如何让HTTP响应报头与正文在同一条消息中返回？

I have no HTTP-related experience and don't understand why this problem occurs. My service is written in Python, here's the response function of my web server:

'''the bytes of sss is 1024'''
def start_burn(self):
    global allburnerdata
    burnPageNumer = round(len(allburnerdata) / 1024)
    self.send_response(200)
    sss = bytearray();
    sss.append(0xF1);
    sss.append(burnPageNumer);
    self.send_header("Content-type", "application/x-www-form-urlencoded")
    self.send_header("Content-Length", str(len(sss)))
    self.end_headers()
    self.wfile.write(sss)
    return

Wireshark shows the server sends two separate messages. How can I make the HTTP headers and body return in the same message?

Alright, let's break down why you're seeing two separate messages and how to fix it.

The Root Cause #

It looks like you're using Python's built-in BaseHTTPRequestHandler (judging by the send_response, send_header, end_headers, and wfile methods). The key issue here is the end_headers() method: by default, it calls self.wfile.flush(), which immediately sends all the header data (status line + headers + trailing blank line) over the network before you write the body. Then when you call self.wfile.write(sss), the body gets sent as a second separate TCP segment.

Fix 1: Override end_headers to skip automatic flushing #

You can modify your request handler class to override the end_headers method, removing the automatic flush. Then you can manually flush after writing the body to send everything in one go:

First, add this method to your handler class:

def end_headers(self):
    # Copy the default behavior but skip the flush
    if self.request_version != 'HTTP/0.9':
        self.wfile.write(b"\r\n")
    # Remove the original self.wfile.flush() call

Then update your start_burn function to flush after writing the body:

def start_burn(self):
    global allburnerdata
    burnPageNumer = round(len(allburnerdata) / 1024)
    # Ensure burnPageNumer is within 0-255 (since bytearray.append expects this range)
    burnPageNumer = max(0, min(255, burnPageNumer))
    sss = bytearray();
    sss.append(0xF1);
    sss.append(burnPageNumer);
    self.send_response(200)
    self.send_header("Content-type", "application/x-www-form-urlencoded")
    self.send_header("Content-Length", str(len(sss)))
    self.end_headers()
    self.wfile.write(sss)
    self.wfile.flush()  # Manually flush all data at once
    return

Fix 2: Manually construct the full HTTP response #

If you prefer to avoid overriding methods, you can build the entire HTTP response (headers + body) as a single byte string and write it all at once. This bypasses the default send_response/send_header flow entirely:

def start_burn(self):
    global allburnerdata
    burnPageNumer = round(len(allburnerdata) / 1024)
    burnPageNumer = max(0, min(255, burnPageNumer))
    sss = bytearray();
    sss.append(0xF1);
    sss.append(burnPageNumer)
    
    # Build the complete HTTP response
    response_headers = (
        f"HTTP/1.1 200 OK\r\n"
        f"Content-Type: application/x-www-form-urlencoded\r\n"
        f"Content-Length: {len(sss)}\r\n"
        f"\r\n"  # Blank line separating headers and body
    ).encode('utf-8')
    
    # Combine headers and body, then write and flush
    full_response = response_headers + sss
    self.wfile.write(full_response)
    self.wfile.flush()
    return

Additional Notes #

• I added a check for burnPageNumer to ensure it's between 0-255 because bytearray.append() only accepts integers in that range—this prevents a potential ValueError if your page count is too large.
• Both fixes ensure all response data (headers + body) is buffered first, then sent in a single TCP segment when you call flush(), which will show up as one message in Wireshark.

内容的提问来源于stack exchange，提问作者BeNice