SINGLE_FILE + locateFile fix (#5952) (#5959)

 * add isDataURI helper

Author: buu700

src/base64Utils.js

@@ -6,13 +6,13 @@
 // public domain.  It would be nice if you left this header intact.
 // Base64 code from Tyler Akins -- http://rumkin.com
 
-var keyStr = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=';
-
+/**
+ * Decodes a base64 string.
+ * @param {String} input The string to decode.
+ */
 var decodeBase64 = typeof atob === 'function' ? atob : function (input) {
-  /**
-   * Decodes a base64 string.
-   * @param {String} input The string to decode.
-   */
+  var keyStr = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=';
+
   var output = '';
   var chr1, chr2, chr3;
   var enc1, enc2, enc3, enc4;
@@ -69,13 +69,7 @@ function intArrayFromBase64(s) {
 // If filename is a base64 data URI, parses and returns data (Buffer on node,
 // Uint8Array otherwise). If filename is not a base64 data URI, returns undefined.
 function tryParseAsDataURI(filename) {
-  var dataURIPrefix = 'data:application/octet-stream;base64,';
-
-  if (!(
-    String.prototype.startsWith ?
-      filename.startsWith(dataURIPrefix) :
-      filename.indexOf(dataURIPrefix) === 0
-  )) {
+  if (!isDataURI(filename)) {
     return;
   }
 

src/postamble.js

@@ -40,10 +40,12 @@ if (memoryInitializer && !ENVIRONMENT_IS_PTHREAD) {
 #else
 if (memoryInitializer) {
 #endif
-  if (typeof Module['locateFile'] === 'function') {
-    memoryInitializer = Module['locateFile'](memoryInitializer);
-  } else if (Module['memoryInitializerPrefixURL']) {
-    memoryInitializer = Module['memoryInitializerPrefixURL'] + memoryInitializer;
+  if (!isDataURI(memoryInitializer)) {
+    if (typeof Module['locateFile'] === 'function') {
+      memoryInitializer = Module['locateFile'](memoryInitializer);
+    } else if (Module['memoryInitializerPrefixURL']) {
+      memoryInitializer = Module['memoryInitializerPrefixURL'] + memoryInitializer;
+    }
   }
   if (ENVIRONMENT_IS_NODE || ENVIRONMENT_IS_SHELL) {
     var data = Module['readBinary'](memoryInitializer);

src/preamble.js

@@ -1987,6 +1987,17 @@ Module['FS_createPreloadedFile'] = FS.createPreloadedFile;
 var cyberDWARFFile = '{{{ BUNDLED_CD_DEBUG_FILE }}}';
 #endif
 
+// Prefix of data URIs emitted by SINGLE_FILE and related options.
+var dataURIPrefix = 'data:application/octet-stream;base64,';
+
+// Indicates whether filename is a base64 data URI.
+function isDataURI(filename) {
+  return String.prototype.startsWith ?
+      filename.startsWith(dataURIPrefix) :
+      filename.indexOf(dataURIPrefix) === 0
+  ;
+}
+
 #if BINARYEN
 function integrateWasmJS() {
   // wasm.js has several methods for creating the compiled code module here:
@@ -2009,9 +2020,15 @@ function integrateWasmJS() {
   var asmjsCodeFile = '{{{ ASMJS_CODE_FILE }}}';
 
   if (typeof Module['locateFile'] === 'function') {
-    wasmTextFile = Module['locateFile'](wasmTextFile);
-    wasmBinaryFile = Module['locateFile'](wasmBinaryFile);
-    asmjsCodeFile = Module['locateFile'](asmjsCodeFile);
+    if (!isDataURI(wasmTextFile)) {
+      wasmTextFile = Module['locateFile'](wasmTextFile);
+    }
+    if (!isDataURI(wasmBinaryFile)) {
+      wasmBinaryFile = Module['locateFile'](wasmBinaryFile);
+    }
+    if (!isDataURI(asmjsCodeFile)) {
+      asmjsCodeFile = Module['locateFile'](asmjsCodeFile);
+    }
   }
 
   // utilities
@@ -2257,7 +2274,7 @@ function integrateWasmJS() {
     // Prefer streaming instantiation if available.
     if (!Module['wasmBinary'] &&
         typeof WebAssembly.instantiateStreaming === 'function' &&
-        wasmBinaryFile.indexOf('data:') !== 0 &&
+        !isDataURI(wasmBinaryFile) &&
         typeof fetch === 'function') {
       WebAssembly.instantiateStreaming(fetch(wasmBinaryFile, { credentials: 'same-origin' }), info)
         .then(receiveInstantiatedSource)

tests/test_browser.py

@@ -3708,6 +3708,35 @@ def test_single_file_html(self):
     self.btest('emscripten_main_loop_setimmediate.cpp', '1', args=['-s', 'SINGLE_FILE=1', '-s', 'WASM=1', '-s', "BINARYEN_METHOD='native-wasm'"], also_proxied=True)
     assert os.path.exists('test.html') and not os.path.exists('test.js') and not os.path.exists('test.worker.js')
 
+  # Tests that SINGLE_FILE works as intended with locateFile
+  def test_single_file_locate_file(self):
+    open('src.cpp', 'w').write(self.with_report_result(open(path_from_root('tests', 'browser_test_hello_world.c')).read()))
+
+    for wasm_enabled in [True, False]:
+      args = [PYTHON, EMCC, 'src.cpp', '-o', 'test.js', '-s', 'SINGLE_FILE=1']
+
+      if wasm_enabled:
+        args += ['-s', 'WASM=1']
+
+      run_process(args)
+
+      open('test.html', 'w').write('''
+        <script>
+          var Module = {
+            locateFile: function (path) {
+              if (path.indexOf('data:') === 0) {
+                throw new Error('Unexpected data URI.');
+              }
+
+              return path;
+            }
+          };
+        </script>
+        <script src="test.js"></script>
+      ''')
+
+      self.run_browser('test.html', None, '/report_result?0')
+
   # Tests that SINGLE_FILE works as intended in a Worker in JS output
   def test_single_file_worker_js(self):
     open('src.cpp', 'w').write(self.with_report_result(open(path_from_root('tests', 'browser_test_hello_world.c')).read()))