在 Cocos Creator 中使用 protobuf.js 库可以方便地进行协议的序列化和反序列化。
下面是使用 protobuf.js 的详细说明:
一、protobuf环境安装
1、安装 npm
protobuf环境安装安装需要使用 npm 命令进行,因此首先需要安装 npm 。
如果你还没安装 npm , 请参考:windows安装npm教程 。
2、全局安装pbjs
打开命令行工具,输入以下命令
npm install -g pbjs
安装成功后 C:\Users\Administrator\AppData\Roaming\npm 显示如下信息:
二、 Cocos Creator 中使用 protobuf.js 库
确定了模块格式和导出方式,就可以在脚本资源里使用 protobufjs
这个模块。
1、创建协议文件
在项目中创建一个 .proto
文件,定义协议结构
例如,创建一个名为 message.proto
的文件,内容如下:
syntax = "proto3"; message Player { string name = 1; int32 level = 2; } ```
这里定义了一个名为 Player
的消息类型,包含 name
和 level
两个字段。
2、将 protobuf 中的 proto 文件编译成TypeScript 文件
pbjs 指令的使用可以查看 官方文档:pbjs地址
我们以打包上面生成的 message.proto
文件为例,进行演示。
命令行执行:
pbjs D:/test/3/proto/message.proto --ts D:/test/3/proto/message.ts
打包完成后,在 D:/test/3/proto/ 目录,生成了 message.ts 文件,如下:
message.ts 文件内容如下:
export interface Player { name?: string; level?: number; } export function encodePlayer(message: Player): Uint8Array { let bb = popByteBuffer(); _encodePlayer(message, bb); return toUint8Array(bb); } function _encodePlayer(message: Player, bb: ByteBuffer): void { // optional string name = 1; let $name = message.name; if ($name !== undefined) { writeVarint32(bb, 10); writeString(bb, $name); } // optional int32 level = 2; let $level = message.level; if ($level !== undefined) { writeVarint32(bb, 16); writeVarint64(bb, intToLong($level)); } } export function decodePlayer(binary: Uint8Array): Player { return _decodePlayer(wrapByteBuffer(binary)); } function _decodePlayer(bb: ByteBuffer): Player { let message: Player = {} as any; end_of_message: while (!isAtEnd(bb)) { let tag = readVarint32(bb); switch (tag >>> 3) { case 0: break end_of_message; // optional string name = 1; case 1: { message.name = readString(bb, readVarint32(bb)); break; } // optional int32 level = 2; case 2: { message.level = readVarint32(bb); break; } default: skipUnknownField(bb, tag & 7); } } return message; } export interface Long { low: number; high: number; unsigned: boolean; } interface ByteBuffer { bytes: Uint8Array; offset: number; limit: number; } function pushTemporaryLength(bb: ByteBuffer): number { let length = readVarint32(bb); let limit = bb.limit; bb.limit = bb.offset + length; return limit; } function skipUnknownField(bb: ByteBuffer, type: number): void { switch (type) { case 0: while (readByte(bb) & 0x80) { } break; case 2: skip(bb, readVarint32(bb)); break; case 5: skip(bb, 4); break; case 1: skip(bb, 8); break; default: throw new Error("Unimplemented type: " + type); } } function stringToLong(value: string): Long { return { low: value.charCodeAt(0) | (value.charCodeAt(1) << 16), high: value.charCodeAt(2) | (value.charCodeAt(3) << 16), unsigned: false, }; } function longToString(value: Long): string { let low = value.low; let high = value.high; return String.fromCharCode( low & 0xFFFF, low >>> 16, high & 0xFFFF, high >>> 16); } // The code below was modified from https://github.com/protobufjs/bytebuffer.js // which is under the Apache License 2.0. let f32 = new Float32Array(1); let f32_u8 = new Uint8Array(f32.buffer); let f64 = new Float64Array(1); let f64_u8 = new Uint8Array(f64.buffer); function intToLong(value: number): Long { value |= 0; return { low: value, high: value >> 31, unsigned: value >= 0, }; } let bbStack: ByteBuffer[] = []; function popByteBuffer(): ByteBuffer { const bb = bbStack.pop(); if (!bb) return { bytes: new Uint8Array(64), offset: 0, limit: 0 }; bb.offset = bb.limit = 0; return bb; } function pushByteBuffer(bb: ByteBuffer): void { bbStack.push(bb); } function wrapByteBuffer(bytes: Uint8Array): ByteBuffer { return { bytes, offset: 0, limit: bytes.length }; } function toUint8Array(bb: ByteBuffer): Uint8Array { let bytes = bb.bytes; let limit = bb.limit; return bytes.length === limit ? bytes : bytes.subarray(0, limit); } function skip(bb: ByteBuffer, offset: number): void { if (bb.offset + offset > bb.limit) { throw new Error('Skip past limit'); } bb.offset += offset; } function isAtEnd(bb: ByteBuffer): boolean { return bb.offset >= bb.limit; } function grow(bb: ByteBuffer, count: number): number { let bytes = bb.bytes; let offset = bb.offset; let limit = bb.limit; let finalOffset = offset + count; if (finalOffset > bytes.length) { let newBytes = new Uint8Array(finalOffset * 2); newBytes.set(bytes); bb.bytes = newBytes; } bb.offset = finalOffset; if (finalOffset > limit) { bb.limit = finalOffset; } return offset; } function advance(bb: ByteBuffer, count: number): number { let offset = bb.offset; if (offset + count > bb.limit) { throw new Error('Read past limit'); } bb.offset += count; return offset; } function readBytes(bb: ByteBuffer, count: number): Uint8Array { let offset = advance(bb, count); return bb.bytes.subarray(offset, offset + count); } function writeBytes(bb: ByteBuffer, buffer: Uint8Array): void { let offset = grow(bb, buffer.length); bb.bytes.set(buffer, offset); } function readString(bb: ByteBuffer, count: number): string { // Sadly a hand-coded UTF8 decoder is much faster than subarray+TextDecoder in V8 let offset = advance(bb, count); let fromCharCode = String.fromCharCode; let bytes = bb.bytes; let invalid = '\uFFFD'; let text = ''; for (let i = 0; i < count; i++) { let c1 = bytes[i + offset], c2: number, c3: number, c4: number, c: number; // 1 byte if ((c1 & 0x80) === 0) { text += fromCharCode(c1); } // 2 bytes else if ((c1 & 0xE0) === 0xC0) { if (i + 1 >= count) text += invalid; else { c2 = bytes[i + offset + 1]; if ((c2 & 0xC0) !== 0x80) text += invalid; else { c = ((c1 & 0x1F) << 6) | (c2 & 0x3F); if (c < 0x80) text += invalid; else { text += fromCharCode(c); i++; } } } } // 3 bytes else if ((c1 & 0xF0) == 0xE0) { if (i + 2 >= count) text += invalid; else { c2 = bytes[i + offset + 1]; c3 = bytes[i + offset + 2]; if (((c2 | (c3 << 8)) & 0xC0C0) !== 0x8080) text += invalid; else { c = ((c1 & 0x0F) << 12) | ((c2 & 0x3F) << 6) | (c3 & 0x3F); if (c < 0x0800 || (c >= 0xD800 && c <= 0xDFFF)) text += invalid; else { text += fromCharCode(c); i += 2; } } } } // 4 bytes else if ((c1 & 0xF8) == 0xF0) { if (i + 3 >= count) text += invalid; else { c2 = bytes[i + offset + 1]; c3 = bytes[i + offset + 2]; c4 = bytes[i + offset + 3]; if (((c2 | (c3 << 8) | (c4 << 16)) & 0xC0C0C0) !== 0x808080) text += invalid; else { c = ((c1 & 0x07) << 0x12) | ((c2 & 0x3F) << 0x0C) | ((c3 & 0x3F) << 0x06) | (c4 & 0x3F); if (c < 0x10000 || c > 0x10FFFF) text += invalid; else { c -= 0x10000; text += fromCharCode((c >> 10) + 0xD800, (c & 0x3FF) + 0xDC00); i += 3; } } } } else text += invalid; } return text; } function writeString(bb: ByteBuffer, text: string): void { // Sadly a hand-coded UTF8 encoder is much faster than TextEncoder+set in V8 let n = text.length; let byteCount = 0; // Write the byte count first for (let i = 0; i < n; i++) { let c = text.charCodeAt(i); if (c >= 0xD800 && c <= 0xDBFF && i + 1 < n) { c = (c << 10) + text.charCodeAt(++i) - 0x35FDC00; } byteCount += c < 0x80 ? 1 : c < 0x800 ? 2 : c < 0x10000 ? 3 : 4; } writeVarint32(bb, byteCount); let offset = grow(bb, byteCount); let bytes = bb.bytes; // Then write the bytes for (let i = 0; i < n; i++) { let c = text.charCodeAt(i); if (c >= 0xD800 && c <= 0xDBFF && i + 1 < n) { c = (c << 10) + text.charCodeAt(++i) - 0x35FDC00; } if (c < 0x80) { bytes[offset++] = c; } else { if (c < 0x800) { bytes[offset++] = ((c >> 6) & 0x1F) | 0xC0; } else { if (c < 0x10000) { bytes[offset++] = ((c >> 12) & 0x0F) | 0xE0; } else { bytes[offset++] = ((c >> 18) & 0x07) | 0xF0; bytes[offset++] = ((c >> 12) & 0x3F) | 0x80; } bytes[offset++] = ((c >> 6) & 0x3F) | 0x80; } bytes[offset++] = (c & 0x3F) | 0x80; } } } function writeByteBuffer(bb: ByteBuffer, buffer: ByteBuffer): void { let offset = grow(bb, buffer.limit); let from = bb.bytes; let to = buffer.bytes; // This for loop is much faster than subarray+set on V8 for (let i = 0, n = buffer.limit; i < n; i++) { from[i + offset] = to[i]; } } function readByte(bb: ByteBuffer): number { return bb.bytes[advance(bb, 1)]; } function writeByte(bb: ByteBuffer, value: number): void { let offset = grow(bb, 1); bb.bytes[offset] = value; } function readFloat(bb: ByteBuffer): number { let offset = advance(bb, 4); let bytes = bb.bytes; // Manual copying is much faster than subarray+set in V8 f32_u8[0] = bytes[offset++]; f32_u8[1] = bytes[offset++]; f32_u8[2] = bytes[offset++]; f32_u8[3] = bytes[offset++]; return f32[0]; } function writeFloat(bb: ByteBuffer, value: number): void { let offset = grow(bb, 4); let bytes = bb.bytes; f32[0] = value; // Manual copying is much faster than subarray+set in V8 bytes[offset++] = f32_u8[0]; bytes[offset++] = f32_u8[1]; bytes[offset++] = f32_u8[2]; bytes[offset++] = f32_u8[3]; } function readDouble(bb: ByteBuffer): number { let offset = advance(bb, 8); let bytes = bb.bytes; // Manual copying is much faster than subarray+set in V8 f64_u8[0] = bytes[offset++]; f64_u8[1] = bytes[offset++]; f64_u8[2] = bytes[offset++]; f64_u8[3] = bytes[offset++]; f64_u8[4] = bytes[offset++]; f64_u8[5] = bytes[offset++]; f64_u8[6] = bytes[offset++]; f64_u8[7] = bytes[offset++]; return f64[0]; } function writeDouble(bb: ByteBuffer, value: number): void { let offset = grow(bb, 8); let bytes = bb.bytes; f64[0] = value; // Manual copying is much faster than subarray+set in V8 bytes[offset++] = f64_u8[0]; bytes[offset++] = f64_u8[1]; bytes[offset++] = f64_u8[2]; bytes[offset++] = f64_u8[3]; bytes[offset++] = f64_u8[4]; bytes[offset++] = f64_u8[5]; bytes[offset++] = f64_u8[6]; bytes[offset++] = f64_u8[7]; } function readInt32(bb: ByteBuffer): number { let offset = advance(bb, 4); let bytes = bb.bytes; return ( bytes[offset] | (bytes[offset + 1] << 8) | (bytes[offset + 2] << 16) | (bytes[offset + 3] << 24) ); } function writeInt32(bb: ByteBuffer, value: number): void { let offset = grow(bb, 4); let bytes = bb.bytes; bytes[offset] = value; bytes[offset + 1] = value >> 8; bytes[offset + 2] = value >> 16; bytes[offset + 3] = value >> 24; } function readInt64(bb: ByteBuffer, unsigned: boolean): Long { return { low: readInt32(bb), high: readInt32(bb), unsigned, }; } function writeInt64(bb: ByteBuffer, value: Long): void { writeInt32(bb, value.low); writeInt32(bb, value.high); } function readVarint32(bb: ByteBuffer): number { let c = 0; let value = 0; let b: number; do { b = readByte(bb); if (c < 32) value |= (b & 0x7F) << c; c += 7; } while (b & 0x80); return value; } function writeVarint32(bb: ByteBuffer, value: number): void { value >>>= 0; while (value >= 0x80) { writeByte(bb, (value & 0x7f) | 0x80); value >>>= 7; } writeByte(bb, value); } function readVarint64(bb: ByteBuffer, unsigned: boolean): Long { let part0 = 0; let part1 = 0; let part2 = 0; let b: number; b = readByte(bb); part0 = (b & 0x7F); if (b & 0x80) { b = readByte(bb); part0 |= (b & 0x7F) << 7; if (b & 0x80) { b = readByte(bb); part0 |= (b & 0x7F) << 14; if (b & 0x80) { b = readByte(bb); part0 |= (b & 0x7F) << 21; if (b & 0x80) { b = readByte(bb); part1 = (b & 0x7F); if (b & 0x80) { b = readByte(bb); part1 |= (b & 0x7F) << 7; if (b & 0x80) { b = readByte(bb); part1 |= (b & 0x7F) << 14; if (b & 0x80) { b = readByte(bb); part1 |= (b & 0x7F) << 21; if (b & 0x80) { b = readByte(bb); part2 = (b & 0x7F); if (b & 0x80) { b = readByte(bb); part2 |= (b & 0x7F) << 7; } } } } } } } } } return { low: part0 | (part1 << 28), high: (part1 >>> 4) | (part2 << 24), unsigned, }; } function writeVarint64(bb: ByteBuffer, value: Long): void { let part0 = value.low >>> 0; let part1 = ((value.low >>> 28) | (value.high << 4)) >>> 0; let part2 = value.high >>> 24; // ref: src/google/protobuf/io/coded_stream.cc let size = part2 === 0 ? part1 === 0 ? part0 < 1 << 14 ? part0 < 1 << 7 ? 1 : 2 : part0 < 1 << 21 ? 3 : 4 : part1 < 1 << 14 ? part1 < 1 << 7 ? 5 : 6 : part1 < 1 << 21 ? 7 : 8 : part2 < 1 << 7 ? 9 : 10; let offset = grow(bb, size); let bytes = bb.bytes; switch (size) { case 10: bytes[offset + 9] = (part2 >>> 7) & 0x01; case 9: bytes[offset + 8] = size !== 9 ? part2 | 0x80 : part2 & 0x7F; case 8: bytes[offset + 7] = size !== 8 ? (part1 >>> 21) | 0x80 : (part1 >>> 21) & 0x7F; case 7: bytes[offset + 6] = size !== 7 ? (part1 >>> 14) | 0x80 : (part1 >>> 14) & 0x7F; case 6: bytes[offset + 5] = size !== 6 ? (part1 >>> 7) | 0x80 : (part1 >>> 7) & 0x7F; case 5: bytes[offset + 4] = size !== 5 ? part1 | 0x80 : part1 & 0x7F; case 4: bytes[offset + 3] = size !== 4 ? (part0 >>> 21) | 0x80 : (part0 >>> 21) & 0x7F; case 3: bytes[offset + 2] = size !== 3 ? (part0 >>> 14) | 0x80 : (part0 >>> 14) & 0x7F; case 2: bytes[offset + 1] = size !== 2 ? (part0 >>> 7) | 0x80 : (part0 >>> 7) & 0x7F; case 1: bytes[offset] = size !== 1 ? part0 | 0x80 : part0 & 0x7F; } } function readVarint32ZigZag(bb: ByteBuffer): number { let value = readVarint32(bb); // ref: src/google/protobuf/wire_format_lite.h return (value >>> 1) ^ -(value & 1); } function writeVarint32ZigZag(bb: ByteBuffer, value: number): void { // ref: src/google/protobuf/wire_format_lite.h writeVarint32(bb, (value << 1) ^ (value >> 31)); } function readVarint64ZigZag(bb: ByteBuffer): Long { let value = readVarint64(bb, /* unsigned */ false); let low = value.low; let high = value.high; let flip = -(low & 1); // ref: src/google/protobuf/wire_format_lite.h return { low: ((low >>> 1) | (high << 31)) ^ flip, high: (high >>> 1) ^ flip, unsigned: false, }; } function writeVarint64ZigZag(bb: ByteBuffer, value: Long): void { let low = value.low; let high = value.high; let flip = high >> 31; // ref: src/google/protobuf/wire_format_lite.h writeVarint64(bb, { low: (low << 1) ^ flip, high: ((high << 1) | (low >>> 31)) ^ flip, unsigned: false, }); }
3、代码中测试打包出来的typescript脚本
因为pbjs打包后protobufjs代码是直接集成在 message.ts 文件中的,因此不需要我们再去引用protobufjs库。
import { _decorator, Component, log } from 'cc'; const { ccclass, property } = _decorator; import{Player,encodePlayer,decodePlayer} from "./message" @ccclass('main') export class main extends Component { onLoad() { const player:Player = { name:"John", level:10, } // 将消息对象序列化为二进制数据 const binaryData:Uint8Array = encodePlayer(player); // 从二进制数据中反序列化为消息对象 const deserializedPlayer = decodePlayer(binaryData); // 访问消息对象的字段 console.log(deserializedPlayer.name); // 输出: John console.log(deserializedPlayer.level); // 输出: 10 } start() { } update(deltaTime: number) { } }
运行打印结果如下:
好了,Cocos Creator 3.8中如何如何将 proto文件打包成typescript 脚本,以及如何使用 protobuf 的教程到此就结束。