Simply test the two different data transmission methods of content length and chunked

Content-Length： Data transmission according to data length

Chunk： Data transmission by block

The sample request shown in the following figure , It's just at this time Content-length Is in the response header

---

One 、 Code testing

This article will use ApiPost Test two different methods of carrying data

1、 Preparation

1、 Prepare one ApiPost（ Their own use postman When testing , Unable to test successfully ）

2、 Write the corresponding test class

public class TestServerSocket {
    
	public static void main(String[] args) throws IOException {
    
		ServerSocket socket = new ServerSocket(8089);
		Socket accept = socket.accept();


		InputStream inputStream = accept.getInputStream();

		while (true) {
    
			byte[] bytes = new byte[100000];
			int readLength = inputStream.read(bytes);
			System.out.println(new String(bytes, 0, readLength));
			System.out.println("---- Complete a data reception over----");

		}
	}
}

2、 test Content-length

1、 Send a message to 8089 port , The carried parameter is name and age

2、 View the data information of the console

We will find that the default parameter carried is content-length, And the length is 269 Bytes .

3、 test Transfer-Encoding：chunked

1、 modify ApiPost Request header for , Add corresponding Transfer-Encoding Parameters

2、 The parameters carried and remain unchanged , Again using ApiPost test , View console data

At this time, the requested parameter information , It includes Content-length, Also contains the Transfer-Encoding：chunked. But the data below seems different , Use chunked after , The data consists of the previous single line string mobian, It becomes multi row data . And mobian Ahead 6 just mobian Length of string ,23 Ahead 2 just 23 The length of , Finally 0 ending .

summary ：

namely content-length It is transmitted based on the total length of data , and chunked It is transmitted in blocks , Each piece is internally transmitted according to the previous length , Until the length is 0 after , Stop data transfer . If two exist at the same time , be content-length invalid .

---

Two 、tomcat Source code

tomcat Chinese vs Content-length and Transfer-Encoding Data processing of .

1、Request The process

A request is in tomcat The internal call logic of is not repeated here , For those who don't know, please refer to my previous blog ： Talking about Tomcat The internal execution process after receiving a request （ Source code ）

It will eventually call prepareRequest() Method

1、prepareRequest Method until after the position in the figure below

1. First, I will get transfer-encoding attribute
2. Then according to the parameters configured later , Pass to addInputFilter Method

2、 If the following parameter is chunked, It will be contentDelimitation Set to true

3、 The analysis is finished transfer-encoding after

1. First of all get contentLength
2. When the content length is greater than 0 after , Will judge contentDelimitation The state of , If we configure it chunked Then change the value to true
3. contentDelimitation by true after , Will remove content-length node , That is, both exist at the same time ,contentlength invalid

2、 other

If you want to know contentlength and chunked There are two ways to read the logic code of data , You can check InputFilter Two implementation classes of the interface , namely IdentityInputFilter（contentlength） and ChunkedInputFilter（chunked） Two kinds of doRead Method .