formdata.c

harvest是一个下载html网页得机器人

            if (filename) {
              if (!(current_form = AddFormInfo(strdup(filename),
                                               NULL, current_form)))
                return_value = CURL_FORMADD_MEMORY;
            }
            else
              return_value = CURL_FORMADD_NULL;
          }
          else
            return_value = CURL_FORMADD_OPTION_TWICE;
        }
        else {
          if (filename)
            current_form->value = strdup(filename);
          else
            return_value = CURL_FORMADD_NULL;
          current_form->flags |= HTTPPOST_BUFFER;
        }
        break;
      }
      
    /* CMC: Added support for buffer uploads */
    case CURLFORM_BUFFERPTR:
        current_form->flags |= HTTPPOST_PTRBUFFER;
      if (current_form->buffer)
        return_value = CURL_FORMADD_OPTION_TWICE;
      else {
        char *buffer =
          array_state?array_value:va_arg(params, char *);
        if (buffer)
          current_form->buffer = buffer; /* store for the moment */
        else
          return_value = CURL_FORMADD_NULL;
      }
      break;

    /* CMC: Added support for buffer uploads */
    case CURLFORM_BUFFERLENGTH:
      if (current_form->bufferlength)
        return_value = CURL_FORMADD_OPTION_TWICE;
      else
        current_form->bufferlength =
          array_state?(long)array_value:va_arg(params, long);
      break;

    case CURLFORM_CONTENTTYPE:
      {
        char *contenttype =
          array_state?array_value:va_arg(params, char *);
        if (current_form->contenttype) {
          if (current_form->flags & HTTPPOST_FILENAME) {
            if (contenttype) {
              if (!(current_form = AddFormInfo(NULL,
                                               strdup(contenttype),
                                               current_form)))
                return_value = CURL_FORMADD_MEMORY;
            }
	    else
	      return_value = CURL_FORMADD_NULL;
          }
          else
            return_value = CURL_FORMADD_OPTION_TWICE;
        }
        else {
	  if (contenttype)
	    current_form->contenttype = strdup(contenttype);
	  else
	    return_value = CURL_FORMADD_NULL;
	}
        break;
      }
    case CURLFORM_CONTENTHEADER:
      {
        /* this "cast increases required alignment of target type" but
           we consider it OK anyway */
        struct curl_slist* list = array_state?
          (struct curl_slist*)array_value:
          va_arg(params, struct curl_slist*);
        
        if( current_form->contentheader )
          return_value = CURL_FORMADD_OPTION_TWICE;
        else
          current_form->contentheader = list;
        
        break;
      }
    case CURLFORM_FILENAME:
      {
        char *filename = array_state?array_value:
          va_arg(params, char *);
        if( current_form->showfilename )
          return_value = CURL_FORMADD_OPTION_TWICE;
        else
          current_form->showfilename = strdup(filename);
        break;
      }
    default:
      return_value = CURL_FORMADD_UNKNOWN_OPTION;
    }
  }

  if(CURL_FORMADD_OK == return_value) {
    /* go through the list, check for copleteness and if everything is
     * alright add the HttpPost item otherwise set return_value accordingly */
    
    post = NULL;
    for(form = first_form;
        form != NULL;
        form = form->more) {
      if ( ((!form->name || !form->value) && !post) ||
           ( (form->contentslength) &&
             (form->flags & HTTPPOST_FILENAME) ) ||
           ( (form->flags & HTTPPOST_FILENAME) &&
             (form->flags & HTTPPOST_PTRCONTENTS) ) ||

           /* CMC: Added support for buffer uploads */
           ( (!form->buffer) &&
             (form->flags & HTTPPOST_BUFFER) &&
             (form->flags & HTTPPOST_PTRBUFFER) ) ||

           ( (form->flags & HTTPPOST_READFILE) &&
             (form->flags & HTTPPOST_PTRCONTENTS) )
           ) {
        return_value = CURL_FORMADD_INCOMPLETE;
        break;
      }
      else {
        if ( ((form->flags & HTTPPOST_FILENAME) ||
              (form->flags & HTTPPOST_BUFFER)) &&
             !form->contenttype ) {
          /* our contenttype is missing */
          form->contenttype
            = strdup(ContentTypeForFilename(form->value, prevtype));
        }
        if ( !(form->flags & HTTPPOST_PTRNAME) &&
             (form == first_form) ) {
          /* copy name (without strdup; possibly contains null characters) */
          if (AllocAndCopy(&form->name, form->namelength)) {
            return_value = CURL_FORMADD_MEMORY;
            break;
          }
        }
        if ( !(form->flags & HTTPPOST_FILENAME) &&
             !(form->flags & HTTPPOST_READFILE) && 
             !(form->flags & HTTPPOST_PTRCONTENTS) &&

             /* CMC: Added support for buffer uploads */
             !(form->flags & HTTPPOST_PTRBUFFER) ) {

          /* copy value (without strdup; possibly contains null characters) */
          if (AllocAndCopy(&form->value, form->contentslength)) {
            return_value = CURL_FORMADD_MEMORY;
            break;
          }
        }
        post = AddHttpPost(form->name, form->namelength,
                           form->value, form->contentslength,

                           /* CMC: Added support for buffer uploads */
                           form->buffer, form->bufferlength,

                           form->contenttype, form->flags,
                           form->contentheader, form->showfilename,
                           post, httppost,
                           last_post);
        
        if(!post)
          return_value = CURL_FORMADD_MEMORY;

        if (form->contenttype)
          prevtype = form->contenttype;
      }
    }
  }

  /* always delete the allocated memory before returning */
  form = first_form;
  while (form != NULL) {
    FormInfo *delete_form;
    
    delete_form = form;
    form = form->more;
    free (delete_form);
  }

  return return_value;
}

CURLFORMcode curl_formadd(struct curl_httppost **httppost,
                 struct curl_httppost **last_post,
                 ...)
{
  va_list arg;
  CURLFORMcode result;
  va_start(arg, last_post);
  result = FormAdd(httppost, last_post, arg);
  va_end(arg);
  return result;
}

static int AddFormData(struct FormData **formp,
                       const void *line,
                       long length)
{
  struct FormData *newform = (struct FormData *)
    malloc(sizeof(struct FormData));
  newform->next = NULL;

  /* we make it easier for plain strings: */
  if(!length)
    length = strlen((char *)line);

  newform->line = (char *)malloc(length+1);
  memcpy(newform->line, line, length);
  newform->length = length;
  newform->line[length]=0; /* zero terminate for easier debugging */
  
  if(*formp) {
    (*formp)->next = newform;
    *formp = newform;
  }
  else
    *formp = newform;

  return length;
}


static int AddFormDataf(struct FormData **formp,
                        const char *fmt, ...)
{
  char s[4096];
  va_list ap;
  va_start(ap, fmt);
  vsprintf(s, fmt, ap);
  va_end(ap);

  return AddFormData(formp, s, 0);
}


char *Curl_FormBoundary(void)
{
  char *retstring;
  static int randomizer=0; /* this is just so that two boundaries within
			      the same form won't be identical */
  int i;

  static char table16[]="abcdef0123456789";

  retstring = (char *)malloc(BOUNDARY_LENGTH+1);

  if(!retstring)
    return NULL; /* failed */

  srand(time(NULL)+randomizer++); /* seed */

  strcpy(retstring, "----------------------------");

  for(i=strlen(retstring); i<BOUNDARY_LENGTH; i++)
    retstring[i] = table16[rand()%16];

  /* 28 dashes and 12 hexadecimal digits makes 12^16 (184884258895036416)
     combinations */
  retstring[BOUNDARY_LENGTH]=0; /* zero terminate */

  return retstring;
}

/* Used from http.c, this cleans a built FormData linked list */ 
void Curl_formclean(struct FormData *form)
{
  struct FormData *next;

  do {
    next=form->next;  /* the following form line */
    free(form->line); /* free the line */
    free(form);       /* free the struct */
  
  } while((form=next)); /* continue */
}

/* external function to free up a whole form post chain */
void curl_formfree(struct curl_httppost *form)
{
  struct curl_httppost *next;

  if(!form)
    /* no form to free, just get out of this */
    return;

  do {
    next=form->next;  /* the following form line */

    /* recurse to sub-contents */
    if(form->more)
      curl_formfree(form->more);

    if( !(form->flags & HTTPPOST_PTRNAME) && form->name)
      free(form->name); /* free the name */
    if( !(form->flags & HTTPPOST_PTRCONTENTS) && form->contents)
      free(form->contents); /* free the contents */
    if(form->contenttype)
      free(form->contenttype); /* free the content type */
    if(form->showfilename)
      free(form->showfilename); /* free the faked file name */
    free(form);       /* free the struct */

  } while((form=next)); /* continue */
}

CURLcode Curl_getFormData(struct FormData **finalform,
                          struct curl_httppost *post,
                          int *sizep)
{
  struct FormData *form = NULL;
  struct FormData *firstform;
  struct curl_httppost *file;
  CURLcode result = CURLE_OK;

  int size =0;
  char *boundary;
  char *fileboundary=NULL;
  struct curl_slist* curList;

  *finalform=NULL; /* default form is empty */

  if(!post)
    return result; /* no input => no output! */

  boundary = Curl_FormBoundary();
  
  /* Make the first line of the output */
  AddFormDataf(&form,
               "Content-Type: multipart/form-data;"
               " boundary=%s\r\n",
               boundary);
  /* we DO NOT count that line since that'll be part of the header! */

  firstform = form;
  
  do {

    if(size)
      size += AddFormDataf(&form, "\r\n");

    /* boundary */
    size += AddFormDataf(&form, "--%s\r\n", boundary);

    size += AddFormData(&form,
                        "Content-Disposition: form-data; name=\"", 0);

    size += AddFormData(&form, post->name, post->namelength);

    size += AddFormData(&form, "\"", 0);

    if(post->more) {
      /* If used, this is a link to more file names, we must then do
         the magic to include several files with the same field name */

      fileboundary = Curl_FormBoundary();

      size += AddFormDataf(&form,
                           "\r\nContent-Type: multipart/mixed,"
                           " boundary=%s\r\n",
                           fileboundary);
    }

    file = post;

    do {

      /* If 'showfilename' is set, that is a faked name passed on to us
         to use to in the formpost. If that is not set, the actually used
         local file name should be added. */

      if(post->more) {
        /* if multiple-file */
        size += AddFormDataf(&form,
                             "\r\n--%s\r\nContent-Disposition: "
                             "attachment; filename=\"%s\"",
                             fileboundary,
                             (file->showfilename?file->showfilename:
                              file->contents));
      }
      else if((post->flags & HTTPPOST_FILENAME) ||

              /* CMC: Added support for buffer uploads */
              (post->flags & HTTPPOST_BUFFER)) {

        size += AddFormDataf(&form,
                             "; filename=\"%s\"",
                             (post->showfilename?post->showfilename:
                              post->contents));
      }
      
      if(file->contenttype) {
        /* we have a specified type */
        size += AddFormDataf(&form,
                             "\r\nContent-Type: %s",