如何从post请求的HTTP主体中提取二进制文件内容？

Question

Currently, I try to implement one simple web server with C++ under windows. To support file upload, POST is implemented in my codes.

Now, when one binary file is uploaded through IE, two extra characters 0D and 0A is added into the end of file.

The file content before uploading

The file content after uploading

Here are some codes pieces

// receive data from http socket
char* orig_buf = buf;
while ( (cnt=recv(m_hSocket, buf, 1, 0)) > 0)
{
    if(*buf++ == '\n')
    {
        *buf = '\0';
        buflen = buf - orig_buf;
        return orig_buf;
    }
}

// save buffer to binary file
std::ofstream ofs(szFilename, std::ofstream::out | std::ofstream::trunc | std::ofstream::binary);
ofs.write(buf, buflen);

Through Fiddler , we find the 0D 0A is added by web browser.

In this specs , we know the file line ending should be added into before Boundary . However, the first question: what if there is no 0D 0A between file content and Boundary? Because if the content of binary file including 0D 0A , it is hard to distinguish whether 0D 0A is belong to the binary file or not.

Per @CodeCaster, add header of post message here

Question 2: How to extract binary file content from http body?

• By Content-Length ? or By Boundary ? with eliminating 0D 0A .

Answer 1

what if there is no 0D 0A between file content and Boundary? Because if the content of binary file including 0D 0A, it is hard to distinguish whether 0D 0A is belong to the binary file or not.

The user agent has to make sure to pick a boundary that isn't present in the payload.