html *, char);
static	int	 print_encode(struct html *, const char *, const char *, int);
static	void	 print_href(struct html *, const char *, const char *, int);
static	void	 print_metaf(struct html *, enum mandoc_esc);


void *
html_alloc(const struct manoutput *outopts)
{
	struct html	*h;

	h = mandoc_calloc(1, sizeof(struct html));

	h->tags.head = NULL;
	h->style = outopts->style;
	h->base_man = outopts->man;
	h->base_includes = outopts->includes;
	if (outopts->fragment)
		h->oflags |= HTML_FRAGMENT;

	return h;
}

void
html_free(void *p)
{
	struct tag	*tag;
	struct html	*h;

	h = (struct html *)p;

	while ((tag = h->tags.head) != NULL) {
		h->tags.head = tag->next;
		free(tag);
	}

	free(h);
}

void
print_gen_head(struct html *h)
{
	struct tag	*t;

	print_otag(h, TAG_META, "?", "charset", "utf-8");

	/*
	 * Print a default style-sheet.
	 */

	t = print_otag(h, TAG_STYLE, "");
	print_text(h, "table.head, table.foot { width: 100%; }");
	print_endline(h);
	print_text(h, "td.head-rtitle, td.foot-os { text-align: right; }");
	print_endline(h);
	print_text(h, "td.head-vol { text-align: center; }");
	print_endline(h);
	print_text(h, "div.Pp { margin: 1ex 0ex; }");
	print_tagq(h, t);

	if (h->style)
		print_otag(h, TAG_LINK, "?h??", "rel", "stylesheet",
		    h->style, "type", "text/css", "media", "all");
}

static void
print_metaf(struct html *h, enum mandoc_esc deco)
{
	enum htmlfont	 font;

	switch (deco) {
	case ESCAPE_FONTPREV:
		font = h->metal;
		break;
	case ESCAPE_FONTITALIC:
		font = HTMLFONT_ITALIC;
		break;
	case ESCAPE_FONTBOLD:
		font = HTMLFONT_BOLD;
		break;
	case ESCAPE_FONTBI:
		font = HTMLFONT_BI;
		break;
	case ESCAPE_FONT:
	case ESCAPE_FONTROMAN:
		font = HTMLFONT_NONE;
		break;
	default:
		abort();
	}

	if (h->metaf) {
		print_tagq(h, h->metaf);
		h->metaf = NULL;
	}

	h->metal = h->metac;
	h->metac = font;

	switch (font) {
	case HTMLFONT_ITALIC:
		h->metaf = print_otag(h, TAG_I, "");
		break;
	case HTMLFONT_BOLD:
		h->metaf = print_otag(h, TAG_B, "");
		break;
	case HTMLFONT_BI:
		h->metaf = print_otag(h, TAG_B, "");
		print_otag(h, TAG_I, "");
		break;
	default:
		break;
	}
}

int
html_strlen(const char *cp)
{
	size_t		 rsz;
	int		 skip, sz;

	/*
	 * Account for escaped sequences within string length
	 * calculations.  This follows the logic in term_strlen() as we
	 * must calculate the width of produced strings.
	 * Assume that characters are always width of "1".  This is
	 * hacky, but it gets the job done for approximation of widths.
	 */

	sz = 0;
	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:
		case ESCAPE_NUMBERED:
		case ESCAPE_SPECIAL:
		case ESCAPE_OVERSTRIKE:
			if (skip)
				skip = 0;
			else
				sz++;
			break;
		case ESCAPE_SKIPCHAR:
			skip = 1;
			break;
		default:
			break;
		}
	}
	return sz;
}

static int
print_escape(struct html *h, char c)
{

	switch (c) {
	case '<':
		print_word(h, "&lt;");
		break;
	case '>':
		print_word(h, "&gt;");
		break;
	case '&':
		print_word(h, "&amp;");
		break;
	case '"':
		print_word(h, "&quot;");
		break;
	case ASCII_NBRSP:
		print_word(h, "&nbsp;");
		break;
	case ASCII_HYPH:
		print_byte(h, '-');
		break;
	case ASCII_BREAK:
		break;
	default:
		return 0;
	}
	return 1;
}

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;
	const char	*seq;
	enum mandoc_esc	 esc;
	static const char rejs[9] = { '\\', '<', '>', '&', '"',
		ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' };

	if (pend == NULL)
		pend = strchr(p, '\0');

	nospace = 0;

	while (p < pend) {
		if (HTML_SKIPCHAR & h->flags && '\\' != *p) {
			h->flags &= ~HTML_SKIPCHAR;
			p++;
			continue;
		}

		for (sz = strcspn(p, rejs); sz-- && p < pend; p++)
			if (*p == ' ')
				print_endword(h);
			else
				print_byte(h, *p);

		if (p >= pend)
			break;

		if (print_escape(h, *p++))
			continue;

		esc = mandoc_escape(&p, &seq, &len);
		if (ESCAPE_ERROR == esc)
			break;

		switch (esc) {
		case ESCAPE_FONT:
		case ESCAPE_FONTPREV:
		case ESCAPE_FONTBOLD:
		case ESCAPE_FONTITALIC:
		case ESCAPE_FONTBI:
		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 past "u" header. */
			c = mchars_num2uc(seq + 1, len - 1);
			break;
		case ESCAPE_NUMBERED:
			c = mchars_num2char(seq, len);
			if (c < 0)
				continue;
			break;
		case ESCAPE_SPECIAL:
			c = mchars_spec2cp(seq, len);
			if (c <= 0)
				continue;
			break;
		case ESCAPE_NOSPACE:
			if ('\0' == *p)
				nospace = 1;
			continue;
		case ESCAPE_OVERSTRIKE:
			if (len == 0)
				continue;
			c = seq[len - 1];
			break;
		default:
			continue;
		}
		if ((c < 0x20 && c != 0x09) ||
		    (c > 0x7E && c < 0xA0))
			c = 0xFFFD;
		if (c > 0x7E) {
			(void)snprintf(numbuf, sizeof(numbuf), "&#%d;", c);
			print_word(h, numbuf);
		} else if (print_escape(h, c) == 0)
			print_byte(h, c);
	}

	return nospace;
}

static void
print_href(struct html *h, const char *name, const char *sec, int man)
{
	const char	*p, *pp;

	pp = man ? h->base_man : h->base_includes;
	while ((p = strchr(pp, '%')) != NULL) {
		print_encode(h, pp, p, 1);
		if (man && p[1] == 'S') {
			if (sec == NULL)
				print_byte(h, '1');
			else
				print_encode(h, sec, NULL, 1);
		} else if ((man && p[1] == 'N') ||
		    (man == 0 && p[1] == 'I'))
			print_encode(h, name, NULL, 1);
		else
			print_encode(h, p, p + 2, 1);
		pp = p + 2;
	}
	if (*pp != '\0')
		print_encode(h, pp, NULL, 1);
}

struct tag *
print_otag(struct html *h, enum htmltag tag, const char *fmt, ...)
{
	va_list		 ap;
	struct roffsu	 mysu, *su;
	char		 numbuf[16];
	struct tag	*t;
	const char	*attr;
	char		*s;
	double		 v;
	int		 i, have_style, tflags;

	tflags = htmltags[tag].flags;

	/* Push this tags onto the stack of open scopes. */

	if ((tflags & HTML_NOSTACK) == 0) {
		t = mandoc_malloc(sizeof(struct tag));
		t->tag = tag;
		t->next = h->tags.head;
		h->tags.head = t;
	} else
		t = NULL;

	if (tflags & HTML_NLBEFORE)
		print_endline(h);
	if (h->col == 0)
		print_indent(h);
	else if ((h->flags & HTML_NOSPACE) == 0) {
		if (h->flags & HTML_KEEP)
			print_word(h, "&#160;");
		else {
			if (h->flags & HTML_PREKEEP)
				h->flags |= HTML_KEEP;
			print_endword(h);
		}
	}

	if ( ! (h->flags & HTML_NONOSPACE))
		h->flags &= ~HTML_NOSPACE;
	else
		h->flags |= HTML_NOSPACE;

	/* Print out the tag name and attributes. */

	print_byte(h, '<');
	print_word(h, htmltags[tag].name);

	va_start(ap, fmt);

	have_style = 0;
	while (*fmt != '\0') {
		if (*fmt == 's') {
			print_word(h, " style=\"");
			have_style = 1;
			fmt++;
			break;
		}
		s = va_arg(ap, char *);
		switch (*fmt++) {
		case 'c':
			attr = "class";
			break;
		case 'h':
			attr = "href";
			break;
		case 'i':
			attr = "id";
			break;
		case '?':
			attr = s;
			s = va_arg(ap, char *);
			break;
		default:
			abort();
		}
		print_byte(h, ' ');
		print_word(h, attr);
		print_byte(h, '=');
		print_byte(h, '"');
		switch (*fmt) {
		case 'M':
			print_href(h, s, va_arg(ap, char *), 1);
			fmt++;
			break;
		case 'I':
			print_href(h, s, NULL, 0);
			fmt++;
			break;
		case 'R':
			print_byte(h, '#');
			fmt++;
			/* FALLTHROUGH */
		default:
			print_encode(h, s, NULL, 1);
			break;
		}
		print_byte(h, '"');
	}

	/* Print out styles. */

	s = NULL;
	su = &mysu;
	while (*fmt != '\0') {

		/* First letter: input argument type. */

		switch (*fmt++) {
		case 'h':
			i = va_arg(ap, int);
			SCALE_HS_INIT(su, i);
			break;
		case 's':
			s = va_arg(ap, char *);
			break;
		case 'u':
			su = va_arg(ap, struct roffsu *);
			break;
		case 'v':
			i = va_arg(ap, int);
			SCALE_VS_INIT(su, i);
			break;
		case 'w':
		case 'W':
			s = va_arg(ap, char *);
			a2width(s, su);
			if (fmt[-1] == 'W')
				su->scale *= -1.0;
			break;
		default:
			abort();
		}

		/* Second letter: style name. */

		switch (*fmt++) {
		case 'b':
			attr = "margin-bottom";
			break;
		case 'h':
			attr = "height";
			break;
		case 'i':
			attr = "text-indent";
			break;
		case 'l':
			attr = "margin-left";
			break;
		case 't':
			attr = "margin-top";
			break;
		case 'w':
			attr = "width";
			break;
		case 'W':
			attr = "min-width";
			break;
		case '?':
			print_word(h, s);
			print_byte(h, ':');
			print_byte(h, ' ');
			print_word(h, va_arg(ap, char *));
			print_byte(h, ';');
			if (*fmt != '\0')
				print_byte(h, ' ');
			continue;
		default:
			abort();
		}
		v = su->scale;
		if (su->unit == SCALE_MM && (v /= 100.0) == 0.0)
			v = 1.0;
		else if (su->unit == SCALE_BU)
			v /= 24.0;
		print_word(h, attr);
		print_byte(h, ':');
		print_byte(h, ' ');
		(void)snprintf(numbuf, sizeof(numbuf), "%.2f", v);
		print_word(h, numbuf);
		print_word(h, roffscales[su->unit]);
		print_byte(h, ';');
		if (*fmt != '\0')
			print_byte(h, ' ');
	}
	if (have_style)
		print_byte(h, '"');

	va_end(ap);

	/* Accommodate for "well-formed" singleton escaping. */

	if (HTML_AUTOCLOSE & htmltags[tag].flags)
		print_byte(h, '/');

	print_byte(h, '>');

	if (tflags & HTML_NLBEGIN)
		print_endline(h);
	else
		h->flags |= HTML_NOSPACE;

	if (tflags & HTML_INDENT)
		h->indent++;
	if (tflags & HTML_NOINDENT)
		h->noindent++;

	return t;
}

static void
print_ctag(struct html *h, struct tag *tag)
{
	int	 tflags;

	/*
	 * Remember to close out and nullify the current
	 * meta-font and table, if applicable.
	 */
	if (tag == h->metaf)
		h->metaf = NULL;
	if (tag == h->tblt)
		h->tblt = NULL;

	tflags = htmltags[tag->tag].flags;

	if (tflags & HTML_INDENT)
		h->indent--;
	if (tflags & HTML_NOINDENT)
		h->noindent--;
	if (tflags & HTML_NLEND)
		print_endline(h);
	print_indent(h);
	print_byte(h, '<');
	print_byte(h, '/');
	print_word(h, htmltags[tag->tag].name);
	print_byte(h, '>');
	if (tflags & HTML_NLAFTER)
		print_endline(h);

	h->tags.head = tag->next;
	free(tag);
}

void
print_gen_decls(struct html *h)
{
	print_word(h, "<!DOCTYPE html>");
	print_endline(h);
}

void
print_text(struct html *h, const char *word)
{
	if (h->col && (h->flags & HTML_NOSPACE) == 0) {
		if ( ! (HTML_KEEP & h->flags)) {
			if (HTML_PREKEEP & h->flags)
				h->flags |= HTML_KEEP;
			print_endword(h);
		} else
			print_word(h, "&#160;");
	}

	assert(NULL == h->metaf);
	switch (h->metac) {
	case HTMLFONT_ITALIC:
		h->metaf = print_otag(h, TAG_I, "");
		break;
	case HTMLFONT_BOLD:
		h->metaf = print_otag(h, TAG_B, "");
		break;
	case HTMLFONT_BI:
		h->metaf = print_otag(h, TAG_B, "");
		print_otag(h, TAG_I, "");
		break;
	default:
		print_indent(h);
		break;
	}

	assert(word);
	if ( ! print_encode(h, word, NULL, 0)) {
		if ( ! (h->flags & HTML_NONOSPACE))
			h->flags &= ~HTML_NOSPACE;
		h->flags &= ~HTML_NONEWLINE;
	} else
		h->flags |= HTML_NOSPACE | HTML_NONEWLINE;

	if (h->metaf) {
		print_tagq(h, h->metaf);
		h->metaf = NULL;
	}

	h->flags &= ~HTML_IGNDELIM;
}

void
print_tagq(struct html *h, const struct tag *until)
{
	struct tag	*tag;

	while ((tag = h->tags.head) != NULL) {
		print_ctag(h, tag);
		if (until && tag == until)
			return;
	}
}

void
print_stagq(struct html *h, const struct tag *suntil)
{
	struct tag	*tag;

	while ((tag = h->tags.head) != NULL) {
		if (suntil && tag == suntil)
			return;
		print_ctag(h, tag);
	}
}

void
print_paragraph(struct html *h)
{
	struct tag	*t;

	t = print_otag(h, TAG_DIV, "c", "Pp");
	print_tagq(h, t);
}


/***********************************************************************
 * Low level output functions.
 * They implement line breaking using a short static buffer.
 ***********************************************************************/

/*
 * Buffer one HTML output byte.
 * If the buffer is full, flush and deactivate it and start a new line.
 * If the buffer is inactive, print directly.
 */
static void
print_byte(struct html *h, char c)
{
	if ((h->flags & HTML_BUFFER) == 0) {
		putchar(c);
		h->col++;
		return;
	}

	if (h->col + h->bufcol < sizeof(h->buf)) {
		h->buf[h->bufcol++] = c;
		return;
	}

	putchar('\n');
	h->col = 0;
	print_indent(h);
	putchar(' ');
	putchar(' ');
	fwrite(h->buf, h->bufcol, 1, stdout);
	putchar(c);
	h->col = (h->indent + 1) * 2 + h->bufcol + 1;
	h->bufcol = 0;
	h->flags &= ~HTML_BUFFER;
}

/*
 * If something was printed on the current output line, end it.
 * Not to be called right after print_indent().
 */
static void
print_endline(struct html *h)
{
	if (h->col == 0)
		return;

	if (h->bufcol) {
		putchar(' ');
		fwrite(h->buf, h->bufcol, 1, stdout);
		h->bufcol = 0;
	}
	putchar('\n');
	h->col = 0;
	h->flags |= HTML_NOSPACE;
	h->flags &= ~HTML_BUFFER;
}

/*
 * Flush the HTML output buffer.
 * If it is inactive, activate it.
 */
static void
print_endword(struct html *h)
{
	if (h->noindent) {
		print_byte(h, ' ');
		return;
	}

	if ((h->flags & HTML_BUFFER) == 0) {
		h->col++;
		h->flags |= HTML_BUFFER;
	} else if (h->bufcol) {
		putchar(' ');
		fwrite(h->buf, h->bufcol, 1, stdout);
		h->col += h->bufcol + 1;
	}
	h->bufcol = 0;
}

/*
 * If at the beginning of a new output line,
 * perform indentation and mark the line as containing output.
 * Make sure to really produce some output right afterwards,
 * but do not use print_otag() for producing it.
 */
static void
print_indent(struct html *h)
{
	size_t	 i;

	if (h->col)
		return;

	if (h->noindent == 0) {
		h->col = h->indent * 2;
		for (i = 0; i < h->col; i++)
			putchar(' ');
	}
	h->flags &= ~HTML_NOSPACE;
}

/*
 * Print or buffer some characters
 * depending on the current HTML output buffer state.
 */
static void
print_word(struct html *h, const char *cp)
{
	while (*cp != '\0')
		print_byte(h, *cp++);
}

/*
 * Calculate the scaling unit passed in a `-width' argument.  This uses
 * either a native scaling unit (e.g., 1i, 2m) or the string length of
 * the value.
 */
static void
a2width(const char *p, struct roffsu *su)
{
	if (a2roffsu(p, su, SCALE_MAX) < 2) {
		su->unit = SCALE_EN;
		su->scale = html_strlen(p);
	} else if (su->scale < 0.0)
		su->scale = 0.0;
}