ServerSocket 文件传输丢失字节

发布于 2024-11-26 11:30:41 字数 1117 浏览 3 评论 0原文

问题

我已经通过数据包嗅探器确认来自客户端的所有数据都已完全发送，并且服务器接收的数据也完整，但是存在一个小问题。

这是服务器内客户端线程相关部分的基本表示。

this.data_in = new DataInputStream(sock.getInputStream());
this.in = new InputStreamReader(this.data_in);

服务器将使用 BufferedReader 读取身份验证数据，一旦身份验证完成，它将从套接字读取数据，如下所示。变量“fos”是 DataInputStream 数据正在写入的 FileOutputStream。

int read = 0;
byte[] buffer = new byte[8192];
while((read = data_in.read(buffer)) != -1) {
    fos.write(buffer, 0, read);
    bytes_received += read;
}

这一切在概念上似乎都是正确的，但实际上，读入的文件的前 1450 个字节丢失了。我尝试使用新的 DataInputStream(sock.getInputStream()) 以防流中的位置很奇怪，但仍然没有运气。感谢您的帮助！

更新 1

因此，我可以使用 BufferedReader 功能，我创建了一种方法来读取字符串行，就像我之前使用它们一样。请参阅下面的代码。

int input;
boolean run = true;
while((input = in.read()) != -1 && run) {
    run = true;
    char[] chars = Character.toChars(input);
    for(char c : chars) {
        if(c == '\n') {
            run = false;
            break;
        } else if(c != '\r') {
            sb.append(c);
        }
    }
}

尽管解决方案看起来很棒，但仍然缺少 1450 字节的信息。也许我在这里走错了方向？

原文

Issue

I have confirmed via a packet sniffer that all data coming from the client is being totally sent, and that the data being received by the server is complete as well, however there is a slight problem.

Here is a basic representation of relevant parts of the client thread within the server.

this.data_in = new DataInputStream(sock.getInputStream());
this.in = new InputStreamReader(this.data_in);

The server will read in authentication data using the BufferedReader, and once authentication is complete it will read data from the socket, like so. The variable 'fos' is a FileOutputStream that the DataInputStream data is being written to.

int read = 0;
byte[] buffer = new byte[8192];
while((read = data_in.read(buffer)) != -1) {
    fos.write(buffer, 0, read);
    bytes_received += read;
}

This all seems correct in concept, however, in practice the first 1450 bytes of the file read in are missing. I have tried using a new DataInputStream(sock.getInputStream()) in case the position in the stream was weird but still no luck. Thanks for any help!

Update 1

So I could use BufferedReader functionality I made a method to read in string lines as I was using them before. See below code.

int input;
boolean run = true;
while((input = in.read()) != -1 && run) {
    run = true;
    char[] chars = Character.toChars(input);
    for(char c : chars) {
        if(c == '\n') {
            run = false;
            break;
        } else if(c != '\r') {
            sb.append(c);
        }
    }
}

As great of a solution it seemed, there is still 1450 bytes of information missing. Maybe I'm going in the wrong direction here?

分享到QQ

分享到微博

如果你对这篇内容有疑问，欢迎到本站社区发帖提问参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。

发布评论

需要登录才能够评论，你可以免费注册一个本站的账号。

北城孤痞 2024-12-03 11:30:41

在我看来， BufferedReader 是问题所在 - 根据 http://ulibgcj.sourceforge.net/javadoc/java/io/BufferedReader.html 的默认缓冲区大小为 8192...无论您的情况是多少，该默认缓冲区大小都是BufferedReader 将以默认缓冲区大小为步长推进 data_in...因此，即使您不使用它们，data_in 通常也会在超出您在身份验证期间消耗的最后一个字节的位置。

回复收藏 0 原文

最偏执的依靠 2024-12-03 11:30:41

您似乎保留了两个引用同一事物的独立流/读取器：this.data_in 和 this.in。从其中任何一个读取都会推进流，然后从另一个读取将出现“错过”在另一个中读取的数据。另外，您正在使用 DataInputStream像常规的旧InputStream一样，使用 read(byte[]) 方法。 DataInputStream 是具有特定用途的专用流。