Merge pull request #648 from ysangkok/chunked-bin-xhr-lazy-loading

Chunked binary webworker xhr lazy loading

3 files changed, 114 insertions, 9 deletions

diff --git a/src/library.js b/src/library.js
index a89f1251..69642151 100644
--- a/src/library.js
+++ b/src/library.js
@@ -289,7 +289,85 @@ LibraryManager.library = {
     // XHR, which is not possible in browsers except in a web worker! Use preloading,
     // either --preload-file in emcc or FS.createPreloadedFile
     createLazyFile: function(parent, name, url, canRead, canWrite) {
-      var properties = {isDevice: false, url: url};
+
+      if (typeof XMLHttpRequest !== 'undefined') {
+        if (!ENVIRONMENT_IS_WORKER) throw 'Cannot do synchronous binary XHRs outside webworkers in modern browsers. Use --embed-file or --preload-file in emcc';
+      
+        // Lazy chunked Uint8Array (implements get and length from Uint8Array). Actual getting is abstracted away for eventual reuse.
+        var LazyUint8Array = function(chunkSize, length) {
+          this.length = length;
+          this.chunkSize = chunkSize;
+          this.chunks = []; // Loaded chunks. Index is the chunk number
+          this.isLazyUint8Array = true;
+        }
+        LazyUint8Array.prototype.get = function(idx) {
+          if (idx > this.length-1 || idx < 0) {
+            return undefined;
+          }
+          var chunkOffset = idx % chunkSize;
+          var chunkNum = Math.floor(idx / chunkSize);
+          return this.getter(chunkNum)[chunkOffset];
+        }
+        LazyUint8Array.prototype.setDataGetter = function(getter) {
+          this.getter = getter;
+        }
+  
+        // Find length
+        var xhr = new XMLHttpRequest();
+        xhr.open('HEAD', url, false);
+        xhr.send(null);
+        if (!(xhr.status >= 200 && xhr.status < 300 || xhr.status === 304)) throw new Error("Couldn't load " + url + ". Status: " + xhr.status);
+        var datalength = Number(xhr.getResponseHeader("Content-length"));
+        var header;
+        var hasByteServing = (header = xhr.getResponseHeader("Accept-Ranges")) && header === "bytes";
+#if SMALL_CHUNKS
+        var chunkSize = 1024; // Chunk size in bytes
+#else
+        var chunkSize = 1024*1024; // Chunk size in bytes
+#endif
+        if (!hasByteServing) chunkSize = datalength;
+  
+        // Function to get a range from the remote URL.
+        var doXHR = (function(from, to) {
+          if (from > to) throw new Error("invalid range (" + from + ", " + to + ") or no bytes requested!");
+          if (to > datalength-1) throw new Error("only " + datalength + " bytes available! programmer error!");
+  
+          // TODO: Use mozResponseArrayBuffer, responseStream, etc. if available.
+          var xhr = new XMLHttpRequest();
+          xhr.open('GET', url, false);
+          if (datalength !== chunkSize) xhr.setRequestHeader("Range", "bytes=" + from + "-" + to);
+  
+          // Some hints to the browser that we want binary data.
+          if (typeof Uint8Array != 'undefined') xhr.responseType = 'arraybuffer';
+          if (xhr.overrideMimeType) {
+            xhr.overrideMimeType('text/plain; charset=x-user-defined');
+          }
+  
+          xhr.send(null);
+          if (!(xhr.status >= 200 && xhr.status < 300 || xhr.status === 304)) throw new Error("Couldn't load " + url + ". Status: " + xhr.status);
+          if (xhr.response !== undefined) {
+            return new Uint8Array(xhr.response || []);
+          } else {
+            return intArrayFromString(xhr.responseText || '', true);
+          }
+        });
+  
+        var lazyArray = new LazyUint8Array(chunkSize, datalength);
+        lazyArray.setDataGetter(function(chunkNum) {
+          var start = chunkNum * lazyArray.chunkSize;
+          var end = (chunkNum+1) * lazyArray.chunkSize - 1; // including this byte
+          end = Math.min(end, datalength-1); // if datalength-1 is selected, this is the last block
+          if (typeof(lazyArray.chunks[chunkNum]) === "undefined") {
+            lazyArray.chunks[chunkNum] = doXHR(start, end);
+          }
+          if (typeof(lazyArray.chunks[chunkNum]) === "undefined") throw new Error("doXHR failed!");
+          return lazyArray.chunks[chunkNum];
+        });
+        var properties = { isDevice: false, contents: lazyArray };
+      } else {
+        var properties = { isDevice: false, url: url };
+      }
+
       return FS.createFile(parent, name, properties, canRead, canWrite);
     },
     // Preloads a file asynchronously. You can call this before run, for example in
@@ -360,8 +438,7 @@ LibraryManager.library = {
       if (obj.isDevice || obj.isFolder || obj.link || obj.contents) return true;
       var success = true;
       if (typeof XMLHttpRequest !== 'undefined') {
-        // Browser.
-        throw 'Cannot do synchronous binary XHRs in modern browsers. Use --embed-file or --preload-file in emcc';
+        throw new Error("Lazy loading should have been performed (contents set) in createLazyFile, but it was not. Lazy loading only works in web workers. Use --embed-file or --preload-file in emcc on the main thread.");
       } else if (Module['read']) {
         // Command-line.
         try {
@@ -1631,8 +1708,14 @@ LibraryManager.library = {
       }
       var contents = stream.object.contents;
       var size = Math.min(contents.length - offset, nbyte);
-      for (var i = 0; i < size; i++) {
-        {{{ makeSetValue('buf', 'i', 'contents[offset + i]', 'i8') }}}
+      if (contents.isLazyUint8Array) {
+        for (var i = 0; i < size; i++) {
+          {{{ makeSetValue('buf', 'i', 'contents.get(offset + i)', 'i8') }}}
+        }
+      } else {
+        for (var i = 0; i < size; i++) {
+          {{{ makeSetValue('buf', 'i', 'contents[offset + i]', 'i8') }}}
+        }
       }
       bytesRead += size;
       return bytesRead;
diff --git a/src/settings.js b/src/settings.js
index 5b1968b9..4017ad98 100644
--- a/src/settings.js
+++ b/src/settings.js
@@ -255,6 +255,10 @@ var WARN_ON_UNDEFINED_SYMBOLS = 0; // If set to 1, we will warn on any undefined
                                    // the existing buildsystem), and (2) functions might be
                                    // implemented later on, say in --pre-js
 
+var SMALL_CHUNKS = 0; // Use small chunk size for binary synchronous XHR's in Web Workers.
+                      // Used for testing.
+                      // See test_chunked_synchronous_xhr in runner.py and library.js.
+
 // Compiler debugging options
 var DEBUG_TAGS_SHOWING = [];
   // Some useful items:
diff --git a/src/shell.js b/src/shell.js
index 891a6328..3fb6cdbb 100644
--- a/src/shell.js
+++ b/src/shell.js
@@ -44,7 +44,9 @@ if (ENVIRONMENT_IS_NODE) {
   if (!Module['arguments']) {
     Module['arguments'] = process['argv'].slice(2);
   }
-} else if (ENVIRONMENT_IS_SHELL) {
+}
+
+if (ENVIRONMENT_IS_SHELL) {
   Module['print'] = print;
   if (typeof printErr != 'undefined') Module['printErr'] = printErr; // not present in v8 or older sm
 
@@ -62,7 +64,9 @@ if (ENVIRONMENT_IS_NODE) {
       Module['arguments'] = arguments;
     }
   }
-} else if (ENVIRONMENT_IS_WEB) {
+}
+
+if (ENVIRONMENT_IS_WEB && !ENVIRONMENT_IS_WORKER) {
   if (!Module['print']) {
     Module['print'] = function(x) {
       console.log(x);
@@ -74,7 +78,9 @@ if (ENVIRONMENT_IS_NODE) {
       console.log(x);
     };
   }
+}
 
+if (ENVIRONMENT_IS_WEB || ENVIRONMENT_IS_WORKER) {
   Module['read'] = function(url) {
     var xhr = new XMLHttpRequest();
     xhr.open('GET', url, false);
@@ -87,12 +93,24 @@ if (ENVIRONMENT_IS_NODE) {
       Module['arguments'] = arguments;
     }
   }
-} else if (ENVIRONMENT_IS_WORKER) {
+}
+
+if (ENVIRONMENT_IS_WORKER) {
   // We can do very little here...
+  var TRY_USE_DUMP = false;
+  if (!Module['print']) {
+    Module['print'] = (TRY_USE_DUMP && (typeof(dump) !== "undefined") ? (function(x) {
+      dump(x);
+    }) : (function(x) {
+      self.postMessage(x);
+    }));
+  }
 
   Module['load'] = importScripts;
+}
 
-} else {
+if (!ENVIRONMENT_IS_WORKER && !ENVIRONMENT_IS_WEB && !ENVIRONMENT_IS_NODE && !ENVIRONMENT_IS_SHELL) {
+  // Unreachable because SHELL is dependant on the others
   throw 'Unknown runtime environment. Where are we?';
 }