[MEMFS] Assume FS.write is takes a TypedArray. (#26413)

We have higher level wrappers `createDataFile` and `writeFile` which
also accept `Array` and `string`, but it simpler if the lower level APIs
can just assume `TypedArray`.

This simplifies the code and reduces code size. We have assertions to
check this in debug builds.

Followup to #26398 and #26411

Author: sbc100

ChangeLog.md

@@ -20,6 +20,9 @@ See docs/process.md for more on how version tagging works.
 
 5.0.3 (in development)
 ----------------------
+- The low level FS.write API now only accepts TypedArray.  The higher level
+  writeFile and createDataFile file still also accept string and Array.
+  (#26413)
 - Warn on usage of the deprecated `EMSCRIPTEN` macro (`__EMSCRIPTEN__` should
   be used instead). (#26381)
 - The `-sRELOCATABLE` setting was effectively removed (moved to legacy

src/lib/libcore.js

@@ -2193,7 +2193,7 @@ addToLibrary({
 #endif
       FS.createPath('/', PATH.dirname(name), true, true);
       // canOwn this data in the filesystem, it is a slice of wasm memory that will never change
-      FS.createDataFile(name, null, HEAP8.subarray(content, content + len), true, true, true);
+      FS.createDataFile(name, null, HEAP8.subarray(content, content + len), true, true, /*canOwn=*/true);
     } while ({{{ makeGetValue('ptr', '0', '*') }}});
 #if RUNTIME_DEBUG
     dbg('done preloading data files');

src/lib/libfs.js

@@ -7,6 +7,7 @@
 var LibraryFS = {
   $FS__deps: ['$randomFill', '$PATH', '$PATH_FS', '$TTY', '$MEMFS',
     '$FS_modeStringToFlags',
+    '$FS_fileDataToTypedArray',
     '$FS_getMode',
     '$intArrayFromString',
 #if LibraryManager.has('libidbfs.js')
@@ -1253,9 +1254,13 @@ FS.staticInit();`;
 #endif
       return bytesRead;
     },
+    /**
+     * @param {TypedArray} buffer
+     */
     write(stream, buffer, offset, length, position, canOwn) {
 #if ASSERTIONS
       assert(offset >= 0);
+      assert(buffer.subarray, 'FS.write expects a TypedArray');
 #endif
       if (length < 0 || position < 0) {
         throw new FS.ErrnoError({{{ cDefs.EINVAL }}});
@@ -1346,17 +1351,14 @@ FS.staticInit();`;
       FS.close(stream);
       return buf;
     },
+    /**
+     * @param {TypedArray|Array|string} data
+     */
     writeFile(path, data, opts = {}) {
       opts.flags = opts.flags || {{{ cDefs.O_TRUNC | cDefs.O_CREAT | cDefs.O_WRONLY }}};
       var stream = FS.open(path, opts.flags, opts.mode);
-      if (typeof data == 'string') {
-        data = new Uint8Array(intArrayFromString(data, true));
-      }
-      if (ArrayBuffer.isView(data)) {
-        FS.write(stream, data, 0, data.byteLength, undefined, opts.canOwn);
-      } else {
-        abort('Unsupported data type');
-      }
+      data = FS_fileDataToTypedArray(data);
+      FS.write(stream, data, 0, data.byteLength, undefined, opts.canOwn);
       FS.close(stream);
     },
 
@@ -1604,6 +1606,9 @@ FS.staticInit();`;
       var mode = FS_getMode(canRead, canWrite);
       return FS.create(path, mode);
     },
+    /**
+     * @param {TypedArray|Array|string=} data
+     */
     createDataFile(parent, name, data, canRead, canWrite, canOwn) {
       var path = name;
       if (parent) {
@@ -1613,11 +1618,7 @@ FS.staticInit();`;
       var mode = FS_getMode(canRead, canWrite);
       var node = FS.create(path, mode);
       if (data) {
-        if (typeof data == 'string') {
-          var arr = new Array(data.length);
-          for (var i = 0, len = data.length; i < len; ++i) arr[i] = data.charCodeAt(i);
-          data = arr;
-        }
+        data = FS_fileDataToTypedArray(data);
         // make sure we can write to the file
         FS.chmod(node, mode | {{{ cDefs.S_IWUGO }}});
         var stream = FS.open(node, {{{ cDefs.O_TRUNC | cDefs.O_CREAT | cDefs.O_WRONLY }}});

src/lib/libfs_shared.js

@@ -107,6 +107,16 @@ addToLibrary({
     return mode;
   },
 
+  $FS_fileDataToTypedArray: (data) => {
+    if (typeof data == 'string') {
+      data = intArrayFromString(data, true);
+    }
+    if (!data.subarray) {
+      data = new Uint8Array(data);
+    }
+    return data;
+  },
+
   $FS_stdin_getChar_buffer: [],
 
   // getChar has 3 particular return values:

src/lib/libmemfs.js

@@ -248,16 +248,23 @@ addToLibrary({
         return size;
       },
 
-      // Writes the byte range (buffer[offset], buffer[offset+length]) to offset 'position' into the file pointed by 'stream'
-      // canOwn: A boolean that tells if this function can take ownership of the passed in buffer from the subbuffer portion
-      //         that the typed array view 'buffer' points to. The underlying ArrayBuffer can be larger than that, but
-      //         canOwn=true will not take ownership of the portion outside the bytes addressed by the view. This means that
-      //         with canOwn=true, creating a copy of the bytes is avoided, but the caller shouldn't touch the passed in range
-      //         of bytes anymore since their contents now represent file data inside the filesystem.
+      /**
+       * Writes the byte range (buffer[offset], buffer[offset+length]) to offset
+       * 'position' into the file pointed by 'stream'.
+       * @param {TypedArray} buffer
+       * @param {boolean=} canOwn - A boolean that tells if this function can
+       *     take ownership of the passed in buffer from the subbuffer portion
+       *     that the typed array view 'buffer' points to. The underlying
+       *     ArrayBuffer can be larger than that, but canOwn=true will not take
+       *     ownership of the portion outside the bytes addressed by the view.
+       *     This means that with canOwn=true, creating a copy of the bytes is
+       *     avoided, but the caller shouldn't touch the passed in range of
+       *     bytes anymore since their contents now represent file data inside
+       *     the filesystem.
+       */
       write(stream, buffer, offset, length, position, canOwn) {
 #if ASSERTIONS
-        // The data buffer should be a typed array view
-        assert(!(buffer instanceof ArrayBuffer));
+        assert(buffer.subarray, 'FS.write expects a TypedArray');
 #endif
 #if ALLOW_MEMORY_GROWTH
         // If the buffer is located in main memory (HEAP), and if
@@ -273,35 +280,21 @@ addToLibrary({
         var node = stream.node;
         node.mtime = node.ctime = Date.now();
 
-        if (buffer.subarray) { // This write is from a typed array to a typed array?
-          if (canOwn) {
+        if (canOwn) {
 #if ASSERTIONS
-            assert(position === 0, 'canOwn must imply no weird position inside the file');
+          assert(position === 0, 'canOwn must imply no weird position inside the file');
 #endif
-            node.contents = buffer.subarray(offset, offset + length);
-            node.usedBytes = length;
-            return length;
-          } else if (node.usedBytes === 0 && position === 0) { // If this is a simple first write to an empty file, do a fast set since we don't need to care about old data.
-            node.contents = buffer.slice(offset, offset + length);
-            node.usedBytes = length;
-            return length;
-          } else if (position + length <= node.usedBytes) { // Writing to an already allocated and used subrange of the file?
-            node.contents.set(buffer.subarray(offset, offset + length), position);
-            return length;
-          }
-        }
-
-        // Appending to an existing file and we need to reallocate, or source data did not come as a typed array.
-        MEMFS.expandFileStorage(node, position+length);
-        if (buffer.subarray) {
+          node.contents = buffer.subarray(offset, offset + length);
+          node.usedBytes = length;
+        } else if (node.usedBytes === 0 && position === 0) { // If this is a simple first write to an empty file, do a fast set since we don't need to care about old data.
+          node.contents = buffer.slice(offset, offset + length);
+          node.usedBytes = length;
+        } else {
+          MEMFS.expandFileStorage(node, position+length);
           // Use typed array write which is available.
           node.contents.set(buffer.subarray(offset, offset + length), position);
-        } else {
-          for (var i = 0; i < length; i++) {
-            node.contents[position + i] = buffer[offset + i]; // Or fall back to manual write if not.
-          }
+          node.usedBytes = Math.max(node.usedBytes, position + length);
         }
-        node.usedBytes = Math.max(node.usedBytes, position + length);
         return length;
       },
 

src/lib/libwasmfs.js

@@ -509,20 +509,17 @@ addToLibrary({
     return FS_mknod(path, mode, 0);
   },
 
-  $FS_writeFile__deps: ['_wasmfs_write_file', '$stackSave', '$stackRestore', 'malloc', 'free'],
+  $FS_writeFile__deps: ['$FS_fileDataToTypedArray', '_wasmfs_write_file', '$stackSave', '$stackRestore', 'malloc', 'free'],
   $FS_writeFile: (path, data) => {
     var sp = stackSave();
     var pathBuffer = stringToUTF8OnStack(path);
-    var len = typeof data == 'string' ? lengthBytesUTF8(data) + 1 : data.length;
+    data = FS_fileDataToTypedArray(data);
+    var len = data.length;
     var dataBuffer = _malloc(len);
 #if ASSERTIONS
     assert(dataBuffer);
 #endif
-    if (typeof data == 'string') {
-      len = stringToUTF8(data, dataBuffer, len);
-    } else {
-      HEAPU8.set(data, dataBuffer);
-    }
+    HEAPU8.set(data, dataBuffer);
     var ret = __wasmfs_write_file(pathBuffer, dataBuffer, len);
     _free(dataBuffer);
     stackRestore(sp);

test/codesize/test_codesize_cxx_ctors1.json

@@ -1,10 +1,10 @@
 {
-  "a.out.js": 19334,
-  "a.out.js.gz": 8012,
+  "a.out.js": 19214,
+  "a.out.js.gz": 7982,
   "a.out.nodebug.wasm": 132865,
   "a.out.nodebug.wasm.gz": 49889,
-  "total": 152199,
-  "total_gz": 57901,
+  "total": 152079,
+  "total_gz": 57871,
   "sent": [
     "__cxa_throw",
     "_abort_js",

test/codesize/test_codesize_cxx_ctors2.json

@@ -1,10 +1,10 @@
 {
-  "a.out.js": 19311,
-  "a.out.js.gz": 7998,
+  "a.out.js": 19191,
+  "a.out.js.gz": 7968,
   "a.out.nodebug.wasm": 132285,
   "a.out.nodebug.wasm.gz": 49542,
-  "total": 151596,
-  "total_gz": 57540,
+  "total": 151476,
+  "total_gz": 57510,
   "sent": [
     "__cxa_throw",
     "_abort_js",

test/codesize/test_codesize_cxx_except.json

@@ -1,10 +1,10 @@
 {
-  "a.out.js": 22995,
-  "a.out.js.gz": 8989,
+  "a.out.js": 22875,
+  "a.out.js.gz": 8959,
   "a.out.nodebug.wasm": 172743,
   "a.out.nodebug.wasm.gz": 57374,
-  "total": 195738,
-  "total_gz": 66363,
+  "total": 195618,
+  "total_gz": 66333,
   "sent": [
     "__cxa_begin_catch",
     "__cxa_end_catch",

test/codesize/test_codesize_cxx_except_wasm.json

@@ -1,10 +1,10 @@
 {
-  "a.out.js": 19145,
-  "a.out.js.gz": 7934,
+  "a.out.js": 19025,
+  "a.out.js.gz": 7906,
   "a.out.nodebug.wasm": 148138,
   "a.out.nodebug.wasm.gz": 55261,
-  "total": 167283,
-  "total_gz": 63195,
+  "total": 167163,
+  "total_gz": 63167,
   "sent": [
     "_abort_js",
     "_tzset_js",