Merge pull request #361 from karthik2804/readd_regex_precompilation

add regex precompilation to build step

Author: karthik2804

packages/build-tools/package-lock.json

@@ -1,10 +1,10 @@
 {
   "name": "@spinframework/build-tools",
-  "version": "1.0.1",
+  "version": "1.0.2",
   "description": "",
   "main": "index.js",
   "scripts": {
-    "build": "make && rm -rf dist && tsc && chmod +x dist/*",
+    "build": "make && rm -rf dist && tsc && chmod +x dist/* && cp src/*.js dist/",
     "fmt": "prettier --write \"src/**/*.{ts,tsx,js,jsx}\"",
     "test": "npm run build && mocha --require ts-node/register test/**/*.spec.ts"
   },
@@ -31,7 +31,11 @@
   "dependencies": {
     "@bytecodealliance/componentize-js": "^0.18.1",
     "@bytecodealliance/jco": "^1.10.2",
-    "yargs": "^17.7.2"
+    "yargs": "^17.7.2",
+    "acorn-walk": "^8.3.4",
+    "acron": "^1.0.5",
+    "magic-string": "^0.30.17",
+    "regexpu-core": "^6.2.0"
   },
   "files": [
     "lib",

packages/build-tools/package.json

@@ -3,15 +3,17 @@
 import { componentize } from '@bytecodealliance/componentize-js';
 import { version as componentizeVersion } from '@bytecodealliance/componentize-js';
 import { getPackagesWithWasiDeps, processWasiDeps } from './wasiDepsParser.js';
-
 import {
   calculateChecksum,
   saveBuildData,
 } from './utils.js';
 import { getCliArgs } from './cli.js';
 import { getBuildDataPath, ShouldComponentize } from './build.js';
-import { writeFile } from 'node:fs/promises';
+import { readFile, writeFile } from 'node:fs/promises';
 import { mergeWit } from '../lib/wit_tools.js';
+//@ts-ignore
+import { precompile } from "./precompile.js"
+import path from 'node:path'
 
 async function main() {
   try {
@@ -61,8 +63,17 @@ async function main() {
       'combined-wit:[email protected]',
     );
 
+    const source = await readFile(src, 'utf8');
+    const precompiledSource = precompile(source, src, true) as string;
+
+    // Write precompiled source to disk for debugging purposes In the future we
+    // will also write a source map to make debugging easier
+    let srcDir = path.dirname(src);
+    let precompiledSourcePath = path.join(srcDir, 'precompiled-source.js');
+    await writeFile(precompiledSourcePath, precompiledSource);
+
     const { component } = await componentize({
-      sourcePath: src,
+      sourcePath: precompiledSourcePath,
       // @ts-ignore
       witWorld: inlineWit,
       runtimeArgs,

packages/build-tools/src/index.ts

@@ -0,0 +1,62 @@
+// Copyright 2024 Fastly Inc.
+// License: the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+import regexpuc from 'regexpu-core';
+import { parse } from 'acorn';
+import MagicString from 'magic-string';
+import { simple as simpleWalk } from 'acorn-walk';
+
+const PREAMBLE = `;{
+   // Precompiled regular expressions
+   const precompile = (r) => { r.exec('a'); r.exec('\\u1000'); };`;
+const POSTAMBLE = '}';
+
+/// Emit a block of javascript that will pre-compile the regular expressions given. As spidermonkey
+/// will intern regular expressions, duplicating them at the top level and testing them with both
+/// an ascii and utf8 string should ensure that they won't be re-compiled when run in the fetch
+/// handler.
+export function precompile(source, filename = '<input>', moduleMode = false) {
+    const magicString = new MagicString(source, {
+        filename,
+    });
+
+    const ast = parse(source, {
+        ecmaVersion: 'latest',
+        sourceType: moduleMode ? 'module' : 'script',
+    });
+
+    const precompileCalls = [];
+    simpleWalk(ast, {
+        Literal(node) {
+            if (!node.regex) return;
+            let transpiledPattern;
+            try {
+                transpiledPattern = regexpuc(node.regex.pattern, node.regex.flags, {
+                    unicodePropertyEscapes: 'transform',
+                });
+            } catch {
+                // swallow regex parse errors here to instead throw them at the engine level
+                // this then also avoids regex parser bugs being thrown unnecessarily
+                transpiledPattern = node.regex.pattern;
+            }
+            const transpiledRegex = `/${transpiledPattern}/${node.regex.flags}`;
+            precompileCalls.push(`precompile(${transpiledRegex});`);
+            magicString.overwrite(node.start, node.end, transpiledRegex);
+        },
+    });
+
+    if (!precompileCalls.length) return source;
+
+    magicString.prepend(`${PREAMBLE}${precompileCalls.join('\n')}${POSTAMBLE}`);
+
+    // When we're ready to pipe in source maps:
+    // const map = magicString.generateMap({
+    //   source: 'source.js',
+    //   file: 'converted.js.map',
+    //   includeContent: true
+    // });
+
+    return magicString.toString();
+}

packages/build-tools/src/precompile.js

@@ -23,7 +23,6 @@ echo "built the test app successfully"
 # Start the spin app in the background
 echo "Starting Spin app"
 spin up &
-SPIN_PID=$!
 
 # wait for app to be up and running
 echo "Waiting for Spin app to be ready"
@@ -36,7 +35,7 @@ echo "\n\nTest completed"
 
 # kill the spin app
 echo "Stopping Spin"
-kill -9 $SPIN_PID
+killall spin
 
 
 if [ "$isFailed" = true ] ; then