worker.js 21 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591
  1. /* Copyright 2017 Mozilla Foundation
  2. *
  3. * Licensed under the Apache License, Version 2.0 (the "License");
  4. * you may not use this file except in compliance with the License.
  5. * You may obtain a copy of the License at
  6. *
  7. * http://www.apache.org/licenses/LICENSE-2.0
  8. *
  9. * Unless required by applicable law or agreed to in writing, software
  10. * distributed under the License is distributed on an "AS IS" BASIS,
  11. * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  12. * See the License for the specific language governing permissions and
  13. * limitations under the License.
  14. */
  15. 'use strict';
  16. Object.defineProperty(exports, "__esModule", {
  17. value: true
  18. });
  19. exports.WorkerMessageHandler = exports.WorkerTask = undefined;
  20. var _typeof = typeof Symbol === "function" && typeof Symbol.iterator === "symbol" ? function (obj) { return typeof obj; } : function (obj) { return obj && typeof Symbol === "function" && obj.constructor === Symbol && obj !== Symbol.prototype ? "symbol" : typeof obj; };
  21. var _util = require('../shared/util');
  22. var _pdf_manager = require('./pdf_manager');
  23. var _primitives = require('./primitives');
  24. var WorkerTask = function WorkerTaskClosure() {
  25. function WorkerTask(name) {
  26. this.name = name;
  27. this.terminated = false;
  28. this._capability = (0, _util.createPromiseCapability)();
  29. }
  30. WorkerTask.prototype = {
  31. get finished() {
  32. return this._capability.promise;
  33. },
  34. finish: function finish() {
  35. this._capability.resolve();
  36. },
  37. terminate: function terminate() {
  38. this.terminated = true;
  39. },
  40. ensureNotTerminated: function ensureNotTerminated() {
  41. if (this.terminated) {
  42. throw new Error('Worker task was terminated');
  43. }
  44. }
  45. };
  46. return WorkerTask;
  47. }();
  48. ;
  49. var PDFWorkerStream = function PDFWorkerStreamClosure() {
  50. function PDFWorkerStream(msgHandler) {
  51. this._msgHandler = msgHandler;
  52. this._contentLength = null;
  53. this._fullRequestReader = null;
  54. this._rangeRequestReaders = [];
  55. }
  56. PDFWorkerStream.prototype = {
  57. getFullReader: function getFullReader() {
  58. (0, _util.assert)(!this._fullRequestReader);
  59. this._fullRequestReader = new PDFWorkerStreamReader(this._msgHandler);
  60. return this._fullRequestReader;
  61. },
  62. getRangeReader: function getRangeReader(begin, end) {
  63. var reader = new PDFWorkerStreamRangeReader(begin, end, this._msgHandler);
  64. this._rangeRequestReaders.push(reader);
  65. return reader;
  66. },
  67. cancelAllRequests: function cancelAllRequests(reason) {
  68. if (this._fullRequestReader) {
  69. this._fullRequestReader.cancel(reason);
  70. }
  71. var readers = this._rangeRequestReaders.slice(0);
  72. readers.forEach(function (reader) {
  73. reader.cancel(reason);
  74. });
  75. }
  76. };
  77. function PDFWorkerStreamReader(msgHandler) {
  78. var _this = this;
  79. this._msgHandler = msgHandler;
  80. this._contentLength = null;
  81. this._isRangeSupported = false;
  82. this._isStreamingSupported = false;
  83. var readableStream = this._msgHandler.sendWithStream('GetReader');
  84. this._reader = readableStream.getReader();
  85. this._headersReady = this._msgHandler.sendWithPromise('ReaderHeadersReady').then(function (data) {
  86. _this._isStreamingSupported = data.isStreamingSupported;
  87. _this._isRangeSupported = data.isRangeSupported;
  88. _this._contentLength = data.contentLength;
  89. });
  90. }
  91. PDFWorkerStreamReader.prototype = {
  92. get headersReady() {
  93. return this._headersReady;
  94. },
  95. get contentLength() {
  96. return this._contentLength;
  97. },
  98. get isStreamingSupported() {
  99. return this._isStreamingSupported;
  100. },
  101. get isRangeSupported() {
  102. return this._isRangeSupported;
  103. },
  104. read: function read() {
  105. return this._reader.read().then(function (_ref) {
  106. var value = _ref.value,
  107. done = _ref.done;
  108. if (done) {
  109. return {
  110. value: undefined,
  111. done: true
  112. };
  113. }
  114. return {
  115. value: value.buffer,
  116. done: false
  117. };
  118. });
  119. },
  120. cancel: function cancel(reason) {
  121. this._reader.cancel(reason);
  122. }
  123. };
  124. function PDFWorkerStreamRangeReader(begin, end, msgHandler) {
  125. this._msgHandler = msgHandler;
  126. this.onProgress = null;
  127. var readableStream = this._msgHandler.sendWithStream('GetRangeReader', {
  128. begin: begin,
  129. end: end
  130. });
  131. this._reader = readableStream.getReader();
  132. }
  133. PDFWorkerStreamRangeReader.prototype = {
  134. get isStreamingSupported() {
  135. return false;
  136. },
  137. read: function read() {
  138. return this._reader.read().then(function (_ref2) {
  139. var value = _ref2.value,
  140. done = _ref2.done;
  141. if (done) {
  142. return {
  143. value: undefined,
  144. done: true
  145. };
  146. }
  147. return {
  148. value: value.buffer,
  149. done: false
  150. };
  151. });
  152. },
  153. cancel: function cancel(reason) {
  154. this._reader.cancel(reason);
  155. }
  156. };
  157. return PDFWorkerStream;
  158. }();
  159. var WorkerMessageHandler = {
  160. setup: function setup(handler, port) {
  161. var testMessageProcessed = false;
  162. handler.on('test', function wphSetupTest(data) {
  163. if (testMessageProcessed) {
  164. return;
  165. }
  166. testMessageProcessed = true;
  167. if (!(data instanceof Uint8Array)) {
  168. handler.send('test', 'main', false);
  169. return;
  170. }
  171. var supportTransfers = data[0] === 255;
  172. handler.postMessageTransfers = supportTransfers;
  173. var xhr = new XMLHttpRequest();
  174. var responseExists = 'response' in xhr;
  175. try {
  176. xhr.responseType;
  177. } catch (e) {
  178. responseExists = false;
  179. }
  180. if (!responseExists) {
  181. handler.send('test', false);
  182. return;
  183. }
  184. handler.send('test', {
  185. supportTypedArray: true,
  186. supportTransfers: supportTransfers
  187. });
  188. });
  189. handler.on('configure', function wphConfigure(data) {
  190. (0, _util.setVerbosityLevel)(data.verbosity);
  191. });
  192. handler.on('GetDocRequest', function wphSetupDoc(data) {
  193. return WorkerMessageHandler.createDocumentHandler(data, port);
  194. });
  195. },
  196. createDocumentHandler: function createDocumentHandler(docParams, port) {
  197. var pdfManager;
  198. var terminated = false;
  199. var cancelXHRs = null;
  200. var WorkerTasks = [];
  201. var docId = docParams.docId;
  202. var docBaseUrl = docParams.docBaseUrl;
  203. var workerHandlerName = docParams.docId + '_worker';
  204. var handler = new _util.MessageHandler(workerHandlerName, docId, port);
  205. handler.postMessageTransfers = docParams.postMessageTransfers;
  206. function ensureNotTerminated() {
  207. if (terminated) {
  208. throw new Error('Worker was terminated');
  209. }
  210. }
  211. function startWorkerTask(task) {
  212. WorkerTasks.push(task);
  213. }
  214. function finishWorkerTask(task) {
  215. task.finish();
  216. var i = WorkerTasks.indexOf(task);
  217. WorkerTasks.splice(i, 1);
  218. }
  219. function loadDocument(recoveryMode) {
  220. var loadDocumentCapability = (0, _util.createPromiseCapability)();
  221. var parseSuccess = function parseSuccess() {
  222. var numPagesPromise = pdfManager.ensureDoc('numPages');
  223. var fingerprintPromise = pdfManager.ensureDoc('fingerprint');
  224. var encryptedPromise = pdfManager.ensureXRef('encrypt');
  225. Promise.all([numPagesPromise, fingerprintPromise, encryptedPromise]).then(function onDocReady(results) {
  226. var doc = {
  227. numPages: results[0],
  228. fingerprint: results[1],
  229. encrypted: !!results[2]
  230. };
  231. loadDocumentCapability.resolve(doc);
  232. }, parseFailure);
  233. };
  234. var parseFailure = function parseFailure(e) {
  235. loadDocumentCapability.reject(e);
  236. };
  237. pdfManager.ensureDoc('checkHeader', []).then(function () {
  238. pdfManager.ensureDoc('parseStartXRef', []).then(function () {
  239. pdfManager.ensureDoc('parse', [recoveryMode]).then(parseSuccess, parseFailure);
  240. }, parseFailure);
  241. }, parseFailure);
  242. return loadDocumentCapability.promise;
  243. }
  244. function getPdfManager(data, evaluatorOptions) {
  245. var pdfManagerCapability = (0, _util.createPromiseCapability)();
  246. var pdfManager;
  247. var source = data.source;
  248. if (source.data) {
  249. try {
  250. pdfManager = new _pdf_manager.LocalPdfManager(docId, source.data, source.password, evaluatorOptions, docBaseUrl);
  251. pdfManagerCapability.resolve(pdfManager);
  252. } catch (ex) {
  253. pdfManagerCapability.reject(ex);
  254. }
  255. return pdfManagerCapability.promise;
  256. }
  257. var pdfStream,
  258. cachedChunks = [];
  259. try {
  260. pdfStream = new PDFWorkerStream(handler);
  261. } catch (ex) {
  262. pdfManagerCapability.reject(ex);
  263. return pdfManagerCapability.promise;
  264. }
  265. var fullRequest = pdfStream.getFullReader();
  266. fullRequest.headersReady.then(function () {
  267. if (!fullRequest.isRangeSupported) {
  268. return;
  269. }
  270. var disableAutoFetch = source.disableAutoFetch || fullRequest.isStreamingSupported;
  271. pdfManager = new _pdf_manager.NetworkPdfManager(docId, pdfStream, {
  272. msgHandler: handler,
  273. url: source.url,
  274. password: source.password,
  275. length: fullRequest.contentLength,
  276. disableAutoFetch: disableAutoFetch,
  277. rangeChunkSize: source.rangeChunkSize
  278. }, evaluatorOptions, docBaseUrl);
  279. for (var i = 0; i < cachedChunks.length; i++) {
  280. pdfManager.sendProgressiveData(cachedChunks[i]);
  281. }
  282. cachedChunks = [];
  283. pdfManagerCapability.resolve(pdfManager);
  284. cancelXHRs = null;
  285. }).catch(function (reason) {
  286. pdfManagerCapability.reject(reason);
  287. cancelXHRs = null;
  288. });
  289. var loaded = 0;
  290. var flushChunks = function flushChunks() {
  291. var pdfFile = (0, _util.arraysToBytes)(cachedChunks);
  292. if (source.length && pdfFile.length !== source.length) {
  293. (0, _util.warn)('reported HTTP length is different from actual');
  294. }
  295. try {
  296. pdfManager = new _pdf_manager.LocalPdfManager(docId, pdfFile, source.password, evaluatorOptions, docBaseUrl);
  297. pdfManagerCapability.resolve(pdfManager);
  298. } catch (ex) {
  299. pdfManagerCapability.reject(ex);
  300. }
  301. cachedChunks = [];
  302. };
  303. var readPromise = new Promise(function (resolve, reject) {
  304. var readChunk = function readChunk(chunk) {
  305. try {
  306. ensureNotTerminated();
  307. if (chunk.done) {
  308. if (!pdfManager) {
  309. flushChunks();
  310. }
  311. cancelXHRs = null;
  312. return;
  313. }
  314. var data = chunk.value;
  315. loaded += (0, _util.arrayByteLength)(data);
  316. if (!fullRequest.isStreamingSupported) {
  317. handler.send('DocProgress', {
  318. loaded: loaded,
  319. total: Math.max(loaded, fullRequest.contentLength || 0)
  320. });
  321. }
  322. if (pdfManager) {
  323. pdfManager.sendProgressiveData(data);
  324. } else {
  325. cachedChunks.push(data);
  326. }
  327. fullRequest.read().then(readChunk, reject);
  328. } catch (e) {
  329. reject(e);
  330. }
  331. };
  332. fullRequest.read().then(readChunk, reject);
  333. });
  334. readPromise.catch(function (e) {
  335. pdfManagerCapability.reject(e);
  336. cancelXHRs = null;
  337. });
  338. cancelXHRs = function cancelXHRs() {
  339. pdfStream.cancelAllRequests('abort');
  340. };
  341. return pdfManagerCapability.promise;
  342. }
  343. function setupDoc(data) {
  344. function onSuccess(doc) {
  345. ensureNotTerminated();
  346. handler.send('GetDoc', { pdfInfo: doc });
  347. }
  348. function onFailure(e) {
  349. ensureNotTerminated();
  350. if (e instanceof _util.PasswordException) {
  351. var task = new WorkerTask('PasswordException: response ' + e.code);
  352. startWorkerTask(task);
  353. handler.sendWithPromise('PasswordRequest', e).then(function (data) {
  354. finishWorkerTask(task);
  355. pdfManager.updatePassword(data.password);
  356. pdfManagerReady();
  357. }).catch(function (ex) {
  358. finishWorkerTask(task);
  359. handler.send('PasswordException', ex);
  360. }.bind(null, e));
  361. } else if (e instanceof _util.InvalidPDFException) {
  362. handler.send('InvalidPDF', e);
  363. } else if (e instanceof _util.MissingPDFException) {
  364. handler.send('MissingPDF', e);
  365. } else if (e instanceof _util.UnexpectedResponseException) {
  366. handler.send('UnexpectedResponse', e);
  367. } else {
  368. handler.send('UnknownError', new _util.UnknownErrorException(e.message, e.toString()));
  369. }
  370. }
  371. function pdfManagerReady() {
  372. ensureNotTerminated();
  373. loadDocument(false).then(onSuccess, function loadFailure(ex) {
  374. ensureNotTerminated();
  375. if (!(ex instanceof _util.XRefParseException)) {
  376. onFailure(ex);
  377. return;
  378. }
  379. pdfManager.requestLoadedStream();
  380. pdfManager.onLoadedStream().then(function () {
  381. ensureNotTerminated();
  382. loadDocument(true).then(onSuccess, onFailure);
  383. });
  384. }, onFailure);
  385. }
  386. ensureNotTerminated();
  387. var evaluatorOptions = {
  388. forceDataSchema: data.disableCreateObjectURL,
  389. maxImageSize: data.maxImageSize === undefined ? -1 : data.maxImageSize,
  390. disableFontFace: data.disableFontFace,
  391. nativeImageDecoderSupport: data.nativeImageDecoderSupport,
  392. ignoreErrors: data.ignoreErrors
  393. };
  394. getPdfManager(data, evaluatorOptions).then(function (newPdfManager) {
  395. if (terminated) {
  396. newPdfManager.terminate();
  397. throw new Error('Worker was terminated');
  398. }
  399. pdfManager = newPdfManager;
  400. handler.send('PDFManagerReady', null);
  401. pdfManager.onLoadedStream().then(function (stream) {
  402. handler.send('DataLoaded', { length: stream.bytes.byteLength });
  403. });
  404. }).then(pdfManagerReady, onFailure);
  405. }
  406. handler.on('GetPage', function wphSetupGetPage(data) {
  407. return pdfManager.getPage(data.pageIndex).then(function (page) {
  408. var rotatePromise = pdfManager.ensure(page, 'rotate');
  409. var refPromise = pdfManager.ensure(page, 'ref');
  410. var userUnitPromise = pdfManager.ensure(page, 'userUnit');
  411. var viewPromise = pdfManager.ensure(page, 'view');
  412. return Promise.all([rotatePromise, refPromise, userUnitPromise, viewPromise]).then(function (results) {
  413. return {
  414. rotate: results[0],
  415. ref: results[1],
  416. userUnit: results[2],
  417. view: results[3]
  418. };
  419. });
  420. });
  421. });
  422. handler.on('GetPageIndex', function wphSetupGetPageIndex(data) {
  423. var ref = new _primitives.Ref(data.ref.num, data.ref.gen);
  424. var catalog = pdfManager.pdfDocument.catalog;
  425. return catalog.getPageIndex(ref);
  426. });
  427. handler.on('GetDestinations', function wphSetupGetDestinations(data) {
  428. return pdfManager.ensureCatalog('destinations');
  429. });
  430. handler.on('GetDestination', function wphSetupGetDestination(data) {
  431. return pdfManager.ensureCatalog('getDestination', [data.id]);
  432. });
  433. handler.on('GetPageLabels', function wphSetupGetPageLabels(data) {
  434. return pdfManager.ensureCatalog('pageLabels');
  435. });
  436. handler.on('GetPageMode', function wphSetupGetPageMode(data) {
  437. return pdfManager.ensureCatalog('pageMode');
  438. });
  439. handler.on('GetAttachments', function wphSetupGetAttachments(data) {
  440. return pdfManager.ensureCatalog('attachments');
  441. });
  442. handler.on('GetJavaScript', function wphSetupGetJavaScript(data) {
  443. return pdfManager.ensureCatalog('javaScript');
  444. });
  445. handler.on('GetOutline', function wphSetupGetOutline(data) {
  446. return pdfManager.ensureCatalog('documentOutline');
  447. });
  448. handler.on('GetMetadata', function wphSetupGetMetadata(data) {
  449. return Promise.all([pdfManager.ensureDoc('documentInfo'), pdfManager.ensureCatalog('metadata')]);
  450. });
  451. handler.on('GetData', function wphSetupGetData(data) {
  452. pdfManager.requestLoadedStream();
  453. return pdfManager.onLoadedStream().then(function (stream) {
  454. return stream.bytes;
  455. });
  456. });
  457. handler.on('GetStats', function wphSetupGetStats(data) {
  458. return pdfManager.pdfDocument.xref.stats;
  459. });
  460. handler.on('GetAnnotations', function wphSetupGetAnnotations(data) {
  461. return pdfManager.getPage(data.pageIndex).then(function (page) {
  462. return pdfManager.ensure(page, 'getAnnotationsData', [data.intent]);
  463. });
  464. });
  465. handler.on('RenderPageRequest', function wphSetupRenderPage(data) {
  466. var pageIndex = data.pageIndex;
  467. pdfManager.getPage(pageIndex).then(function (page) {
  468. var task = new WorkerTask('RenderPageRequest: page ' + pageIndex);
  469. startWorkerTask(task);
  470. var pageNum = pageIndex + 1;
  471. var start = Date.now();
  472. page.getOperatorList({
  473. handler: handler,
  474. task: task,
  475. intent: data.intent,
  476. renderInteractiveForms: data.renderInteractiveForms
  477. }).then(function (operatorList) {
  478. finishWorkerTask(task);
  479. (0, _util.info)('page=' + pageNum + ' - getOperatorList: time=' + (Date.now() - start) + 'ms, len=' + operatorList.totalLength);
  480. }, function (e) {
  481. finishWorkerTask(task);
  482. if (task.terminated) {
  483. return;
  484. }
  485. handler.send('UnsupportedFeature', { featureId: _util.UNSUPPORTED_FEATURES.unknown });
  486. var minimumStackMessage = 'worker.js: while trying to getPage() and getOperatorList()';
  487. var wrappedException;
  488. if (typeof e === 'string') {
  489. wrappedException = {
  490. message: e,
  491. stack: minimumStackMessage
  492. };
  493. } else if ((typeof e === 'undefined' ? 'undefined' : _typeof(e)) === 'object') {
  494. wrappedException = {
  495. message: e.message || e.toString(),
  496. stack: e.stack || minimumStackMessage
  497. };
  498. } else {
  499. wrappedException = {
  500. message: 'Unknown exception type: ' + (typeof e === 'undefined' ? 'undefined' : _typeof(e)),
  501. stack: minimumStackMessage
  502. };
  503. }
  504. handler.send('PageError', {
  505. pageNum: pageNum,
  506. error: wrappedException,
  507. intent: data.intent
  508. });
  509. });
  510. });
  511. }, this);
  512. handler.on('GetTextContent', function wphExtractText(data, sink) {
  513. var pageIndex = data.pageIndex;
  514. sink.onPull = function (desiredSize) {};
  515. sink.onCancel = function (reason) {};
  516. pdfManager.getPage(pageIndex).then(function (page) {
  517. var task = new WorkerTask('GetTextContent: page ' + pageIndex);
  518. startWorkerTask(task);
  519. var pageNum = pageIndex + 1;
  520. var start = Date.now();
  521. page.extractTextContent({
  522. handler: handler,
  523. task: task,
  524. sink: sink,
  525. normalizeWhitespace: data.normalizeWhitespace,
  526. combineTextItems: data.combineTextItems
  527. }).then(function () {
  528. finishWorkerTask(task);
  529. (0, _util.info)('text indexing: page=' + pageNum + ' - time=' + (Date.now() - start) + 'ms');
  530. sink.close();
  531. }, function (reason) {
  532. finishWorkerTask(task);
  533. if (task.terminated) {
  534. return;
  535. }
  536. sink.error(reason);
  537. throw reason;
  538. });
  539. });
  540. });
  541. handler.on('Cleanup', function wphCleanup(data) {
  542. return pdfManager.cleanup();
  543. });
  544. handler.on('Terminate', function wphTerminate(data) {
  545. terminated = true;
  546. if (pdfManager) {
  547. pdfManager.terminate();
  548. pdfManager = null;
  549. }
  550. if (cancelXHRs) {
  551. cancelXHRs();
  552. }
  553. var waitOn = [];
  554. WorkerTasks.forEach(function (task) {
  555. waitOn.push(task.finished);
  556. task.terminate();
  557. });
  558. return Promise.all(waitOn).then(function () {
  559. handler.destroy();
  560. handler = null;
  561. });
  562. });
  563. handler.on('Ready', function wphReady(data) {
  564. setupDoc(docParams);
  565. docParams = null;
  566. });
  567. return workerHandlerName;
  568. },
  569. initializeFromPort: function initializeFromPort(port) {
  570. var handler = new _util.MessageHandler('worker', 'main', port);
  571. WorkerMessageHandler.setup(handler, port);
  572. handler.send('ready', null);
  573. }
  574. };
  575. function isMessagePort(maybePort) {
  576. return typeof maybePort.postMessage === 'function' && 'onmessage' in maybePort;
  577. }
  578. if (typeof window === 'undefined' && !(0, _util.isNodeJS)() && typeof self !== 'undefined' && isMessagePort(self)) {
  579. WorkerMessageHandler.initializeFromPort(self);
  580. }
  581. exports.WorkerTask = WorkerTask;
  582. exports.WorkerMessageHandler = WorkerMessageHandler;