feat: hand written parser

packages/message-parser/jest.config.ts

@@ -1,18 +1,8 @@
-import { dirname, resolve } from 'node:path';
-import { fileURLToPath } from 'node:url';
-
 import server from '@rocket.chat/jest-presets/server';
 import type { Config } from 'jest';
 
-// Jest 30 loads this config via Node's native type stripping (Node 22.18+),
-// which treats the file as ESM where __dirname is not defined as a global.
-const __dirname = dirname(fileURLToPath(import.meta.url));
-
 export default {
 	preset: server.preset,
-	transform: {
-		'\\.pegjs$': resolve(__dirname, './loaders/pegtransform.js'),
-	},
-	moduleFileExtensions: ['js', 'ts', 'pegjs'],
+	moduleFileExtensions: ['js', 'ts'],
 	testPathIgnorePatterns: ['/node_modules/', '\\.bench\\.ts$'],
 } satisfies Config;

packages/message-parser/package.json

@@ -35,7 +35,7 @@
 		".:build:bundle": "webpack-cli",
 		".:build:clean": "rimraf dist",
 		"build": "run-s .:build:clean .:build:bundle",
-		"bench": "ts-node --compiler-options '{\"module\":\"commonjs\"}' -r ./loaders/pegjs-register.js benchmarks/parser.bench.ts",
+		"bench": "ts-node --compiler-options '{\"module\":\"commonjs\"}' benchmarks/parser.bench.ts",
 		"bench:size": "node -e \"const fs=require('fs'),zlib=require('zlib'),p='dist/messageParser.js';try{const s=fs.statSync(p),c=fs.readFileSync(p),g=zlib.gzipSync(c);console.log('Bundle:',p);console.log('Raw:',s.size,'bytes ('+(s.size/1024).toFixed(1),'KB)');console.log('Gzip:',g.length,'bytes ('+(g.length/1024).toFixed(1),'KB)')}catch(e){console.error('Build first: yarn build');process.exit(1)}\"",
 		"lint": "eslint .",
 		"test": "jest",
@@ -47,17 +47,14 @@
 	},
 	"devDependencies": {
 		"@rocket.chat/jest-presets": "workspace:~",
-		"@rocket.chat/peggy-loader": "workspace:~",
 		"@rocket.chat/prettier-config": "~0.31.25",
 		"@types/jest": "~30.0.0",
 		"@types/node": "~22.19.21",
 		"eslint": "~9.39.5",
 		"fast-check": "^4.6.0",
 		"jest": "~30.2.0",
 		"npm-run-all": "^4.1.5",
-		"peggy": "4.1.1",
 		"prettier": "~3.3.3",
-		"prettier-plugin-pegjs": "~0.5.4",
 		"rimraf": "^6.0.1",
 		"tinybench": "^3.0.7",
 		"ts-loader": "~9.5.7",

packages/message-parser/src/chars.ts

@@ -0,0 +1,100 @@
+export function isNewline(ch: string): boolean {
+	return ch === '\n' || ch === '\r';
+}
+
+export function isSpace(ch: string): boolean {
+	return ch === ' ' || ch === '\t';
+}
+
+export function isWhitespace(ch: string): boolean {
+	return isSpace(ch) || isNewline(ch);
+}
+
+export function isDigit(ch: string): boolean {
+	return ch >= '0' && ch <= '9';
+}
+
+export function isAlpha(ch: string): boolean {
+	return (ch >= 'a' && ch <= 'z') || (ch >= 'A' && ch <= 'Z');
+}
+
+export function isAlphaNum(ch: string): boolean {
+	return isAlpha(ch) || isDigit(ch);
+}
+
+export function isMarkupChar(ch: string): boolean {
+	return '*_~`#@:|\\[!<$+()'.includes(ch);
+}
+
+export function isPlainChar(ch: string): boolean {
+	return ch !== '' && !isNewline(ch) && !isMarkupChar(ch) && !isSpace(ch);
+}
+
+export function isUrlStart(ch: string): boolean {
+	return isAlpha(ch) || isDigit(ch);
+}
+
+export function isEmailStart(ch: string): boolean {
+	return isUrlStart(ch) || ch.charCodeAt(0) > 127;
+}
+
+export function isEmojiStart(ch: string): boolean {
+	const code = ch.charCodeAt(0);
+	return (
+		code === 0xa9 ||
+		code === 0xae ||
+		code === 0x203c ||
+		code === 0x2049 ||
+		code === 0x2122 ||
+		code === 0x2139 ||
+		(code >= 0x2194 && code <= 0x21aa) || // Arrows
+		(code >= 0x231a && code <= 0x23fa) || // Misc Technical
+		code === 0x24c2 ||
+		(code >= 0x25aa && code <= 0x27bf) || // Geometric Shapes, Symbols, Dingbats
+		(code >= 0x2934 && code <= 0x2b55) || // Supplemental Arrows
+		code === 0x3030 ||
+		code === 0x303d ||
+		code === 0x3297 ||
+		code === 0x3299 ||
+		(code >= 0xd800 && code <= 0xdbff) || // High surrogates (U+10000+ emoji)
+		ch === '#' || // keycap bases
+		ch === '*' ||
+		(ch >= '0' && ch <= '9')
+	);
+}
+
+export function isHexDigit(ch: string): boolean {
+	return (ch >= '0' && ch <= '9') || (ch >= 'a' && ch <= 'f') || (ch >= 'A' && ch <= 'F');
+}
+
+// ─── Emoticon ──────────────────────────────────────────────────────────────
+export const EMOTICONS: Record<string, string> = {
+	':)': 'slight_smile',
+	':-)': 'slight_smile',
+	':(': 'frowning',
+	':-(': 'frowning',
+	'D:': 'fearful',
+	':D': 'grinning',
+	':-D': 'grinning',
+	':P': 'stuck_out_tongue',
+	':-P': 'stuck_out_tongue',
+	':p': 'stuck_out_tongue',
+	':-p': 'stuck_out_tongue',
+	';)': 'wink',
+	';-)': 'wink',
+	':o': 'open_mouth',
+	':-o': 'open_mouth',
+	':O': 'open_mouth',
+	':-O': 'open_mouth',
+	':|': 'neutral_face',
+	':-|': 'neutral_face',
+	':/': 'confused',
+	':-/': 'confused',
+	':\\': 'confused',
+	':-\\': 'confused',
+	':*': 'kissing_heart',
+	'-_-': 'expressionless',
+};
+
+// Sorted longest-first so e.g. ":-D" wins over ":D"
+export const EMOTICON_KEYS = Object.keys(EMOTICONS).sort((a, b) => b.length - a.length);