大家都说google protobuf很好用,这两天有时间研究了一下,果然很满意;
才开始研究的时候,始终在想一个问题:protobuf里面没有消息头,包长度,如果我有多个消息类型的话,发送方还好说,但是到了接收方,该如何识别接收到的是哪个消息类型呢?苦苦思考良久,发现其实是我的思路原本就不对,其实在发送的时候,把消息头和消息体不要放在一个message里,分开序列化就ok了,费话就不多说了,测试代码如下:
OPlayVideoMsg playVideMsg = OPlayVideoMsg.newBuilder().setPlayurl("http://111.3pg").setPlayspeed(1).build(); byte[] bodybytes = playVideMsg.toByteArray(); OMsgHead head = OMsgHead.newBuilder().setMsgtype(1).setMsglen(bodybytes.length).setMsgseq(2).build(); byte[] headerbytes = head.toByteArray(); ByteArrayOutputStream bos = null; DataOutputStream dos = null; try { bos = new ByteArrayOutputStream(); dos = new DataOutputStream(bos); dos.writeInt(headerbytes.length); dos.write(headerbytes); dos.write(playVideMsg.toByteArray()); // 解析罗 byte[] bytes = bos.toByteArray(); ByteArrayInputStream bis = null; DataInputStream dis = null; byte[] headbytes; try { bis = new ByteArrayInputStream(bytes); dis = new DataInputStream(bis); int headSize = dis.readInt(); System.out.println(String.format("headSize=%d", headSize)); headbytes = new byte[headSize]; dis.read(headbytes); OMsgHead header = OMsgHead.parseFrom(headbytes); System.out.println("header======>" + header); byte[] parsebodybytes = new byte[header.getMsglen()]; dis.readFully(parsebodybytes); if (header.getMsgtype() == 1) { OPlayVideoMsg playVideoMsg = OPlayVideoMsg.parseFrom(parsebodybytes); System.out.println("OPlayVideoMsg======>" + playVideoMsg); } } finally { if (dis != null) dis.close(); if (bis != null) bis.close(); } } finally { if (dos != null) dos.close(); if (bos != null) bos.close(); }
看到35行没,只要解析出头来,就可以继续序列化相应的消息体。
附上我的proto文件:package msginfo;
option java_package = "com.example.tutorial"; message OMsgHead { required int32 msgtype = 1; required int32 msgseq = 2; required int32 msglen = 3; } message OPlayVideoMsg { required string playurl = 2; required int32 playspeed = 3; } message OPlayImageMsg { required string playurl = 2; }