@ohos-ports/hermes-parser 0.37.0-beta.0

package/dist/src/HermesASTAdapter.js

@@ -0,0 +1,192 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = void 0;
+
+var _ParserVisitorKeys = require("./generated/ParserVisitorKeys");
+
+/**
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ *
+ *  strict
+ * @format
+ */
+
+/**
+ * The base class for transforming the Hermes AST to the desired output format.
+ * Extended by concrete adapters which output an ESTree or Babel AST.
+ */
+class HermesASTAdapter {
+  constructor(options) {
+    this.sourceFilename = void 0;
+    this.sourceType = void 0;
+    this.sourceFilename = options.sourceFilename;
+    this.sourceType = options.sourceType;
+  }
+  /**
+   * Transform the input Hermes AST to the desired output format.
+   * This modifies the input AST in place instead of constructing a new AST.
+   */
+
+
+  transform(program) {
+    // Comments are not traversed via visitor keys
+    const comments = program.comments;
+
+    for (let i = 0; i < comments.length; i++) {
+      const comment = comments[i];
+      this.fixSourceLocation(comment);
+      comments[i] = this.mapComment(comment);
+    } // The first comment may be an interpreter directive and is stored directly on the program node
+
+
+    program.interpreter = comments.length > 0 && comments[0].type === 'InterpreterDirective' ? comments.shift() : null; // Tokens are not traversed via visitor keys
+
+    const tokens = program.tokens;
+
+    if (tokens) {
+      for (let i = 0; i < tokens.length; i++) {
+        this.fixSourceLocation(tokens[i]);
+      }
+    }
+
+    const resultNode = this.mapNode(program);
+
+    if (resultNode.type !== 'Program') {
+      throw new Error(`HermesToESTreeAdapter: Must return a Program node, instead of "${resultNode.type}". `);
+    } // $FlowExpectedError[incompatible-type] We know this is a program at this point.
+
+
+    return resultNode;
+  }
+  /**
+   * Transform a Hermes AST node to the output AST format.
+   *
+   * This may modify the input node in-place and return that same node, or a completely
+   * new node may be constructed and returned. Overriden in child classes.
+   */
+
+
+  mapNode(_node) {
+    throw new Error('Implemented in subclasses');
+  }
+
+  mapNodeDefault(node) {
+    const visitorKeys = _ParserVisitorKeys.HERMES_AST_VISITOR_KEYS[node.type];
+
+    for (const key in visitorKeys) {
+      const childType = visitorKeys[key];
+
+      if (childType === _ParserVisitorKeys.NODE_CHILD) {
+        const child = node[key];
+
+        if (child != null) {
+          node[key] = this.mapNode(child);
+        }
+      } else if (childType === _ParserVisitorKeys.NODE_LIST_CHILD) {
+        const children = node[key];
+
+        for (let i = 0; i < children.length; i++) {
+          const child = children[i];
+
+          if (child != null) {
+            children[i] = this.mapNode(child);
+          }
+        }
+      }
+    }
+
+    return node;
+  }
+  /**
+   * Update the source location for this node depending on the output AST format.
+   * This can modify the input node in-place. Overriden in child classes.
+   */
+
+
+  fixSourceLocation(_node) {
+    throw new Error('Implemented in subclasses');
+  }
+
+  getSourceType() {
+    var _this$sourceType;
+
+    return (_this$sourceType = this.sourceType) != null ? _this$sourceType : 'script';
+  }
+
+  setModuleSourceType() {
+    if (this.sourceType == null) {
+      this.sourceType = 'module';
+    }
+  }
+
+  mapComment(node) {
+    return node;
+  }
+
+  mapEmpty(_node) {
+    // $FlowExpectedError[incompatible-type]
+    return null;
+  }
+
+  mapImportDeclaration(node) {
+    if (node.importKind === 'value') {
+      this.setModuleSourceType();
+    }
+
+    return this.mapNodeDefault(node);
+  }
+
+  mapImportSpecifier(node) {
+    if (node.importKind === 'value') {
+      node.importKind = null;
+    }
+
+    return this.mapNodeDefault(node);
+  }
+
+  mapExportDefaultDeclaration(node) {
+    this.setModuleSourceType();
+    return this.mapNodeDefault(node);
+  }
+
+  mapExportNamedDeclaration(node) {
+    if (node.exportKind === 'value') {
+      this.setModuleSourceType();
+    }
+
+    return this.mapNodeDefault(node);
+  }
+
+  mapExportAllDeclaration(node) {
+    if (node.exportKind === 'value') {
+      this.setModuleSourceType();
+    }
+
+    return this.mapNodeDefault(node);
+  }
+
+  formatError(node, message) {
+    return `${message} (${node.loc.start.line}:${node.loc.start.column})`;
+  }
+
+  getBigIntLiteralValue(bigintString) {
+    const bigint = bigintString // estree spec is to not have a trailing `n` on this property
+    // https://github.com/estree/estree/blob/db962bb417a97effcfe9892f87fbb93c81a68584/es2020.md#bigintliteral
+    .replace(/n$/, '') // `BigInt` doesn't accept numeric separator and `bigint` property should not include numeric separator
+    .replaceAll('_', '');
+    return {
+      bigint,
+      // coerce the string to a bigint value if supported by the environment
+      value: typeof BigInt === 'function' ? BigInt(bigint) : null
+    };
+  }
+
+}
+
+exports.default = HermesASTAdapter;

package/dist/src/HermesParser.js

@@ -0,0 +1,108 @@
+/**
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ *
+ *  strict
+ * @format
+ */
+'use strict';
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.parse = parse;
+
+var _HermesParserDeserializer = _interopRequireDefault(require("./HermesParserDeserializer"));
+
+var _HermesParserWASM = _interopRequireDefault(require("./HermesParserWASM"));
+
+function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; }
+
+let HermesParserWASM;
+let hermesParse;
+let hermesParseResult_free;
+let hermesParseResult_getError;
+let hermesParseResult_getErrorLine;
+let hermesParseResult_getErrorColumn;
+let hermesParseResult_getProgramBuffer;
+let hermesParseResult_getPositionBuffer;
+let hermesParseResult_getPositionBufferSize;
+/**
+ * Init the WASM wrapper code generated by `emscripten` to preparse the
+ * HermesParser WASM code.
+ */
+
+function initHermesParserWASM() {
+  if (HermesParserWASM != null) {
+    return;
+  }
+
+  HermesParserWASM = (0, _HermesParserWASM.default)({
+    /**
+     * The emscripten version of `quit` unconditionally assigns the `status` to
+     * `process.exitCode` which overrides any pre-existing code that has been
+     * set, even if it is non zero. For our use case we never want an
+     * `exitCode` to be set so this override removes that functionality.
+     */
+    quit(_status, toThrow) {
+      throw toThrow;
+    }
+
+  });
+  hermesParse = HermesParserWASM.cwrap('hermesParse', 'number', ['number', 'number', 'number', 'number', 'number', 'number', 'number', 'number']);
+  hermesParseResult_free = HermesParserWASM.cwrap('hermesParseResult_free', 'void', ['number']);
+  hermesParseResult_getError = HermesParserWASM.cwrap('hermesParseResult_getError', 'string', ['number']);
+  hermesParseResult_getErrorLine = HermesParserWASM.cwrap('hermesParseResult_getErrorLine', 'number', ['number']);
+  hermesParseResult_getErrorColumn = HermesParserWASM.cwrap('hermesParseResult_getErrorColumn', 'number', ['number']);
+  hermesParseResult_getProgramBuffer = HermesParserWASM.cwrap('hermesParseResult_getProgramBuffer', 'number', ['number']);
+  hermesParseResult_getPositionBuffer = HermesParserWASM.cwrap('hermesParseResult_getPositionBuffer', 'number', ['number']);
+  hermesParseResult_getPositionBufferSize = HermesParserWASM.cwrap('hermesParseResult_getPositionBufferSize', 'number', ['number']);
+} // Copy a string into the WASM heap and null-terminate
+
+
+function copyToHeap(buffer, addr) {
+  HermesParserWASM.HEAP8.set(buffer, addr);
+  HermesParserWASM.HEAP8[addr + buffer.length] = 0;
+}
+
+function parse(source, options) {
+  initHermesParserWASM(); // Allocate space on heap for source text
+
+  const sourceBuffer = Buffer.from(source, 'utf8');
+
+  const sourceAddr = HermesParserWASM._malloc(sourceBuffer.length + 1);
+
+  if (!sourceAddr) {
+    throw new Error('Parser out of memory');
+  }
+
+  try {
+    // Copy source text onto WASM heap
+    copyToHeap(sourceBuffer, sourceAddr);
+    const parseResult = hermesParse(sourceAddr, sourceBuffer.length + 1, options.flow === 'detect', options.enableExperimentalComponentSyntax, options.enableExperimentalFlowMatchSyntax, options.enableExperimentalFlowRecordSyntax, options.tokens, options.allowReturnOutsideFunction);
+
+    try {
+      // Extract and throw error from parse result if parsing failed
+      const err = hermesParseResult_getError(parseResult);
+
+      if (err) {
+        const syntaxError = new SyntaxError(err); // $FlowExpectedError[prop-missing]
+
+        syntaxError.loc = {
+          line: hermesParseResult_getErrorLine(parseResult),
+          column: hermesParseResult_getErrorColumn(parseResult)
+        };
+        throw syntaxError;
+      }
+
+      const deserializer = new _HermesParserDeserializer.default(hermesParseResult_getProgramBuffer(parseResult), hermesParseResult_getPositionBuffer(parseResult), hermesParseResult_getPositionBufferSize(parseResult), HermesParserWASM, options);
+      return deserializer.deserialize();
+    } finally {
+      hermesParseResult_free(parseResult);
+    }
+  } finally {
+    HermesParserWASM._free(sourceAddr);
+  }
+}

package/dist/src/HermesParserDecodeUTF8String.js

@@ -0,0 +1,68 @@
+/**
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ *
+ *  strict
+ * @format
+ */
+'use strict';
+/**
+ * Decode a UTF-8 encoded string from Hermes with a known length.
+ * Based on Emscripten's UTF8ToString with the following differences:
+ * - Always reads all bytes up to the given length, including null bytes. This
+ *   means that we can decode strings that contain null bytes in the middle.
+ * - Allow UTF-8 encoded code points that are part of a surrogate pair, even though
+ *   this is technically invalid UTF-8 that UTF8ToString would convert to 0xfffd.
+ */
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = HermesParserDecodeUTF8String;
+
+function HermesParserDecodeUTF8String(ptrIn, length, heap) {
+  let ptr = ptrIn;
+  const endPtr = ptr + length;
+  let str = '';
+
+  while (ptr < endPtr) {
+    // ASCII characters fit in single byte code point
+    let u0 = heap[ptr++];
+
+    if (!(u0 & 0x80)) {
+      str += String.fromCharCode(u0);
+      continue;
+    } // Two byte code point
+
+
+    const u1 = heap[ptr++] & 0x3f;
+
+    if ((u0 & 0xe0) === 0xc0) {
+      str += String.fromCharCode((u0 & 0x1f) << 6 | u1);
+      continue;
+    }
+
+    const u2 = heap[ptr++] & 0x3f;
+
+    if ((u0 & 0xf0) === 0xe0) {
+      // Three byte code point
+      u0 = (u0 & 0x0f) << 12 | u1 << 6 | u2;
+    } else {
+      // Four byte code point
+      u0 = (u0 & 0x07) << 18 | u1 << 12 | u2 << 6 | heap[ptr++] & 0x3f;
+    }
+
+    if (u0 < 0x10000) {
+      // Code point fits into a single UTF-16 code unit
+      str += String.fromCharCode(u0);
+    } else {
+      // Code point does not fit into single UTF-16 code unit so convert to surrogate pair
+      u0 -= 0x10000;
+      str += String.fromCharCode(0xd800 | u0 >> 10, 0xdc00 | u0 & 0x3ff);
+    }
+  }
+
+  return str;
+}

package/dist/src/HermesParserDeserializer.js

@@ -0,0 +1,242 @@
+/**
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ *
+ *  strict
+ * @format
+ */
+'use strict';
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = void 0;
+
+var _HermesParserDecodeUTF8String = _interopRequireDefault(require("./HermesParserDecodeUTF8String"));
+
+var _HermesParserNodeDeserializers = _interopRequireDefault(require("./HermesParserNodeDeserializers"));
+
+function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; }
+
+class HermesParserDeserializer {
+  // Matches StoredComment::Kind enum in JSLexer.h
+  // Matches TokenType enum in HermesParserJSSerializer.h
+  constructor(programBuffer, positionBuffer, positionBufferSize, wasmParser, options) {
+    this.programBufferIdx = void 0;
+    this.positionBufferIdx = void 0;
+    this.positionBufferSize = void 0;
+    this.locMap = void 0;
+    this.HEAPU8 = void 0;
+    this.HEAPU32 = void 0;
+    this.HEAPF64 = void 0;
+    this.options = void 0;
+    this.commentTypes = ['CommentLine', 'CommentBlock', 'InterpreterDirective'];
+    this.tokenTypes = ['Boolean', 'Identifier', 'Keyword', 'Null', 'Numeric', 'BigInt', 'Punctuator', 'String', 'RegularExpression', 'Template', 'JSXText'];
+    // Program and position buffer are memory addresses, so we must convert
+    // into indices into HEAPU32 (an array of 4-byte integers).
+    this.programBufferIdx = programBuffer / 4;
+    this.positionBufferIdx = positionBuffer / 4;
+    this.positionBufferSize = positionBufferSize;
+    this.locMap = {};
+    this.HEAPU8 = wasmParser.HEAPU8;
+    this.HEAPU32 = wasmParser.HEAPU32;
+    this.HEAPF64 = wasmParser.HEAPF64;
+    this.options = options;
+  }
+  /**
+   * Consume and return the next 4 bytes in the program buffer.
+   */
+
+
+  next() {
+    const num = this.HEAPU32[this.programBufferIdx++];
+    return num;
+  }
+
+  deserialize() {
+    const program = {
+      type: 'Program',
+      loc: this.addEmptyLoc(),
+      body: this.deserializeNodeList(),
+      comments: this.deserializeComments()
+    };
+
+    if (this.options.tokens === true) {
+      program.tokens = this.deserializeTokens();
+    }
+
+    this.fillLocs();
+    return program;
+  }
+  /**
+   * Booleans are serialized as a single 4-byte integer.
+   */
+
+
+  deserializeBoolean() {
+    return Boolean(this.next());
+  }
+  /**
+   * Numbers are serialized directly into program buffer, taking up 8 bytes
+   * preceded by 4 bytes of alignment padding if necessary.
+   */
+
+
+  deserializeNumber() {
+    let floatIdx; // Numbers are aligned on 8-byte boundaries, so skip padding if we are at
+    // an odd index into the 4-byte aligned program buffer.
+
+    if (this.programBufferIdx % 2 === 0) {
+      floatIdx = this.programBufferIdx / 2;
+      this.programBufferIdx += 2;
+    } else {
+      floatIdx = (this.programBufferIdx + 1) / 2;
+      this.programBufferIdx += 3;
+    }
+
+    return this.HEAPF64[floatIdx];
+  }
+  /**
+   * Strings are serialized as a 4-byte pointer into the heap, followed
+   * by their size as a 4-byte integer. The size is only present if the
+   * pointer is non-null.
+   */
+
+
+  deserializeString() {
+    const ptr = this.next();
+
+    if (ptr === 0) {
+      return null;
+    }
+
+    const size = this.next();
+    return (0, _HermesParserDecodeUTF8String.default)(ptr, size, this.HEAPU8);
+  }
+  /**
+   * Nodes are serialized as a 4-byte integer denoting their node kind,
+   * followed by a 4-byte loc ID, followed by serialized node properties.
+   *
+   * If the node kind is 0 the node is null, otherwise the node kind - 1 is an
+   * index into the array of node deserialization functions.
+   */
+
+
+  deserializeNode() {
+    const nodeType = this.next();
+
+    if (nodeType === 0) {
+      return null;
+    }
+
+    const nodeDeserializer = _HermesParserNodeDeserializers.default[nodeType - 1].bind(this);
+
+    return nodeDeserializer();
+  }
+  /**
+   * Node lists are serialized as a 4-byte integer denoting the number of
+   * elements in the list, followed by the serialized elements.
+   */
+
+
+  deserializeNodeList() {
+    const size = this.next();
+    const nodeList = [];
+
+    for (let i = 0; i < size; i++) {
+      nodeList.push(this.deserializeNode());
+    }
+
+    return nodeList;
+  }
+  /**
+   * Comments are serialized as a node list, where each comment is serialized
+   * as a 4-byte integer denoting comment type, followed by a 4-byte value
+   * denoting the loc ID, followed by a serialized string for the comment value.
+   */
+
+
+  deserializeComments() {
+    const size = this.next();
+    const comments = [];
+
+    for (let i = 0; i < size; i++) {
+      const commentType = this.commentTypes[this.next()];
+      const loc = this.addEmptyLoc();
+      const value = this.deserializeString();
+      comments.push({
+        type: commentType,
+        loc,
+        value
+      });
+    }
+
+    return comments;
+  }
+
+  deserializeTokens() {
+    const size = this.next();
+    const tokens = [];
+
+    for (let i = 0; i < size; i++) {
+      const tokenType = this.tokenTypes[this.next()];
+      const loc = this.addEmptyLoc();
+      const value = this.deserializeString();
+      tokens.push({
+        type: tokenType,
+        loc,
+        value
+      });
+    }
+
+    return tokens;
+  }
+  /**
+   * While deserializing the AST locations are represented by
+   * a 4-byte loc ID. This is used to create a map of loc IDs to empty loc
+   * objects that are filled after the AST has been deserialized.
+   */
+
+
+  addEmptyLoc() {
+    const loc = {};
+    this.locMap[this.next()] = loc;
+    return loc;
+  }
+  /**
+   * Positions are serialized as a loc ID which denotes which loc it is associated with,
+   * followed by kind which denotes whether it is a start or end position,
+   * followed by line, column, and offset (4-bytes each).
+   */
+
+
+  fillLocs() {
+    for (let i = 0; i < this.positionBufferSize; i++) {
+      const locId = this.HEAPU32[this.positionBufferIdx++];
+      const kind = this.HEAPU32[this.positionBufferIdx++];
+      const line = this.HEAPU32[this.positionBufferIdx++];
+      const column = this.HEAPU32[this.positionBufferIdx++];
+      const offset = this.HEAPU32[this.positionBufferIdx++];
+      const loc = this.locMap[locId];
+
+      if (kind === 0) {
+        loc.start = {
+          line,
+          column
+        };
+        loc.rangeStart = offset;
+      } else {
+        loc.end = {
+          line,
+          column
+        };
+        loc.rangeEnd = offset;
+      }
+    }
+  }
+
+}
+
+exports.default = HermesParserDeserializer;