Fetch binary CMap data in the worker-thread, when useWorkerFetch is set

This patch uses the new option added in PR 12726 to *also* allow fetching binary CMap data directly in the worker-thread in browsers.
Given that these changes remove the need to transfer data between threads for the default (browser) use-case, we can also revert the changes in PR 11118 since that simplifies the overall implementation.

src/core/evaluator.js

@@ -96,8 +96,9 @@ const DefaultPartialEvaluatorOptions = Object.freeze({
   ignoreErrors: false,
   isEvalSupported: true,
   fontExtraProperties: false,
-  standardFontDataUrl: null,
   useSystemFonts: true,
+  cMapUrl: null,
+  standardFontDataUrl: null,
 });
 
 const PatternType = {
@@ -360,23 +361,25 @@ class PartialEvaluator {
     if (cachedData) {
       return cachedData;
     }
-    const readableStream = this.handler.sendWithStream("FetchBuiltInCMap", {
-      name,
-    });
-    const reader = readableStream.getReader();
+    let data;
 
-    const data = await new Promise(function (resolve, reject) {
-      function pump() {
-        reader.read().then(function ({ value, done }) {
-          if (done) {
-            return;
-          }
-          resolve(value);
-          pump();
-        }, reject);
+    if (this.options.cMapUrl !== null) {
+      // Only compressed CMaps are (currently) supported here.
+      const url = `${this.options.cMapUrl}${name}.bcmap`;
+      const response = await fetch(url);
+      if (!response.ok) {
+        throw new Error(
+          `fetchBuiltInCMap: failed to fetch file "${url}" with "${response.statusText}".`
+        );
       }
-      pump();
-    });
+      data = {
+        cMapData: new Uint8Array(await response.arrayBuffer()),
+        compressionType: CMapCompressionType.BINARY,
+      };
+    } else {
+      // Get the data on the main-thread instead.
+      data = await this.handler.sendWithPromise("FetchBuiltInCMap", { name });
+    }
 
     if (data.compressionType !== CMapCompressionType.NONE) {
       // Given the size of uncompressed CMaps, only cache compressed ones.

src/core/worker.js

@@ -412,6 +412,7 @@ class WorkerMessageHandler {
         isEvalSupported: data.isEvalSupported,
         fontExtraProperties: data.fontExtraProperties,
         useSystemFonts: data.useSystemFonts,
+        cMapUrl: data.cMapUrl,
         standardFontDataUrl: data.standardFontDataUrl,
       };
 

src/display/api.js

@@ -152,19 +152,19 @@ function setPDFNetworkStreamFactory(pdfNetworkStreamFactory) {
  *   reading built-in CMap files. Providing a custom factory is useful for
  *   environments without Fetch API or `XMLHttpRequest` support, such as
  *   Node.js. The default value is {DOMCMapReaderFactory}.
- * @property {boolean} [useSystemFonts] - When true, fonts that aren't embedded
- *   in the PDF will fallback to a system font. Defaults to true for web
- *   environments and false for node.
+ * @property {boolean} [useSystemFonts] - When `true`, fonts that aren't
+ *   embedded in the PDF document will fallback to a system font.
+ *   The default value is `true` in web environments and `false` in Node.js.
  * @property {string} [standardFontDataUrl] - The URL where the standard font
  *   files are located. Include the trailing slash.
- * @property {boolean} [useWorkerFetch] - Enable using fetch in the worker for
- *   resources. This currently only used for fetching the font data from the
- *   worker thread. When `true`, StandardFontDataFactory will be ignored. The
- *   default value is `true` in web environment and `false` for Node.
  * @property {Object} [StandardFontDataFactory] - The factory that will be used
  *   when reading the standard font files. Providing a custom factory is useful
  *   for environments without Fetch API or `XMLHttpRequest` support, such as
  *   Node.js. The default value is {DOMStandardFontDataFactory}.
+ * @property {boolean} [useWorkerFetch] - Enable using the Fetch API in the
+ *   worker-thread when reading CMap and standard font files. When `true`,
+ *   the `CMapReaderFactory` and `StandardFontDataFactory` options are ignored.
+ *   The default value is `true` in web wenvironments and `false` in Node.js.
  * @property {boolean} [stopAtErrors] - Reject certain promises, e.g.
  *   `getOperatorList`, `getTextContent`, and `RenderTask`, when the associated
  *   PDF data cannot be successfully parsed, instead of attempting to recover
@@ -333,6 +333,7 @@ function getDocument(src) {
   }
   if (typeof params.useWorkerFetch !== "boolean") {
     params.useWorkerFetch =
+      params.CMapReaderFactory === DOMCMapReaderFactory &&
       params.StandardFontDataFactory === DOMStandardFontDataFactory;
   }
   if (typeof params.isEvalSupported !== "boolean") {
@@ -487,6 +488,7 @@ function _fetchDocument(worker, source, pdfDataRangeTransport, docId) {
       fontExtraProperties: source.fontExtraProperties,
       enableXfa: source.enableXfa,
       useSystemFonts: source.useSystemFonts,
+      cMapUrl: source.useWorkerFetch ? source.cMapUrl : null,
       standardFontDataUrl: source.useWorkerFetch
         ? source.standardFontDataUrl
         : null,
@@ -2680,36 +2682,19 @@ class WorkerTransport {
       this._onUnsupportedFeature.bind(this)
     );
 
+    messageHandler.on("FetchBuiltInCMap", data => {
+      if (this.destroyed) {
+        return Promise.reject(new Error("Worker was destroyed"));
+      }
+      return this.CMapReaderFactory.fetch(data);
+    });
+
     messageHandler.on("FetchStandardFontData", data => {
       if (this.destroyed) {
         return Promise.reject(new Error("Worker was destroyed"));
       }
       return this.StandardFontDataFactory.fetch(data);
     });
-
-    messageHandler.on("FetchBuiltInCMap", (data, sink) => {
-      if (this.destroyed) {
-        sink.error(new Error("Worker was destroyed"));
-        return;
-      }
-      let fetched = false;
-
-      sink.onPull = () => {
-        if (fetched) {
-          sink.close();
-          return;
-        }
-        fetched = true;
-
-        this.CMapReaderFactory.fetch(data)
-          .then(function (builtInCMap) {
-            sink.enqueue(builtInCMap, 1, [builtInCMap.cMapData.buffer]);
-          })
-          .catch(function (reason) {
-            sink.error(reason);
-          });
-      };
-    });
   }
 
   _onUnsupportedFeature({ featureId }) {

test/driver.js

@@ -19,7 +19,7 @@
 
 const WAITING_TIME = 100; // ms
 const PDF_TO_CSS_UNITS = 96.0 / 72.0;
-const CMAP_URL = "../external/bcmaps/";
+const CMAP_URL = "/build/generic/web/cmaps/";
 const CMAP_PACKED = true;
 const STANDARD_FONT_DATA_URL = "/build/generic/web/standard_fonts/";
 const IMAGE_RESOURCES_PATH = "/web/images/";

test/test_manifest.json

@@ -4046,6 +4046,15 @@
       "lastPage": 1,
       "type": "eq"
     },
+    {  "id": "mao-main_thread_fetch",
+       "file": "pdfs/mao.pdf",
+       "md5": "797093d67c4d4d4231ac6e1fb66bf6c3",
+       "rounds": 1,
+       "link": true,
+       "lastPage": 1,
+       "type": "eq",
+       "useWorkerFetch": false
+    },
     {  "id": "mao-text",
        "file": "pdfs/mao.pdf",
        "md5": "797093d67c4d4d4231ac6e1fb66bf6c3",