java.net.http.HttpClient 是 jdk11 中正式啓用的一個 http 工具類(其實早在 jdk9 的時候就已經存在了,只是處於孵化期),官方寓意爲想要取代 HttpURLConnection 和 Apache HttpClient 等比較古老的開發工具。java
新增的 HttpClient 截止到目前(2019年3月)爲止其實網絡資料還比較少,筆者只是根據一些博文和官方 Demo 本身摸索了一下,作了下總結。json
因爲是 jdk11 中才正式使用的工具類,距離開發者還很遙遠,因此對於源碼筆者暫不打算深挖,淺淺的理解怎麼使用就行緩存
1、HttpClient在 Apache HttpClient 中,通常會建立一個 HttpClient 對象來做爲門面。java.net.http.HttpClient 的邏輯也差很少,只是建立方式更加時髦了:cookie
//建立 builder HttpClient.Builder builder = HttpClient.newBuilder(); //鏈式調用 HttpClient client = builder //http 協議版本 1.1 或者 2 .version(HttpClient.Version.HTTP_2) //.version(HttpClient.Version.HTTP_1_1) //鏈接超時時間,單位爲毫秒 .connectTimeout(Duration.ofMillis(5000)) //.connectTimeout(Duration.ofMinutes(1)) //鏈接完成以後的轉發策略 .followRedirects(HttpClient.Redirect.NEVER) //.followRedirects(HttpClient.Redirect.ALWAYS) //指定線程池 .executor(Executors.newFixedThreadPool(5)) //認證,默認狀況下 Authenticator.getDefault() 是 null 值,會報錯 //.authenticator(Authenticator.getDefault()) //代理地址 //.proxy(ProxySelector.of(new InetSocketAddress("http://www.baidu.com", 8080))) //緩存,默認狀況下 CookieHandler.getDefault() 是 null 值,會報錯 //.cookieHandler(CookieHandler.getDefault()) //建立完成 .build();
在 builder() 方法中,最終會調用到 HttpClientImpl 的構造器,完成 HttpClient 的建立工做:網絡
//HttpClientImpl.class private HttpClientImpl(HttpClientBuilderImpl builder, SingleFacadeFactory facadeFactory) { //CLIENT_IDS 是 AtomicLong 類型的變量,使用 incrementAndGet() 方法實現自增加的 id id = CLIENT_IDS.incrementAndGet(); //記錄下存有 id 的字符串 dbgTag = "HttpClientImpl(" + id + ")"; //ssl 認證 if (builder.sslContext == null) { try { sslContext = SSLContext.getDefault(); } catch (NoSuchAlgorithmException ex) { throw new InternalError(ex); } } else { sslContext = builder.sslContext; } //線程池,沒有的話就默認建立一個 Executor ex = builder.executor; if (ex == null) { ex = Executors.newCachedThreadPool(new DefaultThreadFactory(id)); isDefaultExecutor = true; } else { isDefaultExecutor = false; } delegatingExecutor = new DelegatingExecutor(this::isSelectorThread, ex); facadeRef = new WeakReference<>(facadeFactory.createFacade(this)); //處理 http 2 的 client 類 client2 = new Http2ClientImpl(this);‘ //緩存操做 cookieHandler = builder.cookieHandler; //超時時間 connectTimeout = builder.connectTimeout; //轉發策略,默認爲 NEVER followRedirects = builder.followRedirects == null ? Redirect.NEVER : builder.followRedirects; //代理設置 this.userProxySelector = Optional.ofNullable(builder.proxy); this.proxySelector = userProxySelector .orElseGet(HttpClientImpl::getDefaultProxySelector); if (debug.on()) debug.log("proxySelector is %s (user-supplied=%s)", this.proxySelector, userProxySelector.isPresent()); //認證設置 authenticator = builder.authenticator; //設置 http 協議版本 if (builder.version == null) { version = HttpClient.Version.HTTP_2; } else { version = builder.version; } if (builder.sslParams == null) { sslParams = getDefaultParams(sslContext); } else { sslParams = builder.sslParams; } //鏈接線程池 connections = new ConnectionPool(id); connections.start(); timeouts = new TreeSet<>(); //SelectorManager 本質上是 Thread 類的封裝 //selmgr 會開啓一條線程,HttpClient 的主要邏輯運行在此線程中 //因此說 HttpClient 是非阻塞的,由於並不跑在主線程中 try { selmgr = new SelectorManager(this); } catch (IOException e) { throw new InternalError(e); } //設置爲守護線程 selmgr.setDaemon(true); filters = new FilterFactory(); initFilters(); assert facadeRef.get() != null; }
主要是一些儲存操做,大體理解便可,不細究。app
HttpRequest 是發起請求的主體配置:異步
//建立 builder HttpRequest.Builder reBuilder = HttpRequest.newBuilder(); //鏈式調用 HttpRequest request = reBuilder //存入消息頭 //消息頭是保存在一張 TreeMap 裏的 .header("Content-Type", "application/json") //http 協議版本 .version(HttpClient.Version.HTTP_2) //url 地址 .uri(URI.create("http://openjdk.java.net/")) //超時時間 .timeout(Duration.ofMillis(5009)) //發起一個 post 消息,須要存入一個消息體 .POST(HttpRequest.BodyPublishers.ofString("hello")) //發起一個 get 消息,get 不須要消息體 //.GET() //method(...) 方法是 POST(...) 和 GET(...) 方法的底層,效果同樣 //.method("POST",HttpRequest.BodyPublishers.ofString("hello")) //建立完成 .build();
發起請求:工具
HttpResponse<String> response = client.send(request, HttpResponse.BodyHandlers.ofString());
這是同步式的發起請求方式,先來看一下它的實現:
public <T> HttpResponse<T> send(HttpRequest req, BodyHandler<T> responseHandler) throws IOException, InterruptedException{ CompletableFuture<HttpResponse<T>> cf = null; try { //調用 sendAsync(...) 方法異步地完成主邏輯,並獲取 Future cf = sendAsync(req, responseHandler, null, null); return cf.get(); //這以後的全部代碼都是在進行異常捕捉,因此能夠忽略 } catch (InterruptedException ie) { if (cf != null ) cf.cancel(true); throw ie; } catch (ExecutionException e) { final Throwable throwable = e.getCause(); final String msg = throwable.getMessage(); if (throwable instanceof IllegalArgumentException) { throw new IllegalArgumentException(msg, throwable); } else if (throwable instanceof SecurityException) { throw new SecurityException(msg, throwable); } else if (throwable instanceof HttpConnectTimeoutException) { HttpConnectTimeoutException hcte = new HttpConnectTimeoutException(msg); hcte.initCause(throwable); throw hcte; } else if (throwable instanceof HttpTimeoutException) { throw new HttpTimeoutException(msg); } else if (throwable instanceof ConnectException) { ConnectException ce = new ConnectException(msg); ce.initCause(throwable); throw ce; } else if (throwable instanceof IOException) { throw new IOException(msg, throwable); } else { throw new IOException(msg, throwable); } } }
本質上是使用了異步實現方法 sendAsync(…)。post
在 Demo 中也能夠直接使用:開發工具
//返回的是 future,而後經過 future 來獲取結果 CompletableFuture<String> future = client.sendAsync(request, HttpResponse.BodyHandlers.ofString()) .thenApply(HttpResponse::body); //阻塞線程,從 future 中獲取結果 String body = future.get();