是否有一种方法可以使用Java SE API在Java中创建一个非常基本的HTTP服务器(只支持GET/POST)，而不需要编写代码手动解析HTTP请求和手动格式化HTTP响应?Java SE API在HttpURLConnection中很好地封装了HTTP客户机功能，但是是否有类似的HTTP服务器功能呢?
需要明确的是，我在网上看到的许多ServerSocket示例的问题是，它们自己进行请求解析/响应格式化和错误处理，这很乏味，容易出错，而且不太全面，出于这些原因，我试图避免使用它。
当前回答
这是我简单的web服务器，在JMeter中用于测试webhook(这就是为什么它会在收到请求后关闭并结束自己)。
import java.io.IOException;
import java.io.PrintWriter;
import java.net.ServerSocket;
import java.net.Socket;
public class HttpServer {
private static int extractContentLength(StringBuilder sb) {
int length = 0;
String[] lines = sb.toString().split("\\n");
for (int i = 0; i < lines.length; i++) {
String s = lines[i];
if (s.toLowerCase().startsWith("Content-Length:".toLowerCase()) && i <= lines.length - 2) {
String slength = s.substring(s.indexOf(":") + 1, s.length()).trim();
length = Integer.parseInt(slength);
System.out.println("Length = " + length);
return length;
}
}
return 0;
}
public static void main(String[] args) throws IOException {
int port = Integer.parseInt(args[0]);
System.out.println("starting HTTP Server on port " + port);
StringBuilder outputString = new StringBuilder(1000);
ServerSocket serverSocket = new ServerSocket(port);
serverSocket.setSoTimeout(3 * 60 * 1000); // 3 minutes timeout
while (true) {
outputString.setLength(0); // reset buff
Socket clientSocket = serverSocket.accept(); // blocking
PrintWriter out = new PrintWriter(clientSocket.getOutputStream(), true);
try {
boolean isBodyRead = false;
int dataBuffer;
while ((dataBuffer = clientSocket.getInputStream().read()) != -1) {
if (dataBuffer == 13) { // CR
if (clientSocket.getInputStream().read() == 10) { // LF
outputString.append("\n");
}
} else {
outputString.append((char) dataBuffer);
}
// do we have Content length
int len = extractContentLength(outputString);
if (len > 0) {
int actualLength = len - 1; // we need to substract \r\n
for (int i = 0; i < actualLength; i++) {
int body = clientSocket.getInputStream().read();
outputString.append((char) body);
}
isBodyRead = true;
break;
}
} // end of reading while
if (isBodyRead) {
// response headers
out.println("HTTP/1.1 200 OK");
out.println("Connection: close");
out.println(); // must have empty line for HTTP
out.flush();
out.close(); // close clients connection
}
} catch (IOException ioEx) {
System.out.println(ioEx.getMessage());
}
System.out.println(outputString.toString());
break; // stop server - break while true
} // end of outer while true
serverSocket.close();
} // end of method
}
你可以这样测试:
curl -X POST -H "Content-Type: application/json" -H "Connection: close" -d '{"name": "gustinmi", "email": "gustinmi at google dot com "}' -v http://localhost:8081/
其他回答
httpserver解决方案不能跨jre移植。最好使用javax.xml.ws中的官方webservices API来引导一个最小的HTTP服务器…
import java.io._
import javax.xml.ws._
import javax.xml.ws.http._
import javax.xml.transform._
import javax.xml.transform.stream._
@WebServiceProvider
@ServiceMode(value=Service.Mode.PAYLOAD)
class P extends Provider[Source] {
def invoke(source: Source) = new StreamSource( new StringReader("<p>Hello There!</p>"));
}
val address = "http://127.0.0.1:8080/"
Endpoint.create(HTTPBinding.HTTP_BINDING, new P()).publish(address)
println("Service running at "+address)
println("Type [CTRL]+[C] to quit!")
Thread.sleep(Long.MaxValue)
编辑:这实际上是工作!上面的代码看起来像Groovy之类的。以下是我测试的Java翻译:
import java.io.*;
import javax.xml.ws.*;
import javax.xml.ws.http.*;
import javax.xml.transform.*;
import javax.xml.transform.stream.*;
@WebServiceProvider
@ServiceMode(value = Service.Mode.PAYLOAD)
public class Server implements Provider<Source> {
public Source invoke(Source request) {
return new StreamSource(new StringReader("<p>Hello There!</p>"));
}
public static void main(String[] args) throws InterruptedException {
String address = "http://127.0.0.1:8080/";
Endpoint.create(HTTPBinding.HTTP_BINDING, new Server()).publish(address);
System.out.println("Service running at " + address);
System.out.println("Type [CTRL]+[C] to quit!");
Thread.sleep(Long.MAX_VALUE);
}
}
我玩得很开心，我玩转了一下，拼凑出了这个。我希望这对你有所帮助。 你需要安装Gradle或者使用Maven插件。
build.gradle
plugins {
id 'application'
}
group 'foo.bar'
version '1.0'
repositories {
mavenCentral()
}
application{
mainClass.set("foo.FooServer")
}
dependencies {}
FooServer 主入口点，你的主类。
package foo;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.OutputStream;
import java.net.ServerSocket;
import java.net.Socket;
import java.nio.ByteBuffer;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ConcurrentMap;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
public class FooServer {
public static void main(String[] args) throws IOException {
ServerSocket serverSocket = new ServerSocket(7654);
serverSocket.setPerformancePreferences(0, 1, 2);
/* the higher the numbers, the better the concurrent performance, ha!
we found that a 3:7 ratio to be optimal
3 partitioned executors to 7 network executors */
ExecutorService executors = Executors.newFixedThreadPool(3);
executors.execute(new PartitionedExecutor(serverSocket));
}
public static class PartitionedExecutor implements Runnable {
ServerSocket serverSocket;
public PartitionedExecutor(ServerSocket serverSocket) {
this.serverSocket = serverSocket;
}
@Override
public void run() {
ExecutorService executors = Executors.newFixedThreadPool(30);
executors.execute(new NetworkRequestExecutor(serverSocket, executors));
}
}
public static class NetworkRequestExecutor implements Runnable{
String IGNORE_CHROME = "/favicon.ico";
String BREAK = "\r\n";
String DOUBLEBREAK = "\r\n\r\n";
Integer REQUEST_METHOD = 0;
Integer REQUEST_PATH = 1;
Integer REQUEST_VERSION = 2;
String RENDERER;
Socket socketClient;
ExecutorService executors;
ServerSocket serverSocket;
public NetworkRequestExecutor(ServerSocket serverSocket, ExecutorService executors){
this.serverSocket = serverSocket;
this.executors = executors;
}
@Override
public void run() {
try {
socketClient = serverSocket.accept();
Thread.sleep(19);//do this for safari, its a hack but safari requires something like this.
InputStream requestInputStream = socketClient.getInputStream();
OutputStream clientOutput = socketClient.getOutputStream();
if (requestInputStream.available() == 0) {
requestInputStream.close();
clientOutput.flush();
clientOutput.close();
executors.execute(new NetworkRequestExecutor(serverSocket, executors));
return;
}
ByteArrayOutputStream byteArrayOutputStream = new ByteArrayOutputStream();
ByteBuffer byteBuffer = ByteBuffer.allocate(1024);
int bytesRead;
while ((bytesRead = requestInputStream.read(byteBuffer.array())) != -1) {
byteArrayOutputStream.write(byteBuffer.array(), 0, bytesRead);
if (requestInputStream.available() == 0) break;
}
String completeRequestContent = byteArrayOutputStream.toString();
String[] requestBlocks = completeRequestContent.split(DOUBLEBREAK, 2);
String headerComponent = requestBlocks[0];
String[] methodPathComponentsLookup = headerComponent.split(BREAK);
String methodPathComponent = methodPathComponentsLookup[0];
String[] methodPathVersionComponents = methodPathComponent.split("\\s");
String requestVerb = methodPathVersionComponents[REQUEST_METHOD];
String requestPath = methodPathVersionComponents[REQUEST_PATH];
String requestVersion = methodPathVersionComponents[REQUEST_VERSION];
if (requestPath.equals(IGNORE_CHROME)) {
requestInputStream.close();
clientOutput.flush();
clientOutput.close();
executors.execute(new NetworkRequestExecutor(serverSocket, executors));
return;
}
ConcurrentMap<String, String> headers = new ConcurrentHashMap<>();
String[] headerComponents = headerComponent.split(BREAK);
for (String headerLine : headerComponents) {
String[] headerLineComponents = headerLine.split(":");
if (headerLineComponents.length == 2) {
String fieldKey = headerLineComponents[0].trim();
String content = headerLineComponents[1].trim();
headers.put(fieldKey.toLowerCase(), content);
}
}
clientOutput.write("HTTP/1.1 200 OK".getBytes());
clientOutput.write(BREAK.getBytes());
Integer bytesLength = "hi".length();
String contentLengthBytes = "Content-Length:" + bytesLength;
clientOutput.write(contentLengthBytes.getBytes());
clientOutput.write(BREAK.getBytes());
clientOutput.write("Server: foo server".getBytes());
clientOutput.write(BREAK.getBytes());
clientOutput.write("Content-Type: text/html".getBytes());
clientOutput.write(DOUBLEBREAK.getBytes());
clientOutput.write("hi".getBytes());
clientOutput.close();
socketClient.close();
executors.execute(new NetworkRequestExecutor(serverSocket, executors));
} catch (IOException ex) {
ex.printStackTrace();
} catch (InterruptedException ioException) {
ioException.printStackTrace();
}
}
}
}
运行该程序:
gradle run
浏览:
http://localhost:7654/
从Java SE 6开始，在Sun Oracle JRE中有一个内置的HTTP服务器。Java 9模块名称为jdk.httpserver。httpserver包摘要概述了涉及的类并包含示例。
这里有一个从他们的文档复制粘贴的启动示例。你可以复制，粘贴，然后在Java 6+上运行。 (尽管如此，所有试图编辑它的人，因为它是一段丑陋的代码，请不要，这是一个复制粘贴，不是我的，此外，你不应该编辑引文，除非它们在原始来源中发生了变化)
package com.stackoverflow.q3732109; import java.io.IOException; import java.io.OutputStream; import java.net.InetSocketAddress; import com.sun.net.httpserver.HttpExchange; import com.sun.net.httpserver.HttpHandler; import com.sun.net.httpserver.HttpServer; public class Test { public static void main(String[] args) throws Exception { HttpServer server = HttpServer.create(new InetSocketAddress(8000), 0); server.createContext("/test", new MyHandler()); server.setExecutor(null); // creates a default executor server.start(); } static class MyHandler implements HttpHandler { @Override public void handle(HttpExchange t) throws IOException { String response = "This is the response"; t.sendResponseHeaders(200, response.length()); OutputStream os = t.getResponseBody(); os.write(response.getBytes()); os.close(); } } }
应该注意的是，他们示例中的response.length()部分是坏的，它应该是response.getBytes().length。即使这样，getBytes()方法也必须显式地指定在响应头中指定的字符集。唉，尽管对初学者有误导，但毕竟这只是一个基本的启动示例。
执行它并访问http://localhost:8000/test，你将看到以下响应:
这是反应
As to using com.sun.* classes, do note that this is, in contrary to what some developers think, absolutely not forbidden by the well known FAQ Why Developers Should Not Write Programs That Call 'sun' Packages. That FAQ concerns the sun.* package (such as sun.misc.BASE64Encoder) for internal usage by the Oracle JRE (which would thus kill your application when you run it on a different JRE), not the com.sun.* package. Sun/Oracle also just develop software on top of the Java SE API themselves like as every other company such as Apache and so on. Moreover, this specific HttpServer must be present in every JDK so there is absolutely no means of "portability" issue like as would happen with sun.* package. Using com.sun.* classes is only discouraged (but not forbidden) when it concerns an implementation of a certain Java API, such as GlassFish (Java EE impl), Mojarra (JSF impl), Jersey (JAX-RS impl), etc.
你也可以看看一些NIO应用框架，比如:
网状的:http://jboss.org/netty Apache Mina: http://mina.apache.org/或其子项目AsyncWeb: http://mina.apache.org/asyncweb/
检出简单。它是一个非常简单的嵌入式服务器，内置了对各种操作的支持。我特别喜欢它的线程模型..
神奇的!
推荐文章
- 禁用IntelliJ星(包)导入?
- 面试问题:检查一个字符串是否是另一个字符串的旋转
- 将文件加载为InputStream的不同方法
- 什么是“升级-不安全-请求”HTTP报头?
- 到底是什么导致了堆栈溢出错误?
- 为什么Android工作室说“等待调试器”如果我不调试?
- Java:路径vs文件
- ExecutorService，如何等待所有任务完成
- Maven依赖Servlet 3.0 API?
- 如何在IntelliJ IDEA中添加目录到应用程序运行概要文件中的类路径?
- getter和setter是糟糕的设计吗?相互矛盾的建议
- Android room persistent: AppDatabase_Impl不存在
- Java的String[]在Kotlin中等价于什么?
- Intellij IDEA上的System.out.println()快捷方式
- 使用Spring RestTemplate获取JSON对象列表
