Add option to use source phase imports for wasm module loading

Fixes: emscripten-core#23047

package.json

@@ -1,6 +1,7 @@
 {
   "private": true,
   "devDependencies": {
+    "@babel/plugin-proposal-import-wasm-source": "^7.25.9",
     "@eslint/eslintrc": "^3.1.0",
     "@eslint/js": "^9.9.1",
     "es-check": "^7.2.1",

src/preamble.js

@@ -603,6 +603,37 @@ function instrumentWasmTableWithAbort() {
 }
 #endif
 
+#if LOAD_SOURCE_MAP
+var wasmSourceMap;
+#include "source_map_support.js"
+
+function receiveSourceMapJSON(sourceMap) {
+  wasmSourceMap = new WasmSourceMap(sourceMap);
+  {{{ runIfMainThread("removeRunDependency('source-map');") }}}
+}
+#endif
+
+#if (PTHREADS || WASM_WORKERS) && (LOAD_SOURCE_MAP || USE_OFFSET_CONVERTER)
+// When using postMessage to send an object, it is processed by the structured
+// clone algorithm.  The prototype, and hence methods, on that object is then
+// lost. This function adds back the lost prototype.  This does not work with
+// nested objects that has prototypes, but it suffices for WasmSourceMap and
+// WasmOffsetConverter.
+function resetPrototype(constructor, attrs) {
+  var object = Object.create(constructor.prototype);
+  return Object.assign(object, attrs);
+}
+#endif
+
+#if USE_OFFSET_CONVERTER
+var wasmOffsetConverter;
+#include "wasm_offset_converter.js"
+#endif
+
+#if SOURCE_PHASE_IMPORTS
+// import source wasmModule from './{{{ WASM_BINARY_FILE }}}';
+var wasmModule = await WebAssembly.compileStreaming(readAsync('{{{ WASM_BINARY_FILE }}}'));
+#else
 function findWasmBinary() {
 #if EXPORT_ES6 && USE_ES6_IMPORT_META && !SINGLE_FILE && !AUDIO_WORKLET
   if (Module['locateFile']) {
@@ -669,16 +700,6 @@ async function getWasmBinary(binaryFile) {
   return getBinarySync(binaryFile);
 }
 
-#if LOAD_SOURCE_MAP
-var wasmSourceMap;
-#include "source_map_support.js"
-#endif
-
-#if USE_OFFSET_CONVERTER
-var wasmOffsetConverter;
-#include "wasm_offset_converter.js"
-#endif
-
 #if SPLIT_MODULE
 {{{ makeModuleReceiveWithVar('loadSplitModule', undefined, 'instantiateSync',  true) }}}
 var splitModuleProxyHandler = {
@@ -707,13 +728,6 @@ var splitModuleProxyHandler = {
 };
 #endif
 
-#if LOAD_SOURCE_MAP
-function receiveSourceMapJSON(sourceMap) {
-  wasmSourceMap = new WasmSourceMap(sourceMap);
-  {{{ runIfMainThread("removeRunDependency('source-map');") }}}
-}
-#endif
-
 #if SPLIT_MODULE || !WASM_ASYNC_COMPILATION
 function instantiateSync(file, info) {
   var module;
@@ -761,18 +775,6 @@ function instantiateSync(file, info) {
 }
 #endif
 
-#if (PTHREADS || WASM_WORKERS) && (LOAD_SOURCE_MAP || USE_OFFSET_CONVERTER)
-// When using postMessage to send an object, it is processed by the structured
-// clone algorithm.  The prototype, and hence methods, on that object is then
-// lost. This function adds back the lost prototype.  This does not work with
-// nested objects that has prototypes, but it suffices for WasmSourceMap and
-// WasmOffsetConverter.
-function resetPrototype(constructor, attrs) {
-  var object = Object.create(constructor.prototype);
-  return Object.assign(object, attrs);
-}
-#endif
-
 #if WASM_ASYNC_COMPILATION
 async function instantiateArrayBuffer(binaryFile, imports) {
   try {
@@ -873,6 +875,7 @@ async function instantiateAsync(binary, binaryFile, imports) {
   return instantiateArrayBuffer(binaryFile, imports);
 }
 #endif // WASM_ASYNC_COMPILATION
+#endif // SOURCE_PHASE_IMPORTS
 
 function getWasmImports() {
 #if PTHREADS
@@ -1075,8 +1078,10 @@ function getWasmImports() {
   }
 #endif
 
+#if SOURCE_PHASE_IMPORTS
+  return WebAssembly.instantiate(wasmModule, info);
+#else
   wasmBinaryFile ??= findWasmBinary();
-
 #if WASM_ASYNC_COMPILATION
 #if RUNTIME_DEBUG
   dbg('asynchronously preparing wasm');
@@ -1108,6 +1113,7 @@ function getWasmImports() {
   return receiveInstance(result[0]);
 #endif
 #endif // WASM_ASYNC_COMPILATION
+#endif // SOURCE_PHASE_IMPORTS
 }
 
 #if !WASM_BIGINT

src/settings.js

@@ -2186,6 +2186,8 @@ var LEGACY_RUNTIME = false;
 // [link]
 var SIGNATURE_CONVERSIONS = [];
 
+var SOURCE_PHASE_IMPORTS = false;
+
 // For renamed settings the format is:
 // [OLD_NAME, NEW_NAME]
 // For removed settings (which now effectively have a fixed value and can no

test/test_other.py

@@ -359,8 +359,11 @@ def test_emcc_generate_config(self, compiler):
   @parameterized({
     '': ([],),
     'node': (['-sENVIRONMENT=node'],),
+    # load a worker before startup to check ES6 modules there as well
+    'pthreads': (['-pthread', '-sPTHREAD_POOL_SIZE=1'],),
+    'source_phase_imports': (['-sSOURCE_PHASE_IMPORTS', '-sENVIRONMENT=node'],),
   })
-  def test_emcc_output_mjs(self, args):
+  def test_esm(self, args):
     self.run_process([EMCC, '-o', 'hello_world.mjs',
                       '--extern-post-js', test_file('modularize_post_js.js'),
                       test_file('hello_world.c')] + args)
@@ -373,7 +376,7 @@ def test_emcc_output_mjs(self, args):
     'node': (['-sENVIRONMENT=node'],),
   })
   @node_pthreads
-  def test_emcc_output_worker_mjs(self, args):
+  def test_esm_worker(self, args):
     os.mkdir('subdir')
     self.run_process([EMCC, '-o', 'subdir/hello_world.mjs',
                       '-sEXIT_RUNTIME', '-sPROXY_TO_PTHREAD', '-pthread', '-O1',
@@ -386,7 +389,7 @@ def test_emcc_output_worker_mjs(self, args):
     self.assertContained('hello, world!', self.run_js('subdir/hello_world.mjs'))
 
   @node_pthreads
-  def test_emcc_output_worker_mjs_single_file(self):
+  def test_esm_worker_single_file(self):
     self.run_process([EMCC, '-o', 'hello_world.mjs', '-pthread',
                       '--extern-post-js', test_file('modularize_post_js.js'),
                       test_file('hello_world.c'), '-sSINGLE_FILE'])
@@ -395,15 +398,15 @@ def test_emcc_output_worker_mjs_single_file(self):
     self.assertContained("new Worker(new URL('hello_world.mjs', import.meta.url), workerOptions)", src)
     self.assertContained('hello, world!', self.run_js('hello_world.mjs'))
 
-  def test_emcc_output_mjs_closure(self):
+  def test_esm_closure(self):
     self.run_process([EMCC, '-o', 'hello_world.mjs',
                       '--extern-post-js', test_file('modularize_post_js.js'),
                       test_file('hello_world.c'), '--closure=1'])
     src = read_file('hello_world.mjs')
     self.assertContained('new URL("hello_world.wasm", import.meta.url)', src)
     self.assertContained('hello, world!', self.run_js('hello_world.mjs'))
 
-  def test_emcc_output_mjs_web_no_import_meta(self):
+  def test_esm_web_no_import_meta(self):
     # Ensure we don't emit import.meta.url at all for:
     # ENVIRONMENT=web + EXPORT_ES6 + USE_ES6_IMPORT_META=0
     self.run_process([EMCC, '-o', 'hello_world.mjs',
@@ -413,46 +416,25 @@ def test_emcc_output_mjs_web_no_import_meta(self):
     self.assertNotContained('import.meta.url', src)
     self.assertContained('export default Module;', src)
 
-  def test_export_es6_implies_modularize(self):
+  def test_esm_implies_modularize(self):
     self.run_process([EMCC, test_file('hello_world.c'), '-sEXPORT_ES6'])
     src = read_file('a.out.js')
     self.assertContained('export default Module;', src)
 
-  def test_export_es6_requires_modularize(self):
+  def test_esm_requires_modularize(self):
     err = self.expect_fail([EMCC, test_file('hello_world.c'), '-sEXPORT_ES6', '-sMODULARIZE=0'])
     self.assertContained('EXPORT_ES6 requires MODULARIZE to be set', err)
 
-  def test_export_es6_node_requires_import_meta(self):
+  def test_esm_node_requires_import_meta(self):
     err = self.expect_fail([EMCC, test_file('hello_world.c'),
                             '-sENVIRONMENT=node', '-sEXPORT_ES6', '-sUSE_ES6_IMPORT_META=0'])
     self.assertContained('EXPORT_ES6 and ENVIRONMENT=*node* requires USE_ES6_IMPORT_META to be set', err)
 
-  def test_export_es6_allows_export_in_post_js(self):
+  def test_esm_allows_export_in_post_js(self):
     self.run_process([EMCC, test_file('hello_world.c'), '-O3', '-sEXPORT_ES6', '--post-js', test_file('export_module.js')])
     src = read_file('a.out.js')
     self.assertContained('export{doNothing};', src)
 
-  @parameterized({
-    '': (False,),
-    'package_json': (True,),
-  })
-  @parameterized({
-    '': ([],),
-    # load a worker before startup to check ES6 modules there as well
-    'pthreads': (['-pthread', '-sPTHREAD_POOL_SIZE=1'],),
-  })
-  def test_export_es6(self, package_json, args):
-    self.run_process([EMCC, test_file('hello_world.c'), '-sEXPORT_ES6',
-                      '-o', 'hello.mjs'] + args)
-    # In ES6 mode we use MODULARIZE, so we must instantiate an instance of the
-    # module to run it.
-    create_file('runner.mjs', '''
-      import Hello from "./hello.mjs";
-      Hello();
-    ''')
-
-    self.assertContained('hello, world!', self.run_js('runner.mjs'))
-
   @parameterized({
     '': ([],),
     'pthreads': (['-pthread'],),

tools/building.py

@@ -516,8 +516,13 @@ def version_split(v):
 def transpile(filename):
   config = {
     'sourceType': 'script',
-    'targets': {}
+    'targets': {},
+    #'plugins': [
+    #  '@babel/plugin-proposal-import-wasm-source'
+    #],
   }
+  if settings.EXPORT_ES6:
+    config['sourceType'] = 'module'
   if settings.MIN_CHROME_VERSION != UNSUPPORTED:
     config['targets']['chrome'] = str(settings.MIN_CHROME_VERSION)
   if settings.MIN_FIREFOX_VERSION != UNSUPPORTED:
@@ -533,7 +538,8 @@ def transpile(filename):
   config_file = shared.get_temp_files().get('babel_config.json').name
   logger.debug(config_json)
   utils.write_file(config_file, config_json)
-  cmd = shared.get_npm_cmd('babel') + [filename, '-o', outfile, '--presets', '@babel/preset-env', '--config-file', config_file]
+  cmd = shared.get_npm_cmd('babel') + [filename, '-o', outfile,
+  '--plugins=@babel/plugin-proposal-import-wasm-source', '--presets', '@babel/preset-env', '--config-file', config_file]
   check_call(cmd, cwd=path_from_root())
   return outfile
 

tools/link.py

@@ -1178,7 +1178,8 @@ def phase_linker_setup(options, state, newargs):  # noqa: C901, PLR0912, PLR0915
     settings.TRANSPILE = (settings.MIN_FIREFOX_VERSION < 79 or
                           settings.MIN_CHROME_VERSION < 85 or
                           settings.MIN_SAFARI_VERSION < 140000 or
-                          settings.MIN_NODE_VERSION < 160000)
+                          settings.MIN_NODE_VERSION < 160000 or
+                          settings.SOURCE_PHASE_IMPORTS)
 
   # https://caniuse.com/class: FF:45 CHROME:49 SAFARI:9
   supports_es6_classes = (settings.MIN_FIREFOX_VERSION >= 45 and
@@ -2396,15 +2397,11 @@ def modularize():
 
   # Multi-environment ES6 builds require an async function
   async_emit = ''
-  if settings.EXPORT_ES6 and \
+  if settings.EXPORT_ES6 and (settings.SOURCE_PHASE_IMPORTS or \
      settings.ENVIRONMENT_MAY_BE_NODE and \
      settings.ENVIRONMENT_MAY_BE_WEB:
     async_emit = 'async '
 
-  # TODO: Remove when https://bugs.webkit.org/show_bug.cgi?id=223533 is resolved.
-  if async_emit != '' and settings.EXPORT_NAME == 'config':
-    diagnostics.warning('emcc', 'EXPORT_NAME should not be named "config" when targeting Safari')
-
   if settings.MODULARIZE == 'instance':
     src = '''
 export default async function init(moduleArg = {}) {