gatherd.c

harvest是一个下载html网页得机器人

	}
	if (access(cffile, R_OK)) {
		Log("WARNING: Configuration file not readable: %s\n", cffile);
		cffile = NULL;
	}
	/* check to see if allzipped is up-to-date */
	if (indexfile && allzipped) {
		struct stat sba, sbi;

		stat(allzipped, &sba);
		stat(indexfile, &sbi);
		if (sba.st_mtime < sbi.st_mtime) {
			Log("WARNING: Cache file out-of-date: %s\n", allzipped);
			allzipped = NULL;
		}
	}
	if (cffile) {
		load_configuration(cffile);
	} else {
		allow_all = 1;
		allow_hosts[0] = deny_hosts[0] = NULL;
	}

	/* Check to see if the program is being run via inetd */
	if (strstr(pgm, "in.gatherd") != NULL) {
		write_pid();
		exit(serve_client(0));
	}
	/* Stand-alone daemon needs a port number */
	if (argc == 0) {
		(void) unlink(pidfile);
		usage();
	}
	master_slave_mode(atoi(*argv));
	exit(0);
}


/*
 *  master_slave_mode() - Enters a Master-Slave mode to process client
 *  request.  For each client connection, it spawns a process to
 *  handle the request.
 */
static void master_slave_mode(portnumber)
     int portnumber;
{
	struct sockaddr_in sin, cin;
	int s, cs, clen, pid, on = 1, i;
	FILE *logfp;
#ifdef USE_TIMEOUT
	fd_set R;
	struct timeval to;
	int x;
	struct stat sb;
#endif

	if (portnumber < 1) {
		fatal("Illegal port number: %d\n", portnumber);
	}
	disconnect();		/* be a friendly daemon */

	logfp = fopen(logfile, "a+");
	if (logfp != NULL) {
		setbuf(logfp, NULL);
		init_log(logfp, logfp);
	} else {
		init_log(NULL, NULL);	/* we have no choice */
	}

	memset(&sin, '\0', sizeof(sin));
	sin.sin_family = AF_INET;
	sin.sin_addr.s_addr = htonl(INADDR_ANY);
	sin.sin_port = htons(portnumber);
	if ((s = socket(PF_INET, SOCK_STREAM, 0)) < 0) {
		log_errno("socket");
		exit(1);
	}
	if ((setsockopt(s, SOL_SOCKET, SO_REUSEADDR, (char *) &on,
		    sizeof(on))) < 0) {
		log_errno("setsockopt (SO_REUSEADDR = on)");
		exit(1);
	}
	if (bind(s, (struct sockaddr *) &sin, sizeof(sin)) < 0) {
		log_errno("bind");
		(void) close(s);
		exit(1);
	}
	if (listen(s, 5) < 0) {
		log_errno("listen");
		exit(1);
	}
	clen = sizeof(cin);

	write_pid();

#ifdef NSIG
	for (i = 1; i < NSIG; i++) {
#else
	for (i = 1; i < _sys_nsig; i++) {
#endif
		if (i == SIGCHLD || i == SIGWINCH || i == SIGSEGV)
			continue;
		signal(i, sigcleanup);
	}

	Log("Ready to serve information...\n");

	while (1) {
#ifdef USE_TIMEOUT
		to.tv_sec = 10;	/* Lunix modifies 'to' on return */
		to.tv_usec = 0;
		FD_ZERO(&R);
		FD_SET(s, &R);
		x = select(s + 1, &R, 0, 0, &to);
		if (x < 0) {
			if (errno == EINTR)
				continue;
			log_errno("select");
			unlink(pidfile);
			exit(1);
		} else
			/*      If select times out, check for the existance of the     */
			/*      pidfile.  If gone, somebody wants us dead, so die.      */

		if (x == 0) {
			if (stat(pidfile, &sb) < 0)
				exit(1);
			continue;
		}
#endif
		if ((cs = accept(s, (struct sockaddr *) &cin, &clen)) < 0) {
			if (errno == EINTR)
				continue;
			log_errno("accept");
			unlink(pidfile);
			exit(1);
		}
		setsocket_linger(cs, 30);	/* 30 sec linger time */

		if ((pid = fork()) < 0) {
			log_errno("fork");
			unlink(pidfile);
			exit(1);
		}
		if (pid == 0) {	/* child */
			logfp = fopen(logfile, "a+");
			if (logfp != NULL) {
				setbuf(logfp, NULL);
				init_log(logfp, logfp);
			} else {
				init_log(NULL, NULL);	/* we have no choice */
			}
			close(s);
			exit(serve_client(cs));
		}
		close(cs);	/* parent */
		while (waitpid(-1, NULL, WNOHANG) > 0);		/* catch any zombies  */
	}

}

/* 
 *  disconnect() - Disconnect the process from the controlling terminal.
 *  Adapted from Harvest's floodd daemon in the replicator.
 */
static void disconnect()
{
	int pid, fd;

#ifdef SIGTTOU
	signal(SIGTTOU, SIG_IGN);
#endif
#ifdef SIGTTIN
	signal(SIGTTIN, SIG_IGN);
#endif
#ifdef SIGTSTP
	signal(SIGTSTP, SIG_IGN);
#endif

	if ((pid = fork()) < 0) {
		log_errno("fork");
		return;
	}
	if (pid) {		/* parent */
		exit(0);	/* quietly exit to let child do the work */
	}
	(void) setsid();
	/* Logging beneath here will not work */
	/* Close all file descriptors */
	close_all_fds(0);

	/* Redirect the stdin, stdout, and stderr to /dev/null */
	if ((fd = open("/dev/null", O_RDWR)) < 0) {
		exit(1);
	}
	(void) dup2(fd, 0);
	(void) dup2(fd, 1);
	(void) dup2(fd, 2);
}

static void load_configuration(filename)
     char *filename;
{
	FILE *fp;
	char buf[BUFSIZ], *s;
	int a = 0, d = 0;
	extern void Tolower();

	memset(allow_hosts, '\0', BUFSIZ * sizeof(char *));
	memset(deny_hosts, '\0', BUFSIZ * sizeof(char *));
	allow_all = deny_all = 0;

	if ((fp = fopen(filename, "r")) == NULL) {
		log_errno(filename);
		return;
	}
	while (fgets(buf, BUFSIZ, fp)) {
		if (buf[0] == '#' || buf[0] == '\n')
			continue;
		if (!strncmp(buf, "Allow", 5)) {
			s = strtok(buf, " \t\n");
			while ((s = strtok(NULL, " \t\n")) != NULL) {
				if (!strcmp(s, "all")) {
					allow_all = 1;
#ifdef DEBUG
					Log("Allowing access for all hosts\n");
#endif
					break;
				}
				allow_hosts[a] = strdup(s);
				Tolower(allow_hosts[a]);
#ifdef DEBUG
				Log("Allowing access for %s\n", allow_hosts[a]);
#endif
				a++;
			}
		} else if (!strncmp(buf, "Deny", 4)) {
			s = strtok(buf, " \t\n");
			while ((s = strtok(NULL, " \t\n")) != NULL) {
				if (!strcmp(s, "all")) {
					deny_all = 1;
#ifdef DEBUG
					Log("Denying access for all hosts\n");
#endif
					break;
				}
				deny_hosts[d] = strdup(s);
				Tolower(deny_hosts[d]);
#ifdef DEBUG
				Log("Denying access for %s\n", deny_hosts[d]);
#endif
				d++;
			}
		} else if (!strncmp(buf, "Gzip", 4)) {
			(void) strtok(buf, " \t\n");
			if ((s = strtok(NULL, " \t\n")) != NULL)
				cmd_gzip = xstrdup(s);
		}
	}
	fclose(fp);
	allow_hosts[a] = NULL;
	deny_hosts[d] = NULL;
	if (allow_all && deny_all) {
		errorlog("Illegal Configuration!\n");
		errorlog("Cannot have both \"Allow all\" and \"Deny all\"\n");
		exit(1);
	}
}

static void write_pid()
{
	FILE *fp;

	/* Write the pid to a file */
	if ((fp = fopen(pidfile, "w")) != NULL) {
		fprintf(fp, "%d\n", master_pid = getpid());
		fclose(fp);
	} else {
		log_errno(pidfile);
	}
}

int read_pid()
{
	FILE *fp;
	int pid = -1;

	if ((fp = fopen(pidfile, "r")) != NULL) {
		fscanf(fp, "%d\n", &pid);
		fclose(fp);
	} else {
		log_errno(pidfile);
	}
	return pid;
}