enum.c

harvest是一个下载html网页得机器人

static char rcsid[] = "$Id: enum.c,v 2.2 1997/08/27 18:06:28 sxw Exp $";
/*
 *  enum.c - Enumerate RootNode URLs into LeafNode URLs.  Also produce
 *  Timestamps or MD5s for each LeafNode URL to support incremental
 *  Gatherering (using the dbcheck program).
 *
 *  Input:
 *    URL Option1 ... Option N
 *
 *  Output:
 *    URL\tMD5:xxx
 *    URL\tLast-Modification-Time:xxx
 *
 *  Usage:  enum [(-db | -del | -log | -tmpdb) file] [-Ds,l] [-delete] < url_enum_list
 *
 *  DEBUG: section  40, level 1, 5, 9   Gatherer URL enumeration
 *  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 <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <unistd.h>
#include <memory.h>
#include <limits.h>
#include <fcntl.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <sys/errno.h>
#include <signal.h>
#include <gdbm.h>
#include "util.h"
#include "url.h"
#include "template.h"

/* Local variables */

/* Easily define how to enumerate with the various access methods */
#define Arg_URL         1	/* just give the url as the arg */
#define Arg_Host_File   2	/* 'host path' are the args */
#define Arg_Path        3	/* 'path' is the only arg */

struct enumerators {		/* table is currently ignored */
	int url_type;		/* type this enumerator handles */
	char *url_enum_breadth;	/* the pgm for breadth first searches */
	char *url_enum_depth;	/* the pgm for depth first searches */
	char *special;		/* special flags if any */
	char *stamp_kind;	/* proddb label for the stamp */
	int arg_types;		/* sigh, need until all conform */
} enums[] = {

	{
	URL_FILE, "fileenum", "fileenum", "", T_TIMESTAMP, Arg_URL}, {
	URL_FTP, "ftpenum", "ftpenum", "", T_TIMESTAMP, Arg_URL}, {
	URL_GOPHER, "gopherenum-breadth", "gopherenum-depth", "", T_MD5,
		    Arg_URL}, {
	URL_HTTP, "httpenum-breadth", "httpenum-depth", "", T_MD5, Arg_URL},
	{
	URL_NEWS, "newsenum", "newsenum", "", T_MD5, Arg_URL}
};

#define DBFILE  	"PRODUCTION.gdbm"	/* default proddb */
#define DBNEWFILE 	"tmpdb.gdbm"	/* tmp log of all urls enumed */
#define DELFILE 	"url.del"	/* Stale URLs in proddb */
#define LOGFILE   	"enum.log"	/* default debug logfile */

static char *tree_root = NULL;
static char *addfile = NULL;
static char *delfile = DELFILE;
static char *proddbfile = DBFILE;
static char *logfile = LOGFILE;
static char *tmpdbfile = DBNEWFILE;
static char *cur_attr;		/* set to md5 or time's label */
static char *root;		/* set to root url of enum space */
static int find_deleted = 0;	/* find any deleted objects */

static char url[BUFSIZ];	/* URL */
static int url_max;		/* Option 1 */
static char url_filter[BUFSIZ];	/* Option 2 */
static int host_max;		/* Option 3 */
static char host_filter[BUFSIZ];	/* Option 4 */
static int delay;		/* Option 5 */
static int max_depth;		/* Option 6 */
static int cur_depth;		/* Option 6 */
static char access_types[64];	/* Option 7 */
static char user_enum_pgm[BUFSIZ];	/* Option 8 */
static char search_type[BUFSIZ];	/* Option 9 */

FILE *fadd;			/* the add filehandles (stdout) */
FILE *fdel;			/* the deleted urls */
FILE *flog;			/* log output */

GDBM_FILE proddb;		/* The Gatherer's production database */
GDBM_FILE newdb;		/* A Temporary db to store timestamps */

typedef struct _nodespec {
	char *buf;
	struct _nodespec *next;
} nodespec;

static nodespec *ns_head = 0, *ns_tail = 0;

typedef struct _cpi {
	int childpid;
	char *result_file;
	FILE *wfp;
} cpi;

/* Local functions */
static int do_an_enumeration ();
static int get_data ();
static int do_kid ();
static void usage ();
static void init ();
static void finish ();
static void process_url_stamp ();
static void do_add ();
static void add_to_newdb ();
static void add_to_todo ();
static void enum_to_get_deleted ();
static void set_envs ();
static cpi *create_user_enum_pgm_pipes ();

/*
 *  init() - Initializes the overall Enumeration process
 */
static void
init (argc, argv)
int argc;
char *argv[];
{
	debug_init ();
	for (argc--, argv++; argc > 0 && **argv == '-'; argc--, argv++) {
		if (strncmp (*argv, "-D", 2) == 0) {
			debug_flag (*argv);
		} else if (strcmp (*argv, "-delete") == 0) {
			find_deleted = 1;
		} else if (strcmp (*argv, "-db") == 0) {
			if (--argc < 1)
				usage ();
			proddbfile = xstrdup (*++argv);
		} else if (strcmp (*argv, "-add") == 0) {
			if (--argc < 1)
				usage ();
			addfile = xstrdup (*++argv);
		} else if (strcmp (*argv, "-tmpdb") == 0) {
			if (--argc < 1)
				usage ();
			tmpdbfile = xstrdup (*++argv);
		} else if (strcmp (*argv, "-del") == 0) {
			if (--argc < 1)
				usage ();
			delfile = xstrdup (*++argv);
		} else if (strcmp (*argv, "-log") == 0) {
			if (--argc < 1)
				usage ();
			logfile = xstrdup (*++argv);
		} else {
			usage ();
		}
	}

	if ((flog = fopen (logfile, "a+")) == NULL) {
		if (getenv ("HARVEST_GATHERER_LOGFILE") != (char *) NULL)
			flog =
			    fopen (getenv ("HARVEST_GATHERER_LOGFILE"), "a+");
		if (flog == (FILE *) NULL)
			flog = stderr;
	}
	init_log3 ("enum", flog, stderr);

	/*
	 *  open the main proddb and output files; If proddb file is not found,
	 *  we assume a new proddb
	 */
	Debug (40, 9, ("using: %s: for production db\n", proddbfile));
	proddb = gdbm_open (proddbfile, 0, GDBM_READER, 0644, 0);

	fadd = stdout;
	if (addfile != NULL) {
		Debug (40, 9, ("using: %s: for addfile\n", addfile));
		if ((fadd = fopen (addfile, "w")) == NULL) {
			log_errno (addfile);
			fatal ("Internal enum error 2.\n");
		}
	}
	fdel = NULL;
	if (find_deleted) {
		Debug (40, 9, ("using: %s: for delfile\n", delfile));
		if ((fdel = fopen (delfile, "w")) == NULL) {
			log_errno (delfile);
			fatal ("Internal enum error 3.\n");
		}
	}
}

static void
finish ()
{
	if (proddb != NULL)
		gdbm_close (proddb);
	if (fdel != NULL)
		fclose (fdel);
	fclose (fadd);
	fclose (flog);
}

/*
 *  process_url_stamp() - Check in the database for the URL name with
 *  timestamp/md5 stamp.  If the URL is registered in the database
 *  and the timestamps are the same, then ignore.  If they're different
 *  then add the name to the todo list.
 */
static void
process_url_stamp (name, stamp)
char *name, *stamp;
{
	datum key, d;
	Template *template;
	AVPair *avp;

	key.dptr = name;
	key.dsize = strlen (name);

	d = gdbm_fetch (proddb, key);
	if (d.dptr != NULL) {
		init_parse_template_string (d.dptr, d.dsize);
		if ((template = parse_template ()) == NULL)
			fatal ("Cannot parse template with %s\n", key.dptr);
		finish_parse_template ();
		free (d.dptr);

		avp = extract_AVPair (template->list, cur_attr);
		if ((avp != NULL) && !strcmp (stamp, avp->value)) {
			free_template (template);
			return;
		}
		free_template (template);
	}
	do_add (name, stamp);
}

/*
 *  do_add() - save the URL/timestamp/md5 data in the to-add pile
 */
static void
do_add (name, stamp)
char *name;			/* the url to 'add' */
char *stamp;			/* the md5/timestamp from enum */
{
	Debug (40, 9, ("do_add: Saving %s, %s\n", name, stamp));
	fprintf (fadd, "%s\t%s:%s\n", name, cur_attr, stamp);
	fflush (fadd);		/* MUST flush */
}

static void
add_to_newdb (name, timestamp)	/* save to my proddb */
char *name, *timestamp;
{
	datum k;		/* the key  */
	datum d;		/* the data */

	k.dptr = name;		/* the key is url */
	k.dsize = strlen (name) + 1;	/* save the null too */

	d.dptr = timestamp;
	d.dsize = strlen (timestamp) + 1;

	Debug (40, 9, ("add_to_newdb: Marking %s\n", name));
	if (gdbm_store (newdb, k, d, GDBM_INSERT))
		Log ("Warning: Ignoring re-visited URL: %s\n", name);
}

static void
enum_to_get_deleted ()
{				/* get urls to delete */
	datum k, nk;		/* current key in both proddb's */
	datum d;		/* data (if any) in new proddb */
	int rootlen;

	rootlen = strlen (root);	/* don't include null here */

	k = gdbm_firstkey (proddb);	/* get first key in ess proddb */
	while (k.dptr != NULL) {	/* while there are keys */
		if (strncmp (k.dptr, root, rootlen) == 0) {
			/* if in root url */
			d = gdbm_fetch (newdb, k);
			/* see if I've seen this url */
			if (d.dptr == NULL && fdel != NULL)
				/* if not, then it's old! */
				fprintf (fdel, "%s\n", k.dptr);
			if (d.dptr != NULL)
				free (d.dptr);
		}
		nk = gdbm_nextkey (proddb, k);
		free (k.dptr);
		k = nk;
	}
}

static cpi *
create_user_enum_pgm_pipes (enum_arg)
char *enum_arg;			/* the URL from which we are doing this enumeration */
{
	int pfd[2], pid, fd;
	char *argv[64], buf[BUFSIZ], *tmpfile;
	static cpi *cpinfo;	/* need static to return var */

	tmpfile = xstrdup (tempnam (NULL, "uep"));
	Debug (40, 1, ("RUNNING enumerator: %s %s > %s\n", user_enum_pgm,
		       enum_arg, tmpfile));

	cpinfo = NULL;
	if (pipe (pfd) < 0) {
		fatal_errno ("pipe");
	}
	/* need to use fork() since parse_argv() will cause mem leak */
	if ((pid = fork ()) < 0) {
		fatal_errno ("fork");
	}
	if (pid == 0) {
		/* child */
		if ((fd = creat (tmpfile, 0664)) < 0) {
			fatal_errno (tmpfile);
		}