worker.js 17 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558
  1. /**
  2. * @licstart The following is the entire license notice for the
  3. * Javascript code in this page
  4. *
  5. * Copyright 2020 Mozilla Foundation
  6. *
  7. * Licensed under the Apache License, Version 2.0 (the "License");
  8. * you may not use this file except in compliance with the License.
  9. * You may obtain a copy of the License at
  10. *
  11. * http://www.apache.org/licenses/LICENSE-2.0
  12. *
  13. * Unless required by applicable law or agreed to in writing, software
  14. * distributed under the License is distributed on an "AS IS" BASIS,
  15. * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  16. * See the License for the specific language governing permissions and
  17. * limitations under the License.
  18. *
  19. * @licend The above is the entire license notice for the
  20. * Javascript code in this page
  21. */
  22. "use strict";
  23. Object.defineProperty(exports, "__esModule", {
  24. value: true
  25. });
  26. exports.WorkerMessageHandler = exports.WorkerTask = void 0;
  27. var _util = require("../shared/util.js");
  28. var _primitives = require("./primitives.js");
  29. var _pdf_manager = require("./pdf_manager.js");
  30. var _is_node = require("../shared/is_node.js");
  31. var _message_handler = require("../shared/message_handler.js");
  32. var _worker_stream = require("./worker_stream.js");
  33. var _core_utils = require("./core_utils.js");
  34. var WorkerTask = function WorkerTaskClosure() {
  35. function WorkerTask(name) {
  36. this.name = name;
  37. this.terminated = false;
  38. this._capability = (0, _util.createPromiseCapability)();
  39. }
  40. WorkerTask.prototype = {
  41. get finished() {
  42. return this._capability.promise;
  43. },
  44. finish() {
  45. this._capability.resolve();
  46. },
  47. terminate() {
  48. this.terminated = true;
  49. },
  50. ensureNotTerminated() {
  51. if (this.terminated) {
  52. throw new Error("Worker task was terminated");
  53. }
  54. }
  55. };
  56. return WorkerTask;
  57. }();
  58. exports.WorkerTask = WorkerTask;
  59. var WorkerMessageHandler = {
  60. setup(handler, port) {
  61. var testMessageProcessed = false;
  62. handler.on("test", function wphSetupTest(data) {
  63. if (testMessageProcessed) {
  64. return;
  65. }
  66. testMessageProcessed = true;
  67. if (!(data instanceof Uint8Array)) {
  68. handler.send("test", null);
  69. return;
  70. }
  71. const supportTransfers = data[0] === 255;
  72. handler.postMessageTransfers = supportTransfers;
  73. handler.send("test", {
  74. supportTransfers
  75. });
  76. });
  77. handler.on("configure", function wphConfigure(data) {
  78. (0, _util.setVerbosityLevel)(data.verbosity);
  79. });
  80. handler.on("GetDocRequest", function wphSetupDoc(data) {
  81. return WorkerMessageHandler.createDocumentHandler(data, port);
  82. });
  83. },
  84. createDocumentHandler(docParams, port) {
  85. var pdfManager;
  86. var terminated = false;
  87. var cancelXHRs = null;
  88. var WorkerTasks = [];
  89. const verbosity = (0, _util.getVerbosityLevel)();
  90. const apiVersion = docParams.apiVersion;
  91. const workerVersion = '2.5.207';
  92. if (apiVersion !== workerVersion) {
  93. throw new Error(`The API version "${apiVersion}" does not match ` + `the Worker version "${workerVersion}".`);
  94. }
  95. const enumerableProperties = [];
  96. for (const property in []) {
  97. enumerableProperties.push(property);
  98. }
  99. if (enumerableProperties.length) {
  100. throw new Error("The `Array.prototype` contains unexpected enumerable properties: " + enumerableProperties.join(", ") + "; thus breaking e.g. `for...in` iteration of `Array`s.");
  101. }
  102. if (typeof ReadableStream === "undefined" || typeof Promise.allSettled === "undefined") {
  103. throw new Error("The browser/environment lacks native support for critical " + "functionality used by the PDF.js library (e.g. " + "`ReadableStream` and/or `Promise.allSettled`); " + "please use an ES5-compatible build instead.");
  104. }
  105. var docId = docParams.docId;
  106. var docBaseUrl = docParams.docBaseUrl;
  107. var workerHandlerName = docParams.docId + "_worker";
  108. var handler = new _message_handler.MessageHandler(workerHandlerName, docId, port);
  109. handler.postMessageTransfers = docParams.postMessageTransfers;
  110. function ensureNotTerminated() {
  111. if (terminated) {
  112. throw new Error("Worker was terminated");
  113. }
  114. }
  115. function startWorkerTask(task) {
  116. WorkerTasks.push(task);
  117. }
  118. function finishWorkerTask(task) {
  119. task.finish();
  120. var i = WorkerTasks.indexOf(task);
  121. WorkerTasks.splice(i, 1);
  122. }
  123. async function loadDocument(recoveryMode) {
  124. await pdfManager.ensureDoc("checkHeader");
  125. await pdfManager.ensureDoc("parseStartXRef");
  126. await pdfManager.ensureDoc("parse", [recoveryMode]);
  127. if (!recoveryMode) {
  128. await pdfManager.ensureDoc("checkFirstPage");
  129. }
  130. const [numPages, fingerprint] = await Promise.all([pdfManager.ensureDoc("numPages"), pdfManager.ensureDoc("fingerprint")]);
  131. return {
  132. numPages,
  133. fingerprint
  134. };
  135. }
  136. function getPdfManager(data, evaluatorOptions) {
  137. var pdfManagerCapability = (0, _util.createPromiseCapability)();
  138. let newPdfManager;
  139. var source = data.source;
  140. if (source.data) {
  141. try {
  142. newPdfManager = new _pdf_manager.LocalPdfManager(docId, source.data, source.password, evaluatorOptions, docBaseUrl);
  143. pdfManagerCapability.resolve(newPdfManager);
  144. } catch (ex) {
  145. pdfManagerCapability.reject(ex);
  146. }
  147. return pdfManagerCapability.promise;
  148. }
  149. var pdfStream,
  150. cachedChunks = [];
  151. try {
  152. pdfStream = new _worker_stream.PDFWorkerStream(handler);
  153. } catch (ex) {
  154. pdfManagerCapability.reject(ex);
  155. return pdfManagerCapability.promise;
  156. }
  157. var fullRequest = pdfStream.getFullReader();
  158. fullRequest.headersReady.then(function () {
  159. if (!fullRequest.isRangeSupported) {
  160. return;
  161. }
  162. var disableAutoFetch = source.disableAutoFetch || fullRequest.isStreamingSupported;
  163. newPdfManager = new _pdf_manager.NetworkPdfManager(docId, pdfStream, {
  164. msgHandler: handler,
  165. password: source.password,
  166. length: fullRequest.contentLength,
  167. disableAutoFetch,
  168. rangeChunkSize: source.rangeChunkSize
  169. }, evaluatorOptions, docBaseUrl);
  170. for (let i = 0; i < cachedChunks.length; i++) {
  171. newPdfManager.sendProgressiveData(cachedChunks[i]);
  172. }
  173. cachedChunks = [];
  174. pdfManagerCapability.resolve(newPdfManager);
  175. cancelXHRs = null;
  176. }).catch(function (reason) {
  177. pdfManagerCapability.reject(reason);
  178. cancelXHRs = null;
  179. });
  180. var loaded = 0;
  181. var flushChunks = function () {
  182. var pdfFile = (0, _util.arraysToBytes)(cachedChunks);
  183. if (source.length && pdfFile.length !== source.length) {
  184. (0, _util.warn)("reported HTTP length is different from actual");
  185. }
  186. try {
  187. newPdfManager = new _pdf_manager.LocalPdfManager(docId, pdfFile, source.password, evaluatorOptions, docBaseUrl);
  188. pdfManagerCapability.resolve(newPdfManager);
  189. } catch (ex) {
  190. pdfManagerCapability.reject(ex);
  191. }
  192. cachedChunks = [];
  193. };
  194. var readPromise = new Promise(function (resolve, reject) {
  195. var readChunk = function ({
  196. value,
  197. done
  198. }) {
  199. try {
  200. ensureNotTerminated();
  201. if (done) {
  202. if (!newPdfManager) {
  203. flushChunks();
  204. }
  205. cancelXHRs = null;
  206. return;
  207. }
  208. loaded += (0, _util.arrayByteLength)(value);
  209. if (!fullRequest.isStreamingSupported) {
  210. handler.send("DocProgress", {
  211. loaded,
  212. total: Math.max(loaded, fullRequest.contentLength || 0)
  213. });
  214. }
  215. if (newPdfManager) {
  216. newPdfManager.sendProgressiveData(value);
  217. } else {
  218. cachedChunks.push(value);
  219. }
  220. fullRequest.read().then(readChunk, reject);
  221. } catch (e) {
  222. reject(e);
  223. }
  224. };
  225. fullRequest.read().then(readChunk, reject);
  226. });
  227. readPromise.catch(function (e) {
  228. pdfManagerCapability.reject(e);
  229. cancelXHRs = null;
  230. });
  231. cancelXHRs = function (reason) {
  232. pdfStream.cancelAllRequests(reason);
  233. };
  234. return pdfManagerCapability.promise;
  235. }
  236. function setupDoc(data) {
  237. function onSuccess(doc) {
  238. ensureNotTerminated();
  239. handler.send("GetDoc", {
  240. pdfInfo: doc
  241. });
  242. }
  243. function onFailure(ex) {
  244. ensureNotTerminated();
  245. if (ex instanceof _util.PasswordException) {
  246. var task = new WorkerTask(`PasswordException: response ${ex.code}`);
  247. startWorkerTask(task);
  248. handler.sendWithPromise("PasswordRequest", ex).then(function ({
  249. password
  250. }) {
  251. finishWorkerTask(task);
  252. pdfManager.updatePassword(password);
  253. pdfManagerReady();
  254. }).catch(function () {
  255. finishWorkerTask(task);
  256. handler.send("DocException", ex);
  257. });
  258. } else if (ex instanceof _util.InvalidPDFException || ex instanceof _util.MissingPDFException || ex instanceof _util.UnexpectedResponseException || ex instanceof _util.UnknownErrorException) {
  259. handler.send("DocException", ex);
  260. } else {
  261. handler.send("DocException", new _util.UnknownErrorException(ex.message, ex.toString()));
  262. }
  263. }
  264. function pdfManagerReady() {
  265. ensureNotTerminated();
  266. loadDocument(false).then(onSuccess, function loadFailure(ex) {
  267. ensureNotTerminated();
  268. if (!(ex instanceof _core_utils.XRefParseException)) {
  269. onFailure(ex);
  270. return;
  271. }
  272. pdfManager.requestLoadedStream();
  273. pdfManager.onLoadedStream().then(function () {
  274. ensureNotTerminated();
  275. loadDocument(true).then(onSuccess, onFailure);
  276. });
  277. }, onFailure);
  278. }
  279. ensureNotTerminated();
  280. var evaluatorOptions = {
  281. maxImageSize: data.maxImageSize,
  282. disableFontFace: data.disableFontFace,
  283. ignoreErrors: data.ignoreErrors,
  284. isEvalSupported: data.isEvalSupported,
  285. fontExtraProperties: data.fontExtraProperties
  286. };
  287. getPdfManager(data, evaluatorOptions).then(function (newPdfManager) {
  288. if (terminated) {
  289. newPdfManager.terminate(new _util.AbortException("Worker was terminated."));
  290. throw new Error("Worker was terminated");
  291. }
  292. pdfManager = newPdfManager;
  293. pdfManager.onLoadedStream().then(function (stream) {
  294. handler.send("DataLoaded", {
  295. length: stream.bytes.byteLength
  296. });
  297. });
  298. }).then(pdfManagerReady, onFailure);
  299. }
  300. handler.on("GetPage", function wphSetupGetPage(data) {
  301. return pdfManager.getPage(data.pageIndex).then(function (page) {
  302. return Promise.all([pdfManager.ensure(page, "rotate"), pdfManager.ensure(page, "ref"), pdfManager.ensure(page, "userUnit"), pdfManager.ensure(page, "view")]).then(function ([rotate, ref, userUnit, view]) {
  303. return {
  304. rotate,
  305. ref,
  306. userUnit,
  307. view
  308. };
  309. });
  310. });
  311. });
  312. handler.on("GetPageIndex", function wphSetupGetPageIndex(data) {
  313. var ref = _primitives.Ref.get(data.ref.num, data.ref.gen);
  314. var catalog = pdfManager.pdfDocument.catalog;
  315. return catalog.getPageIndex(ref);
  316. });
  317. handler.on("GetDestinations", function wphSetupGetDestinations(data) {
  318. return pdfManager.ensureCatalog("destinations");
  319. });
  320. handler.on("GetDestination", function wphSetupGetDestination(data) {
  321. return pdfManager.ensureCatalog("getDestination", [data.id]);
  322. });
  323. handler.on("GetPageLabels", function wphSetupGetPageLabels(data) {
  324. return pdfManager.ensureCatalog("pageLabels");
  325. });
  326. handler.on("GetPageLayout", function wphSetupGetPageLayout(data) {
  327. return pdfManager.ensureCatalog("pageLayout");
  328. });
  329. handler.on("GetPageMode", function wphSetupGetPageMode(data) {
  330. return pdfManager.ensureCatalog("pageMode");
  331. });
  332. handler.on("GetViewerPreferences", function (data) {
  333. return pdfManager.ensureCatalog("viewerPreferences");
  334. });
  335. handler.on("GetOpenAction", function (data) {
  336. return pdfManager.ensureCatalog("openAction");
  337. });
  338. handler.on("GetAttachments", function wphSetupGetAttachments(data) {
  339. return pdfManager.ensureCatalog("attachments");
  340. });
  341. handler.on("GetJavaScript", function wphSetupGetJavaScript(data) {
  342. return pdfManager.ensureCatalog("javaScript");
  343. });
  344. handler.on("GetOutline", function wphSetupGetOutline(data) {
  345. return pdfManager.ensureCatalog("documentOutline");
  346. });
  347. handler.on("GetPermissions", function (data) {
  348. return pdfManager.ensureCatalog("permissions");
  349. });
  350. handler.on("GetMetadata", function wphSetupGetMetadata(data) {
  351. return Promise.all([pdfManager.ensureDoc("documentInfo"), pdfManager.ensureCatalog("metadata")]);
  352. });
  353. handler.on("GetData", function wphSetupGetData(data) {
  354. pdfManager.requestLoadedStream();
  355. return pdfManager.onLoadedStream().then(function (stream) {
  356. return stream.bytes;
  357. });
  358. });
  359. handler.on("GetStats", function wphSetupGetStats(data) {
  360. return pdfManager.pdfDocument.xref.stats;
  361. });
  362. handler.on("GetAnnotations", function ({
  363. pageIndex,
  364. intent
  365. }) {
  366. return pdfManager.getPage(pageIndex).then(function (page) {
  367. return page.getAnnotationsData(intent);
  368. });
  369. });
  370. handler.on("GetOperatorList", function wphSetupRenderPage(data, sink) {
  371. var pageIndex = data.pageIndex;
  372. pdfManager.getPage(pageIndex).then(function (page) {
  373. var task = new WorkerTask(`GetOperatorList: page ${pageIndex}`);
  374. startWorkerTask(task);
  375. const start = verbosity >= _util.VerbosityLevel.INFOS ? Date.now() : 0;
  376. page.getOperatorList({
  377. handler,
  378. sink,
  379. task,
  380. intent: data.intent,
  381. renderInteractiveForms: data.renderInteractiveForms
  382. }).then(function (operatorListInfo) {
  383. finishWorkerTask(task);
  384. if (start) {
  385. (0, _util.info)(`page=${pageIndex + 1} - getOperatorList: time=` + `${Date.now() - start}ms, len=${operatorListInfo.length}`);
  386. }
  387. sink.close();
  388. }, function (reason) {
  389. finishWorkerTask(task);
  390. if (task.terminated) {
  391. return;
  392. }
  393. handler.send("UnsupportedFeature", {
  394. featureId: _util.UNSUPPORTED_FEATURES.errorOperatorList
  395. });
  396. sink.error(reason);
  397. });
  398. });
  399. }, this);
  400. handler.on("GetTextContent", function wphExtractText(data, sink) {
  401. var pageIndex = data.pageIndex;
  402. sink.onPull = function (desiredSize) {};
  403. sink.onCancel = function (reason) {};
  404. pdfManager.getPage(pageIndex).then(function (page) {
  405. var task = new WorkerTask("GetTextContent: page " + pageIndex);
  406. startWorkerTask(task);
  407. const start = verbosity >= _util.VerbosityLevel.INFOS ? Date.now() : 0;
  408. page.extractTextContent({
  409. handler,
  410. task,
  411. sink,
  412. normalizeWhitespace: data.normalizeWhitespace,
  413. combineTextItems: data.combineTextItems
  414. }).then(function () {
  415. finishWorkerTask(task);
  416. if (start) {
  417. (0, _util.info)(`page=${pageIndex + 1} - getTextContent: time=` + `${Date.now() - start}ms`);
  418. }
  419. sink.close();
  420. }, function (reason) {
  421. finishWorkerTask(task);
  422. if (task.terminated) {
  423. return;
  424. }
  425. sink.error(reason);
  426. });
  427. });
  428. });
  429. handler.on("FontFallback", function (data) {
  430. return pdfManager.fontFallback(data.id, handler);
  431. });
  432. handler.on("Cleanup", function wphCleanup(data) {
  433. return pdfManager.cleanup(true);
  434. });
  435. handler.on("Terminate", function wphTerminate(data) {
  436. terminated = true;
  437. const waitOn = [];
  438. if (pdfManager) {
  439. pdfManager.terminate(new _util.AbortException("Worker was terminated."));
  440. const cleanupPromise = pdfManager.cleanup();
  441. waitOn.push(cleanupPromise);
  442. pdfManager = null;
  443. } else {
  444. (0, _primitives.clearPrimitiveCaches)();
  445. }
  446. if (cancelXHRs) {
  447. cancelXHRs(new _util.AbortException("Worker was terminated."));
  448. }
  449. WorkerTasks.forEach(function (task) {
  450. waitOn.push(task.finished);
  451. task.terminate();
  452. });
  453. return Promise.all(waitOn).then(function () {
  454. handler.destroy();
  455. handler = null;
  456. });
  457. });
  458. handler.on("Ready", function wphReady(data) {
  459. setupDoc(docParams);
  460. docParams = null;
  461. });
  462. return workerHandlerName;
  463. },
  464. initializeFromPort(port) {
  465. var handler = new _message_handler.MessageHandler("worker", "main", port);
  466. WorkerMessageHandler.setup(handler, port);
  467. handler.send("ready", null);
  468. }
  469. };
  470. exports.WorkerMessageHandler = WorkerMessageHandler;
  471. function isMessagePort(maybePort) {
  472. return typeof maybePort.postMessage === "function" && "onmessage" in maybePort;
  473. }
  474. if (typeof window === "undefined" && !_is_node.isNodeJS && typeof self !== "undefined" && isMessagePort(self)) {
  475. WorkerMessageHandler.initializeFromPort(self);
  476. }