curl/lib/transfer.c

/***************************************************************************
 *                                  _   _ ____  _     
 *  Project                     ___| | | |  _ \| |    
 *                             / __| | | | |_) | |    
 *                            | (__| |_| |  _ <| |___ 
 *                             \___|\___/|_| \_\_____|
 *
 * Copyright (C) 1998 - 2002, Daniel Stenberg, <daniel@haxx.se>, et al.
 *
 * This software is licensed as described in the file COPYING, which
 * you should have received as part of this distribution. The terms
 * are also available at http://curl.haxx.se/docs/copyright.html.
 * 
 * You may opt to use, copy, modify, merge, publish, distribute and/or sell
 * copies of the Software, and permit persons to whom the Software is
 * furnished to do so, under the terms of the COPYING file.
 *
 * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
 * KIND, either express or implied.
 *
 * $Id$
 ***************************************************************************/

#include "setup.h"

/* -- WIN32 approved -- */
#include <stdio.h>
#include <string.h>
#include <stdarg.h>
#include <stdlib.h>
#include <ctype.h>
#include <sys/types.h>
#include <sys/stat.h>

#include <errno.h>

#include "strequal.h"

#if defined(WIN32) && !defined(__GNUC__) || defined(__MINGW32__)
#include <winsock.h>
#include <time.h>
#include <io.h>
#else
#ifdef HAVE_SYS_SOCKET_H
#include <sys/socket.h>
#endif
#include <netinet/in.h>
#include <sys/time.h>
#include <sys/resource.h>
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif
#include <netdb.h>
#ifdef HAVE_ARPA_INET_H
#include <arpa/inet.h>
#endif
#ifdef HAVE_NET_IF_H
#include <net/if.h>
#endif
#include <sys/ioctl.h>
#include <signal.h>

#ifdef HAVE_SYS_PARAM_H
#include <sys/param.h>
#endif

#ifdef HAVE_SYS_SELECT_H
#include <sys/select.h>
#endif

#ifndef HAVE_SELECT
#error "We can't compile without select() support!"
#endif
#ifndef HAVE_SOCKET
#error "We can't compile without socket() support!"
#endif

#endif

#include "urldata.h"
#include <curl/curl.h>
#include <curl/types.h>
#include "netrc.h"

#include "content_encoding.h"   /* content encoding support. 08/27/02 jhrg */

#include "hostip.h"
#include "transfer.h"
#include "sendf.h"
#include "speedcheck.h"
#include "getpass.h"
#include "progress.h"
#include "getdate.h"
#include "http.h"
#include "url.h"
#include "getinfo.h"
#include "ssluse.h"

#define _MPRINTF_REPLACE /* use our functions only */
#include <curl/mprintf.h>

/* The last #include file should be: */
#ifdef MALLOCDEBUG
#include "memdebug.h"
#endif

#ifndef min
#define min(a, b)   ((a) < (b) ? (a) : (b))
#endif

enum {
  KEEP_NONE,
  KEEP_READ,
  KEEP_WRITE
};


/*
 * compareheader()
 *
 * Returns TRUE if 'headerline' contains the 'header' with given 'content'.
 * Pass headers WITH the colon.
 */
static bool
compareheader(char *headerline, /* line to check */
              const char *header,     /* header keyword _with_ colon */
              const char *content)    /* content string to find */
{
  /* RFC2616, section 4.2 says: "Each header field consists of a name followed
   * by a colon (":") and the field value. Field names are case-insensitive.
   * The field value MAY be preceded by any amount of LWS, though a single SP
   * is preferred." */

  size_t hlen = strlen(header);
  size_t clen;
  size_t len;
  char *start;
  char *end;

  if(!strnequal(headerline, header, hlen))
    return FALSE; /* doesn't start with header */

  /* pass the header */
  start = &headerline[hlen];

  /* pass all white spaces */
  while(*start && isspace((int)*start))
    start++;

  /* find the end of the header line */
  end = strchr(start, '\r'); /* lines end with CRLF */
  if(!end) {
    /* in case there's a non-standard compliant line here */
    end = strchr(start, '\n');

    if(!end)
      /* hm, there's no line ending here, return false and bail out! */
      return FALSE;
  }

  len = end-start; /* length of the content part of the input line */
  clen = strlen(content); /* length of the word to find */

  /* find the content string in the rest of the line */
  for(;len>=clen;len--, start++) {
    if(strnequal(start, content, clen))
      return TRUE; /* match! */
  }

  return FALSE; /* no match */
}

/* We keep this static and global since this is read-only and NEVER
   changed. It should just remain a blanked-out timeout value. */
static struct timeval notimeout={0,0};

CURLcode Curl_readwrite(struct connectdata *conn,
                        bool *done)
{
  struct Curl_transfer_keeper *k = &conn->keep;
  struct SessionHandle *data = conn->data;
  int result;
  ssize_t nread; /* number of bytes read */
  int didwhat=0;
  
  /* These two are used only if no other select() or _fdset() have been
     invoked before this. This typicly happens if you use the multi interface
     and call curl_multi_perform() without calling curl_multi_fdset()
     first. */
  fd_set extrareadfd;
  fd_set extrawritefd;

  fd_set *readfdp = k->readfdp;
  fd_set *writefdp = k->writefdp;
  
  if((k->keepon & KEEP_READ) && !readfdp) {
    /* reading is requested, but no socket descriptor pointer was set */
    FD_ZERO(&extrareadfd);
    FD_SET(conn->sockfd, &extrareadfd);
    readfdp = &extrareadfd;

    /* no write, no exceptions, no timeout */
    select(conn->sockfd+1, readfdp, NULL, NULL, &notimeout);
  }
  if((k->keepon & KEEP_WRITE) && !writefdp) {
    /* writing is requested, but no socket descriptor pointer was set */
    FD_ZERO(&extrawritefd);
    FD_SET(conn->writesockfd, &extrawritefd);
    writefdp = &extrawritefd;

    /* no read, no exceptions, no timeout */
    select(conn->writesockfd+1, NULL, writefdp, NULL, &notimeout);
  }

  do {
    /* If we still have reading to do, we check if we have a readable
       socket. Sometimes the reafdp is NULL, it no fd_set was done using
       the multi interface and then we can do nothing but to attempt a
       read to be sure. */
    if((k->keepon & KEEP_READ) &&
       (FD_ISSET(conn->sockfd, readfdp))) {

      /* read! */
      result = Curl_read(conn, conn->sockfd, k->buf,
                         data->set.buffer_size?
                         data->set.buffer_size:BUFSIZE -1,
                         &nread);

      if(0>result)
        break; /* get out of loop */
      if(result>0)
        return result;

      if ((k->bytecount == 0) && (k->writebytecount == 0))
        Curl_pgrsTime(data, TIMER_STARTTRANSFER);

      didwhat |= KEEP_READ;

      /* NULL terminate, allowing string ops to be used */
      if (0 < nread)
        k->buf[nread] = 0;

      /* if we receive 0 or less here, the server closed the connection and
         we bail out from this! */
      else if (0 >= nread) {
        k->keepon &= ~KEEP_READ;
        FD_ZERO(&k->rkeepfd);
        break;
      }

      /* Default buffer to use when we write the buffer, it may be changed
         in the flow below before the actual storing is done. */
      k->str = k->buf;

      /* Since this is a two-state thing, we check if we are parsing
         headers at the moment or not. */          
      if (k->header) {
        /* we are in parse-the-header-mode */
        bool stop_reading = FALSE;

        /* header line within buffer loop */
        do {
          int hbufp_index;
              
          /* str_start is start of line within buf */
          k->str_start = k->str;
              
          k->end_ptr = strchr (k->str_start, '\n');
              
          if (!k->end_ptr) {
            /* Not a complete header line within buffer, append the data to
               the end of the headerbuff. */

            if (k->hbuflen + nread >= data->state.headersize) {
              /* We enlarge the header buffer as it is too small */
              char *newbuff;
              long newsize=MAX((k->hbuflen+nread)*3/2,
                               data->state.headersize*2);
              hbufp_index = k->hbufp - data->state.headerbuff;
              newbuff = (char *)realloc(data->state.headerbuff, newsize);
              if(!newbuff) {
                failf (data, "Failed to alloc memory for big header!");
                return CURLE_OUT_OF_MEMORY;
              }
              data->state.headersize=newsize;
              data->state.headerbuff = newbuff;
              k->hbufp = data->state.headerbuff + hbufp_index;
            }
            memcpy(k->hbufp, k->str, nread);
            k->hbufp += nread;
            k->hbuflen += nread;
            if (!k->headerline && (k->hbuflen>5)) {
              /* make a first check that this looks like a HTTP header */
              if(!strnequal(data->state.headerbuff, "HTTP/", 5)) {
                /* this is not the beginning of a HTTP first header line */
                k->header = FALSE;
                k->badheader = TRUE;
                break;
              }
            }

            break;		/* read more and try again */
          }

          k->str = k->end_ptr + 1; /* move past new line */

          /*
           * We're about to copy a chunk of data to the end of the
           * already received header. We make sure that the full string
           * fit in the allocated header buffer, or else we enlarge 
           * it.
           */
          if (k->hbuflen + (k->str - k->str_start) >=
              data->state.headersize) {
            char *newbuff;
            long newsize=MAX((k->hbuflen+
                              (k->str-k->str_start))*3/2,
                             data->state.headersize*2);
            hbufp_index = k->hbufp - data->state.headerbuff;
            newbuff = (char *)realloc(data->state.headerbuff, newsize);
            if(!newbuff) {
              failf (data, "Failed to alloc memory for big header!");
              return CURLE_OUT_OF_MEMORY;
            }
            data->state.headersize= newsize;
            data->state.headerbuff = newbuff;
            k->hbufp = data->state.headerbuff + hbufp_index;
          }

          /* copy to end of line */
          strncpy (k->hbufp, k->str_start, k->str - k->str_start);
          k->hbufp += k->str - k->str_start;
          k->hbuflen += k->str - k->str_start;
          *k->hbufp = 0;
              
          k->p = data->state.headerbuff;
              
          /****
           * We now have a FULL header line that p points to
           *****/

          if (('\n' == *k->p) || ('\r' == *k->p)) {
            int headerlen;
            /* Zero-length header line means end of headers! */

            if ('\r' == *k->p)
              k->p++; /* pass the \r byte */
            if ('\n' == *k->p)
              k->p++; /* pass the \n byte */

            if(100 == k->httpcode) {
              /*
               * we have made a HTTP PUT or POST and this is 1.1-lingo
               * that tells us that the server is OK with this and ready
               * to receive our stuff.
               * However, we'll get more headers now so we must get
               * back into the header-parsing state!
               */
              k->header = TRUE;
              k->headerline = 0; /* restart the header line counter */
              /* if we did wait for this do enable write now! */
              if (k->write_after_100_header) {

                k->write_after_100_header = FALSE;
                FD_SET (conn->writesockfd, &k->writefd); /* write */
                k->keepon |= KEEP_WRITE;
                k->wkeepfd = k->writefd;
              }
            }
            else
              k->header = FALSE; /* no more header to parse! */

            if (417 == k->httpcode) {
              /*
               * we got: "417 Expectation Failed" this means:
               * we have made a HTTP call and our Expect Header
               * seems to cause a problem => abort the write operations
               * (or prevent them from starting).
               */
              k->write_after_100_header = FALSE;
              k->keepon &= ~KEEP_WRITE;
              FD_ZERO(&k->wkeepfd);
            }

            /* now, only output this if the header AND body are requested:
             */
            k->writetype = CLIENTWRITE_HEADER;
            if (data->set.http_include_header)
              k->writetype |= CLIENTWRITE_BODY;

            headerlen = k->p - data->state.headerbuff;

            result = Curl_client_write(data, k->writetype,
                                       data->state.headerbuff,
                                       headerlen);
            if(result)
              return result;

            data->info.header_size += headerlen;
            conn->headerbytecount += headerlen;

            if(!k->header) {
              /*
               * really end-of-headers.
               *
               * If we requested a "no body", this is a good time to get
               * out and return home.
               */
              if(data->set.no_body)
                stop_reading = TRUE;
              else if(!conn->bits.close) {
                /* If this is not the last request before a close, we must
                   set the maximum download size to the size of the
                   expected document or else, we won't know when to stop
                   reading! */
                if(-1 != conn->size)
                  conn->maxdownload = conn->size;
              }
              /* If max download size is *zero* (nothing) we already
                 have nothing and can safely return ok now! */
              if(0 == conn->maxdownload)
                stop_reading = TRUE;
                    
              if(stop_reading) {
                /* we make sure that this socket isn't read more now */
                k->keepon &= ~KEEP_READ;
                FD_ZERO(&k->rkeepfd);
              }

              break;		/* exit header line loop */
            }

            /* We continue reading headers, so reset the line-based
               header parsing variables hbufp && hbuflen */
            k->hbufp = data->state.headerbuff;
            k->hbuflen = 0;
            continue;
          }

          /*
           * Checks for special headers coming up.
           */
              
          if (!k->headerline++) {
            /* This is the first header, it MUST be the error code line
               or else we consiser this to be the body right away! */
            int httpversion_major;
            int nc=sscanf (k->p, " HTTP/%d.%d %3d",
                           &httpversion_major,
                           &k->httpversion,
                           &k->httpcode);
            if (nc==3) {
              k->httpversion += 10 * httpversion_major;
            }
            else {
              /* this is the real world, not a Nirvana
                 NCSA 1.5.x returns this crap when asked for HTTP/1.1
              */
              nc=sscanf (k->p, " HTTP %3d", &k->httpcode);
              k->httpversion = 10;
            }

            if (nc) {
              data->info.httpcode = k->httpcode;
              data->info.httpversion = k->httpversion;

              /* 404 -> URL not found! */
              if (data->set.http_fail_on_error &&
                  (k->httpcode >= 400)) {
                /* If we have been told to fail hard on HTTP-errors,
                   here is the check for that: */
                /* serious error, go home! */
                failf (data, "The requested file was not found");
                return CURLE_HTTP_NOT_FOUND;
              }

              if(k->httpversion == 10)
                /* Default action for HTTP/1.0 must be to close, unless
                   we get one of those fancy headers that tell us the
                   server keeps it open for us! */
                conn->bits.close = TRUE;

              switch(k->httpcode) {
              case 204:
                /* (quote from RFC2616, section 10.2.5): The server has
                 * fulfilled the request but does not need to return an
                 * entity-body ... The 204 response MUST NOT include a
                 * message-body, and thus is always terminated by the first
                 * empty line after the header fields. */
                /* FALLTHROUGH */
              case 304:
                /* (quote from RFC2616, section 10.3.5): The 304 response MUST
                 * NOT contain a message-body, and thus is always terminated
                 * by the first empty line after the header fields.  */
                conn->size=0;
                conn->maxdownload=0;
                break;
              default:
                /* nothing */
                break;
              }
            }
            else {
              k->header = FALSE;   /* this is not a header line */
              k->badheader = TRUE; /* this was a bad header */
              break;
            }
          }

          /* check for Content-Length: header lines to get size */
          if (strnequal("Content-Length:", k->p, 15) &&
              sscanf (k->p+15, " %ld", &k->contentlength)) {
            conn->size = k->contentlength;
            Curl_pgrsSetDownloadSize(data, k->contentlength);
            }
          /* check for Content-Type: header lines to get the mime-type */
          else if (strnequal("Content-Type:", k->p, 13)) {
            char *start;
            char *end;
            int len;
              
            /* Find the first non-space letter */
            for(start=k->p+14;
                *start && isspace((int)*start);
                start++);

            /* count all non-space letters following */
            for(end=start, len=0;
                *end && !isspace((int)*end);
                end++, len++);

            /* allocate memory of a cloned copy */
            data->info.contenttype = malloc(len + 1);
            if (NULL == data->info.contenttype)
              return CURLE_OUT_OF_MEMORY;

            /* copy the content-type string */
            memcpy(data->info.contenttype, start, len);
            data->info.contenttype[len] = 0; /* zero terminate */
          }
          else if((k->httpversion == 10) &&
                  conn->bits.httpproxy &&
                  compareheader(k->p, "Proxy-Connection:", "keep-alive")) {
            /*
             * When a HTTP/1.0 reply comes when using a proxy, the
             * 'Proxy-Connection: keep-alive' line tells us the
             * connection will be kept alive for our pleasure.
             * Default action for 1.0 is to close.
             */
            conn->bits.close = FALSE; /* don't close when done */
            infof(data, "HTTP/1.0 proxy connection set to keep alive!\n");
          }
          else if((k->httpversion == 10) &&
                  compareheader(k->p, "Connection:", "keep-alive")) {
            /*
             * A HTTP/1.0 reply with the 'Connection: keep-alive' line
             * tells us the connection will be kept alive for our
             * pleasure.  Default action for 1.0 is to close.
             *
             * [RFC2068, section 19.7.1] */
            conn->bits.close = FALSE; /* don't close when done */
            infof(data, "HTTP/1.0 connection set to keep alive!\n");
          }
          else if (compareheader(k->p, "Connection:", "close")) {
            /*
             * [RFC 2616, section 8.1.2.1]
             * "Connection: close" is HTTP/1.1 language and means that
             * the connection will close when this request has been
             * served.
             */
            conn->bits.close = TRUE; /* close when done */
          }
          else if (compareheader(k->p, "Transfer-Encoding:", "chunked")) {
            /*
             * [RFC 2616, section 3.6.1] A 'chunked' transfer encoding
             * means that the server will send a series of "chunks". Each
             * chunk starts with line with info (including size of the
             * coming block) (terminated with CRLF), then a block of data
             * with the previously mentioned size. There can be any amount
             * of chunks, and a chunk-data set to zero signals the
             * end-of-chunks. */
            conn->bits.chunk = TRUE; /* chunks coming our way */

            /* init our chunky engine */
            Curl_httpchunk_init(conn);
          }
          else if (strnequal("Content-Encoding:", k->p, 17) &&
                   data->set.encoding) {
            /*
             * Process Content-Encoding. Look for the values: identity, gzip,
             * defalte, compress, x-gzip and x-compress. x-gzip and
             * x-compress are the same as gzip and compress. (Sec 3.5 RFC
             * 2616). zlib cannot handle compress, and gzip is not currently
             * implemented. However, errors are handled further down when the
             * response body is processed 08/27/02 jhrg */
            char *start;

            /* Find the first non-space letter */
            for(start=k->p+18;
                *start && isspace((int)*start);
                start++);

            /* Record the content-encoding for later use. 08/27/02 jhrg */
            if (strnequal("identity", start, 8))
              k->content_encoding = IDENTITY;
            else if (strnequal("deflate", start, 7))
              k->content_encoding = DEFLATE;
            else if (strnequal("gzip", start, 4) 
                     || strnequal("x-gzip", start, 6))
              k->content_encoding = GZIP;
            else if (strnequal("compress", start, 8) 
                     || strnequal("x-compress", start, 10))
              k->content_encoding = COMPRESS;
          }
          else if (strnequal("Content-Range:", k->p, 14)) {
            if (sscanf (k->p+14, " bytes %d-", &k->offset) ||
                sscanf (k->p+14, " bytes: %d-", &k->offset)) {
              /* This second format was added August 1st 2000 by Igor
                 Khristophorov since Sun's webserver JavaWebServer/1.1.1
                 obviously sends the header this way! :-( */
              if (conn->resume_from == k->offset) {
                /* we asked for a resume and we got it */
                k->content_range = TRUE;
              }
            }
          }
          else if(data->cookies &&
                  strnequal("Set-Cookie:", k->p, 11)) {
            Curl_cookie_add(data->cookies, TRUE, k->p+11, conn->name);
          }
          else if(strnequal("Last-Modified:", k->p,
                            strlen("Last-Modified:")) &&
                  (data->set.timecondition || data->set.get_filetime) ) {
            time_t secs=time(NULL);
            k->timeofdoc = curl_getdate(k->p+strlen("Last-Modified:"),
                                        &secs);
            if(data->set.get_filetime)
              data->info.filetime = k->timeofdoc;
          }
          else if ((k->httpcode >= 300 && k->httpcode < 400) &&
                   (data->set.http_follow_location) &&
                   strnequal("Location:", k->p, 9)) {
            /* this is the URL that the server advices us to get instead */
            char *ptr;
            char *start=k->p;
            char backup;

            start += 9; /* pass "Location:" */

            /* Skip spaces and tabs. We do this to support multiple
               white spaces after the "Location:" keyword. */
            while(*start && isspace((int)*start ))
              start++;
            ptr = start; /* start scanning here */

            /* scan through the string to find the end */
            while(*ptr && !isspace((int)*ptr))
              ptr++;
            backup = *ptr; /* store the ending letter */
            *ptr = '\0';   /* zero terminate */
            conn->newurl = strdup(start); /* clone string */
            *ptr = backup; /* restore ending letter */
          }

          /*
           * End of header-checks. Write them to the client.
           */

          k->writetype = CLIENTWRITE_HEADER;
          if (data->set.http_include_header)
            k->writetype |= CLIENTWRITE_BODY;

          if(data->set.verbose)
            Curl_debug(data, CURLINFO_HEADER_IN,
                       k->p, k->hbuflen);

          result = Curl_client_write(data, k->writetype, k->p,
                                     k->hbuflen);
          if(result)
            return result;

          data->info.header_size += k->hbuflen;
          conn->headerbytecount += k->hbuflen;
              
          /* reset hbufp pointer && hbuflen */
          k->hbufp = data->state.headerbuff;
          k->hbuflen = 0;
        }
        while (!stop_reading && *k->str); /* header line within buffer */

        if(stop_reading)
          /* We've stopped dealing with input, get out of the do-while loop */
          break;

        /* We might have reached the end of the header part here, but
           there might be a non-header part left in the end of the read
           buffer. */

        if (!k->header) {
          /* starting here, this is not part of the header! */

          /* we subtract the remaining header size from the buffer */
          nread -= (k->str - k->buf);
        }

      }			/* end if header mode */

      /* This is not an 'else if' since it may be a rest from the header
         parsing, where the beginning of the buffer is headers and the end
         is non-headers. */
      if (k->str && !k->header && (nread > 0)) {
            
        if(0 == k->bodywrites) {
          /* These checks are only made the first time we are about to
             write a piece of the body */
          if(conn->protocol&PROT_HTTP) {
            /* HTTP-only checks */
            if (conn->newurl) {
              /* abort after the headers if "follow Location" is set */
              infof (data, "Follow to new URL: %s\n", conn->newurl);
              k->keepon &= ~KEEP_READ;
              FD_ZERO(&k->rkeepfd);
              return CURLE_OK;
            }
            else if (conn->resume_from &&
                     !k->content_range &&
                     (data->set.httpreq==HTTPREQ_GET)) {
              /* we wanted to resume a download, although the server
                 doesn't seem to support this and we did this with a GET
                 (if it wasn't a GET we did a POST or PUT resume) */
              failf (data, "HTTP server doesn't seem to support "
                     "byte ranges. Cannot resume.");
              return CURLE_HTTP_RANGE_ERROR;
            }
            else if(data->set.timecondition && !conn->range) {
              /* A time condition has been set AND no ranges have been
                 requested. This seems to be what chapter 13.3.4 of
                 RFC 2616 defines to be the correct action for a
                 HTTP/1.1 client */
              if((k->timeofdoc > 0) && (data->set.timevalue > 0)) {
                switch(data->set.timecondition) {
                case TIMECOND_IFMODSINCE:
                default:
                  if(k->timeofdoc < data->set.timevalue) {
                    infof(data,
                          "The requested document is not new enough\n");
                    return CURLE_OK;
                  }
                  break;
                case TIMECOND_IFUNMODSINCE:
                  if(k->timeofdoc > data->set.timevalue) {
                    infof(data,
                          "The requested document is not old enough\n");
                    return CURLE_OK;
                  }
                  break;
                } /* switch */
              } /* two valid time strings */
            } /* we have a time condition */

          } /* this is HTTP */
        } /* this is the first time we write a body part */
        k->bodywrites++;

        /* pass data to the debug function before it gets "dechunked" */
        if(data->set.verbose)
          Curl_debug(data, CURLINFO_DATA_IN, k->str, nread);

        if(conn->bits.chunk) {
          /*
           * Bless me father for I have sinned. Here comes a chunked
           * transfer flying and we need to decode this properly.  While
           * the name says read, this function both reads and writes away
           * the data. The returned 'nread' holds the number of actual
           * data it wrote to the client.  */
          CHUNKcode res =
            Curl_httpchunk_read(conn, k->str, nread, &nread);

          if(CHUNKE_OK < res) {
            if(CHUNKE_WRITE_ERROR == res) {
              failf(data, "Failed writing data");
              return CURLE_WRITE_ERROR;
            }
            failf(data, "Received problem in the chunky parser");
            return CURLE_RECV_ERROR;
          }
          else if(CHUNKE_STOP == res) {
            /* we're done reading chunks! */
            k->keepon &= ~KEEP_READ; /* read no more */
            FD_ZERO(&k->rkeepfd);

            /* There are now possibly N number of bytes at the end of the
               str buffer that weren't written to the client, but we don't
               care about them right now. */
          }
          /* If it returned OK, we just keep going */
        }

        if((-1 != conn->maxdownload) &&
           (k->bytecount + nread >= conn->maxdownload)) {
          nread = conn->maxdownload - k->bytecount;
          if(nread < 0 ) /* this should be unusual */
            nread = 0;

          k->keepon &= ~KEEP_READ; /* we're done reading */
          FD_ZERO(&k->rkeepfd);
        }

        k->bytecount += nread;

        Curl_pgrsSetDownloadCounter(data, (double)k->bytecount);
            
        if(!conn->bits.chunk && (nread || k->badheader)) {
          /* If this is chunky transfer, it was already written */

          if(k->badheader) {
            /* we parsed a piece of data wrongly assuming it was a header
               and now we output it as body instead */
            result = Curl_client_write(data, CLIENTWRITE_BODY,
                                       data->state.headerbuff,
                                       k->hbuflen);
            k->badheader = FALSE; /* taken care of now */
          }

          /* This switch handles various content encodings. If there's an
             error here, be sure to check over the almost identical code in
             http_chunk.c. 08/29/02 jhrg */
#ifdef HAVE_LIBZ
          switch (k->content_encoding) {
            case IDENTITY:
#endif
              /* This is the default when the server sends no
                 Content-Encoding header. See Curl_readwrite_init; the
                 memset() call initializes k->content_encoding to zero.
                 08/28/02 jhrg */
              result = Curl_client_write(data, CLIENTWRITE_BODY, k->str, 
                                         nread);
#ifdef HAVE_LIBZ
              break;

            case DEFLATE: 
              /* Assume CLIENTWRITE_BODY; headers are not encoded. */
              result = Curl_unencode_deflate_write(data, k, nread);
              break;

            case GZIP:          /* FIXME 08/27/02 jhrg */
            case COMPRESS:
            default:
              failf (data, "Unrecognized content encoding type. "
                     "libcurl understands `identity' and `deflate' "
                     "content encodings.");
              result = CURLE_BAD_CONTENT_ENCODING;
              break;
          }
#endif

          if(result)
            return result;
        }

      } /* if (! header and data to read ) */
    } /* if( read from socket ) */

    /* If we still have writing to do, we check if we have a writable
       socket. Sometimes the writefdp is NULL, it no fd_set was done using
       the multi interface and then we can do nothing but to attempt a
       write to be sure. */
    if((k->keepon & KEEP_WRITE) &&
       (FD_ISSET(conn->writesockfd, writefdp)) ) {
      /* write */

      int i, si;
      ssize_t bytes_written;

      if ((k->bytecount == 0) && (k->writebytecount == 0))
        Curl_pgrsTime(data, TIMER_STARTTRANSFER);

      didwhat |= KEEP_WRITE;

      /* only read more data if there's no upload data already
         present in the upload buffer */
      if(0 == conn->upload_present) {
        /* init the "upload from here" pointer */
        conn->upload_fromhere = k->uploadbuf;

        nread = data->set.fread(conn->upload_fromhere, 1,
                                BUFSIZE, data->set.in);

        /* the signed int typecase of nread of for systems that has
           unsigned size_t */
        if (nread<=0) {
          /* done */
          k->keepon &= ~KEEP_WRITE; /* we're done writing */
          FD_ZERO(&k->wkeepfd);
          break;
        }

        /* store number of bytes available for upload */
        conn->upload_present = nread;

        /* convert LF to CRLF if so asked */
        if (data->set.crlf) {
          for(i = 0, si = 0; i < nread; i++, si++) {
            if (conn->upload_fromhere[i] == 0x0a) {
              data->state.scratch[si++] = 0x0d;
              data->state.scratch[si] = 0x0a;
            }
            else
              data->state.scratch[si] = conn->upload_fromhere[i];
          }
          if(si != nread) {
            /* only perform the special operation if we really did replace
               anything */
            nread = si;

            /* upload from the new (replaced) buffer instead */
            conn->upload_fromhere = data->state.scratch;

            /* set the new amount too */
            conn->upload_present = nread;
          }
        }
      }
      else {
        /* We have a partial buffer left from a previous "round". Use
           that instead of reading more data */
      }

      /* write to socket */
      result = Curl_write(conn,
                          conn->writesockfd,
                          conn->upload_fromhere,
                          conn->upload_present,
                          &bytes_written);
      if(result)
        return result;
      else if(conn->upload_present != bytes_written) {
        /* we only wrote a part of the buffer (if anything), deal with it! */

        /* store the amount of bytes left in the buffer to write */
        conn->upload_present -= bytes_written;

        /* advance the pointer where to find the buffer when the next send
           is to happen */
        conn->upload_fromhere += bytes_written;
      }
      else {
        /* we've uploaded that buffer now */
        conn->upload_fromhere = k->uploadbuf;
        conn->upload_present = 0; /* no more bytes left */
      }

      if(data->set.verbose)
        Curl_debug(data, CURLINFO_DATA_OUT, conn->upload_fromhere,
                   bytes_written);
      

      k->writebytecount += bytes_written;
      Curl_pgrsSetUploadCounter(data, (double)k->writebytecount);

    }

  } while(0); /* just to break out from! */

  if(didwhat) {
    /* Update read/write counters */
    if(conn->bytecountp)
      *conn->bytecountp = k->bytecount; /* read count */
    if(conn->writebytecountp)
      *conn->writebytecountp = k->writebytecount; /* write count */
  }
  else {
    /* no read no write, this is a timeout? */
    if (k->write_after_100_header) {
      /* This should allow some time for the header to arrive, but only a
         very short time as otherwise it'll be too much wasted times too
         often. */
      k->write_after_100_header = FALSE;
      FD_SET (conn->writesockfd, &k->writefd); /* write socket */
      k->keepon |= KEEP_WRITE;
      k->wkeepfd = k->writefd;
    }    
  }

  k->now = Curl_tvnow();
  if(Curl_pgrsUpdate(conn))
    result = CURLE_ABORTED_BY_CALLBACK;
  else
    result = Curl_speedcheck (data, k->now);
  if (result)
    return result;
    
  if (data->set.timeout &&
      ((Curl_tvdiff(k->now, k->start)/1000) >= data->set.timeout)) {
    failf (data, "Operation timed out with %d out of %d bytes received",
           k->bytecount, conn->size);
    return CURLE_OPERATION_TIMEOUTED;
  }

  if(!k->keepon) {
    /*
     * The transfer has been performed. Just make some general checks before
     * returning.
     */

    if(!(data->set.no_body) && k->contentlength &&
       (k->bytecount != k->contentlength) &&
       !conn->newurl) {
      failf(data, "transfer closed with %d bytes remaining to read",
            k->contentlength-k->bytecount);
      return CURLE_PARTIAL_FILE;
    }
    else if(conn->bits.chunk && conn->proto.http->chunk.datasize) {
      failf(data, "transfer closed with at least %d bytes remaining",
            conn->proto.http->chunk.datasize);
      return CURLE_PARTIAL_FILE;
    }
    if(Curl_pgrsUpdate(conn))
      return CURLE_ABORTED_BY_CALLBACK;
  }

  /* Now update the "done" boolean we return */
  *done = !k->keepon;

  return CURLE_OK;
}

CURLcode Curl_readwrite_init(struct connectdata *conn)
{
  struct SessionHandle *data = conn->data;
  struct Curl_transfer_keeper *k = &conn->keep;

  /* NB: the content encoding software depends on this initialization of
     Curl_transfer_keeper. 08/28/02 jhrg */
  memset(k, 0, sizeof(struct Curl_transfer_keeper));

  k->start = Curl_tvnow(); /* start time */
  k->now = k->start;   /* current time is now */
  k->header = TRUE; /* assume header */
  k->httpversion = -1; /* unknown at this point */

  data = conn->data; /* there's the root struct */
  k->buf = data->state.buffer;
  k->uploadbuf = data->state.uploadbuffer;
  k->maxfd = (conn->sockfd>conn->writesockfd?
              conn->sockfd:conn->writesockfd)+1;
  k->hbufp = data->state.headerbuff;

  Curl_pgrsTime(data, TIMER_PRETRANSFER);
  Curl_speedinit(data);

  Curl_pgrsSetUploadCounter(data, 0);
  Curl_pgrsSetDownloadCounter(data, 0);

  if (!conn->getheader) {
    k->header = FALSE;
    if(conn->size > 0)
      Curl_pgrsSetDownloadSize(data, conn->size);
  }
  /* we want header and/or body, if neither then don't do this! */
  if(conn->getheader || !data->set.no_body) {

    FD_ZERO (&k->readfd);		/* clear it */
    if(conn->sockfd != -1) {
      FD_SET (conn->sockfd, &k->readfd); /* read socket */
      k->keepon |= KEEP_READ;
    }

    FD_ZERO (&k->writefd);		/* clear it */
    if(conn->writesockfd != -1) {
      if (data->set.expect100header)
        /* wait with write until we either got 100-continue or a timeout */
        k->write_after_100_header = TRUE;
      else {
        FD_SET (conn->writesockfd, &k->writefd); /* write socket */
        k->keepon |= KEEP_WRITE;
      }
    }

    /* get these in backup variables to be able to restore them on each lap in
       the select() loop */
    k->rkeepfd = k->readfd;
    k->wkeepfd = k->writefd;

  }

  return CURLE_OK;
}

void Curl_single_fdset(struct connectdata *conn,
                       fd_set *read_fd_set,
                       fd_set *write_fd_set,
                       fd_set *exc_fd_set,
                       int *max_fd)
{
  *max_fd = -1; /* init */
  if(conn->keep.keepon & KEEP_READ) {
    FD_SET(conn->sockfd, read_fd_set);
    *max_fd = conn->sockfd;
    conn->keep.readfdp = read_fd_set; /* store the address of the set */
  }
  if(conn->keep.keepon & KEEP_WRITE) {
    FD_SET(conn->writesockfd, write_fd_set);
    if(conn->writesockfd > *max_fd)
      *max_fd = conn->writesockfd;
    conn->keep.writefdp = write_fd_set; /* store the address of the set */
  }
  /* we don't use exceptions, only touch that one to prevent compiler
     warnings! */
  *exc_fd_set = *exc_fd_set;
}


/*
 * Transfer()
 *
 * This function is what performs the actual transfer. It is capable of
 * doing both ways simultaneously.
 * The transfer must already have been setup by a call to Curl_Transfer().
 *
 * Note that headers are created in a preallocated buffer of a default size.
 * That buffer can be enlarged on demand, but it is never shrinken again.
 *
 * Parts of this function was once written by the friendly Mark Butler
 * <butlerm@xmission.com>.
 */

static CURLcode
Transfer(struct connectdata *conn)
{
  struct SessionHandle *data = conn->data;
  CURLcode result;
  struct Curl_transfer_keeper *k = &conn->keep;
  bool done=FALSE;

  Curl_readwrite_init(conn);

  if((conn->sockfd == -1) && (conn->writesockfd == -1))
    /* nothing to read, nothing to write, we're already OK! */
    return CURLE_OK;

  /* we want header and/or body, if neither then don't do this! */
  if(!conn->getheader && data->set.no_body)
    return CURLE_OK;

  k->writefdp = &k->writefd; /* store the address of the set */
  k->readfdp = &k->readfd;   /* store the address of the set */

  while (!done) {
    struct timeval interval;
    k->readfd = k->rkeepfd;  /* set these every lap in the loop */
    k->writefd = k->wkeepfd;
    interval.tv_sec = 1;
    interval.tv_usec = 0;

    switch (select (k->maxfd, k->readfdp, k->writefdp, NULL, &interval)) {
    case -1: /* select() error, stop reading */
#ifdef EINTR
      /* The EINTR is not serious, and it seems you might get this more
         ofen when using the lib in a multi-threaded environment! */
      if(errno == EINTR)
        ;
      else
#endif
        done = TRUE; /* no more read or write */
      continue;
    case 0:  /* timeout */
      result = Curl_readwrite(conn, &done);
      break;

    default: /* readable descriptors */
      result = Curl_readwrite(conn, &done);
      break;
    }
    if(result)
      return result;
    
    /* "done" signals to us if the transfer(s) are ready */
  }

  return CURLE_OK;
}

CURLcode Curl_pretransfer(struct SessionHandle *data)
{
  if(!data->change.url)
    /* we can't do anything wihout URL */
    return CURLE_URL_MALFORMAT;

#ifdef USE_SSLEAY
  /* Init the SSL session ID cache here. We do it here since we want to
     do it after the *_setopt() calls (that could change the size) but
     before any transfer. */
  Curl_SSL_InitSessions(data, data->set.ssl.numsessions);
#endif

  data->set.followlocation=0; /* reset the location-follow counter */
  data->state.this_is_a_follow = FALSE; /* reset this */
  data->state.errorbuf = FALSE; /* no error has occurred */

 /* Allow data->set.use_port to set which port to use. This needs to be
  * disabled for example when we follow Location: headers to URLs using
  * different ports! */
  data->state.allow_port = TRUE;

#if defined(HAVE_SIGNAL) && defined(SIGPIPE)
  /*************************************************************
   * Tell signal handler to ignore SIGPIPE
   *************************************************************/
  if(!data->set.no_signal)
    data->state.prev_signal = signal(SIGPIPE, SIG_IGN);
#endif  

  Curl_initinfo(data); /* reset session-specific information "variables" */
  Curl_pgrsStartNow(data);

  return CURLE_OK;
}

CURLcode Curl_posttransfer(struct SessionHandle *data)
{
#if defined(HAVE_SIGNAL) && defined(SIGPIPE)
  /* restore the signal handler for SIGPIPE before we get back */
  if(!data->set.no_signal)
    signal(SIGPIPE, data->state.prev_signal);
#endif  

  return CURLE_OK;
}

CURLcode Curl_follow(struct SessionHandle *data,
                     char *newurl) /* this 'newurl' is the Location: string,
                                      and it must be malloc()ed before passed
                                      here */
{
  /* Location: redirect */
  char prot[16]; /* URL protocol string storage */
  char letter;   /* used for a silly sscanf */
  
  if (data->set.maxredirs &&
      (data->set.followlocation >= data->set.maxredirs)) {
    failf(data,"Maximum (%d) redirects followed", data->set.maxredirs);
    return CURLE_TOO_MANY_REDIRECTS;
  }

  /* mark the next request as a followed location: */
  data->state.this_is_a_follow = TRUE;

  data->set.followlocation++; /* count location-followers */

  if(data->set.http_auto_referer) {
    /* We are asked to automatically set the previous URL as the
       referer when we get the next URL. We pick the ->url field,
       which may or may not be 100% correct */

    if(data->change.referer_alloc)
      /* If we already have an allocated referer, free this first */
      free(data->change.referer);

    data->change.referer = strdup(data->change.url);
    data->change.referer_alloc = TRUE; /* yes, free this later */
  }

  if(2 != sscanf(newurl, "%15[^?&/:]://%c", prot, &letter)) {
    /***
     *DANG* this is an RFC 2068 violation. The URL is supposed
     to be absolute and this doesn't seem to be that!
     ***
     Instead, we have to TRY to append this new path to the old URL
     to the right of the host part. Oh crap, this is doomed to cause
     problems in the future...
    */
    char *protsep;
    char *pathsep;
    char *newest;

    char *useurl = newurl;

    /* we must make our own copy of the URL to play with, as it may
       point to read-only data */
    char *url_clone=strdup(data->change.url);

    if(!url_clone)
      return CURLE_OUT_OF_MEMORY; /* skip out of this NOW */

    /* protsep points to the start of the host name */
    protsep=strstr(url_clone, "//");
    if(!protsep)
      protsep=url_clone;
    else
      protsep+=2; /* pass the slashes */

    if('/' != newurl[0]) {
      int level=0;

      /* First we need to find out if there's a ?-letter in the URL,
         and cut it and the right-side of that off */
      pathsep = strrchr(protsep, '?');
      if(pathsep)
        *pathsep=0;
      
      /* we have a relative path to append to the last slash if
         there's one available */
      pathsep = strrchr(protsep, '/');
      if(pathsep)
        *pathsep=0;

      /* Check if there's any slash after the host name, and if so,
         remember that position instead */
      pathsep = strchr(protsep, '/');
      if(pathsep)
        protsep = pathsep+1;
      else
        protsep = NULL;

      /* now deal with one "./" or any amount of "../" in the newurl
         and act accordingly */

      if((useurl[0] == '.') && (useurl[1] == '/'))
        useurl+=2; /* just skip the "./" */
            
      while((useurl[0] == '.') &&
            (useurl[1] == '.') &&
            (useurl[2] == '/')) {
        level++; 
        useurl+=3; /* pass the "../" */
      }

      if(protsep) {
        while(level--) {
          /* cut off one more level from the right of the original URL */
          pathsep = strrchr(protsep, '/');
          if(pathsep)
            *pathsep=0;
          else {
            *protsep=0;
            break;
          }
        }
      }
    }
    else {
      /* We got a new absolute path for this server, cut off from the
         first slash */
      pathsep = strchr(protsep, '/');
      if(pathsep)
        *pathsep=0;
    }

    newest=(char *)malloc( strlen(url_clone) +
                           1 + /* possible slash */
                           strlen(useurl) + 1/* zero byte */);
    
    if(!newest)
      return CURLE_OUT_OF_MEMORY; /* go out from this */

    sprintf(newest, "%s%s%s", url_clone,
            (('/' == useurl[0]) || !*protsep)?"":"/",
            useurl);
    free(newurl); /* newurl is the allocated pointer */
    free(url_clone);
    newurl = newest;
  }
  else
    /* This is an absolute URL, don't allow the custom port number */
    data->state.allow_port = FALSE;

  if(data->change.url_alloc)
    free(data->change.url);
  else
    data->change.url_alloc = TRUE; /* the URL is allocated */
      
  /* TBD: set the URL with curl_setopt() */
  data->change.url = newurl;
  newurl = NULL; /* don't free! */

  infof(data, "Follows Location: to new URL: '%s'\n", data->change.url);

  /*
   * We get here when the HTTP code is 300-399. We need to perform
   * differently based on exactly what return code there was.
   * Discussed on the curl mailing list and posted about on the 26th
   * of January 2001.
   */
  switch(data->info.httpcode) {
  case 300: /* Multiple Choices */
  case 306: /* Not used */
  case 307: /* Temporary Redirect */
  default:  /* for all unknown ones */
    /* These are explicitly mention since I've checked RFC2616 and they
     * seem to be OK to POST to.
     */
    break;
  case 301: /* Moved Permanently */
    /* (quote from RFC2616, section 10.3.2):
     * 
     *  Note: When automatically redirecting a POST request after
     *  receiving a 301 status code, some existing HTTP/1.0 user agents
     *  will erroneously change it into a GET request.
     *
     * ----
     * Warning: Because most of importants user agents do this clear
     * RFC2616 violation, many webservers expect this misbehavior. So
     * these servers often answers to a POST request with an error page.
     * To be sure that libcurl gets the page that most user agents
     * would get, libcurl has to force GET:
     */
    if( data->set.httpreq == HTTPREQ_POST
        || data->set.httpreq == HTTPREQ_POST_FORM) {
      infof(data,
            "Violate RFC 2616/10.3.2 and switch from POST to GET\n");
      data->set.httpreq = HTTPREQ_GET;
    }
    break;
  case 302: /* Found */
    /* (From 10.3.3)
       
    Note: RFC 1945 and RFC 2068 specify that the client is not allowed
    to change the method on the redirected request.  However, most
    existing user agent implementations treat 302 as if it were a 303
    response, performing a GET on the Location field-value regardless
    of the original request method. The status codes 303 and 307 have
    been added for servers that wish to make unambiguously clear which
    kind of reaction is expected of the client.
    
    (From 10.3.4)
    
    Note: Many pre-HTTP/1.1 user agents do not understand the 303
    status. When interoperability with such clients is a concern, the
    302 status code may be used instead, since most user agents react
    to a 302 response as described here for 303.             
    */
  case 303: /* See Other */
    /* Disable both types of POSTs, since doing a second POST when
     * following isn't what anyone would want! */
    if(data->set.httpreq != HTTPREQ_GET) {
      data->set.httpreq = HTTPREQ_GET; /* enforce GET request */
      infof(data, "Disables POST, goes with %s\n",
            data->set.no_body?"HEAD":"GET");
    }
    break;
  case 304: /* Not Modified */
    /* 304 means we did a conditional request and it was "Not modified".
     * We shouldn't get any Location: header in this response!
     */
    break;
  case 305: /* Use Proxy */
    /* (quote from RFC2616, section 10.3.6):
     * "The requested resource MUST be accessed through the proxy given
     * by the Location field. The Location field gives the URI of the
     * proxy.  The recipient is expected to repeat this single request
     * via the proxy. 305 responses MUST only be generated by origin
     * servers."
     */
    break;
  }
  Curl_pgrsTime(data, TIMER_REDIRECT);
  Curl_pgrsResetTimes(data);

  return CURLE_OK;
}

CURLcode Curl_perform(struct SessionHandle *data)
{
  CURLcode res;
  CURLcode res2;
  struct connectdata *conn=NULL;
  char *newurl = NULL; /* possibly a new URL to follow to! */

  data->state.used_interface = Curl_if_easy;

  res = Curl_pretransfer(data);
  if(res)
    return res;

  /*
   * It is important that there is NO 'return' from this function at any other
   * place than falling down to the end of the function! This is because we
   * have cleanup stuff that must be done before we get back, and that is only
   * performed after this do-while loop.
   */

  do {
    Curl_pgrsTime(data, TIMER_STARTSINGLE);
    res = Curl_connect(data, &conn);
    if(res == CURLE_OK) {
      res = Curl_do(&conn);

      if(res == CURLE_OK) {
        CURLcode res2; /* just a local extra result container */

        if(conn->protocol&PROT_FTPS)
          /* FTPS, disable ssl while transfering data */
          conn->ssl.use = FALSE;
        res = Transfer(conn); /* now fetch that URL please */
        if(conn->protocol&PROT_FTPS)
          /* FTPS, enable ssl again after havving transferred data */
          conn->ssl.use = TRUE;

        if(res == CURLE_OK)
          /*
           * We must duplicate the new URL here as the connection data
           * may be free()ed in the Curl_done() function.
           */
          newurl = conn->newurl?strdup(conn->newurl):NULL;
        else {
          /* The transfer phase returned error, we mark the connection to get
           * closed to prevent being re-used. This is becasue we can't
           * possibly know if the connection is in a good shape or not now. */
          conn->bits.close = TRUE;

          if(-1 !=conn->secondarysocket) {
            /* if we failed anywhere, we must clean up the secondary socket if
               it was used */
            sclose(conn->secondarysocket);
            conn->secondarysocket=-1;
          }
        }

        /* Always run Curl_done(), even if some of the previous calls
           failed, but return the previous (original) error code */
        res2 = Curl_done(conn);

        if(CURLE_OK == res)
          res = res2;
      }

      /*
       * Important: 'conn' cannot be used here, since it may have been closed
       * in 'Curl_done' or other functions.
       */

      if((res == CURLE_OK) && newurl) {
        res = Curl_follow(data, newurl);
        if(CURLE_OK == res) {
          newurl = NULL;
          continue;
        }
      }
    }
    break; /* it only reaches here when this shouldn't loop */

  } while(1); /* loop if Location: */

  if(newurl)
    free(newurl);

  /* run post-transfer uncondionally, but don't clobber the return code if
     we already have an error code recorder */
  res2 = Curl_posttransfer(data);
  if(!res && res2)
    res = res2;

  return res;
}

CURLcode 
Curl_Transfer(struct connectdata *c_conn, /* connection data */
              int sockfd,	/* socket to read from or -1 */
              int size,		/* -1 if unknown at this point */
              bool getheader,	/* TRUE if header parsing is wanted */
              long *bytecountp,	/* return number of bytes read or NULL */
              int writesockfd,  /* socket to write to, it may very well be
                                   the same we read from. -1 disables */
              long *writebytecountp /* return number of bytes written or
                                       NULL */
              )
{
  struct connectdata *conn = (struct connectdata *)c_conn;
  if(!conn)
    return CURLE_BAD_FUNCTION_ARGUMENT;

  /* now copy all input parameters */
  conn->sockfd = sockfd;
  conn->size = size;
  conn->getheader = getheader;
  conn->bytecountp = bytecountp;
  conn->writesockfd = writesockfd;
  conn->writebytecountp = writebytecountp;

  return CURLE_OK;

}
          
/*
 * local variables:
 * eval: (load-file "../curl-mode.el")
 * end:
 * vim600: fdm=marker
 * vim: et sw=2 ts=2 sts=2 tw=78
 */