Netty的基本使用

Netty的基本使用

目录

[TOC]

引用

    implementation 'io.netty:netty-all:4.1.6.Final'

Netty 4最后一个release版本是4.1.28,但是亲测这个版本启动就会报异常...........所以降级回4.1.6

使用

TCP

服务端

    EventLoopGroup boss = new NioEventLoopGroup();
    EventLoopGroup worker = new NioEventLoopGroup();
            try {
        ServerBootstrap bootstrap = new ServerBootstrap();
        bootstrap.group(boss, worker);
        bootstrap.channel(NioServerSocketChannel.class);
        bootstrap.option(ChannelOption.TCP_NODELAY, true);
        bootstrap.childOption(ChannelOption.SO_KEEPALIVE, true);
        bootstrap.childHandler(new ChannelInitializer() {
            @Override
            protected void initChannel(SocketChannel socketChannel) throws Exception {
                ChannelPipeline p = socketChannel.pipeline();
                p.addLast("framedecoder", new LengthFieldBasedFrameDecoder(1024 * 1024 * 10, 0, 4, 0, 4));
                p.addLast(new NettyServerHandler());
            }
        });
        mChannelFuture = bootstrap.bind(mPort).sync();
        if (mChannelFuture.isSuccess()) {
            System.out.printf("start tcp server succeed,mPort : " + mPort);
        }
        mChannelFuture.channel().closeFuture().sync();
    } catch (Exception e) {
        System.out.printf("a exception happened while tcp server running");
        e.printStackTrace();
    } finally {
        boss.shutdownGracefully();
        worker.shutdownGracefully();
        mIsRunning = false;
    }

其中

mChannelFuture.channel().closeFuture().sync();

会阻塞线程,所以这段代码应该放在一个线程中.

Channel是连接服务端和客户端的通道.

ChannelPipeline p = socketChannel.pipeline();

ChannelPipeline是一个处理器容器,当有新数据接收时会调用容器中的处理器.

NettyServerHandler是处理最终接收到数据的地方

    private class NettyServerHandler extends ChannelInboundHandlerAdapter {

        @Override
        public void channelActive(ChannelHandlerContext ctx) throws Exception {
            initTransport(ctx.channel());
        }

        @Override
        public void channelRead(ChannelHandlerContext ctx, Object msg) throws Exception {
            WrapPacket packet = SocketUtils.decodeWrapPacketFromByteBufAfterSplit((ByteBuf) msg);
            String channelId = ctx.channel().id().asLongText();
            SubTransport transport = mChannelTransportMap.get(channelId);
            if (transport == null) {
                throw new RuntimeException("can not get transport form map , create a transport and put to map");
            }
            transport.onNewPacketReceived(packet);
        }

    }

LengthFieldBasedFrameDecoder是基于包长度的包解析器,用来拆包和处理半包.还有其他的自带解析器,比如分割,固定包长度等,不过个人偏向于喜欢这个,这个可以将资源利用最大化,而且没有字符不能使用的限制,如果是分割符的解析器,则会占用掉一些字符组合不能使用.当然这个也会翻车,当别人恶意攻击时,或者协议设计有问题时,一个错了,那就接下来所有都错了.错一步,满盘皆输,也是渗人.当然也可以自定义包解析器喽.

客户端

    EventLoopGroup eventLoopGroup = new NioEventLoopGroup();
            try {
        Bootstrap bootstrap = new Bootstrap();
        bootstrap.channel(NioSocketChannel.class);
        bootstrap.option(ChannelOption.SO_KEEPALIVE, true);
        bootstrap.group(eventLoopGroup);
        bootstrap.remoteAddress(mAddress, mPort);
        bootstrap.handler(new ChannelInitializer() {
            @Override
            protected void initChannel(SocketChannel socketChannel) throws Exception {
                socketChannel.pipeline().addLast("framedecoder", new LengthFieldBasedFrameDecoder(1024 * 1024 * 10, 0, 4, 0, 0));
                socketChannel.pipeline().addLast(new NettyClientHandler());
            }
        });
        mChannelFuture = bootstrap.connect(mAddress, mPort).sync();
        if (mChannelFuture.isSuccess()) {
            Log.d("connect to mAddress : " + mAddress + ", mPort : " + mPort + " succeed");
        }
        mChannelFuture.channel().closeFuture().sync();
    } catch (InterruptedException e) {
        e.printStackTrace();
        Log.d("can not connect to mAddress : " + mAddress + " ,mPort : " + mPort);
    } finally {
        eventLoopGroup.shutdownGracefully();
    }

同样也是阻塞的,和服务端差不多,便不再说明.

未完待续 ......