Return to mdocterm.c CVS log

Up to [cvsweb.bsd.lv] / mandoc

Many more ASCII escapes added.
Stripped nroff punt from mdocterm.c.
Fixed multiple-file handling for mdocterm.c.
Cleared up mdoc.h API.

/* $Id: mdocterm.c,v 1.49 2009/03/19 11:49:00 kristaps Exp $ */
/*
 * Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@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 AUTHOR DISCLAIMS ALL
 * WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE
 * AUTHOR 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 <sys/types.h>
#include <sys/stat.h>

#include <assert.h>
#include <err.h>
#include <fcntl.h>
#include <getopt.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>

#include "term.h"

#define	WARN_WALL	  0x03		/* All-warnings mask. */
#define	WARN_WCOMPAT	 (1 << 0)	/* Compatibility warnings. */
#define	WARN_WSYNTAX	 (1 << 1)	/* Syntax warnings. */
#define	WARN_WERR	 (1 << 2)	/* Warnings->errors. */

enum	termt {
	TERMT_ASCII,
	TERMT_LINT,
	TERMT_TREE
};

extern	char		 *__progname;

__dead	static void	  version(void);
__dead	static void	  usage(void);
#if 0
__dead	static void	  punt(struct cmdargs *, char **);
#endif
static	int		  foptions(int *, char *);
static	int		  toptions(enum termt *, char *);
static	int		  woptions(int *, char *);
static	int		  merr(void *, int, int, const char *);
static	int		  mwarn(void *, int, int, 
				enum mdoc_warn, const char *);
static	void		  body(struct termp *, struct termpair *,
				const struct mdoc_meta *,
				const struct mdoc_node *);
static	void		  header(struct termp *,
				const struct mdoc_meta *);
static	void		  footer(struct termp *,
				const struct mdoc_meta *);
static	int		  file(char **, size_t *, char **, size_t *, 
				const char *, struct mdoc *);
static	int		  fdesc(char **, size_t *, char **, size_t *, 
				const char *, int, struct mdoc *);
static	void		  pword(struct termp *, const char *, int);
static	void		  pescape(struct termp *, 
				const char *, int *, int);
static	void		  nescape(struct termp *,
				const char *, size_t);
static	void		  chara(struct termp *, char);
static	void		  stringa(struct termp *, 
				const char *, size_t);
static	void		  sanity(const struct mdoc_node *);


int
main(int argc, char *argv[])
{
	struct termp	 termp;
	int		 c, fflags, wflags;
	struct mdoc_cb	 cb;
	struct mdoc	*mdoc;
	char		*buf, *line;
	size_t		 bufsz, linesz;
	enum termt	 termt;

	bzero(&termp, sizeof(struct termp));
	bzero(&cb, sizeof(struct mdoc_cb));

	termt = TERMT_ASCII;
	fflags = wflags = 0;

	/* LINTED */
	while (-1 != (c = getopt(argc, argv, "f:VW:T:")))
		switch (c) {
		case ('f'):
			if ( ! foptions(&fflags, optarg))
				return(0);
			break;
		case ('T'):
			if ( ! toptions(&termt, optarg))
				return(0);
			break;
		case ('W'):
			if ( ! woptions(&wflags, optarg))
				return(0);
			break;
		case ('V'):
			version();
			/* NOTREACHED */
		default:
			usage();
			/* NOTREACHED */
		}

	argc -= optind;
	argv += optind;

	termp.maxrmargin = 78;			/* FIXME */

	cb.mdoc_err = merr;
	cb.mdoc_warn = mwarn;

	/* Line and block buffers persist between parses. */

	buf = line = NULL;
	bufsz = linesz = 0;

	/* Overall mdoc persists between parses. */

	mdoc = mdoc_alloc(&wflags, fflags, &cb);

	while (*argv) {
		if ( ! file(&line, &linesz, &buf, &bufsz, *argv, mdoc))
			break;

		switch (termt) {
		case (TERMT_ASCII):
			if (NULL == termp.symtab)
				termp.symtab = ascii2htab();
			header(&termp, mdoc_meta(mdoc));
			body(&termp, NULL, mdoc_meta(mdoc), 
					mdoc_node(mdoc));
			footer(&termp, mdoc_meta(mdoc));
			break;
		default:
			break;
		}

		mdoc_reset(mdoc);
		argv++;
	}

	if (buf)
		free(buf);
	if (line)
		free(line);
	if (termp.buf)
		free(termp.buf);
	if (termp.symtab)
		asciifree(termp.symtab);

	mdoc_free(mdoc);

	return(0);
}


__dead static void
version(void)
{

	(void)printf("%s %s\n", __progname, VERSION);
	exit(0);
	/* NOTREACHED */
}


__dead static void
usage(void)
{

	(void)fprintf(stderr, "usage: %s\n", __progname);
	exit(1);
	/* NOTREACHED */
}


static int
file(char **ln, size_t *lnsz, char **buf, size_t *bufsz, 
		const char *file, struct mdoc *mdoc)
{
	int		 fd, c;

	if (-1 == (fd = open(file, O_RDONLY, 0))) {
		warn("%s", file);
		return(0);
	}

	c = fdesc(ln, lnsz, buf, bufsz, file, fd, mdoc);

	if (-1 == close(fd))
		warn("%s", file);

	return(c);
}


static int
fdesc(char **lnp, size_t *lnsz, char **bufp, size_t *bufsz, 
		const char *f, int fd, struct mdoc *mdoc)
{
	size_t		 sz;
	ssize_t		 ssz;
	struct stat	 st;
	int		 j, i, pos, lnn;
	char		*ln, *buf;

	buf = *bufp;
	ln = *lnp;

	/*
	 * Two buffers: ln and buf.  buf is the input buffer, optimised
	 * for each file's block size.  ln is a line buffer.  Both
	 * growable, hence passed in by ptr-ptr.
	 */

	if (-1 == fstat(fd, &st)) {
		warnx("%s", f);
		sz = BUFSIZ;
	} else 
		sz = (unsigned)BUFSIZ > st.st_blksize ?
			(size_t)BUFSIZ : st.st_blksize;

	if (sz > *bufsz) {
		if (NULL == (buf = realloc(buf, sz)))
			err(1, "realloc");
		*bufp = buf;
		*bufsz = sz;
	}

	/*
	 * Fill buf with file blocksize and parse newlines into ln.
	 */

	for (lnn = 1, pos = 0; ; ) {
		if (-1 == (ssz = read(fd, buf, sz))) {
			warn("%s", f);
			return(0);
		} else if (0 == ssz) 
			break;

		for (i = 0; i < (int)ssz; i++) {
			if (pos >= (int)*lnsz) {
				*lnsz += 256; /* Step-size. */
				ln = realloc(ln, *lnsz);
				if (NULL == ln)
					err(1, "realloc");
				*lnp = ln;
			}

			if ('\n' != buf[i]) {
				ln[pos++] = buf[i];
				continue;
			}

			/* Check for CPP-escaped newline.  */

			if (pos > 0 && '\\' == ln[pos - 1]) {
				for (j = pos - 1; j >= 0; j--)
					if ('\\' != ln[j])
						break;

				if ( ! ((pos - j) % 2)) {
					pos--;
					lnn++;
					continue;
				}
			}

			ln[pos] = 0;
			if ( ! mdoc_parseln(mdoc, lnn, ln))
				return(0);
			lnn++;
			pos = 0;
		}
	}

	return(mdoc_endparse(mdoc));
}


static int
toptions(enum termt *tflags, char *arg)
{

	if (0 == strcmp(arg, "ascii"))
		*tflags = TERMT_ASCII;
	else if (0 == strcmp(arg, "lint"))
		*tflags = TERMT_LINT;
	else if (0 == strcmp(arg, "tree"))
		*tflags = TERMT_TREE;
	else {
		warnx("bad argument: -T%s", arg);
		return(0);
	}

	return(1);
}


/*
 * Parse out the options for [-fopt...] setting compiler options.  These
 * can be comma-delimited or called again.
 */
static int
foptions(int *fflags, char *arg)
{
	char		*v;
	char		*toks[] = { "ign-scope", "ign-escape", 
				    "ign-macro", NULL };

	while (*arg) 
		switch (getsubopt(&arg, toks, &v)) {
		case (0):
			*fflags |= MDOC_IGN_SCOPE;
			break;
		case (1):
			*fflags |= MDOC_IGN_ESCAPE;
			break;
		case (2):
			*fflags |= MDOC_IGN_MACRO;
			break;
		default:
			warnx("bad argument: -f%s", arg);
			return(0);
		}

	return(1);
}


/* 
 * Parse out the options for [-Werr...], which sets warning modes.
 * These can be comma-delimited or called again.  XXX - should this be
 * using -w like troff?
 */
static int
woptions(int *wflags, char *arg)
{
	char		*v;
	char		*toks[] = { "all", "compat", 
			"syntax", "error", NULL };

	while (*arg) 
		switch (getsubopt(&arg, toks, &v)) {
		case (0):
			*wflags |= WARN_WALL;
			break;
		case (1):
			*wflags |= WARN_WCOMPAT;
			break;
		case (2):
			*wflags |= WARN_WSYNTAX;
			break;
		case (3):
			*wflags |= WARN_WERR;
			break;
		default:
			warnx("bad argument: -W%s", arg);
			return(0);
		}

	return(1);
}


/*
 * Flush a line of text.  A "line" is loosely defined as being something
 * that should be followed by a newline, regardless of whether it's
 * broken apart by newlines getting there.  A line can also be a
 * fragment of a columnar list.
 *
 * Specifically, a line is whatever's in p->buf of length p->col, which
 * is zeroed after this function returns.
 *
 * The variables TERMP_NOLPAD, TERMP_LITERAL and TERMP_NOBREAK are of
 * critical importance here.  Their behaviour follows:
 *
 *  - TERMP_NOLPAD: when beginning to write the line, don't left-pad the
 *    offset value.  This is useful when doing columnar lists where the
 *    prior column has right-padded.
 *
 *  - TERMP_NOBREAK: this is the most important and is used when making
 *    columns.  In short: don't print a newline and instead pad to the
 *    right margin.  Used in conjunction with TERMP_NOLPAD.
 *
 *  - TERMP_NONOBREAK: don't newline when TERMP_NOBREAK is specified.
 *
 *  In-line line breaking:
 *
 *  If TERMP_NOBREAK is specified and the line overruns the right
 *  margin, it will break and pad-right to the right margin after
 *  writing.  If maxrmargin is violated, it will break and continue
 *  writing from the right-margin, which will lead to the above
 *  scenario upon exit.
 *
 *  Otherwise, the line will break at the right margin.  Extremely long
 *  lines will cause the system to emit a warning (TODO: hyphenate, if
 *  possible).
 */
void
flushln(struct termp *p)
{
	int		 i, j;
	size_t		 vsz, vis, maxvis, mmax, bp;

	/*
	 * First, establish the maximum columns of "visible" content.
	 * This is usually the difference between the right-margin and
	 * an indentation, but can be, for tagged lists or columns, a
	 * small set of values.
	 */

	assert(p->offset < p->rmargin);
	maxvis = p->rmargin - p->offset;
	mmax = p->maxrmargin - p->offset;
	bp = TERMP_NOBREAK & p->flags ? mmax : maxvis;
	vis = 0;

	/*
	 * If in the standard case (left-justified), then begin with our
	 * indentation, otherwise (columns, etc.) just start spitting
	 * out text.
	 */

	if ( ! (p->flags & TERMP_NOLPAD))
		/* LINTED */
		for (j = 0; j < (int)p->offset; j++)
			putchar(' ');

	for (i = 0; i < (int)p->col; i++) {
		/*
		 * Count up visible word characters.  Control sequences
		 * (starting with the CSI) aren't counted.  A space
		 * generates a non-printing word, which is valid (the
		 * space is printed according to regular spacing rules).
		 */

		/* LINTED */
		for (j = i, vsz = 0; j < (int)p->col; j++) {
			if (' ' == p->buf[j])
				break;
			else if (8 == p->buf[j])
				j += 1;
			else
				vsz++;
		}

		/*
		 * Do line-breaking.  If we're greater than our
		 * break-point and already in-line, break to the next
		 * line and start writing.  If we're at the line start,
		 * then write out the word (TODO: hyphenate) and break
		 * in a subsequent loop invocation.
		 */

		if ( ! (TERMP_NOBREAK & p->flags)) {
			if (vis && vis + vsz > bp) {
				putchar('\n');
				for (j = 0; j < (int)p->offset; j++)
					putchar(' ');
				vis = 0;
			} else if (vis + vsz > bp)
				warnx("word breaks right margin");

			/* TODO: hyphenate. */

		} else {
			if (vis && vis + vsz > bp) {
				putchar('\n');
				for (j = 0; j < (int)p->rmargin; j++)
					putchar(' ');
				vis = p->rmargin - p->offset;
			} else if (vis + vsz > bp) 
				warnx("word breaks right margin");

			/* TODO: hyphenate. */
		}

		/* 
		 * Write out the word and a trailing space.  Omit the
		 * space if we're the last word in the line or beyond
		 * our breakpoint.
		 */

		for ( ; i < (int)p->col; i++) {
			if (' ' == p->buf[i])
				break;
			putchar(p->buf[i]);
		}
		vis += vsz;
		if (i < (int)p->col && vis <= bp) {
			putchar(' ');
			vis++;
		}
	}

	/*
	 * If we've overstepped our maximum visible no-break space, then
	 * cause a newline and offset at the right margin.
	 */

	if ((TERMP_NOBREAK & p->flags) && vis >= maxvis) {
		if ( ! (TERMP_NONOBREAK & p->flags)) {
			putchar('\n');
			for (i = 0; i < (int)p->rmargin; i++)
				putchar(' ');
		}
		p->col = 0;
		return;
	}

	/*
	 * If we're not to right-marginalise it (newline), then instead
	 * pad to the right margin and stay off.
	 */

	if (p->flags & TERMP_NOBREAK) {
		if ( ! (TERMP_NONOBREAK & p->flags))
			for ( ; vis < maxvis; vis++)
				putchar(' ');
	} else
		putchar('\n');

	p->col = 0;
}


/* 
 * A newline only breaks an existing line; it won't assert vertical
 * space.  All data in the output buffer is flushed prior to the newline
 * assertion.
 */
void
newln(struct termp *p)
{

	p->flags |= TERMP_NOSPACE;
	if (0 == p->col) {
		p->flags &= ~TERMP_NOLPAD;
		return;
	}
	flushln(p);
	p->flags &= ~TERMP_NOLPAD;
}


/*
 * Asserts a vertical space (a full, empty line-break between lines).
 * Note that if used twice, this will cause two blank spaces and so on.
 * All data in the output buffer is flushed prior to the newline
 * assertion.
 */
void
vspace(struct termp *p)
{

	newln(p);
	putchar('\n');
}


/*
 * Break apart a word into "pwords" (partial-words, usually from
 * breaking up a phrase into individual words) and, eventually, put them
 * into the output buffer.  If we're a literal word, then don't break up
 * the word and put it verbatim into the output buffer.
 */
void
word(struct termp *p, const char *word)
{
	int 		 i, j, len;

	if (p->flags & TERMP_LITERAL) {
		pword(p, word, (int)strlen(word));
		return;
	}

	if (0 == (len = (int)strlen(word)))
		errx(1, "blank line not in literal context");

	if (mdoc_isdelim(word)) {
		if ( ! (p->flags & TERMP_IGNDELIM))
			p->flags |= TERMP_NOSPACE;
		p->flags &= ~TERMP_IGNDELIM;
	}

	/* LINTED */
	for (j = i = 0; i < len; i++) {
		if (' ' != word[i]) {
			j++;
			continue;
		} 
		
		/* Escaped spaces don't delimit... */
		if (i && ' ' == word[i] && '\\' == word[i - 1]) {
			j++;
			continue;
		}

		if (0 == j)
			continue;
		assert(i >= j);
		pword(p, &word[i - j], j);
		j = 0;
	}
	if (j > 0) {
		assert(i >= j);
		pword(p, &word[i - j], j);
	}
}


/*
 * This is the main function for printing out nodes.  It's constituted
 * of PRE and POST functions, which correspond to prefix and infix
 * processing.  The termpair structure allows data to persist between
 * prefix and postfix invocations.
 */
static void
body(struct termp *p, struct termpair *ppair,
		const struct mdoc_meta *meta,
		const struct mdoc_node *node)
{
	int		 dochild;
	struct termpair	 pair;

	/* Some quick sanity-checking. */

	sanity(node);

	/* Pre-processing. */

	dochild = 1;
	pair.ppair = ppair;
	pair.type = 0;
	pair.offset = pair.rmargin = 0;
	pair.flag = 0;
	pair.count = 0;

	if (MDOC_TEXT != node->type) {
		if (termacts[node->tok].pre)
			if ( ! (*termacts[node->tok].pre)(p, &pair, meta, node))
				dochild = 0;
	} else /* MDOC_TEXT == node->type */
		word(p, node->string);

	/* Children. */

	if (TERMPAIR_FLAG & pair.type)
		p->flags |= pair.flag;

	if (dochild && node->child)
		body(p, &pair, meta, node->child);

	if (TERMPAIR_FLAG & pair.type)
		p->flags &= ~pair.flag;

	/* Post-processing. */

	if (MDOC_TEXT != node->type)
		if (termacts[node->tok].post)
			(*termacts[node->tok].post)(p, &pair, meta, node);

	/* Siblings. */

	if (node->next)
		body(p, ppair, meta, node->next);
}


static void
footer(struct termp *p, const struct mdoc_meta *meta)
{
	struct tm	*tm;
	char		*buf, *os;

	if (NULL == (buf = malloc(p->rmargin)))
		err(1, "malloc");
	if (NULL == (os = malloc(p->rmargin)))
		err(1, "malloc");

	tm = localtime(&meta->date);

#ifdef __OpenBSD__
	if (NULL == strftime(buf, p->rmargin, "%B %d, %Y", tm))
#else
	if (0 == strftime(buf, p->rmargin, "%B %d, %Y", tm))
#endif
		err(1, "strftime");

	(void)strlcpy(os, meta->os, p->rmargin);

	/*
	 * This is /slightly/ different from regular groff output
	 * because we don't have page numbers.  Print the following:
	 *
	 * OS                                            MDOCDATE
	 */

	vspace(p);

	p->flags |= TERMP_NOSPACE | TERMP_NOBREAK;
	p->rmargin = p->maxrmargin - strlen(buf);
	p->offset = 0;

	word(p, os);
	flushln(p);

	p->flags |= TERMP_NOLPAD | TERMP_NOSPACE;
	p->offset = p->rmargin;
	p->rmargin = p->maxrmargin;
	p->flags &= ~TERMP_NOBREAK;

	word(p, buf);
	flushln(p);

	free(buf);
	free(os);
}


static void
header(struct termp *p, const struct mdoc_meta *meta)
{
	char		*buf, *title;

	p->rmargin = p->maxrmargin;
	p->offset = 0;

	if (NULL == (buf = malloc(p->rmargin)))
		err(1, "malloc");
	if (NULL == (title = malloc(p->rmargin)))
		err(1, "malloc");

	/*
	 * The header is strange.  It has three components, which are
	 * really two with the first duplicated.  It goes like this:
	 *
	 * IDENTIFIER              TITLE                   IDENTIFIER
	 *
	 * The IDENTIFIER is NAME(SECTION), which is the command-name
	 * (if given, or "unknown" if not) followed by the manual page
	 * section.  These are given in `Dt'.  The TITLE is a free-form
	 * string depending on the manual volume.  If not specified, it
	 * switches on the manual section.
	 */

	assert(meta->vol);
	(void)strlcpy(buf, meta->vol, p->rmargin);

	if (meta->arch) {
		(void)strlcat(buf, " (", p->rmargin);
		(void)strlcat(buf, meta->arch, p->rmargin);
		(void)strlcat(buf, ")", p->rmargin);
	}

	(void)snprintf(title, p->rmargin, "%s(%d)", 
			meta->title, meta->msec);

	p->offset = 0;
	p->rmargin = (p->maxrmargin - strlen(buf)) / 2;
	p->flags |= TERMP_NOBREAK | TERMP_NOSPACE;

	word(p, title);
	flushln(p);

	p->flags |= TERMP_NOLPAD | TERMP_NOSPACE;
	p->offset = p->rmargin;
	p->rmargin = p->maxrmargin - strlen(title);

	word(p, buf);
	flushln(p);

	p->offset = p->rmargin;
	p->rmargin = p->maxrmargin;
	p->flags &= ~TERMP_NOBREAK;
	p->flags |= TERMP_NOLPAD | TERMP_NOSPACE;

	word(p, title);
	flushln(p);

	p->rmargin = p->maxrmargin;
	p->offset = 0;
	p->flags &= ~TERMP_NOSPACE;

	free(title);
	free(buf);
}


/*
 * Determine the symbol indicated by an escape sequences, that is, one
 * starting with a backslash.  Once done, we pass this value into the
 * output buffer by way of the symbol table.
 */
static void
nescape(struct termp *p, const char *word, size_t len)
{
	const char	*rhs;
	size_t		 sz;

	if (NULL == (rhs = a2ascii(p->symtab, word, len, &sz))) 
		return;
	stringa(p, rhs, sz);
}


/*
 * Handle an escape sequence: determine its length and pass it to the
 * escape-symbol look table.  Note that we assume mdoc(3) has validated
 * the escape sequence (we assert upon badly-formed escape sequences).
 */
static void
pescape(struct termp *p, const char *word, int *i, int len)
{
	int		 j;

	if (++(*i) >= len)
		return;

	if ('(' == word[*i]) {
		(*i)++;
		if (*i + 1 >= len)
			return;

		nescape(p, &word[*i], 2);
		(*i)++;
		return;

	} else if ('*' == word[*i]) { 
		(*i)++;
		if (*i >= len)
			return;

		switch (word[*i]) {
		case ('('):
			(*i)++;
			if (*i + 1 >= len)
				return;

			nescape(p, &word[*i], 2);
			(*i)++;
			return;
		case ('['):
			break;
		default:
			nescape(p, &word[*i], 1);
			return;
		}

	} else if ('[' != word[*i]) {
		nescape(p, &word[*i], 1);
		return;
	}

	(*i)++;
	for (j = 0; word[*i] && ']' != word[*i]; (*i)++, j++)
		/* Loop... */ ;

	if (0 == word[*i])
		return;

	nescape(p, &word[*i - j], (size_t)j);
}


/*
 * Handle pwords, partial words, which may be either a single word or a
 * phrase that cannot be broken down (such as a literal string).  This
 * handles word styling.
 */
static void
pword(struct termp *p, const char *word, int len)
{
	int		 i;

	if ( ! (TERMP_NOSPACE & p->flags) && 
			! (TERMP_LITERAL & p->flags))
		chara(p, ' ');

	if ( ! (p->flags & TERMP_NONOSPACE))
		p->flags &= ~TERMP_NOSPACE;

	/* 
	 * If ANSI (word-length styling), then apply our style now,
	 * before the word.
	 */

	for (i = 0; i < len; i++) {
		if ('\\' == word[i]) {
			pescape(p, word, &i, len);
			continue;
		}

		if (TERMP_STYLE & p->flags) {
			if (TERMP_BOLD & p->flags) {
				chara(p, word[i]);
				chara(p, 8);
			}
			if (TERMP_UNDER & p->flags) {
				chara(p, '_');
				chara(p, 8);
			}
		}

		chara(p, word[i]);
	}
}


/*
 * Like chara() but for arbitrary-length buffers.  Resize the buffer by
 * a factor of two (if the buffer is less than that) or the buffer's
 * size.
 */
static void
stringa(struct termp *p, const char *c, size_t sz)
{
	size_t		 s;

	if (0 == sz)
		return;

	assert(c);
	if (p->col + sz >= p->maxcols) {
		if (0 == p->maxcols)
			p->maxcols = 256;
		s = sz > p->maxcols * 2 ? sz : p->maxcols * 2;
		p->buf = realloc(p->buf, s);
		if (NULL == p->buf)
			err(1, "realloc");
		p->maxcols = s;
	}

	(void)memcpy(&p->buf[(int)p->col], c, sz);
	p->col += sz;
}


/*
 * Insert a single character into the line-buffer.  If the buffer's
 * space is exceeded, then allocate more space by doubling the buffer
 * size.
 */
static void
chara(struct termp *p, char c)
{
	size_t		 s;

	if (p->col + 1 >= p->maxcols) {
		if (0 == p->maxcols)
			p->maxcols = 256;
		s = p->maxcols * 2;
		p->buf = realloc(p->buf, s);
		if (NULL == p->buf)
			err(1, "realloc");
		p->maxcols = s;
	}
	p->buf[(int)(p->col)++] = c;
}


static void
sanity(const struct mdoc_node *n)
{

	switch (n->type) {
	case (MDOC_TEXT):
		if (n->child) 
			errx(1, "regular form violated (1)");
		if (NULL == n->parent) 
			errx(1, "regular form violated (2)");
		if (NULL == n->string)
			errx(1, "regular form violated (3)");
		switch (n->parent->type) {
		case (MDOC_TEXT):
			/* FALLTHROUGH */
		case (MDOC_ROOT):
			errx(1, "regular form violated (4)");
			/* NOTREACHED */
		default:
			break;
		}
		break;
	case (MDOC_ELEM):
		if (NULL == n->parent)
			errx(1, "regular form violated (5)");
		switch (n->parent->type) {
		case (MDOC_TAIL):
			/* FALLTHROUGH */
		case (MDOC_BODY):
			/* FALLTHROUGH */
		case (MDOC_HEAD):
			break;
		default:
			errx(1, "regular form violated (6)");
			/* NOTREACHED */
		}
		if (n->child) switch (n->child->type) {
		case (MDOC_TEXT):
			break;
		default:
			errx(1, "regular form violated (7(");
			/* NOTREACHED */
		}
		break;
	case (MDOC_HEAD):
		/* FALLTHROUGH */
	case (MDOC_BODY):
		/* FALLTHROUGH */
	case (MDOC_TAIL):
		if (NULL == n->parent)
			errx(1, "regular form violated (8)");
		if (MDOC_BLOCK != n->parent->type)
			errx(1, "regular form violated (9)");
		if (n->child) switch (n->child->type) {
		case (MDOC_BLOCK):
			/* FALLTHROUGH */
		case (MDOC_ELEM):
			/* FALLTHROUGH */
		case (MDOC_TEXT):
			break;
		default:
			errx(1, "regular form violated (a)");
			/* NOTREACHED */
		}
		break;
	case (MDOC_BLOCK):
		if (NULL == n->parent)
			errx(1, "regular form violated (b)");
		if (NULL == n->child)
			errx(1, "regular form violated (c)");
		switch (n->parent->type) {
		case (MDOC_ROOT):
			/* FALLTHROUGH */
		case (MDOC_HEAD):
			/* FALLTHROUGH */
		case (MDOC_BODY):
			/* FALLTHROUGH */
		case (MDOC_TAIL):
			break;
		default:
			errx(1, "regular form violated (d)");
			/* NOTREACHED */
		}
		switch (n->child->type) {
		case (MDOC_ROOT):
			/* FALLTHROUGH */
		case (MDOC_ELEM):
			errx(1, "regular form violated (e)");
			/* NOTREACHED */
		default:
			break;
		}
		break;
	case (MDOC_ROOT):
		if (n->parent)
			errx(1, "regular form violated (f)");
		if (NULL == n->child)
			errx(1, "regular form violated (10)");
		switch (n->child->type) {
		case (MDOC_BLOCK):
			break;
		default:
			errx(1, "regular form violated (11)");
			/* NOTREACHED */
		}
		break;
	}
}


static int
merr(void *arg, int line, int col, const char *msg)
{

	warnx("error: %s (line %d, column %d)", msg, line, col);
	return(0);
}


static int
mwarn(void *arg, int line, int col, 
		enum mdoc_warn type, const char *msg)
{
	int		 flags;
	char		*wtype;

	flags = *(int *)arg;
	wtype = NULL;

	switch (type) {
	case (WARN_COMPAT):
		wtype = "compat";
		if (flags & WARN_WCOMPAT)
			break;
		return(1);
	case (WARN_SYNTAX):
		wtype = "syntax";
		if (flags & WARN_WSYNTAX)
			break;
		return(1);
	}

	assert(wtype);
	warnx("%s warning: %s (line %d, column %d)", 
			wtype, msg, line, col);

	if ( ! (flags & WARN_WERR))
		return(1);

	warnx("%s: considering warnings as errors", 
			__progname);
	return(0);
}