Add SINGLE_FILE option to embed all subresources into emitted JS

As discussed in #5279, subresource paths are converted into base64 data URIs.

Author: buu700

emcc.py

@@ -27,7 +27,7 @@
 if __name__ == '__main__':
   ToolchainProfiler.record_process_start()
 
-import os, sys, shutil, tempfile, subprocess, shlex, time, re, logging, urllib
+import os, sys, shutil, tempfile, subprocess, shlex, time, re, logging, urllib, base64
 from subprocess import PIPE
 from tools import shared, jsrun, system_libs
 from tools.shared import execute, suffix, unsuffixed, unsuffixed_basename, WINDOWS, safe_move
@@ -513,6 +513,16 @@ def filter_emscripten_options(argv):
 
   # ---------------- Utilities ---------------
 
+  # Returns the run-time subresource location for accessing by such means as XHR
+  def get_subresource_location(path, media_type):
+    if shared.Settings.SINGLE_FILE:
+      f = open(path, 'rb')
+      data = base64.b64encode(f.read())
+      f.close()
+      return 'data:' + media_type + ';base64,' + data
+    else:
+      return os.path.basename(path)
+
   seen_names = {}
   def uniquename(name):
     if name not in seen_names:
@@ -799,7 +809,7 @@ def detect_fixed_language_mode(args):
         options.separate_asm = True
         logging.warning('forcing separate asm output (--separate-asm), because -s PRECISE_F32=2 or -s USE_PTHREADS=2 was passed.')
       if options.separate_asm:
-        shared.Settings.SEPARATE_ASM = os.path.basename(asm_target)
+        shared.Settings.SEPARATE_ASM = get_subresource_location(asm_target, 'application/javascript')
 
       if 'EMCC_STRICT' in os.environ:
         shared.Settings.STRICT = os.environ.get('EMCC_STRICT') != '0'
@@ -1116,9 +1126,9 @@ def check(input_file):
 
       if shared.Settings.BINARYEN:
         # set file locations, so that JS glue can find what it needs
-        shared.Settings.WASM_TEXT_FILE = os.path.basename(wasm_text_target)
-        shared.Settings.WASM_BINARY_FILE = os.path.basename(wasm_binary_target)
-        shared.Settings.ASMJS_CODE_FILE = os.path.basename(asm_target)
+        shared.Settings.WASM_TEXT_FILE = get_subresource_location(wasm_text_target, 'text/plain')
+        shared.Settings.WASM_BINARY_FILE = get_subresource_location(wasm_binary_target, 'application/octet-stream')
+        shared.Settings.ASMJS_CODE_FILE = get_subresource_location(asm_target, 'application/javascript')
 
         shared.Settings.ASM_JS = 2 # when targeting wasm, we use a wasm Memory, but that is not compatible with asm.js opts
         shared.Settings.GLOBAL_BASE = 1024 # leave some room for mapping global vars
@@ -1624,12 +1634,12 @@ def repl(m):
             # Copy into temp dir as well, so can be run there too
             shared.safe_copy(memfile, os.path.join(shared.get_emscripten_temp_dir(), os.path.basename(memfile)))
           if not shared.Settings.BINARYEN:
-            return 'memoryInitializer = "%s";' % os.path.basename(memfile)
+            return 'memoryInitializer = "%s";' % get_subresource_location(memfile, 'application/octet-stream')
           else:
             # with wasm, we may have the mem init file in the wasm binary already
             return ('memoryInitializer = Module["wasmJSMethod"].indexOf("asmjs") >= 0 || '
                     'Module["wasmJSMethod"].indexOf("interpret-asm2wasm") >= 0 ? "%s" : null;'
-                    % os.path.basename(memfile))
+                    % get_subresource_location(memfile, 'application/octet-stream'))
         src = re.sub(shared.JS.memory_initializer_pattern, repl, open(final).read(), count=1)
         open(final + '.mem.js', 'w').write(src)
         final += '.mem.js'
@@ -2397,6 +2407,8 @@ def generate_html(target, options, js_target, target_basename,
 ''' % (shared.Settings.EMTERPRETIFY_FILE, script.inline)
 
   if options.memory_init_file:
+    memfile_location = get_subresource_location(memfile, 'application/octet-stream')
+
     # start to load the memory init file in the HTML, in parallel with the JS
     script.un_src()
     script.inline = ('''
@@ -2412,13 +2424,15 @@ def generate_html(target, options, js_target, target_basename,
             meminitXHR.responseType = 'arraybuffer';
             meminitXHR.send(null);
           })();
-''' % os.path.basename(memfile)) + script.inline
+''' % memfile_location) + script.inline
 
   # Download .asm.js if --separate-asm was passed in an asm.js build, or if 'asmjs' is one
   # of the wasm run methods.
   if not options.separate_asm or (shared.Settings.BINARYEN and 'asmjs' not in shared.Settings.BINARYEN_METHOD):
     assert len(asm_mods) == 0, 'no --separate-asm means no client code mods are possible'
   else:
+    asm_target_location = get_subresource_location(asm_target, 'application/javascript')
+
     script.un_src()
     if len(asm_mods) == 0:
       # just load the asm, then load the rest
@@ -2431,7 +2445,7 @@ def generate_html(target, options, js_target, target_basename,
       }, 1); // delaying even 1ms is enough to allow compilation memory to be reclaimed
     };
     document.body.appendChild(script);
-''' % (os.path.basename(asm_target), script.inline)
+''' % (asm_target_location, script.inline)
     else:
       # may need to modify the asm code, load it as text, modify, and load asynchronously
       script.inline = '''
@@ -2454,9 +2468,11 @@ def generate_html(target, options, js_target, target_basename,
       document.body.appendChild(script);
     };
     codeXHR.send(null);
-''' % (os.path.basename(asm_target), '\n'.join(asm_mods), script.inline)
+''' % (asm_target_location, '\n'.join(asm_mods), script.inline)
 
   if shared.Settings.BINARYEN and not shared.Settings.BINARYEN_ASYNC_COMPILATION:
+    wasm_binary_target_location = get_subresource_location(wasm_binary_target, 'application/octet-stream')
+
     # We need to load the wasm file before anything else, it has to be synchronously ready TODO: optimize
     script.un_src()
     script.inline = '''
@@ -2468,7 +2484,7 @@ def generate_html(target, options, js_target, target_basename,
 %s
           };
           wasmXHR.send(null);
-''' % (os.path.basename(wasm_binary_target), script.inline)
+''' % (wasm_binary_target_location, script.inline)
 
   html = open(target, 'wb')
   html_contents = shell.replace('{{{ SCRIPT }}}', script.replacement())

src/settings.js

@@ -856,6 +856,13 @@ var FETCH = 0; // If nonzero, enables emscripten_fetch API.
 
 var ASMFS = 0; // If set to 1, uses the multithreaded filesystem that is implemented within the asm.js module, using emscripten_fetch. Implies -s FETCH=1.
 
+var SINGLE_FILE = 0; // If set to 1, embeds all subresources in the emitted JS file
+                     // by converting their file names into base64 data URIs.
+                     //
+                     // Note that using this option may require a change to consuming
+                     // pages' Content Security Policies, specifically adding data:
+                     // to their connect-src directives.
+
 var WASM_TEXT_FILE = ''; // name of the file containing wasm text, if relevant
 var WASM_BINARY_FILE = ''; // name of the file containing wasm binary, if relevant
 var ASMJS_CODE_FILE = ''; // name of the file containing asm.js, if relevant