Merge pull request #326 from karthik2804/add_regex_precompilation

add precompilation for regex during the build step

Author: tschneidereit

bin/j2w.mjs

@@ -7,6 +7,7 @@ import { resolve, basename } from 'node:path';
 import yargs from 'yargs';
 import { hideBin } from 'yargs/helpers';
 import path from 'path';
+import { precompile } from "./precompile.mjs";
 
 const componentizeVersion = '0.16.0';
 const __filename = new URL(import.meta.url).pathname;
@@ -103,14 +104,15 @@ async function saveBuildData(buildDataPath, checksum, version) {
         }
 
         const source = await readFile(src, 'utf8');
+        const precompiledSource = precompile(source, src, true);
 
         // Check if a non-default wit directory is supplied
         const witPath = args.witPath ? resolve(args.witPath) : path.join(__dirname, 'wit');
         if (args.witPath) {
             console.log(`Using user-provided wit in: ${witPath}`);
         }
 
-        const { component } = await componentize(source, {
+        const { component } = await componentize(precompiledSource, {
             sourceName: basename(src),
             witPath,
             worldName: args.triggerType,

bin/precompile.mjs

@@ -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();
+}

package-lock.json

@@ -1,6 +1,6 @@
 {
   "name": "@fermyon/spin-sdk",
-  "version": "3.0.0",
+  "version": "3.1.0",
   "description": "",
   "main": "lib/index.js",
   "typings": "lib/index.d.ts",
@@ -28,7 +28,11 @@
     "@bytecodealliance/componentize-js": "^0.16.0",
     "yargs": "^17.7.2",
     "typedoc-plugin-missing-exports": "^3.0.0",
-    "@fermyon/knitwit": "^0.3.0"
+    "@fermyon/knitwit": "^0.3.0",
+    "acorn-walk": "^8.3.4",
+    "acron": "^1.0.5",
+    "magic-string": "^0.30.17",
+    "regexpu-core": "^6.2.0"
   },
   "files": [
     "lib",