From 12779b3e02308b42a2b5aaa22b02fd8b4e68542f Mon Sep 17 00:00:00 2001
From: legendecas <legendecas@gmail.com>
Date: Tue, 27 Sep 2022 00:55:59 +0800
Subject: [PATCH] src,lib: retrieve parsed source map url from v8

V8 already parses the source map magic comments. Currently, only scripts
and functions expose the parsed source map URLs. It is unnecessary to
parse the source map magic comments again when the parsed information is
available.

PR-URL: https://github.com/nodejs/node/pull/44798
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com>
Reviewed-By: Jan Krems <jan.krems@gmail.com>
---
 doc/api/vm.md                               |  35 ++++++
 lib/internal/modules/cjs/loader.js          |  26 ++++-
 lib/internal/process/pre_execution.js       |   4 +-
 lib/internal/source_map/source_map_cache.js |  98 +++++++++--------
 lib/internal/vm.js                          | 113 ++++++++++++++++++++
 lib/vm.js                                   |  92 +---------------
 src/env_properties.h                        |   1 +
 src/node_contextify.cc                      |  28 +++--
 test/parallel/test-bootstrap-modules.js     |   1 +
 test/parallel/test-vm-source-map-url.js     |  27 +++++
 10 files changed, 282 insertions(+), 143 deletions(-)
 create mode 100644 lib/internal/vm.js
 create mode 100644 test/parallel/test-vm-source-map-url.js

diff --git a/doc/api/vm.md b/doc/api/vm.md
index f99f4966db82b6..91f77cfc044a9f 100644
--- a/doc/api/vm.md
+++ b/doc/api/vm.md
@@ -344,6 +344,41 @@ console.log(globalVar);
 // 1000
 ```
 
+### `script.sourceMapURL`
+
+<!-- YAML
+added: REPLACEME
+-->
+
+* {string|undefined}
+
+When the script is compiled from a source that contains a source map magic
+comment, this property will be set to the URL of the source map.
+
+```mjs
+import vm from 'node:vm';
+
+const script = new vm.Script(`
+function myFunc() {}
+//# sourceMappingURL=sourcemap.json
+`);
+
+console.log(script.sourceMapURL);
+// Prints: sourcemap.json
+```
+
+```cjs
+const vm = require('node:vm');
+
+const script = new vm.Script(`
+function myFunc() {}
+//# sourceMappingURL=sourcemap.json
+`);
+
+console.log(script.sourceMapURL);
+// Prints: sourcemap.json
+```
+
 ## Class: `vm.Module`
 
 <!-- YAML
diff --git a/lib/internal/modules/cjs/loader.js b/lib/internal/modules/cjs/loader.js
index 68edd781e33a3e..7e6cad0f6f3388 100644
--- a/lib/internal/modules/cjs/loader.js
+++ b/lib/internal/modules/cjs/loader.js
@@ -86,7 +86,8 @@ const {
   filterOwnProperties,
   setOwnProperty,
 } = require('internal/util');
-const vm = require('vm');
+const { Script } = require('vm');
+const { internalCompileFunction } = require('internal/vm');
 const assert = require('internal/assert');
 const fs = require('fs');
 const internalFS = require('internal/fs/utils');
@@ -1073,19 +1074,28 @@ let hasPausedEntry = false;
 function wrapSafe(filename, content, cjsModuleInstance) {
   if (patched) {
     const wrapper = Module.wrap(content);
-    return vm.runInThisContext(wrapper, {
+    const script = new Script(wrapper, {
       filename,
       lineOffset: 0,
-      displayErrors: true,
       importModuleDynamically: async (specifier, _, importAssertions) => {
         const loader = asyncESM.esmLoader;
         return loader.import(specifier, normalizeReferrerURL(filename),
                              importAssertions);
       },
     });
+
+    // Cache the source map for the module if present.
+    if (script.sourceMapURL) {
+      maybeCacheSourceMap(filename, content, this, false, undefined, script.sourceMapURL);
+    }
+
+    return script.runInThisContext({
+      displayErrors: true,
+    });
   }
+
   try {
-    return vm.compileFunction(content, [
+    const result = internalCompileFunction(content, [
       'exports',
       'require',
       'module',
@@ -1099,6 +1109,13 @@ function wrapSafe(filename, content, cjsModuleInstance) {
                              importAssertions);
       },
     });
+
+    // Cache the source map for the module if present.
+    if (result.sourceMapURL) {
+      maybeCacheSourceMap(filename, content, this, false, undefined, result.sourceMapURL);
+    }
+
+    return result.function;
   } catch (err) {
     if (process.mainModule === cjsModuleInstance)
       enrichCJSError(err, content);
@@ -1119,7 +1136,6 @@ Module.prototype._compile = function(content, filename) {
     policy.manifest.assertIntegrity(moduleURL, content);
   }
 
-  maybeCacheSourceMap(filename, content, this);
   const compiledWrapper = wrapSafe(filename, content, this);
 
   let inspectorWrapper = null;
diff --git a/lib/internal/process/pre_execution.js b/lib/internal/process/pre_execution.js
index c264ec6af655c8..756041e5044083 100644
--- a/lib/internal/process/pre_execution.js
+++ b/lib/internal/process/pre_execution.js
@@ -596,9 +596,11 @@ function initializeESMLoader() {
 }
 
 function initializeSourceMapsHandlers() {
-  const { setSourceMapsEnabled } =
+  const { setSourceMapsEnabled, getSourceMapsEnabled } =
     require('internal/source_map/source_map_cache');
   process.setSourceMapsEnabled = setSourceMapsEnabled;
+  // Initialize the environment flag of source maps.
+  getSourceMapsEnabled();
 }
 
 function initializeFrozenIntrinsics() {
diff --git a/lib/internal/source_map/source_map_cache.js b/lib/internal/source_map/source_map_cache.js
index ee8de441ba9316..656c4e3f296e94 100644
--- a/lib/internal/source_map/source_map_cache.js
+++ b/lib/internal/source_map/source_map_cache.js
@@ -97,7 +97,21 @@ function extractSourceURLMagicComment(content) {
   return sourceURL;
 }
 
-function maybeCacheSourceMap(filename, content, cjsModuleInstance, isGeneratedSource, sourceURL) {
+function extractSourceMapURLMagicComment(content) {
+  let match;
+  let lastMatch;
+  // A while loop is used here to get the last occurrence of sourceMappingURL.
+  // This is needed so that we don't match sourceMappingURL in string literals.
+  while ((match = RegExpPrototypeExec(kSourceMappingURLMagicComment, content))) {
+    lastMatch = match;
+  }
+  if (lastMatch == null) {
+    return null;
+  }
+  return lastMatch.groups.sourceMappingURL;
+}
+
+function maybeCacheSourceMap(filename, content, cjsModuleInstance, isGeneratedSource, sourceURL, sourceMapURL) {
   const sourceMapsEnabled = getSourceMapsEnabled();
   if (!(process.env.NODE_V8_COVERAGE || sourceMapsEnabled)) return;
   try {
@@ -108,52 +122,52 @@ function maybeCacheSourceMap(filename, content, cjsModuleInstance, isGeneratedSo
     return;
   }
 
-  let match;
-  let lastMatch;
-  // A while loop is used here to get the last occurrence of sourceMappingURL.
-  // This is needed so that we don't match sourceMappingURL in string literals.
-  while ((match = RegExpPrototypeExec(kSourceMappingURLMagicComment, content))) {
-    lastMatch = match;
+  if (sourceMapURL === undefined) {
+    sourceMapURL = extractSourceMapURLMagicComment(content);
+  }
+
+  // Bail out when there is no source map url.
+  if (typeof sourceMapURL !== 'string') {
+    return;
   }
 
   if (sourceURL === undefined) {
     sourceURL = extractSourceURLMagicComment(content);
   }
-  if (lastMatch) {
-    const data = dataFromUrl(filename, lastMatch.groups.sourceMappingURL);
-    const url = data ? null : lastMatch.groups.sourceMappingURL;
-    if (cjsModuleInstance) {
-      cjsSourceMapCache.set(cjsModuleInstance, {
-        filename,
-        lineLengths: lineLengths(content),
-        data,
-        url,
-        sourceURL,
-      });
-    } else if (isGeneratedSource) {
-      const entry = {
-        lineLengths: lineLengths(content),
-        data,
-        url,
-        sourceURL
-      };
-      generatedSourceMapCache.set(filename, entry);
-      if (sourceURL) {
-        generatedSourceMapCache.set(sourceURL, entry);
-      }
-    } else {
-      // If there is no cjsModuleInstance and is not generated source assume we are in a
-      // "modules/esm" context.
-      const entry = {
-        lineLengths: lineLengths(content),
-        data,
-        url,
-        sourceURL,
-      };
-      esmSourceMapCache.set(filename, entry);
-      if (sourceURL) {
-        esmSourceMapCache.set(sourceURL, entry);
-      }
+
+  const data = dataFromUrl(filename, sourceMapURL);
+  const url = data ? null : sourceMapURL;
+  if (cjsModuleInstance) {
+    cjsSourceMapCache.set(cjsModuleInstance, {
+      filename,
+      lineLengths: lineLengths(content),
+      data,
+      url,
+      sourceURL,
+    });
+  } else if (isGeneratedSource) {
+    const entry = {
+      lineLengths: lineLengths(content),
+      data,
+      url,
+      sourceURL
+    };
+    generatedSourceMapCache.set(filename, entry);
+    if (sourceURL) {
+      generatedSourceMapCache.set(sourceURL, entry);
+    }
+  } else {
+    // If there is no cjsModuleInstance and is not generated source assume we are in a
+    // "modules/esm" context.
+    const entry = {
+      lineLengths: lineLengths(content),
+      data,
+      url,
+      sourceURL,
+    };
+    esmSourceMapCache.set(filename, entry);
+    if (sourceURL) {
+      esmSourceMapCache.set(sourceURL, entry);
     }
   }
 }
diff --git a/lib/internal/vm.js b/lib/internal/vm.js
new file mode 100644
index 00000000000000..32bcdf06234b67
--- /dev/null
+++ b/lib/internal/vm.js
@@ -0,0 +1,113 @@
+'use strict';
+
+const {
+  ArrayPrototypeForEach,
+} = primordials;
+
+const {
+  compileFunction,
+  isContext: _isContext,
+} = internalBinding('contextify');
+const {
+  validateArray,
+  validateBoolean,
+  validateBuffer,
+  validateFunction,
+  validateObject,
+  validateString,
+  validateUint32,
+} = require('internal/validators');
+const {
+  ERR_INVALID_ARG_TYPE,
+} = require('internal/errors').codes;
+
+function isContext(object) {
+  validateObject(object, 'object', { __proto__: null, allowArray: true });
+
+  return _isContext(object);
+}
+
+function internalCompileFunction(code, params, options) {
+  validateString(code, 'code');
+  if (params !== undefined) {
+    validateArray(params, 'params');
+    ArrayPrototypeForEach(params,
+                          (param, i) => validateString(param, `params[${i}]`));
+  }
+
+  const {
+    filename = '',
+    columnOffset = 0,
+    lineOffset = 0,
+    cachedData = undefined,
+    produceCachedData = false,
+    parsingContext = undefined,
+    contextExtensions = [],
+    importModuleDynamically,
+  } = options;
+
+  validateString(filename, 'options.filename');
+  validateUint32(columnOffset, 'options.columnOffset');
+  validateUint32(lineOffset, 'options.lineOffset');
+  if (cachedData !== undefined)
+    validateBuffer(cachedData, 'options.cachedData');
+  validateBoolean(produceCachedData, 'options.produceCachedData');
+  if (parsingContext !== undefined) {
+    if (
+      typeof parsingContext !== 'object' ||
+      parsingContext === null ||
+      !isContext(parsingContext)
+    ) {
+      throw new ERR_INVALID_ARG_TYPE(
+        'options.parsingContext',
+        'Context',
+        parsingContext
+      );
+    }
+  }
+  validateArray(contextExtensions, 'options.contextExtensions');
+  ArrayPrototypeForEach(contextExtensions, (extension, i) => {
+    const name = `options.contextExtensions[${i}]`;
+    validateObject(extension, name, { __proto__: null, nullable: true });
+  });
+
+  const result = compileFunction(
+    code,
+    filename,
+    lineOffset,
+    columnOffset,
+    cachedData,
+    produceCachedData,
+    parsingContext,
+    contextExtensions,
+    params
+  );
+
+  if (produceCachedData) {
+    result.function.cachedDataProduced = result.cachedDataProduced;
+  }
+
+  if (result.cachedData) {
+    result.function.cachedData = result.cachedData;
+  }
+
+  if (importModuleDynamically !== undefined) {
+    validateFunction(importModuleDynamically,
+                     'options.importModuleDynamically');
+    const { importModuleDynamicallyWrap } =
+      require('internal/vm/module');
+    const { callbackMap } = internalBinding('module_wrap');
+    const wrapped = importModuleDynamicallyWrap(importModuleDynamically);
+    const func = result.function;
+    callbackMap.set(result.cacheKey, {
+      importModuleDynamically: (s, _k, i) => wrapped(s, func, i),
+    });
+  }
+
+  return result;
+}
+
+module.exports = {
+  internalCompileFunction,
+  isContext,
+};
diff --git a/lib/vm.js b/lib/vm.js
index 1bb199363afbe1..ec91c8a2651775 100644
--- a/lib/vm.js
+++ b/lib/vm.js
@@ -32,9 +32,7 @@ const {
   ContextifyScript,
   MicrotaskQueue,
   makeContext,
-  isContext: _isContext,
   constants,
-  compileFunction: _compileFunction,
   measureMemory: _measureMemory,
 } = internalBinding('contextify');
 const {
@@ -45,9 +43,7 @@ const {
   isArrayBufferView,
 } = require('internal/util/types');
 const {
-  validateArray,
   validateBoolean,
-  validateBuffer,
   validateFunction,
   validateInt32,
   validateObject,
@@ -60,6 +56,10 @@ const {
   kEmptyObject,
   kVmBreakFirstLineSymbol,
 } = require('internal/util');
+const {
+  internalCompileFunction,
+  isContext,
+} = require('internal/vm');
 const kParsingContext = Symbol('script parsing context');
 
 class Script extends ContextifyScript {
@@ -213,12 +213,6 @@ function getContextOptions(options) {
   return contextOptions;
 }
 
-function isContext(object) {
-  validateObject(object, 'object', { allowArray: true });
-
-  return _isContext(object);
-}
-
 let defaultContextNameIndex = 1;
 function createContext(contextObject = {}, options = kEmptyObject) {
   if (isContext(contextObject)) {
@@ -314,83 +308,7 @@ function runInThisContext(code, options) {
 }
 
 function compileFunction(code, params, options = kEmptyObject) {
-  validateString(code, 'code');
-  if (params !== undefined) {
-    validateArray(params, 'params');
-    ArrayPrototypeForEach(params,
-                          (param, i) => validateString(param, `params[${i}]`));
-  }
-
-  const {
-    filename = '',
-    columnOffset = 0,
-    lineOffset = 0,
-    cachedData = undefined,
-    produceCachedData = false,
-    parsingContext = undefined,
-    contextExtensions = [],
-    importModuleDynamically,
-  } = options;
-
-  validateString(filename, 'options.filename');
-  validateUint32(columnOffset, 'options.columnOffset');
-  validateUint32(lineOffset, 'options.lineOffset');
-  if (cachedData !== undefined)
-    validateBuffer(cachedData, 'options.cachedData');
-  validateBoolean(produceCachedData, 'options.produceCachedData');
-  if (parsingContext !== undefined) {
-    if (
-      typeof parsingContext !== 'object' ||
-      parsingContext === null ||
-      !isContext(parsingContext)
-    ) {
-      throw new ERR_INVALID_ARG_TYPE(
-        'options.parsingContext',
-        'Context',
-        parsingContext
-      );
-    }
-  }
-  validateArray(contextExtensions, 'options.contextExtensions');
-  ArrayPrototypeForEach(contextExtensions, (extension, i) => {
-    const name = `options.contextExtensions[${i}]`;
-    validateObject(extension, name, { nullable: true });
-  });
-
-  const result = _compileFunction(
-    code,
-    filename,
-    lineOffset,
-    columnOffset,
-    cachedData,
-    produceCachedData,
-    parsingContext,
-    contextExtensions,
-    params
-  );
-
-  if (produceCachedData) {
-    result.function.cachedDataProduced = result.cachedDataProduced;
-  }
-
-  if (result.cachedData) {
-    result.function.cachedData = result.cachedData;
-  }
-
-  if (importModuleDynamically !== undefined) {
-    validateFunction(importModuleDynamically,
-                     'options.importModuleDynamically');
-    const { importModuleDynamicallyWrap } =
-      require('internal/vm/module');
-    const { callbackMap } = internalBinding('module_wrap');
-    const wrapped = importModuleDynamicallyWrap(importModuleDynamically);
-    const func = result.function;
-    callbackMap.set(result.cacheKey, {
-      importModuleDynamically: (s, _k, i) => wrapped(s, func, i),
-    });
-  }
-
-  return result.function;
+  return internalCompileFunction(code, params, options).function;
 }
 
 const measureMemoryModes = {
diff --git a/src/env_properties.h b/src/env_properties.h
index c708518efa293f..fe065c172b3957 100644
--- a/src/env_properties.h
+++ b/src/env_properties.h
@@ -275,6 +275,7 @@
   V(sni_context_err_string, "Invalid SNI context")                             \
   V(sni_context_string, "sni_context")                                         \
   V(source_string, "source")                                                   \
+  V(source_map_url_string, "sourceMapURL")                                     \
   V(stack_string, "stack")                                                     \
   V(standard_name_string, "standardName")                                      \
   V(start_time_string, "startTime")                                            \
diff --git a/src/node_contextify.cc b/src/node_contextify.cc
index efc09a7a0faa53..1bfa01c19059b4 100644
--- a/src/node_contextify.cc
+++ b/src/node_contextify.cc
@@ -843,12 +843,11 @@ void ContextifyScript::New(const FunctionCallbackInfo<Value>& args) {
   ShouldNotAbortOnUncaughtScope no_abort_scope(env);
   Context::Scope scope(parsing_context);
 
-  MaybeLocal<UnboundScript> v8_script = ScriptCompiler::CompileUnboundScript(
-      isolate,
-      &source,
-      compile_options);
+  MaybeLocal<UnboundScript> maybe_v8_script =
+      ScriptCompiler::CompileUnboundScript(isolate, &source, compile_options);
 
-  if (v8_script.IsEmpty()) {
+  Local<UnboundScript> v8_script;
+  if (!maybe_v8_script.ToLocal(&v8_script)) {
     errors::DecorateErrorStack(env, try_catch);
     no_abort_scope.Close();
     if (!try_catch.HasTerminated())
@@ -857,7 +856,7 @@ void ContextifyScript::New(const FunctionCallbackInfo<Value>& args) {
                      "ContextifyScript::New");
     return;
   }
-  contextify_script->script_.Reset(isolate, v8_script.ToLocalChecked());
+  contextify_script->script_.Reset(isolate, v8_script);
 
   Local<Context> env_context = env->context();
   if (compile_options == ScriptCompiler::kConsumeCodeCache) {
@@ -866,8 +865,8 @@ void ContextifyScript::New(const FunctionCallbackInfo<Value>& args) {
         env->cached_data_rejected_string(),
         Boolean::New(isolate, source.GetCachedData()->rejected)).Check();
   } else if (produce_cached_data) {
-    std::unique_ptr<ScriptCompiler::CachedData> cached_data {
-      ScriptCompiler::CreateCodeCache(v8_script.ToLocalChecked()) };
+    std::unique_ptr<ScriptCompiler::CachedData> cached_data{
+        ScriptCompiler::CreateCodeCache(v8_script)};
     bool cached_data_produced = cached_data != nullptr;
     if (cached_data_produced) {
       MaybeLocal<Object> buf = Buffer::Copy(
@@ -883,6 +882,13 @@ void ContextifyScript::New(const FunctionCallbackInfo<Value>& args) {
         env->cached_data_produced_string(),
         Boolean::New(isolate, cached_data_produced)).Check();
   }
+
+  args.This()
+      ->Set(env_context,
+            env->source_map_url_string(),
+            v8_script->GetSourceMappingURL())
+      .Check();
+
   TRACE_EVENT_END0(TRACING_CATEGORY_NODE2(vm, script), "ContextifyScript::New");
 }
 
@@ -1230,6 +1236,12 @@ void ContextifyContext::CompileFunction(
   if (result->Set(parsing_context, env->cache_key_string(), cache_key)
           .IsNothing())
     return;
+  if (result
+          ->Set(parsing_context,
+                env->source_map_url_string(),
+                fn->GetScriptOrigin().SourceMapUrl())
+          .IsNothing())
+    return;
 
   if (produce_cached_data) {
     const std::unique_ptr<ScriptCompiler::CachedData> cached_data(
diff --git a/test/parallel/test-bootstrap-modules.js b/test/parallel/test-bootstrap-modules.js
index d4c03dad0953eb..84bfbba8f03feb 100644
--- a/test/parallel/test-bootstrap-modules.js
+++ b/test/parallel/test-bootstrap-modules.js
@@ -148,6 +148,7 @@ const expectedModules = new Set([
   'NativeModule internal/util/parse_args/parse_args',
   'NativeModule internal/util/types',
   'NativeModule internal/validators',
+  'NativeModule internal/vm',
   'NativeModule internal/vm/module',
   'NativeModule internal/wasm_web_api',
   'NativeModule internal/webstreams/adapters',
diff --git a/test/parallel/test-vm-source-map-url.js b/test/parallel/test-vm-source-map-url.js
new file mode 100644
index 00000000000000..22bfc55550e456
--- /dev/null
+++ b/test/parallel/test-vm-source-map-url.js
@@ -0,0 +1,27 @@
+'use strict';
+
+require('../common');
+const assert = require('assert');
+const vm = require('vm');
+
+function checkSourceMapUrl(source, expectedSourceMapURL) {
+  const script = new vm.Script(source);
+  assert.strictEqual(script.sourceMapURL, expectedSourceMapURL);
+}
+
+// No magic comment
+checkSourceMapUrl(`
+function myFunc() {}
+`, undefined);
+
+// Malformed magic comment
+checkSourceMapUrl(`
+function myFunc() {}
+// sourceMappingURL=sourcemap.json
+`, undefined);
+
+// Expected magic comment
+checkSourceMapUrl(`
+function myFunc() {}
+//# sourceMappingURL=sourcemap.json
+`, 'sourcemap.json');