xref: /src/distrib/utils/more/linenum.c

/*	$NetBSD: linenum.c,v 1.5 2003/10/13 14:34:25 agc Exp $	*/

/*
 * Copyright (c) 1988 Mark Nudelman
 * Copyright (c) 1988, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

#include <sys/cdefs.h>
#ifndef lint
#if 0
static char sccsid[] = "@(#)linenum.c	8.1 (Berkeley) 6/6/93";
#else
__RCSID("$NetBSD: linenum.c,v 1.5 2003/10/13 14:34:25 agc Exp $");
#endif
#endif /* not lint */

/*
 * Code to handle displaying line numbers.
 *
 * Finding the line number of a given file position is rather tricky.
 * We don't want to just start at the beginning of the file and
 * count newlines, because that is slow for large files (and also
 * wouldn't work if we couldn't get to the start of the file; e.g.
 * if input is a long pipe).
 *
 * So we use the function add_lnum to cache line numbers.
 * We try to be very clever and keep only the more interesting
 * line numbers when we run out of space in our table.  A line
 * number is more interesting than another when it is far from
 * other line numbers.   For example, we'd rather keep lines
 * 100,200,300 than 100,101,300.  200 is more interesting than
 * 101 because 101 can be derived very cheaply from 100, while
 * 200 is more expensive to derive from 100.
 *
 * The function currline() returns the line number of a given
 * position in the file.  As a side effect, it calls add_lnum
 * to cache the line number.  Therefore currline is occasionally
 * called to make sure we cache line numbers often enough.
 */

#include <sys/types.h>
#include <stdio.h>
#include <time.h>

#include "less.h"
#include "extern.h"

/*
 * Structure to keep track of a line number and the associated file position.
 * A doubly-linked circular list of line numbers is kept ordered by line number.
 */
struct linenum
{
	struct linenum *next;		/* Link to next in the list */
	struct linenum *prev;		/* Line to previous in the list */
	off_t pos;			/* File position */
	off_t gap;			/* Gap between prev and next */
	int line;			/* Line number */
};
/*
 * "gap" needs some explanation: the gap of any particular line number
 * is the distance between the previous one and the next one in the list.
 * ("Distance" means difference in file position.)  In other words, the
 * gap of a line number is the gap which would be introduced if this
 * line number were deleted.  It is used to decide which one to replace
 * when we have a new one to insert and the table is full.
 */

#define	NPOOL	50			/* Size of line number pool */

#define	LONGTIME	(2)		/* In seconds */

int lnloop = 0;				/* Are we in the line num loop? */

static struct linenum anchor;		/* Anchor of the list */
static struct linenum *freelist;	/* Anchor of the unused entries */
static struct linenum pool[NPOOL];	/* The pool itself */
static struct linenum *spare;		/* We always keep one spare entry */

static void calcgap __P((struct linenum *));
static void longloopmessage __P((void));
/*
 * Initialize the line number structures.
 */
void
clr_linenum()
{
	struct linenum *p;

	/*
	 * Put all the entries on the free list.
	 * Leave one for the "spare".
	 */
	for (p = pool;  p < &pool[NPOOL-2];  p++)
		p->next = p+1;
	pool[NPOOL-2].next = NULL;
	freelist = pool;

	spare = &pool[NPOOL-1];

	/*
	 * Initialize the anchor.
	 */
	anchor.next = anchor.prev = &anchor;
	anchor.gap = 0;
	anchor.pos = (off_t)0;
	anchor.line = 1;
}

/*
 * Calculate the gap for an entry.
 */
static void
calcgap(p)
	struct linenum *p;
{
	/*
	 * Don't bother to compute a gap for the anchor.
	 * Also don't compute a gap for the last one in the list.
	 * The gap for that last one should be considered infinite,
	 * but we never look at it anyway.
	 */
	if (p == &anchor || p->next == &anchor)
		return;
	p->gap = p->next->pos - p->prev->pos;
}

/*
 * Add a new line number to the cache.
 * The specified position (pos) should be the file position of the
 * FIRST character in the specified line.
 */
void
add_lnum(line, pos)
	int line;
	off_t pos;
{
	struct linenum *p;
	struct linenum *new;
	struct linenum *nextp;
	struct linenum *prevp;
	off_t mingap;

	/*
	 * Find the proper place in the list for the new one.
	 * The entries are sorted by position.
	 */
	for (p = anchor.next;  p != &anchor && p->pos < pos;  p = p->next)
		if (p->line == line)
			/* We already have this one. */
			return;
	nextp = p;
	prevp = p->prev;

	if (freelist != NULL)
	{
		/*
		 * We still have free (unused) entries.
		 * Use one of them.
		 */
		new = freelist;
		freelist = freelist->next;
	} else
	{
		/*
		 * No free entries.
		 * Use the "spare" entry.
		 */
		new = spare;
		spare = NULL;
	}

	/*
	 * Fill in the fields of the new entry,
	 * and insert it into the proper place in the list.
	 */
	new->next = nextp;
	new->prev = prevp;
	new->pos = pos;
	new->line = line;

	nextp->prev = new;
	prevp->next = new;

	/*
	 * Recalculate gaps for the new entry and the neighboring entries.
	 */
	calcgap(new);
	calcgap(nextp);
	calcgap(prevp);

	if (spare == NULL)
	{
		/*
		 * We have used the spare entry.
		 * Scan the list to find the one with the smallest
		 * gap, take it out and make it the spare.
		 * We should never remove the last one, so stop when
		 * we get to p->next == &anchor.  This also avoids
		 * looking at the gap of the last one, which is
		 * not computed by calcgap.
		 */
		mingap = anchor.next->gap;
		for (p = anchor.next;  p->next != &anchor;  p = p->next)
		{
			if (p->gap <= mingap)
			{
				spare = p;
				mingap = p->gap;
			}
		}
		spare->next->prev = spare->prev;
		spare->prev->next = spare->next;
	}
}

/*
 * If we get stuck in a long loop trying to figure out the
 * line number, print a message to tell the user what we're doing.
 */
static void
longloopmessage()
{
	ierror("Calculating line numbers");
	/*
	 * Set the lnloop flag here, so if the user interrupts while
	 * we are calculating line numbers, the signal handler will
	 * turn off line numbers (linenums=0).
	 */
	lnloop = 1;
}

/*
 * Find the line number associated with a given position.
 * Return 0 if we can't figure it out.
 */
int
find_linenum(pos)
	off_t pos;
{
	struct linenum *p;
	int lno;
	int loopcount;
	off_t cpos;
	time_t startime;

	if (!linenums)
		/*
		 * We're not using line numbers.
		 */
		return (0);
	if (pos == NULL_POSITION)
		/*
		 * Caller doesn't know what he's talking about.
		 */
		return (0);
	if (pos == (off_t)0)
		/*
		 * Beginning of file is always line number 1.
		 */
		return (1);

	/*
	 * Find the entry nearest to the position we want.
	 */
	for (p = anchor.next;  p != &anchor && p->pos < pos;  p = p->next)
		continue;
	if (p->pos == pos)
		/* Found it exactly. */
		return (p->line);

	/*
	 * This is the (possibly) time-consuming part.
	 * We start at the line we just found and start
	 * reading the file forward or backward till we
	 * get to the place we want.
	 *
	 * First decide whether we should go forward from the
	 * previous one or backwards from the next one.
	 * The decision is based on which way involves
	 * traversing fewer bytes in the file.
	 */
	flush();
	(void)time(&startime);
	if (p == &anchor || pos - p->prev->pos < p->pos - pos)
	{
		/*
		 * Go forward.
		 */
		p = p->prev;
		if (ch_seek(p->pos))
			return (0);
		loopcount = 0;
		for (lno = p->line, cpos = p->pos;  cpos < pos;  lno++)
		{
			/*
			 * Allow a signal to abort this loop.
			 */
			cpos = forw_raw_line(cpos);
			if (sigs || cpos == NULL_POSITION)
				return (0);
			if (loopcount >= 0 && ++loopcount > 100) {
				loopcount = 0;
				if (time((time_t *)NULL)
				    >= startime + LONGTIME) {
					longloopmessage();
					loopcount = -1;
				}
			}
		}
		lnloop = 0;
		/*
		 * If the given position is not at the start of a line,
		 * make sure we return the correct line number.
		 */
		if (cpos > pos)
			lno--;
	} else
	{
		/*
		 * Go backward.
		 */
		if (ch_seek(p->pos))
			return (0);
		loopcount = 0;
		for (lno = p->line, cpos = p->pos;  cpos > pos;  lno--)
		{
			/*
			 * Allow a signal to abort this loop.
			 */
			cpos = back_raw_line(cpos);
			if (sigs || cpos == NULL_POSITION)
				return (0);
			if (loopcount >= 0 && ++loopcount > 100) {
				loopcount = 0;
				if (time((time_t *)NULL)
				    >= startime + LONGTIME) {
					longloopmessage();
					loopcount = -1;
				}
			}
		}
		lnloop = 0;
	}

	/*
	 * We might as well cache it.
	 */
	add_lnum(lno, cpos);
	return (lno);
}

/*
 * Return the line number of the "current" line.
 * The argument "where" tells which line is to be considered
 * the "current" line (e.g. TOP, BOTTOM, MIDDLE, etc).
 */
int
currline(where)
	int where;
{
	off_t pos;

	if ((pos = position(where)) == NULL_POSITION)
		pos = ch_length();
	return(find_linenum(pos));
}