httpenum-breadth.c

harvest是一个下载html网页得机器人

static char rcsid[] =
    "$Id: httpenum-breadth.c,v 2.9 2000/02/03 12:45:56 sxw Exp $";
/*
 *  httpenum-breadth.c - Breadth first RootNode URL enumerator for HTTP URLs
 *
 *  Usage: httpenum-breadth http-URL
 *
 *  Outputs the following format:
 *
 *      URL of tree root
 *      URL <tab> md5
 *      ...
 *      URL <tab> md5
 *
 *  DEBUG: section  42, level 1, 5, 9   Gatherer enumeration for HTTP
 *  AUTHOR: Harvest derived
 *
 *  Harvest Indexer http://harvest.sourceforge.net/
 *  -----------------------------------------------
 *
 *  The Harvest Indexer is a continued development of code developed by
 *  the Harvest Project. Development is carried out by numerous individuals
 *  in the Internet community, and is not officially connected with the
 *  original Harvest Project or its funding sources.
 *
 *  Please mail [e-mail] if you are interested in participating
 *  in the development effort.
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, write to the Free Software
 *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */

/*  ----------------------------------------------------------------------
 *  Copyright (c) 1994, 1995.  All rights reserved.
 *
 *    The Harvest software was developed by the Internet Research Task
 *    Force Research Group on Resource Discovery (IRTF-RD):
 *
 *          Mic Bowman of Transarc Corporation.
 *          Peter Danzig of the University of Southern California.
 *          Darren R. Hardy of the University of Colorado at Boulder.
 *          Udi Manber of the University of Arizona.
 *          Michael F. Schwartz of the University of Colorado at Boulder.
 *          Duane Wessels of the University of Colorado at Boulder.
 *
 *    This copyright notice applies to software in the Harvest
 *    ``src/'' directory only.  Users should consult the individual
 *    copyright notices in the ``components/'' subdirectories for
 *    copyright information about other software bundled with the
 *    Harvest source code distribution.
 *
 *  TERMS OF USE
 *
 *    The Harvest software may be used and re-distributed without
 *    charge, provided that the software origin and research team are
 *    cited in any use of the system.  Most commonly this is
 *    accomplished by including a link to the Harvest Home Page
 *    (http://harvest.cs.colorado.edu/) from the query page of any
 *    Broker you deploy, as well as in the query result pages.  These
 *    links are generated automatically by the standard Broker
 *    software distribution.
 *
 *    The Harvest software is provided ``as is'', without express or
 *    implied warranty, and with no support nor obligation to assist
 *    in its use, correction, modification or enhancement.  We assume
 *    no liability with respect to the infringement of copyrights,
 *    trade secrets, or any patents, and are not responsible for
 *    consequential damages.  Proper use of the Harvest software is
 *    entirely the responsibility of the user.
 *
 *  DERIVATIVE WORKS
 *
 *    Users may make derivative works from the Harvest software, subject
 *    to the following constraints:
 *
 *      - You must include the above copyright notice and these
 *        accompanying paragraphs in all forms of derivative works,
 *        and any documentation and other materials related to such
 *        distribution and use acknowledge that the software was
 *        developed at the above institutions.
 *
 *      - You must notify IRTF-RD regarding your distribution of
 *        the derivative work.
 *
 *      - You must clearly notify users that your are distributing
 *        a modified version and not the original Harvest software.
 *
 *      - Any derivative product is also subject to these copyright
 *        and use restrictions.
 *
 *    Note that the Harvest software is NOT in the public domain.  We
 *    retain copyright, as specified above.
 *
 *  HISTORY OF FREE SOFTWARE STATUS
 *
 *    Originally we required sites to license the software in cases
 *    where they were going to build commercial products/services
 *    around Harvest.  In June 1995 we changed this policy.  We now
 *    allow people to use the core Harvest software (the code found in
 *    the Harvest ``src/'' directory) for free.  We made this change
 *    in the interest of encouraging the widest possible deployment of
 *    the technology.  The Harvest software is really a reference
 *    implementation of a set of protocols and formats, some of which
 *    we intend to standardize.  We encourage commercial
 *    re-implementations of code complying to this set of standards.
 *
 */

#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <memory.h>
#include <string.h>
#include <signal.h>
#include <gdbm.h>
#include "util.h"
#include "url.h"
#define PUBLIC extern
#include "filter.h"

typedef struct _list_t {
	void *ptr;
	int depth;
	struct _list_t *next;
} list_t;

list_t *head = NULL;
list_t **Tail = NULL;

/* List of candidate URL's */
char *filename_candidates;
int fd_candidates_read = -1, fd_candidates_write = -1;

/*
 *  define HOST_COUNT_IP to 'count' visited hosts based on IP, not the
 *  given hostname.  This way aliased machines will be properly
 *  enumerated
 *  If you define HOST_COUNT_IP then support for HTTP/1.1 virtual hosts
 *  and for hosts that have more than one IP address (for round-robin
 *  load balancing) may not work
 */
#undef HOST_COUNT_IP

/* Global variables */
int max_depth = 0;
int cur_depth = 0;
int depth_hist[100];

/* Local variables */
static int url_max = 0;
static int nurls = 0;
static int host_max = 0;
static int nhosts = 0;
static char *tree_root = NULL;
static char *urldb_filename = NULL;
static char *hostdb_filename = NULL;
static char *md5db_filename = NULL;
static GDBM_FILE urldbf = NULL;
static GDBM_FILE hostdbf = NULL;
static GDBM_FILE md5dbf = NULL;

static FILE *not_visited = NULL;

/* Local functions */
static void usage ();
static void mark_failed ();
static void mark_retrieved ();
static void sigdie ();
static int url_in_db ();
static int md5_in_db ();
static int http_enum ();

extern int RobotsTxtCheck _PARAMS ((URL *));

list_t *
free_from_list (l)
list_t *l;
{
	list_t *r = NULL;
	int len;

	if (filename_candidates) {
		if (fd_candidates_read == -1)
			fd_candidates_read =
			    open (filename_candidates, O_RDONLY);
		if (fd_candidates_read > -1) {
			r = (list_t *) xmalloc (sizeof (list_t));
			if (read
			    (fd_candidates_read, &(r->depth), sizeof (r->depth))
			    <= 0) {
				xfree (r);
				return NULL;
			}
			read (fd_candidates_read, &(len), sizeof (len));
			r->ptr = (void *) xmalloc (len + 1);
			read (fd_candidates_read, r->ptr, len);
			((char *) r->ptr)[len] = '\0';
		}
	} else
		r = l->next;

	if (l) {
		xfree (l->ptr);
		xfree (l);
	}
	return r;
}

void
add_to_list (url, depth)
char *url;
int depth;
{
	list_t *l = NULL;
	int len;

	if (filename_candidates) {
		if (fd_candidates_write == -1)
			fd_candidates_write = open (filename_candidates,
						    O_WRONLY | O_CREAT |
						    O_TRUNC, 0600);
		if (fd_candidates_write > -1) {
			len = strlen (url);
			write (fd_candidates_write, &depth, sizeof (depth));
			write (fd_candidates_write, &len, sizeof (len));
			write (fd_candidates_write, url, len);
		}
		if (!head)
			head = free_from_list (0);
		return;
	}

	/* No candidates file - keep list in memory. */
	l = (list_t *) xmalloc (sizeof (list_t));
	l->ptr = (void *) xstrdup (url);
	l->next = (list_t *) NULL;
	l->depth = depth;
	*Tail = l;
	Tail = &(l->next);
	return;
}

/* ---------------------------------------------------------------------- */

/*
 *  mark_failed() - Mark that a URL failed to be retrieved, so that the
 *  enumerator doesn't try it again. This option may not be wanted by
 *  some users and so should be configurable.
 */
static void
mark_failed (URL * up)
{
	datum k, d;

	Debug (42, 9, ("mark_failed: url='%s'", up->url));

	k.dptr = xstrdup (up->url);
	k.dsize = strlen (k.dptr) + 1;
	d.dptr = xstrdup ("FailedAccess");
	d.dsize = strlen (d.dptr) + 1;

	if (!gdbm_exists (urldbf, k) && gdbm_store (urldbf, k, d, GDBM_INSERT))
		fatal ("GDBM URLDB: %s: %s", k.dptr,
		       gdbm_strerror (gdbm_errno));

	xfree (k.dptr);
	xfree (d.dptr);
}

/*
 *  mark_retrieved() - Mark that the given URL was successfully retrieved,
 *  so that the URL is not retrieved again.  This prevents cycles in the
 *  enumeration.
 */
static void
mark_retrieved (up)
URL *up;
{
	datum k, d;

	Debug (42, 9,
	       ("mark_retrieved: url='%s', md5='%s'\n", up->url, up->md5));

	k.dptr = xstrdup (up->url);
	k.dsize = strlen (k.dptr) + 1;
	d.dptr = xstrdup (up->md5);
	d.dsize = strlen (d.dptr) + 1;

	if (!gdbm_exists (urldbf, k) && gdbm_store (urldbf, k, d, GDBM_INSERT))
		fatal ("GDBM URLDB: %s: %s", k.dptr,
		       gdbm_strerror (gdbm_errno));
	if (!gdbm_exists (md5dbf, d) && gdbm_store (md5dbf, d, k, GDBM_INSERT))
		fatal ("GDBM MD5DB: %s: %s", k.dptr,
		       gdbm_strerror (gdbm_errno));

	xfree (k.dptr);
	xfree (d.dptr);

	if (up->redir_from_url != (char *) NULL) {
		Debug (42, 9, ("mark_retrieved: url='%s', md5='%s'\n",
			       up->redir_from_url, up->md5));

		k.dptr = xstrdup (up->redir_from_url);
		k.dsize = strlen (k.dptr) + 1;
		d.dptr = xstrdup (up->md5);
		d.dsize = strlen (d.dptr) + 1;

		if (!gdbm_exists (urldbf, k)
		    && gdbm_store (urldbf, k, d, GDBM_INSERT))
			fatal ("GDBM URLDB: %s: %s", k.dptr,
			       gdbm_strerror (gdbm_errno));

		xfree (k.dptr);
		xfree (d.dptr);
	}

	if (nurls++ >= url_max) {
		Log ("Truncating RootNode %s at %d LeafNode URLs\n",
		     tree_root, url_max);
		url_close (up);
		up = NULL;
		sigdie (0);
	}
}

/*
 *  url_in_db() - check to see if the URL is in the database
 */
static int
url_in_db (url)
char *url;
{
	datum k;
	int r;

	Debug (42, 9, ("url_in_db: checking for url='%s'\n", url));

	k.dptr = xstrdup (url);
	k.dsize = strlen (k.dptr) + 1;
	r = gdbm_exists (urldbf, k);
	xfree (k.dptr);
	return (r);
}

/*
 *  md5_in_db() - check to see if the MD5 is in the database
 */
static int
md5_in_db (md5)
char *md5;
{
	datum k;
	int r;

	k.dptr = xstrdup (md5);
	k.dsize = strlen (k.dptr) + 1;
	r = gdbm_exists (md5dbf, k);
	xfree (k.dptr);
	return (r);
}

/*
 *  host_in_db() - check to see if the host is in the database
 */
static int
host_in_db (host)
char *host;
{
	datum k;
	int r;
#ifdef HOST_COUNT_IP
	Host *h;

	h = get_host (host);
	if (!h)
		return 0;
	k.dptr = xstrdup (h->dotaddr);
#else
	k.dptr = xstrdup (host);
#endif
	k.dsize = strlen (k.dptr) + 1;
	r = gdbm_exists (hostdbf, k);
	xfree (k.dptr);
	return (r);
}

/*
 *  visit_server() - Determine if we should visit the server.  Return
 *  zero if we should not process the URL; otherwise, return non-zero.
 */
static int
visit_server (up)
URL *up;
{
	datum k, d;
#ifdef HOST_COUNT_IP
	Host *h = NULL;
#endif

	if (host_in_db (up->host))	/* Host is already in the db */
		return (1);
	if (++nhosts > host_max)
		return (0);

#ifdef HOST_COUNT_IP
	h = get_host (up->host);
	if (!h)
		return (0);
	k.dptr = xstrdup (h->dotaddr);
#else
	k.dptr = xstrdup (up->host);
#endif
	k.dsize = strlen (k.dptr) + 1;
	d.dptr = xstrdup (up->url);
	d.dsize = strlen (d.dptr) + 1;

	if (gdbm_store (hostdbf, k, d, GDBM_INSERT))
		fatal ("GDBM HOSTDB: %s: %s", k.dptr,
		       gdbm_strerror (gdbm_errno));
	xfree (k.dptr);
	xfree (d.dptr);
	return (1);
}

int
url_is_allowed (url)
char *url;
{
	URL *tup = NULL;