2016-02-29 26 views
-2

我有一個用C語言編寫的簡單代理服務器程序。我遇到了一些數據不一致的問題。當我請求網站時,我將結果數據保存在服務器端的文件中,然後將其發送到客戶端並保存在客戶端。客戶端和服務器上的結果將具有不同的大小,並且似乎某些HTML正在被複制。通常情況下,保存在服務器上的文件將小於客戶端上保存的文件,儘管這兩個文件仍然大於實際網頁(即,如果我要右鍵單擊該頁面並「另存爲」,則生成的頁面比我的代碼返回的小)。我已經嘗試了各種方法來解決這個問題,似乎沒有任何工作。結果甚至在同一個網站上嘗試不同。例如,我可以請求同一網站兩次,但文件大小不同於兩次嘗試。在非常罕見的情況下,特別是在小型網站上,客戶端和服務器程序都會返回正確的網頁,並且這兩個文件都是合適的大小。C - 來自套接字的數據不一致

注意:我知道代碼仍然相當混亂。在我進一步行動之前,我更擔心解決這個問題。一旦此問題得到解決,我會解決問題(例如檢查套接字未能打開),因此請僅解決我已列出的問題。

Server.c

#include <sys/types.h> 
#include <sys/socket.h> 
#include <netdb.h> 
#include <stdio.h> 
#include <string.h> 
#include <arpa/inet.h> 
#include <regex.h> 
#include <time.h> 

/*Code used to resolve URL into IP address adapted from following URL: 
http://www.binarytides.com/hostname-to-ip-address-c-sockets-linux/ 
*/ 

//This code has been adapted from the server code provided in class 

int main(int argc, char** argv) 
{ 
    char str[655360]; 
    char recvline[655360]; 
    char parsedRecv[655360]; 
    char domain[1025]; 
    char directory[1025]; 
    char absoluteURL[1025]; 
    char temp[1025]; 


    int listen_fd, conn_fd, n, tempCount; 
    struct sockaddr_in servaddr; 
    int bytesRead; 
    int stringCounter; 

    int port; 

    FILE *fp; 

    //Variables used for second socket and resolving of host 
    char ip[100]; 
    int sockfd, secondSocketCount; 
    int len = sizeof(struct sockaddr); 
    struct sockaddr_in secondServaddr; 
    struct addrinfo *servinfo, *p; 
    struct sockaddr_in *h; 
    int rv; 
    char simpleMatch[10]; 
    int flag = 0; 
    //End 

    //Used for HTTP GET request 
    char request[2049]; 

    listen_fd = socket(AF_INET, SOCK_STREAM, 0); 

    bzero(&servaddr, sizeof(servaddr)); 

    servaddr.sin_family = AF_INET; 
    servaddr.sin_addr.s_addr = htons(INADDR_ANY); 
    if(argc < 2) 
    { 
     printf("Error! Enter a port number to run this server on.\n\tEx: ./server 22000\n\r\0"); 
     return 0; 
    } 
    else 
    { 
     port = atoi(argv[1]); 
    } 

    servaddr.sin_port = htons(port); 

    printf("\n"); 
    printf("Awaiting connections...\n"); 

    bind(listen_fd, (struct sockaddr*)&servaddr, sizeof(servaddr)); 
    listen(listen_fd, 10); 

    //Once the server is listening, enter an infinite loop to keep listening 
    while(1) 
    { 
     conn_fd = accept(listen_fd, (struct sockaddr*) NULL, NULL); 
     bytesRead = read(conn_fd, recvline, sizeof(recvline)); 
     if(bytesRead > 0) //data read 
     { 
      recvline[bytesRead] = '\0'; 

      bzero(absoluteURL, 1025); 
      strcpy(absoluteURL, recvline); 

      //Extract host and page from recvline 
      //For loop used to check if URL begins with HTTP or HTTPS 
      for(stringCounter = 0; stringCounter < 5; stringCounter++) 
      { 
       simpleMatch[stringCounter] = tolower(recvline[stringCounter]); 
      } 
      simpleMatch[strlen(simpleMatch)] = '\0'; 
      if(strcmp("http:", simpleMatch) == 0) 
      { 
       for(stringCounter = 7, tempCount = 0; stringCounter < strlen(recvline); stringCounter++, tempCount++) 
       { 
        temp[tempCount] = recvline[stringCounter]; 
       } 
       temp[strlen(temp)] = '\0'; 
       strcpy(recvline, temp); 
      } 
      else if(strcmp("https", simpleMatch) == 0) 
      { 
       for(stringCounter = 8, tempCount = 0; stringCounter < strlen(recvline); stringCounter++, tempCount++) 
       { 
        temp[tempCount] = recvline[stringCounter]; 
       } 
       temp[strlen(temp)] = '\0'; 
       strcpy(recvline, temp); 
      } 

      //printf("\n\nAfter stripping HTTP, we are left with: %s\n\n", recvline); 

      //Now that HTTP:// or HTTPS:// has been stripped, can parse for domain 
      for(stringCounter = 0, tempCount = 0; stringCounter < strlen(recvline); stringCounter++) 
      { 
       //moving domain into the domain string 
       if(flag == 0) 
       { 
        if(recvline[stringCounter] != '/') 
        { 
         domain[stringCounter] = recvline[stringCounter]; 
        } 
        else 
        { 
         domain[stringCounter + 1] = '\0'; 
         //directory[tempCount] = recvline[stringCounter]; 
         flag = 1; 
         tempCount++; 
        } 
       } 
       else 
       { 
        directory[tempCount] = recvline[stringCounter]; 
        tempCount++; 
       } 
      } 

      //printf("\n\nDirectory is: %s\n\n", directory); 

      //reset flag and append '\0' to directory and domain 
      flag = 0; 
      if(tempCount < 1025) 
      { 
       directory[tempCount] = '\0'; 
      } 
      //directory[strlen(directory)] = '\0'; 
      //domain[strlen(domain)] = '\0'; 

      //Done extracting 

      //Resolve hostname to IP 

      if((rv = getaddrinfo(domain, NULL, NULL, &servinfo)) != 0) 
      { 
       printf("Error: an IP address cannot be resolved for %s\n", domain); 
       return 0; 
       //fprintf(stderr, "getaddrinfo: %s\n", gai_strerror(rv)); 
      } 
      else 
      { 
       for(p = servinfo; p != NULL; p = p->ai_next) 
       { 
        h = (struct sockaddr_in *) p->ai_addr; 
        strcpy(ip, inet_ntoa(h->sin_addr)); 
       } 

       freeaddrinfo (servinfo); 

       printf("%s resolved to: %s\n", domain, ip); 
      } 
      //End Resolve 

      //Now that the IP is resolved, open a socket and connect to the IP 

      //Open socket 
      sockfd = socket(AF_INET, SOCK_STREAM, 0); 
      bzero(&secondServaddr, sizeof(secondServaddr)); 

      secondServaddr.sin_family = AF_INET; 
      secondServaddr.sin_port = htons(80); 

      inet_pton(AF_INET, ip, &(secondServaddr.sin_addr)); //IP taken from earlier resolution 
      connect(sockfd, (struct sockaddr*) &secondServaddr, sizeof(secondServaddr)); 

      //socket is open, can create and send request, finally 
      bzero(request, 2049); 
      //sprintf(request, "GET %s HTTP/1.1\r\nHost: %s\r\n\r\n", directory, domain); 
      //sprintf(request, "GET %s HTTP/1.1\r\n\r\n", absoluteURL); 
      //sprintf(request, "GET %s HTTP/1.1\r\nHost: %s\r\n\r\n", absoluteURL, domain); 
      sprintf(request, "GET /%s HTTP/1.1\r\nHost: %s\r\n\r\n", directory, domain); 

      write(sockfd, request, strlen(request)); 

      printf("\tAttempting to retrieve data: this may be slow.\n"); 

      bzero(recvline, 655360); 
      bzero(parsedRecv, 655360); 

      //Old method used to retrieve data 
      //This was changed when I began to run into issues 
      /*while(1) 
      { 
       secondSocketCount = read(sockfd, parsedRecv, sizeof(parsedRecv)); 
       if(secondSocketCount == -1) 
       { 
        printf("Error receiving data: server terminating.\n"); 
        return 0; 
       } 
       else if(secondSocketCount == 0) 
       { 
        //no more data 
        break; 
       } 
       strcat(recvline, parsedRecv); 
      }*/ 

      //This while loop is used to read in data (the response from the server) 
      bzero(str, 655360); 
      while(secondSocketCount = read(sockfd, recvline, sizeof(recvline)) > 0) 
      { 
       strcat(str, recvline); 
      } 
      //bzero(parsedRecv, 655360); 
      //recvline[strlen(recvline)] = '\0'; 

      printf("\tData retrieved from main server.\n"); 

      //This for loop finds the end of the HTTP header and copies everything after into parsedRecv 
      for(stringCounter = 0, tempCount = 0; stringCounter < strlen(str); stringCounter++) 
       { 
        //lazy if statement to find two \r\n in a row to mark the end of the header 
        if(str[stringCounter] == '\r' && str[stringCounter + 1] == '\n' && str[stringCounter + 2] == '\r' && str[stringCounter + 3] == '\n' && flag == 0) 
        { 
         flag = 1; 
         stringCounter += 3; 
        } 
        if(flag == 1) 
        { 
         parsedRecv[tempCount] = str[stringCounter]; 
         tempCount++; 
        } 
       } 
      flag = 0; 
      parsedRecv[strlen(parsedRecv)] = '\0'; 
      fp = fopen("ReturnedPageServer.html", "w"); 
      if(fp != NULL) 
      { 
       fprintf(fp, "%s", parsedRecv); 
       //fprintf(fp, "%s", recvline); 
      } 
      fclose(fp); 
      printf("\tData saved to ReturnedPageServer.html\n"); 

     } 

     //strcpy(str, "This is a test of the Hello World Broadcast System.\n"); 
     bzero(str, 655360); 
     strcpy(str, parsedRecv); 
     write(conn_fd, str, strlen(str)); 
     close(conn_fd); 
     printf("\tData sent to client.\n"); 
     printf("Awaiting further connections...\n"); 

     //strcpy(directory, ""); 
     //strcpy(domain, ""); 
     //strcpy(recvline, ""); 
     bzero(directory, 1025); 
     bzero(domain, 1025); 
     bzero(temp, 1025); 
     bzero(recvline, 655360); 
    } 
    return 0; 
} 

Client.c

#include <sys/types.h> 
#include <sys/socket.h> 
#include <netdb.h> 
#include <stdio.h> 
#include <stdlib.h> 
#include <string.h> 

//Code adapted from client code provided in class 

int main(int argc, char** argv) 
{ 
    int sockfd, n, port; 
    int len = sizeof(struct sockaddr); 
    char sendline[10000]; 
    char recvline[655360]; 
    struct sockaddr_in servaddr; 

    FILE *fp; 

    sockfd = socket(AF_INET, SOCK_STREAM, 0); 
    bzero(&servaddr, sizeof(servaddr)); 

    servaddr.sin_family = AF_INET; 
    if(argc < 2) 
    { 
     printf("Error! Enter the port number for the server.\n\tEx: ./client 22000\n\r\0"); 
     return 0; 
    } 
    else 
    { 
     port = atoi(argv[1]); 
    } 
    servaddr.sin_port = htons(port); 

    inet_pton(AF_INET, "129.120.151.94", &(servaddr.sin_addr)); //CSE01 IP 
    connect(sockfd, (struct sockaddr*) &servaddr, sizeof(servaddr)); 

    printf("url: "); 
    scanf("%s", sendline); 

    //strcpy(sendline, "The server should display this text.\n\0"); 
    //printf("\nLength of string: %d\n", strlen(sendline)); 
    //printf("\t%s\n", sendline); 
    write(sockfd, sendline, strlen(sendline)); 

    fp = fopen("ReturnedPageClient.html", "w"); 
    bzero(recvline, 655360); 
    while(n = read(sockfd, recvline, sizeof(recvline)) > 0) 
    { 
     //printf("%s", recvline); 
     if(fp != NULL) 
     { 
      fprintf(fp, "%s", recvline); 
     } 
     else 
     { 
      printf("\tError saving file: client terminating.\n"); 
      fclose(fp); 
      return 0; 
     } 

    } 
    fclose(fp); 
    printf("\tResponse received from proxy server.\n\tFile saved as \"ReturnedPageClient.html\"\n"); 
    close(sockfd); 
    return 0; 
} 
+3

'simpleMatch [strlen(simpleMatch)] ='\ 0';'不會終止字符串! – chqrlie

+2

您對非C風格字符串的東西使用'strlen'。如果他們是C風格的字符串,他們已經終止了。如果它們不是C風格的字符串,則不能將它們傳遞給'strlen'。 –

+0

語法錯誤:'while(n = read(sockfd,recvline,sizeof(recvline))> 0)' – chqrlie

回答

0

要排序的客戶端和服務器與插座之間的通信,需要某種形式的協議,以便在連接每一方都可以確定當一條消息完成時。通過TCP套接字發送的消息可以以不同於發送方寫入的大小的塊接收。

你不會在你的程序中考慮到這一點,你假設一個成功的read請求返回發送端的write請求所寫的數據。這是不正確的:您應該繼續讀取套接字並將讀取的塊存儲到請求緩衝區中,直到完成請求或套接字關閉。

一個簡單的協議就是一次寫一行。使用\n字符排序請求和答案。這或多或少是HTTP,SMTP和POP協議所做的。

代碼諸多問題:

  • 它具有防止編譯語法錯誤。

  • 您將非常大的緩衝區定義爲自動變量,超過2MB時,您可能會在某些系統上發生堆棧溢出。

  • 由於您沒有傳輸終止符,因此讀入緩衝區的數據應該手動終止。你似乎意識到這個問題,但你的方法parsedRecv[strlen(parsedRecv)] = '\0';沒有做任何有用的事情。根據定義,strlen(parsedRecv)'\0'字節的偏移量。如果字符串未正確空終止,則strlen將在緩衝區的末尾掃描並調用未定義的行爲。您應該在成功read後手動設置'\0'字節與recvline[n] = '\0';

  • 你嘗試和匹配字符串片段與strcmp:這是不是最有效的,因爲你需要先將片段複製到一個單獨的緩衝和空終止它,以便strcmp可以比較,作爲一個完整的字符串。使用memcmp代替指定的片段長度。

2

一個問題 - 不確定是否它的'問題' - 是你期待TCP以消息爲導向。即您發送500字節的消息,並期望收到一個500字節的數據塊。這不是TCP的工作原理。 TCP是面向數據流的,500字節的發送可以在服務器上讀取250個2字節,或者1,500字節或100,50,100,2,2,2,2,2,240。您必須在服務器上循環直到你收到所有的'消息'。這會導致不一致的行爲,特別是事情會在本地工作,但不會在「真實」網絡上工作。

這反過來引發瞭如何知道您已收到整個「消息」的問題。你需要有一些更高級別的協議,允許消息幀(例如發送一個固定長度的長度然後發送)

+0

難道我們沒有一個好的重複嗎?這是今天同樣性質的第三個問題。 – SergeyA

相關問題