worker.js 21 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696
  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. var sharedUtil = require('../shared/util.js');
  17. var corePrimitives = require('./primitives.js');
  18. var corePdfManager = require('./pdf_manager.js');
  19. var UNSUPPORTED_FEATURES = sharedUtil.UNSUPPORTED_FEATURES;
  20. var InvalidPDFException = sharedUtil.InvalidPDFException;
  21. var MessageHandler = sharedUtil.MessageHandler;
  22. var MissingPDFException = sharedUtil.MissingPDFException;
  23. var UnexpectedResponseException = sharedUtil.UnexpectedResponseException;
  24. var PasswordException = sharedUtil.PasswordException;
  25. var UnknownErrorException = sharedUtil.UnknownErrorException;
  26. var XRefParseException = sharedUtil.XRefParseException;
  27. var arrayByteLength = sharedUtil.arrayByteLength;
  28. var arraysToBytes = sharedUtil.arraysToBytes;
  29. var assert = sharedUtil.assert;
  30. var createPromiseCapability = sharedUtil.createPromiseCapability;
  31. var info = sharedUtil.info;
  32. var warn = sharedUtil.warn;
  33. var setVerbosityLevel = sharedUtil.setVerbosityLevel;
  34. var isNodeJS = sharedUtil.isNodeJS;
  35. var Ref = corePrimitives.Ref;
  36. var LocalPdfManager = corePdfManager.LocalPdfManager;
  37. var NetworkPdfManager = corePdfManager.NetworkPdfManager;
  38. var WorkerTask = function WorkerTaskClosure() {
  39. function WorkerTask(name) {
  40. this.name = name;
  41. this.terminated = false;
  42. this._capability = createPromiseCapability();
  43. }
  44. WorkerTask.prototype = {
  45. get finished() {
  46. return this._capability.promise;
  47. },
  48. finish: function () {
  49. this._capability.resolve();
  50. },
  51. terminate: function () {
  52. this.terminated = true;
  53. },
  54. ensureNotTerminated: function () {
  55. if (this.terminated) {
  56. throw new Error('Worker task was terminated');
  57. }
  58. }
  59. };
  60. return WorkerTask;
  61. }();
  62. var PDFWorkerStream = function PDFWorkerStreamClosure() {
  63. function PDFWorkerStream(params, msgHandler) {
  64. this._queuedChunks = [];
  65. var initialData = params.initialData;
  66. if (initialData && initialData.length > 0) {
  67. this._queuedChunks.push(initialData);
  68. }
  69. this._msgHandler = msgHandler;
  70. this._isRangeSupported = !params.disableRange;
  71. this._isStreamingSupported = !params.disableStream;
  72. this._contentLength = params.length;
  73. this._fullRequestReader = null;
  74. this._rangeReaders = [];
  75. msgHandler.on('OnDataRange', this._onReceiveData.bind(this));
  76. msgHandler.on('OnDataProgress', this._onProgress.bind(this));
  77. }
  78. PDFWorkerStream.prototype = {
  79. _onReceiveData: function PDFWorkerStream_onReceiveData(args) {
  80. if (args.begin === undefined) {
  81. if (this._fullRequestReader) {
  82. this._fullRequestReader._enqueue(args.chunk);
  83. } else {
  84. this._queuedChunks.push(args.chunk);
  85. }
  86. } else {
  87. var found = this._rangeReaders.some(function (rangeReader) {
  88. if (rangeReader._begin !== args.begin) {
  89. return false;
  90. }
  91. rangeReader._enqueue(args.chunk);
  92. return true;
  93. });
  94. assert(found);
  95. }
  96. },
  97. _onProgress: function PDFWorkerStream_onProgress(evt) {
  98. if (this._rangeReaders.length > 0) {
  99. var firstReader = this._rangeReaders[0];
  100. if (firstReader.onProgress) {
  101. firstReader.onProgress({ loaded: evt.loaded });
  102. }
  103. }
  104. },
  105. _removeRangeReader: function PDFWorkerStream_removeRangeReader(reader) {
  106. var i = this._rangeReaders.indexOf(reader);
  107. if (i >= 0) {
  108. this._rangeReaders.splice(i, 1);
  109. }
  110. },
  111. getFullReader: function PDFWorkerStream_getFullReader() {
  112. assert(!this._fullRequestReader);
  113. var queuedChunks = this._queuedChunks;
  114. this._queuedChunks = null;
  115. return new PDFWorkerStreamReader(this, queuedChunks);
  116. },
  117. getRangeReader: function PDFWorkerStream_getRangeReader(begin, end) {
  118. var reader = new PDFWorkerStreamRangeReader(this, begin, end);
  119. this._msgHandler.send('RequestDataRange', {
  120. begin: begin,
  121. end: end
  122. });
  123. this._rangeReaders.push(reader);
  124. return reader;
  125. },
  126. cancelAllRequests: function PDFWorkerStream_cancelAllRequests(reason) {
  127. if (this._fullRequestReader) {
  128. this._fullRequestReader.cancel(reason);
  129. }
  130. var readers = this._rangeReaders.slice(0);
  131. readers.forEach(function (rangeReader) {
  132. rangeReader.cancel(reason);
  133. });
  134. }
  135. };
  136. function PDFWorkerStreamReader(stream, queuedChunks) {
  137. this._stream = stream;
  138. this._done = false;
  139. this._queuedChunks = queuedChunks || [];
  140. this._requests = [];
  141. this._headersReady = Promise.resolve();
  142. stream._fullRequestReader = this;
  143. this.onProgress = null;
  144. }
  145. PDFWorkerStreamReader.prototype = {
  146. _enqueue: function PDFWorkerStreamReader_enqueue(chunk) {
  147. if (this._done) {
  148. return;
  149. }
  150. if (this._requests.length > 0) {
  151. var requestCapability = this._requests.shift();
  152. requestCapability.resolve({
  153. value: chunk,
  154. done: false
  155. });
  156. return;
  157. }
  158. this._queuedChunks.push(chunk);
  159. },
  160. get headersReady() {
  161. return this._headersReady;
  162. },
  163. get isRangeSupported() {
  164. return this._stream._isRangeSupported;
  165. },
  166. get isStreamingSupported() {
  167. return this._stream._isStreamingSupported;
  168. },
  169. get contentLength() {
  170. return this._stream._contentLength;
  171. },
  172. read: function PDFWorkerStreamReader_read() {
  173. if (this._queuedChunks.length > 0) {
  174. var chunk = this._queuedChunks.shift();
  175. return Promise.resolve({
  176. value: chunk,
  177. done: false
  178. });
  179. }
  180. if (this._done) {
  181. return Promise.resolve({
  182. value: undefined,
  183. done: true
  184. });
  185. }
  186. var requestCapability = createPromiseCapability();
  187. this._requests.push(requestCapability);
  188. return requestCapability.promise;
  189. },
  190. cancel: function PDFWorkerStreamReader_cancel(reason) {
  191. this._done = true;
  192. this._requests.forEach(function (requestCapability) {
  193. requestCapability.resolve({
  194. value: undefined,
  195. done: true
  196. });
  197. });
  198. this._requests = [];
  199. }
  200. };
  201. function PDFWorkerStreamRangeReader(stream, begin, end) {
  202. this._stream = stream;
  203. this._begin = begin;
  204. this._end = end;
  205. this._queuedChunk = null;
  206. this._requests = [];
  207. this._done = false;
  208. this.onProgress = null;
  209. }
  210. PDFWorkerStreamRangeReader.prototype = {
  211. _enqueue: function PDFWorkerStreamRangeReader_enqueue(chunk) {
  212. if (this._done) {
  213. return;
  214. }
  215. if (this._requests.length === 0) {
  216. this._queuedChunk = chunk;
  217. } else {
  218. var requestsCapability = this._requests.shift();
  219. requestsCapability.resolve({
  220. value: chunk,
  221. done: false
  222. });
  223. this._requests.forEach(function (requestCapability) {
  224. requestCapability.resolve({
  225. value: undefined,
  226. done: true
  227. });
  228. });
  229. this._requests = [];
  230. }
  231. this._done = true;
  232. this._stream._removeRangeReader(this);
  233. },
  234. get isStreamingSupported() {
  235. return false;
  236. },
  237. read: function PDFWorkerStreamRangeReader_read() {
  238. if (this._queuedChunk) {
  239. return Promise.resolve({
  240. value: this._queuedChunk,
  241. done: false
  242. });
  243. }
  244. if (this._done) {
  245. return Promise.resolve({
  246. value: undefined,
  247. done: true
  248. });
  249. }
  250. var requestCapability = createPromiseCapability();
  251. this._requests.push(requestCapability);
  252. return requestCapability.promise;
  253. },
  254. cancel: function PDFWorkerStreamRangeReader_cancel(reason) {
  255. this._done = true;
  256. this._requests.forEach(function (requestCapability) {
  257. requestCapability.resolve({
  258. value: undefined,
  259. done: true
  260. });
  261. });
  262. this._requests = [];
  263. this._stream._removeRangeReader(this);
  264. }
  265. };
  266. return PDFWorkerStream;
  267. }();
  268. var PDFNetworkStream;
  269. function setPDFNetworkStreamClass(cls) {
  270. PDFNetworkStream = cls;
  271. }
  272. var WorkerMessageHandler = {
  273. setup: function wphSetup(handler, port) {
  274. var testMessageProcessed = false;
  275. handler.on('test', function wphSetupTest(data) {
  276. if (testMessageProcessed) {
  277. return;
  278. }
  279. testMessageProcessed = true;
  280. if (!(data instanceof Uint8Array)) {
  281. handler.send('test', 'main', false);
  282. return;
  283. }
  284. var supportTransfers = data[0] === 255;
  285. handler.postMessageTransfers = supportTransfers;
  286. var xhr = new XMLHttpRequest();
  287. var responseExists = 'response' in xhr;
  288. try {
  289. xhr.responseType;
  290. } catch (e) {
  291. responseExists = false;
  292. }
  293. if (!responseExists) {
  294. handler.send('test', false);
  295. return;
  296. }
  297. handler.send('test', {
  298. supportTypedArray: true,
  299. supportTransfers: supportTransfers
  300. });
  301. });
  302. handler.on('configure', function wphConfigure(data) {
  303. setVerbosityLevel(data.verbosity);
  304. });
  305. handler.on('GetDocRequest', function wphSetupDoc(data) {
  306. return WorkerMessageHandler.createDocumentHandler(data, port);
  307. });
  308. },
  309. createDocumentHandler: function wphCreateDocumentHandler(docParams, port) {
  310. var pdfManager;
  311. var terminated = false;
  312. var cancelXHRs = null;
  313. var WorkerTasks = [];
  314. var docId = docParams.docId;
  315. var docBaseUrl = docParams.docBaseUrl;
  316. var workerHandlerName = docParams.docId + '_worker';
  317. var handler = new MessageHandler(workerHandlerName, docId, port);
  318. handler.postMessageTransfers = docParams.postMessageTransfers;
  319. function ensureNotTerminated() {
  320. if (terminated) {
  321. throw new Error('Worker was terminated');
  322. }
  323. }
  324. function startWorkerTask(task) {
  325. WorkerTasks.push(task);
  326. }
  327. function finishWorkerTask(task) {
  328. task.finish();
  329. var i = WorkerTasks.indexOf(task);
  330. WorkerTasks.splice(i, 1);
  331. }
  332. function loadDocument(recoveryMode) {
  333. var loadDocumentCapability = createPromiseCapability();
  334. var parseSuccess = function parseSuccess() {
  335. var numPagesPromise = pdfManager.ensureDoc('numPages');
  336. var fingerprintPromise = pdfManager.ensureDoc('fingerprint');
  337. var encryptedPromise = pdfManager.ensureXRef('encrypt');
  338. Promise.all([
  339. numPagesPromise,
  340. fingerprintPromise,
  341. encryptedPromise
  342. ]).then(function onDocReady(results) {
  343. var doc = {
  344. numPages: results[0],
  345. fingerprint: results[1],
  346. encrypted: !!results[2]
  347. };
  348. loadDocumentCapability.resolve(doc);
  349. }, parseFailure);
  350. };
  351. var parseFailure = function parseFailure(e) {
  352. loadDocumentCapability.reject(e);
  353. };
  354. pdfManager.ensureDoc('checkHeader', []).then(function () {
  355. pdfManager.ensureDoc('parseStartXRef', []).then(function () {
  356. pdfManager.ensureDoc('parse', [recoveryMode]).then(parseSuccess, parseFailure);
  357. }, parseFailure);
  358. }, parseFailure);
  359. return loadDocumentCapability.promise;
  360. }
  361. function getPdfManager(data, evaluatorOptions) {
  362. var pdfManagerCapability = createPromiseCapability();
  363. var pdfManager;
  364. var source = data.source;
  365. if (source.data) {
  366. try {
  367. pdfManager = new LocalPdfManager(docId, source.data, source.password, evaluatorOptions, docBaseUrl);
  368. pdfManagerCapability.resolve(pdfManager);
  369. } catch (ex) {
  370. pdfManagerCapability.reject(ex);
  371. }
  372. return pdfManagerCapability.promise;
  373. }
  374. var pdfStream;
  375. try {
  376. if (source.chunkedViewerLoading) {
  377. pdfStream = new PDFWorkerStream(source, handler);
  378. } else {
  379. assert(PDFNetworkStream, 'pdfjs/core/network module is not loaded');
  380. pdfStream = new PDFNetworkStream(data);
  381. }
  382. } catch (ex) {
  383. pdfManagerCapability.reject(ex);
  384. return pdfManagerCapability.promise;
  385. }
  386. var fullRequest = pdfStream.getFullReader();
  387. fullRequest.headersReady.then(function () {
  388. if (!fullRequest.isStreamingSupported || !fullRequest.isRangeSupported) {
  389. fullRequest.onProgress = function (evt) {
  390. handler.send('DocProgress', {
  391. loaded: evt.loaded,
  392. total: evt.total
  393. });
  394. };
  395. }
  396. if (!fullRequest.isRangeSupported) {
  397. return;
  398. }
  399. var disableAutoFetch = source.disableAutoFetch || fullRequest.isStreamingSupported;
  400. pdfManager = new NetworkPdfManager(docId, pdfStream, {
  401. msgHandler: handler,
  402. url: source.url,
  403. password: source.password,
  404. length: fullRequest.contentLength,
  405. disableAutoFetch: disableAutoFetch,
  406. rangeChunkSize: source.rangeChunkSize
  407. }, evaluatorOptions, docBaseUrl);
  408. pdfManagerCapability.resolve(pdfManager);
  409. cancelXHRs = null;
  410. }).catch(function (reason) {
  411. pdfManagerCapability.reject(reason);
  412. cancelXHRs = null;
  413. });
  414. var cachedChunks = [], loaded = 0;
  415. var flushChunks = function () {
  416. var pdfFile = arraysToBytes(cachedChunks);
  417. if (source.length && pdfFile.length !== source.length) {
  418. warn('reported HTTP length is different from actual');
  419. }
  420. try {
  421. pdfManager = new LocalPdfManager(docId, pdfFile, source.password, evaluatorOptions, docBaseUrl);
  422. pdfManagerCapability.resolve(pdfManager);
  423. } catch (ex) {
  424. pdfManagerCapability.reject(ex);
  425. }
  426. cachedChunks = [];
  427. };
  428. var readPromise = new Promise(function (resolve, reject) {
  429. var readChunk = function (chunk) {
  430. try {
  431. ensureNotTerminated();
  432. if (chunk.done) {
  433. if (!pdfManager) {
  434. flushChunks();
  435. }
  436. cancelXHRs = null;
  437. return;
  438. }
  439. var data = chunk.value;
  440. loaded += arrayByteLength(data);
  441. if (!fullRequest.isStreamingSupported) {
  442. handler.send('DocProgress', {
  443. loaded: loaded,
  444. total: Math.max(loaded, fullRequest.contentLength || 0)
  445. });
  446. }
  447. if (pdfManager) {
  448. pdfManager.sendProgressiveData(data);
  449. } else {
  450. cachedChunks.push(data);
  451. }
  452. fullRequest.read().then(readChunk, reject);
  453. } catch (e) {
  454. reject(e);
  455. }
  456. };
  457. fullRequest.read().then(readChunk, reject);
  458. });
  459. readPromise.catch(function (e) {
  460. pdfManagerCapability.reject(e);
  461. cancelXHRs = null;
  462. });
  463. cancelXHRs = function () {
  464. pdfStream.cancelAllRequests('abort');
  465. };
  466. return pdfManagerCapability.promise;
  467. }
  468. function setupDoc(data) {
  469. function onSuccess(doc) {
  470. ensureNotTerminated();
  471. handler.send('GetDoc', { pdfInfo: doc });
  472. }
  473. function onFailure(e) {
  474. if (e instanceof PasswordException) {
  475. var task = new WorkerTask('PasswordException: response ' + e.code);
  476. startWorkerTask(task);
  477. handler.sendWithPromise('PasswordRequest', e).then(function (data) {
  478. finishWorkerTask(task);
  479. pdfManager.updatePassword(data.password);
  480. pdfManagerReady();
  481. }).catch(function (ex) {
  482. finishWorkerTask(task);
  483. handler.send('PasswordException', ex);
  484. }.bind(null, e));
  485. } else if (e instanceof InvalidPDFException) {
  486. handler.send('InvalidPDF', e);
  487. } else if (e instanceof MissingPDFException) {
  488. handler.send('MissingPDF', e);
  489. } else if (e instanceof UnexpectedResponseException) {
  490. handler.send('UnexpectedResponse', e);
  491. } else {
  492. handler.send('UnknownError', new UnknownErrorException(e.message, e.toString()));
  493. }
  494. }
  495. function pdfManagerReady() {
  496. ensureNotTerminated();
  497. loadDocument(false).then(onSuccess, function loadFailure(ex) {
  498. ensureNotTerminated();
  499. if (!(ex instanceof XRefParseException)) {
  500. onFailure(ex);
  501. return;
  502. }
  503. pdfManager.requestLoadedStream();
  504. pdfManager.onLoadedStream().then(function () {
  505. ensureNotTerminated();
  506. loadDocument(true).then(onSuccess, onFailure);
  507. });
  508. }, onFailure);
  509. }
  510. ensureNotTerminated();
  511. var evaluatorOptions = {
  512. forceDataSchema: data.disableCreateObjectURL,
  513. maxImageSize: data.maxImageSize === undefined ? -1 : data.maxImageSize,
  514. disableFontFace: data.disableFontFace,
  515. disableNativeImageDecoder: data.disableNativeImageDecoder
  516. };
  517. getPdfManager(data, evaluatorOptions).then(function (newPdfManager) {
  518. if (terminated) {
  519. newPdfManager.terminate();
  520. throw new Error('Worker was terminated');
  521. }
  522. pdfManager = newPdfManager;
  523. handler.send('PDFManagerReady', null);
  524. pdfManager.onLoadedStream().then(function (stream) {
  525. handler.send('DataLoaded', { length: stream.bytes.byteLength });
  526. });
  527. }).then(pdfManagerReady, onFailure);
  528. }
  529. handler.on('GetPage', function wphSetupGetPage(data) {
  530. return pdfManager.getPage(data.pageIndex).then(function (page) {
  531. var rotatePromise = pdfManager.ensure(page, 'rotate');
  532. var refPromise = pdfManager.ensure(page, 'ref');
  533. var userUnitPromise = pdfManager.ensure(page, 'userUnit');
  534. var viewPromise = pdfManager.ensure(page, 'view');
  535. return Promise.all([
  536. rotatePromise,
  537. refPromise,
  538. userUnitPromise,
  539. viewPromise
  540. ]).then(function (results) {
  541. return {
  542. rotate: results[0],
  543. ref: results[1],
  544. userUnit: results[2],
  545. view: results[3]
  546. };
  547. });
  548. });
  549. });
  550. handler.on('GetPageIndex', function wphSetupGetPageIndex(data) {
  551. var ref = new Ref(data.ref.num, data.ref.gen);
  552. var catalog = pdfManager.pdfDocument.catalog;
  553. return catalog.getPageIndex(ref);
  554. });
  555. handler.on('GetDestinations', function wphSetupGetDestinations(data) {
  556. return pdfManager.ensureCatalog('destinations');
  557. });
  558. handler.on('GetDestination', function wphSetupGetDestination(data) {
  559. return pdfManager.ensureCatalog('getDestination', [data.id]);
  560. });
  561. handler.on('GetPageLabels', function wphSetupGetPageLabels(data) {
  562. return pdfManager.ensureCatalog('pageLabels');
  563. });
  564. handler.on('GetAttachments', function wphSetupGetAttachments(data) {
  565. return pdfManager.ensureCatalog('attachments');
  566. });
  567. handler.on('GetJavaScript', function wphSetupGetJavaScript(data) {
  568. return pdfManager.ensureCatalog('javaScript');
  569. });
  570. handler.on('GetOutline', function wphSetupGetOutline(data) {
  571. return pdfManager.ensureCatalog('documentOutline');
  572. });
  573. handler.on('GetMetadata', function wphSetupGetMetadata(data) {
  574. return Promise.all([
  575. pdfManager.ensureDoc('documentInfo'),
  576. pdfManager.ensureCatalog('metadata')
  577. ]);
  578. });
  579. handler.on('GetData', function wphSetupGetData(data) {
  580. pdfManager.requestLoadedStream();
  581. return pdfManager.onLoadedStream().then(function (stream) {
  582. return stream.bytes;
  583. });
  584. });
  585. handler.on('GetStats', function wphSetupGetStats(data) {
  586. return pdfManager.pdfDocument.xref.stats;
  587. });
  588. handler.on('GetAnnotations', function wphSetupGetAnnotations(data) {
  589. return pdfManager.getPage(data.pageIndex).then(function (page) {
  590. return pdfManager.ensure(page, 'getAnnotationsData', [data.intent]);
  591. });
  592. });
  593. handler.on('RenderPageRequest', function wphSetupRenderPage(data) {
  594. var pageIndex = data.pageIndex;
  595. pdfManager.getPage(pageIndex).then(function (page) {
  596. var task = new WorkerTask('RenderPageRequest: page ' + pageIndex);
  597. startWorkerTask(task);
  598. var pageNum = pageIndex + 1;
  599. var start = Date.now();
  600. page.getOperatorList(handler, task, data.intent, data.renderInteractiveForms).then(function (operatorList) {
  601. finishWorkerTask(task);
  602. info('page=' + pageNum + ' - getOperatorList: time=' + (Date.now() - start) + 'ms, len=' + operatorList.totalLength);
  603. }, function (e) {
  604. finishWorkerTask(task);
  605. if (task.terminated) {
  606. return;
  607. }
  608. handler.send('UnsupportedFeature', { featureId: UNSUPPORTED_FEATURES.unknown });
  609. var minimumStackMessage = 'worker.js: while trying to getPage() and getOperatorList()';
  610. var wrappedException;
  611. if (typeof e === 'string') {
  612. wrappedException = {
  613. message: e,
  614. stack: minimumStackMessage
  615. };
  616. } else if (typeof e === 'object') {
  617. wrappedException = {
  618. message: e.message || e.toString(),
  619. stack: e.stack || minimumStackMessage
  620. };
  621. } else {
  622. wrappedException = {
  623. message: 'Unknown exception type: ' + typeof e,
  624. stack: minimumStackMessage
  625. };
  626. }
  627. handler.send('PageError', {
  628. pageNum: pageNum,
  629. error: wrappedException,
  630. intent: data.intent
  631. });
  632. });
  633. });
  634. }, this);
  635. handler.on('GetTextContent', function wphExtractText(data) {
  636. var pageIndex = data.pageIndex;
  637. var normalizeWhitespace = data.normalizeWhitespace;
  638. var combineTextItems = data.combineTextItems;
  639. return pdfManager.getPage(pageIndex).then(function (page) {
  640. var task = new WorkerTask('GetTextContent: page ' + pageIndex);
  641. startWorkerTask(task);
  642. var pageNum = pageIndex + 1;
  643. var start = Date.now();
  644. return page.extractTextContent(task, normalizeWhitespace, combineTextItems).then(function (textContent) {
  645. finishWorkerTask(task);
  646. info('text indexing: page=' + pageNum + ' - time=' + (Date.now() - start) + 'ms');
  647. return textContent;
  648. }, function (reason) {
  649. finishWorkerTask(task);
  650. if (task.terminated) {
  651. return;
  652. }
  653. throw reason;
  654. });
  655. });
  656. });
  657. handler.on('Cleanup', function wphCleanup(data) {
  658. return pdfManager.cleanup();
  659. });
  660. handler.on('Terminate', function wphTerminate(data) {
  661. terminated = true;
  662. if (pdfManager) {
  663. pdfManager.terminate();
  664. pdfManager = null;
  665. }
  666. if (cancelXHRs) {
  667. cancelXHRs();
  668. }
  669. var waitOn = [];
  670. WorkerTasks.forEach(function (task) {
  671. waitOn.push(task.finished);
  672. task.terminate();
  673. });
  674. return Promise.all(waitOn).then(function () {
  675. handler.destroy();
  676. handler = null;
  677. });
  678. });
  679. handler.on('Ready', function wphReady(data) {
  680. setupDoc(docParams);
  681. docParams = null;
  682. });
  683. return workerHandlerName;
  684. }
  685. };
  686. function initializeWorker() {
  687. var handler = new MessageHandler('worker', 'main', self);
  688. WorkerMessageHandler.setup(handler, self);
  689. handler.send('ready', null);
  690. }
  691. if (typeof window === 'undefined' && !isNodeJS()) {
  692. initializeWorker();
  693. }
  694. exports.setPDFNetworkStreamClass = setPDFNetworkStreamClass;
  695. exports.WorkerTask = WorkerTask;
  696. exports.WorkerMessageHandler = WorkerMessageHandler;