Support new PDF.js fingerprint API

In https://github.com/mozilla/pdf.js/pull/13661 the API for retrieving
PDF fingerprints (aka. File Identifiers) changed. In generic builds of
PDF.js the old API remains available, but not in the non-generic one
that Firefox's built-in viewer uses.

This commit makes Hypothesis use the new API if available or the old API
otherwise. The fingerprint value should be the same in both cases.

Fixes #3673

src/annotator/integrations/pdf-metadata.js

@@ -16,6 +16,11 @@ import { normalizeURI } from '../util/url';
  * @prop {string} documentFingerprint - The fingerprint of this PDF. This is
  *   referred to as the "File Identifier" in the PDF spec. It may be a hash of
  *   part of the content if the PDF file does not have a File Identifier.
+ *
+ *   PDFs may have two file identifiers. The first is the "original" identifier
+ *   which is not supposed to change if the file is updated and the second
+ *   one is the "last modified" identifier. This property is the original
+ *   identifier.
  */
 
 /**
@@ -117,7 +122,7 @@ export class PDFMetadata {
     return this._loaded.then(app => {
       let uri = getPDFURL(app);
       if (!uri) {
-        uri = fingerprintToURN(app.pdfDocument.fingerprint);
+        uri = fingerprintToURN(getFingerprint(app));
       }
       return uri;
     });
@@ -138,8 +143,8 @@ export class PDFMetadata {
       contentDispositionFilename,
       metadata,
     } = await app.pdfDocument.getMetadata();
-    const documentFingerprint = app.pdfDocument.fingerprint;
 
+    const documentFingerprint = getFingerprint(app);
     const url = getPDFURL(app);
 
     // Return the title metadata embedded in the PDF if available, otherwise
@@ -176,8 +181,27 @@ export class PDFMetadata {
   }
 }
 
+/**
+ * Get the fingerprint/file identifier of the currently loaded PDF.
+ *
+ * @param {PDFViewerApplication} app
+ */
+function getFingerprint(app) {
+  if (Array.isArray(app.pdfDocument.fingerprints)) {
+    return app.pdfDocument.fingerprints[0];
+  } else {
+    return /** @type {string} */ (app.pdfDocument.fingerprint);
+  }
+}
+
+/**
+ * Generate a URI from a PDF fingerprint suitable for storing as the main
+ * or associated URI of an annotation.
+ *
+ * @param {string} fingerprint
+ */
 function fingerprintToURN(fingerprint) {
-  return 'urn:x-pdf:' + String(fingerprint);
+  return `urn:x-pdf:${fingerprint}`;
 }
 
 /**

src/annotator/integrations/test/pdf-metadata-test.js

@@ -33,12 +33,20 @@ class FakePDFDocumentProxy {
     fingerprint,
     info,
     metadata = null,
-  }) {
-    this.fingerprint = fingerprint;
 
+    // Whether to expose fingerprints via the new API (after
+    // https://github.com/mozilla/pdf.js/pull/13661) or the old one.
+    newFingerprintAPI = true,
+  }) {
     this._contentDispositionFilename = contentDispositionFilename;
     this._info = info;
     this._metadata = metadata;
+
+    if (newFingerprintAPI) {
+      this.fingerprints = [fingerprint, null];
+    } else {
+      this.fingerprint = fingerprint;
+    }
   }
 
   async getMetadata() {
@@ -67,10 +75,16 @@ class FakePDFViewerApplication {
    *   @prop {boolean} domEvents - Whether events are emitted on the DOM
    *   @prop {boolean} eventBusEvents - Whether the `eventBus` API is enabled
    *   @prop {boolean} initializedPromise - Whether the `initializedPromise` API is enabled
+   *   @prop {boolean} newFingerprintAPI - Whether to emulate the new fingerprints API
    */
   constructor(
     url = '',
-    { domEvents = false, eventBusEvents = true, initializedPromise = true } = {}
+    {
+      domEvents = false,
+      eventBusEvents = true,
+      initializedPromise = true,
+      newFingerprintAPI = true,
+    } = {}
   ) {
     this.url = url;
     this.documentInfo = undefined;
@@ -78,6 +92,7 @@ class FakePDFViewerApplication {
     this.pdfDocument = null;
     this.dispatchDOMEvents = domEvents;
     this.initialized = false;
+    this.newFingerprintAPI = newFingerprintAPI;
 
     // Use `EventEmitter` as a fake version of PDF.js's `EventBus` class as the
     // API for subscribing to events is the same.
@@ -118,9 +133,10 @@ class FakePDFViewerApplication {
 
     this.pdfDocument = new FakePDFDocumentProxy({
       contentDispositionFilename,
+      fingerprint,
       info,
       metadata: metadata ? new FakeMetadata(metadata) : null,
-      fingerprint,
+      newFingerprintAPI: this.newFingerprintAPI,
     });
 
     if (this.dispatchDOMEvents) {
@@ -235,8 +251,11 @@ describe('PDFMetadata', () => {
     url: 'http://fake.com/',
   };
 
-  function createPDFMetadata(metadata = testMetadata) {
-    const fakePDFViewerApplication = new FakePDFViewerApplication();
+  function createPDFMetadata(metadata = testMetadata, viewerOptions) {
+    const fakePDFViewerApplication = new FakePDFViewerApplication(
+      '',
+      viewerOptions
+    );
     fakePDFViewerApplication.completeInit();
     fakePDFViewerApplication.finishLoading(metadata);
     return {
@@ -254,13 +273,18 @@ describe('PDFMetadata', () => {
       });
     });
 
-    it('returns the fingerprint as a URN when the PDF URL is a file:// URL', async () => {
-      const { pdfMetadata } = createPDFMetadata({
-        url: 'file:///test.pdf',
-        fingerprint: 'fakeFingerprint',
+    [true, false].forEach(newFingerprintAPI => {
+      it('returns the fingerprint as a URN when the PDF URL is a file:// URL', async () => {
+        const { pdfMetadata } = createPDFMetadata(
+          {
+            url: 'file:///test.pdf',
+            fingerprint: 'fakeFingerprint',
+          },
+          { newFingerprintAPI }
+        );
+        const uri = await pdfMetadata.getUri();
+        assert.equal(uri, 'urn:x-pdf:fakeFingerprint');
       });
-      const uri = await pdfMetadata.getUri();
-      assert.equal(uri, 'urn:x-pdf:fakeFingerprint');
     });
 
     it('resolves relative URLs', async () => {
@@ -280,10 +304,14 @@ describe('PDFMetadata', () => {
   });
 
   describe('#getMetadata', () => {
-    it('returns the document fingerprint in the `documentFingerprint` property', async () => {
-      const { pdfMetadata } = createPDFMetadata();
-      const metadata = await pdfMetadata.getMetadata();
-      assert.equal(metadata.documentFingerprint, testMetadata.fingerprint);
+    [true, false].forEach(newFingerprintAPI => {
+      it('returns the document fingerprint in the `documentFingerprint` property', async () => {
+        const { pdfMetadata } = createPDFMetadata(testMetadata, {
+          newFingerprintAPI,
+        });
+        const metadata = await pdfMetadata.getMetadata();
+        assert.equal(metadata.documentFingerprint, testMetadata.fingerprint);
+      });
     });
 
     it('returns the PDF URL in the `links` array', async () => {

src/types/pdfjs.js

@@ -51,7 +51,13 @@
 
 /**
  * @typedef PDFDocument
- * @prop {string} fingerprint
+ * @prop {string} [fingerprint] - PDF fingerprint in PDF.js before v2.10.377.
+ *   May exist in later versions depending on the PDF.js build.
+ * @prop {[string, string|null]} [fingerprints] - PDF fingerprints in PDF.js after
+ *   v2.10.377. See https://github.com/mozilla/pdf.js/pull/13661. The first
+ *   entry of this array is the "original" fingerprint and the same as the
+ *   `fingerprint` property in older versions. The second entry is the "modified"
+ *   fingerprint. See "File Identifiers" section in the PDF spec.
  * @prop {() => Promise<PDFDocumentMetadata>} getMetadata
  */