From ba2edeae18739e767f10087df521f6355f917c45 Mon Sep 17 00:00:00 2001
From: Jonas Jenwald <jonas.jenwald@gmail.com>
Date: Tue, 18 Dec 2018 23:26:02 +0100
Subject: [PATCH] [api-minor] Add support, in `getMetadata`, for custom
 information dictionary entries (issue 5970, issue 10344) (#10346)

The custom entries, provided that they exist *and* that their types are safe to include, are exposed through a new `Custom` infoDict entry to clearly separate them from the standard ones.

Fixes 5970.
Fixes 10344.
---
 src/core/document.js  | 36 ++++++++++++++++++++++++++++--------
 test/unit/api_spec.js | 30 ++++++++++++++++++++++++++++++
 2 files changed, 58 insertions(+), 8 deletions(-)

diff --git a/src/core/document.js b/src/core/document.js
index b9db08ec8..9069a0cd7 100644
--- a/src/core/document.js
+++ b/src/core/document.js
@@ -14,8 +14,8 @@
  */
 
 import {
-  assert, FormatError, getInheritableProperty, info, isArrayBuffer, isNum,
-  isSpace, isString, MissingDataException, OPS, shadow, stringToBytes,
+  assert, FormatError, getInheritableProperty, info, isArrayBuffer, isBool,
+  isNum, isSpace, isString, MissingDataException, OPS, shadow, stringToBytes,
   stringToPDFString, Util, warn
 } from '../shared/util';
 import { Catalog, ObjectLoader, XRef } from './obj';
@@ -544,17 +544,37 @@ var PDFDocument = (function PDFDocumentClosure() {
         info('The document information dictionary is invalid.');
       }
       if (isDict(infoDict)) {
-        // Only fill the document info with valid entries from the spec.
-        for (let key in DocumentInfoValidators) {
-          if (infoDict.has(key)) {
-            const value = infoDict.get(key);
-            // Make sure the value conforms to the spec.
+        // Fill the document info with valid entries from the specification,
+        // as well as any existing well-formed custom entries.
+        for (let key of infoDict.getKeys()) {
+          const value = infoDict.get(key);
+
+          if (DocumentInfoValidators[key]) {
+            // Make sure the (standard) value conforms to the specification.
             if (DocumentInfoValidators[key](value)) {
               docInfo[key] = (typeof value !== 'string' ?
                               value : stringToPDFString(value));
             } else {
-              info('Bad value in document info for "' + key + '"');
+              info(`Bad value in document info for "${key}".`);
             }
+          } else if (typeof key === 'string') {
+            // For custom values, only accept white-listed types to prevent
+            // errors that would occur when trying to send non-serializable
+            // objects to the main-thread (for example `Dict` or `Stream`).
+            let customValue;
+            if (isString(value)) {
+              customValue = stringToPDFString(value);
+            } else if (isName(value) || isNum(value) || isBool(value)) {
+              customValue = value;
+            } else {
+              info(`Unsupported value in document info for (custom) "${key}".`);
+              continue;
+            }
+
+            if (!docInfo['Custom']) {
+              docInfo['Custom'] = Object.create(null);
+            }
+            docInfo['Custom'][key] = customValue;
           }
         }
       }
diff --git a/test/unit/api_spec.js b/test/unit/api_spec.js
index ead5e5d79..4ae03fbf0 100644
--- a/test/unit/api_spec.js
+++ b/test/unit/api_spec.js
@@ -829,6 +829,8 @@ describe('api', function() {
       var promise = doc.getMetadata();
       promise.then(function({ info, metadata, contentDispositionFilename, }) {
         expect(info['Title']).toEqual('Basic API Test');
+        // Custom, non-standard, information dictionary entries.
+        expect(info['Custom']).toEqual(undefined);
         // The following are PDF.js specific, non-standard, properties.
         expect(info['PDFFormatVersion']).toEqual('1.7');
         expect(info['IsLinearized']).toEqual(false);
@@ -842,6 +844,34 @@ describe('api', function() {
         done();
       }).catch(done.fail);
     });
+    it('gets metadata, with custom info dict entries', function(done) {
+      var loadingTask = getDocument(buildGetDocumentParams('tracemonkey.pdf'));
+
+      loadingTask.promise.then(function(pdfDocument) {
+        return pdfDocument.getMetadata();
+      }).then(function({ info, metadata, contentDispositionFilename, }) {
+        expect(info['Creator']).toEqual('TeX');
+        expect(info['Producer']).toEqual('pdfeTeX-1.21a');
+        expect(info['CreationDate']).toEqual('D:20090401163925-07\'00\'');
+        // Custom, non-standard, information dictionary entries.
+        const custom = info['Custom'];
+        expect(typeof custom === 'object' && custom !== null).toEqual(true);
+
+        expect(custom['PTEX.Fullbanner']).toEqual('This is pdfeTeX, ' +
+          'Version 3.141592-1.21a-2.2 (Web2C 7.5.4) kpathsea version 3.5.6');
+        // The following are PDF.js specific, non-standard, properties.
+        expect(info['PDFFormatVersion']).toEqual('1.4');
+        expect(info['IsLinearized']).toEqual(false);
+        expect(info['IsAcroFormPresent']).toEqual(false);
+        expect(info['IsXFAPresent']).toEqual(false);
+
+        expect(metadata).toEqual(null);
+        expect(contentDispositionFilename).toEqual(null);
+
+        loadingTask.destroy().then(done);
+      }).catch(done.fail);
+    });
+
     it('gets data', function(done) {
       var promise = doc.getData();
       promise.then(function (data) {