Root-me TCP返校挑战C++实现求助：结果发送失败排查

问题背景 #

• 挑战任务：通过IPv6 TCP连接到2001:bc8:35b0:c166::151:52002，解析服务器发送的数学题计算num1的平方根并乘以num2，将结果四舍五入到两位小数，在2秒内以二进制int/float类型发送给服务器（注意：不是字符串形式）
• 遇到的问题：
  • 发送字符串格式的结果时，服务器返回[!] Please only send int/float !
  • 直接发送float精度不足，发送double无响应，多数情况下服务器无反馈

当前代码的核心问题 #

你的代码错误地将计算结果转换为ASCII字符串发送，而服务器要求的是二进制格式的数值数据（即直接发送float或int类型的内存字节）。此外：

• 使用ceil实现向上取整，不符合挑战要求的四舍五入
• 发送字符串时存在拼接错误（如测试输出中出现8807.008807这类异常内容）

修复方案 #

关键改动点 #

1. 用round()函数实现正确的四舍五入到两位小数
2. 将计算结果转换为32位float类型，直接发送其二进制内存数据
3. 确保发送的字节长度为sizeof(float)，而非字符串长度

修改后的完整代码 #

// TCP - Retour au collège
#include <sys/socket.h> 
#include <iostream>
#include <cmath>
#include <cstring>
#include <netinet/in.h>
#include <unistd.h>
#include <iomanip>
#include <arpa/inet.h>
#include <string>

int main() {
    // Connect to root-me.org with tcp (this script is client side)
    int client = socket(AF_INET6, SOCK_STREAM, 0);
    if (client == -1) {
        std::cerr << "Socket creation failed" << std::endl;
        return 1;
    }
    
    // set socket address type to network type 
    struct sockaddr_in6 serv;
    std::memset(&serv, 0, sizeof(serv));
    serv.sin6_family = AF_INET6;
    serv.sin6_port = htons(52002);
    if (inet_pton(AF_INET6, "2001:bc8:35b0:c166::151", &serv.sin6_addr) != 1) {
        std::cerr << "Invalid IPv6 address" << std::endl;
        close(client);
        return 1;
    }
    
    // Connect to server
    if (connect(client, (struct sockaddr*)&serv, sizeof(serv)) == -1) {
        std::cerr << "Connection failed" << std::endl;
        close(client);
        return 1;
    }
    std::cout << "Connected to port 52002" << std::endl;
    
    //Receive data and processing
    char buffer[1024] = { 0 }; 
    ssize_t recv_len = recv(client, buffer, sizeof(buffer)-1, 0); 
    if (recv_len <= 0) {
        std::cerr << "Receive failed or server closed connection" << std::endl;
        close(client);
        return 1;
    }
    buffer[recv_len] = '\0'; // Ensure null termination
    std::cout << "Message from serv:\n" << buffer << std::endl;
    
    std::string str = buffer;
    size_t sqrt_pos = str.find("square root of ");
    size_t multiply_pos = str.find(" and multiply by ");
    size_t equal_pos = str.find(" = ");
    
    // Parse numbers with error checking
    if (sqrt_pos == std::string::npos || multiply_pos == std::string::npos || equal_pos == std::string::npos) {
        std::cerr << "Failed to parse server message" << std::endl;
        close(client);
        return 1;
    }
    
    std::string number1 = str.substr(sqrt_pos + strlen("square root of "), multiply_pos - (sqrt_pos + strlen("square root of ")));
    int number1_ = std::stoi(number1);
    std::string number2 = str.substr(multiply_pos + strlen(" and multiply by "), equal_pos - (multiply_pos + strlen(" and multiply by ")));
    int number2_ = std::stoi(number2);
    
    double res = std::sqrt(number1_) * number2_;
    std::cout << "Raw result: " << std::setprecision(15) << res << std::endl;
    
    // Correct rounding to two decimal places
    res = std::round(res * 100.0) / 100.0;
    std::cout << "Rounded result: " << std::fixed << std::setprecision(2) << res << std::endl;
    
    // Convert to float and send as binary
    float res_float = static_cast<float>(res);
    ssize_t send_len = send(client, &res_float, sizeof(res_float), 0);
    if (send_len != sizeof(res_float)) {
        std::cerr << "Failed to send full data" << std::endl;
        close(client);
        return 1;
    }
    std::cout << "Sent binary float (size: " << sizeof(res_float) << " bytes)" << std::endl;
    
    // Receive server response
    char buffer1[1024] = { 0 }; 
    ssize_t recv_len1 = recv(client, buffer1, sizeof(buffer1)-1, 0); 
    if (recv_len1 > 0) {
        buffer1[recv_len1] = '\0';
        std::cout << "\nServer response:\n" << buffer1 << std::endl;
    } else {
        std::cerr << "No response from server or connection closed" << std::endl;
    }
   
    //closing&ending
    close(client); 
    return 0;
}

改动说明 #

• 增加了错误检查：socket创建、连接、接收、解析步骤的错误处理，避免程序崩溃
• 替换ceil为round：实现标准四舍五入（ceil是向上取整，会导致结果偏大）
• 发送逻辑修改：直接发送float类型的内存数据，长度为sizeof(float)（通常4字节），而非字符串
• 确保字符串解析的正确性：调整了子串截取的计算方式，避免越界

测试验证 #

运行修改后的代码后，服务器应该会返回正确的响应（包含挑战的flag）。如果仍然无响应，检查：

• 是否在2秒内完成计算和发送（当前代码的计算逻辑足够快，不会超时）
• 网络环境是否允许访问IPv6地址
• 是否正确处理了服务器的完整消息（增加的错误检查会提示解析失败的情况）

内容的提问来源于stack exchange，提问作者Get