httpenum-depth.c

harvest是一个下载html网页得机器人

static char rcsid[] =
    "$Id: httpenum-depth.c,v 2.8 2000/02/03 12:45:56 sxw Exp $";
/*
 *  httpenum-depth.c - Depth First RootNode URL enumerator for HTTP URLs
 *
 *  Usage: httpenum-depth 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.
 */

#include <stdio.h>
#include <string.h>
#include <signal.h>
#include <stdlib.h>
#include <gdbm.h>
#include "util.h"
#include "url.h"
#define PUBLIC extern
#include "filter.h"

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

/*
 *  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 process_url ();
static void usage ();
static void mark_retrieved ();
static void mark_failed ();
static void sigdie ();
static int url_in_db ();
static int md5_in_db ();
static list_t *http_enum ();

/* From robots-txt.c */
extern int RobotsTxtCheck _PARAMS ((URL *));

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

/*
 *  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 (up)
URL *up;
{
	int y;

	Debug (42, 1, ("Checking URL: %s\nHost: %s\n", up->url, up->host));

	if (url_in_db (up->url)) {	/* Have we been here? */
		Debug (42, 1, ("Already Visited URL: %s\n", up->url));
		return 0;
	}
	if (y = filter_selection (up)) {	/* Match the URL based on REs */
		Debug (42, 1, ("Removing Candidate: [%s] %s\n",
			       Filter_Type_Name[y], up->url));
		if (not_visited)
			fprintf (not_visited, "[FILTER] %s\n", up->url);
		return 0;
	}
	if (!visit_server (up)) {	/* Can we visit this server? */
		Debug (42, 1, ("Disallowed to Visit Server: %s\n", up->url));
		if (not_visited)
			fprintf (not_visited, "[SERVER] %s\n", up->url);
		return 0;
	}
	if (!RobotsTxtCheck (up)) {	/* Is it disallowed by the robots.txt file */
		Debug (42, 1, ("Disallowed by robots.txt file: %s\n", up->url));
		if (not_visited)
			fprintf (not_visited, "[ROBOTS.TXT] %s\n", up->url);
		return 0;
	}
	return 1;
}

/*
 *  http_enum() - Returns a linked list of all the URLs in this object,
 *  or NULL on error.  Checks for "text/html" and "text/x-soif" in MIME
 *  headers and then runs "HTMLurls" on the file or urldb_getrefs() on the URL.
 */
static list_t *
http_enum (up)
URL *up;
{
	list_t *head = NULL;
	list_t **Tail = NULL;
	list_t *l = NULL;
	FILE *fp = NULL;
	char *s = NULL;
	char *t0 = NULL;
	char *t1 = NULL;
	char *t2 = NULL;
	URL *tup = NULL;
	char buf[BUFSIZ];
	int pipefds[2];
	int pid;
	int err;
	int status;
	int count = 0;
	int nurls = 0;

	/* Check to see if we're allowed to visit the URL */
	if (!url_is_allowed (up))
		return (NULL);

	/*
	 *  Ack.  Check for symbolic link loops in server generated HTML listings
	 *  Do this by comparing the last two pathname components.  If they are
	 *  the same then guess its a loop.
	 */
	s = xstrdup (up->pathname);
	t0 = t1 = t2 = NULL;
	for (t0 = strtok (s, "/"); t0; t0 = strtok (NULL, "/")) {
		t2 = t1;
		t1 = t0;
	}
	if (t1 != NULL && t2 != NULL) {
		if (strcmp (t1, t2) == 0) {
			Debug (42, 1, ("Possible symlink loop: %s\n", up->url));
			xfree (s);
			s = NULL;
			return (NULL);
		}
	}
	xfree (s);
	s = NULL;

	/* Recurse over the redirect chain */
	while (((status = url_retrieve (up)) == -1)
	       && count < HTTP_MAX_REDIRECTS) {
		count++;
		if (!url_is_allowed (up))
			return (NULL);
	}

	if (status) {		/* Grab the URL; success? */
		Debug (42, 1, ("Cannot Retrieve URL: %s\n", up->url));