main.c

harvest是一个下载html网页得机器人

	break;

      case 'n': /* --basic for completeness */
	config->authtype = CURLAUTH_BASIC;
	break;

      case 'o': /* --anyauth, let libcurl pick it */
	config->authtype = CURLAUTH_ANY;
	break;

#ifdef __DJGPP__
      case 'p': /* --wdebug */
        dbug_init();
        break;
#endif
      case 'q': /* --ftp-create-dirs */
        config->ftp_create_dirs ^= TRUE;
        break;

      case 'r': /* --create-dirs */
        config->create_dirs = TRUE;
        break;

      case 's': /* --max-redirs */
        /* specified max no of redirects (http(s)) */
        if(str2num(&config->maxredirs, nextarg))
          return PARAM_BAD_NUMERIC;
        break;

      case 't': /* --proxy-ntlm */
        config->proxyntlm ^= TRUE;
        break;

      case 'u': /* --crlf */
        /* LF -> CRLF conversinon? */
        config->crlf = TRUE;
        break;

      case 'v': /* --stderr */
        if(strcmp(nextarg, "-")) {
          config->errors = fopen(nextarg, "wt");
          config->errors_fopened = TRUE;
        }
        else
          config->errors = stdout;
      break;
      case 'w': /* --interface */
        /* interface */
        GetStr(&config->iface, nextarg);
        break;
      case 'x': /* --krb4 */
        /* krb4 level string */
        GetStr(&config->krb4level, nextarg);
        break;
      case 'y': /* --max-filesize */
        if(str2num(&config->max_filesize, nextarg))
          return PARAM_BAD_NUMERIC;
        break;
      case 'z': /* --disable-eprt */
        config->disable_eprt ^= TRUE;
        break;

      default: /* the URL! */
        {
          struct getout *url;
          if(config->url_get || (config->url_get=config->url_list)) {
            /* there's a node here, if it already is filled-in continue to find
               an "empty" node */
            while(config->url_get && (config->url_get->flags&GETOUT_URL))
              config->url_get = config->url_get->next;
          }

          /* now there might or might not be an available node to fill in! */

          if(config->url_get)
            /* existing node */
            url = config->url_get;
          else
            /* there was no free node, create one! */
            url=new_getout(config);
          
          if(url) {
            /* fill in the URL */
            GetStr(&url->url, nextarg);
            url->flags |= GETOUT_URL;
          }
        }
      }
      break;
    case '#': /* added 19990617 larsa */
      config->progressmode ^= CURL_PROGRESS_BAR;
      break;
    case '0': 
      /* HTTP version 1.0 */
      config->httpversion = CURL_HTTP_VERSION_1_0;
      break;
    case '1':
      /* TLS version 1 */
      config->ssl_version = CURL_SSLVERSION_TLSv1;
      break;
    case '2': 
      /* SSL version 2 */
      config->ssl_version = CURL_SSLVERSION_SSLv2;
      break;
    case '3': 
      /* SSL version 3 */
      config->ssl_version = CURL_SSLVERSION_SSLv3;
      break;
    case '4': 
      /* IPv4 */
      config->ip_version = 4;
      break;
    case '6': 
      /* IPv6 */
      config->ip_version = 6;
      break;
    case 'a':
      /* This makes the FTP sessions use APPE instead of STOR */
      config->conf ^= CONF_FTPAPPEND;
      break;
    case 'A':
      /* This specifies the User-Agent name */
      GetStr(&config->useragent, nextarg);
      break;
    case 'b': /* cookie string coming up: */
      if(nextarg[0] == '@') {
        nextarg++;
      }
      else if(strchr(nextarg, '=')) {
        /* A cookie string must have a =-letter */
        GetStr(&config->cookie, nextarg);
        break;
      }
      /* We have a cookie file to read from! */
      GetStr(&config->cookiefile, nextarg);
      break;
    case 'B':
      /* use ASCII/text when transfering */
      config->conf ^= CONF_GETTEXT;
      break;
    case 'c':
      /* get the file name to dump all cookies in */
      GetStr(&config->cookiejar, nextarg);
      break;
    case 'C':
      /* This makes us continue an ftp transfer at given position */
      if(!curl_strequal(nextarg, "-")) {
        if(str2num(&config->resume_from, nextarg))
          return PARAM_BAD_NUMERIC;
        config->resume_from_current = FALSE;
      }
      else {
        config->resume_from_current = TRUE;
        config->resume_from = 0;
      }
      config->use_resume=TRUE;
      break;
    case 'd':
      /* postfield data */
      {
        char *postdata=NULL;

        if('@' == *nextarg) {
          /* the data begins with a '@' letter, it means that a file name
             or - (stdin) follows */
          FILE *file;

          nextarg++; /* pass the @ */

          if(curl_strequal("-", nextarg))
            file = stdin;
          else 
            file = fopen(nextarg, "rb");

          if(subletter == 'b') /* forced binary */
            postdata = file2memory(file, &config->postfieldsize);
          else
            postdata = file2string(file);
          if(file && (file != stdin))
            fclose(file);
        }
        else {
          GetStr(&postdata, nextarg);
        }

        if(config->postfields) {
          /* we already have a string, we append this one
             with a separating &-letter */
          char *oldpost=config->postfields;
          config->postfields=aprintf("%s&%s", oldpost, postdata);
          free(oldpost);
          free(postdata);
        }
        else
          config->postfields=postdata;
      }
      /*
        We can't set the request type here, as this data might be used in
        a simple GET if -G is used. Already or soon.

        if(SetHTTPrequest(HTTPREQ_SIMPLEPOST, &config->httpreq))
          return PARAM_BAD_USE;
      */
      break;
    case 'D':
      /* dump-header to given file name */
      GetStr(&config->headerfile, nextarg);
      break;
    case 'e':
      {
        char *ptr = strstr(nextarg, ";auto");
        if(ptr) {
          /* Automatic referer requested, this may be combined with a
             set initial one */
          config->conf |= CONF_AUTO_REFERER;
          *ptr = 0; /* zero terminate here */
        }
        GetStr(&config->referer, nextarg);
      }
      break;
    case 'E':
      switch(subletter) {
      case 'a': /* CA info PEM file */
        /* CA info PEM file */
        GetStr(&config->cacert, nextarg);
        break;
      case 'b': /* cert file type */
        GetStr(&config->cert_type, nextarg);
        break;
      case 'c': /* private key file */
        GetStr(&config->key, nextarg);
        break;
      case 'd': /* private key file type */
        GetStr(&config->key_type, nextarg);
        break;
      case 'e': /* private key passphrase */
        GetStr(&config->key_passwd, nextarg);
        cleanarg(nextarg);
        break;
      case 'f': /* crypto engine */
        GetStr(&config->engine, nextarg);
        break;
      case 'g': /* CA info PEM file */
        /* CA cert directory */
        GetStr(&config->capath, nextarg);
        break;
      default: /* certificate file */
        {
          char *ptr = strchr(nextarg, ':');
          /* Since we live in a world of weirdness and confusion, the win32
             dudes can use : when using drive letters and thus
             c:\file:password needs to work. In order not to break
             compatibility, we still use : as separator, but we try to detect
             when it is used for a file name! On windows. */
#ifdef WIN32
          if(ptr &&
             (ptr == &nextarg[1]) &&
             (nextarg[2] == '\\') &&
             (isalpha((int)nextarg[0])) )
             /* colon in the second column, followed by a backslash, and the
                first character is an alphabetic letter:

                this is a drive letter colon */
            ptr = strchr(&nextarg[3], ':'); /* find the next one instead */
#endif
          if(ptr) {
            /* we have a password too */
            *ptr=0;
            ptr++;
            GetStr(&config->key_passwd, ptr);
          }
          GetStr(&config->cert, nextarg);
          cleanarg(nextarg);
        }
      }
      break;
    case 'f':
      /* fail hard on errors  */
      config->conf ^= CONF_FAILONERROR;
      break;
    case 'F':
      /* "form data" simulation, this is a little advanced so lets do our best
	 to sort this out slowly and carefully */
      if(formparse(nextarg,
                   &config->httppost,
                   &config->last_post))
	return PARAM_BAD_USE;
      if(SetHTTPrequest(HTTPREQ_POST, &config->httpreq))
        return PARAM_BAD_USE;
      break;

    case 'g': /* g disables URLglobbing */
      config->globoff ^= TRUE;
      break;

    case 'G': /* HTTP GET */
      config->use_httpget = TRUE;
      break;

    case 'h': /* h for help */
      help();
      return PARAM_HELP_REQUESTED;
    case 'H':
      /* A custom header to append to a list */
      config->headers = curl_slist_append(config->headers, nextarg);
      break;
    case 'i':
      config->conf ^= CONF_HEADER; /* include the HTTP header as well */
      break;
    case 'j':
      config->cookiesession ^= TRUE;
      break;
    case 'I':
      /*
       * This is a bit tricky. We either SET both bits, or we clear both
       * bits. Let's not make any other outcomes from this.
       */
      if((CONF_HEADER|CONF_NOBODY) !=
         (config->conf&(CONF_HEADER|CONF_NOBODY)) ) {
        /* one of them weren't set, set both */
        config->conf |= (CONF_HEADER|CONF_NOBODY);
        if(SetHTTPrequest(HTTPREQ_HEAD, &config->httpreq))
          return PARAM_BAD_USE;
      }
      else {
        /* both were set, clear both */
        config->conf &= ~(CONF_HEADER|CONF_NOBODY);
        if(SetHTTPrequest(HTTPREQ_GET, &config->httpreq))
          return PARAM_BAD_USE;
      }
      break;
    case 'k': /* allow insecure SSL connects */
      config->insecure_ok ^= TRUE;
      break;
    case 'K': /* parse config file */
      res = parseconfig(nextarg, config);
      config->configread = TRUE;
      if(res)
	return res;
      break;
    case 'l':
      config->conf ^= CONF_FTPLISTONLY; /* only list the names of the FTP dir */
      break;
    case 'L':
      config->conf ^= CONF_FOLLOWLOCATION; /* Follow Location: HTTP headers */
      switch (subletter) {
      case 't':
        /* Continue to send authentication (user+password) when following
         * locations, even when hostname changed */
        config->conf ^= CONF_UNRESTRICTED_AUTH;
        break;
      }
      break;
    case 'm':
      /* specified max time */
      if(str2num(&config->timeout, nextarg))
        return PARAM_BAD_NUMERIC;
      break;
    case 'M': /* M for manual, huge help */
      hugehelp();
      return PARAM_HELP_REQUESTED;
    case 'n':
      switch(subletter) {
      case 'o': /* CA info PEM file */
        /* use .netrc or URL */
        config->conf ^= CONF_NETRC_OPT;
        break;
      default:
        /* pick info from .netrc, if this is used for http, curl will
           automatically enfore user+password with the request */
        config->conf ^= CONF_NETRC;
        break;
      }
      break;
    case 'N':
      /* disable the output I/O buffering */
      config->nobuffer ^= 1;
      break;
    case 'o':
    case 'O':
      /* output file */
      {
        struct getout *url;
        if(config->url_out || (config->url_out=config->url_list)) {
          /* there's a node here, if it already is filled-in continue to find
             an "empty" node */
          while(config->url_out && (config->url_out->flags&GETOUT_OUTFILE))
            config->url_out = config->url_out->next;
        }

        /* now there might or might not be an available node to fill in! */

        if(config->url_out)
          /* existing node */
          url = config->url_out;
        else
          /* there was no free node, create one! */
          url=new_getout(config);

        if(url) {
          /* fill in the outfile */
          if('o' == letter)
            GetStr(&url->outfile, nextarg);
          else {
            url->outfile=NULL; /* leave it */
            url->flags |= GETOUT_USEREMOTE;
          }
          url->flags |= GETOUT_OUTFILE;
        }
      }
      break;
    case 'P':
      /* This makes the FTP sessions use PORT instead of PASV */
      /* use <eth0> or <192.168.10.10> style addresses. Anything except
	 this will make us try to get the "default" address.
	 NOTE: this is a changed behaviour since the released 4.1!
	 */
      GetStr(&config->ftpport, nextarg);
      break;
    case 'p':
      /* proxy tunnel for non-http protocols */
      config->proxytunnel ^= TRUE;
      break;

    case 'q': /* if used first, already taken care of, we do it like
		 this so we don't cause an error! */
      break;
    case 'Q':
      /* QUOTE command to send to FTP server */
      switch(nextarg[0]) {
      case '-':
        /* prefixed with a dash makes it a POST TRANSFER one */
        nextarg++;
        config->postquote = curl_slist_append(config->postquote, nextarg);
        break;
      case '+':
        /* prefixed with a plus makes it a just-before-transfer one */
        nextarg++;
        config->prequote = curl_slist_append(config->prequote, nextarg);
        break;
      default:
        config->quote = curl_slist_append(config->quote, nextarg);
      }
      break;
    case 'r':
      /* byte range requested */
      GetStr(&config->range, nextarg);
      break;
    case 'R':
      /* use remote file's time */
      config->remote_time ^= TRUE;
      break;
    case 's':
      /* don't show progress meter, don't show errors : */
      config->conf |= (CONF_MUTE|CONF_NOPROGRESS);
      config->showerror ^= TRUE; /* toggle off */
      break;
    case 'S':