Merge pull request #7782 from Snuffleupagus/page-labels-more-validation

Use `stringToPDFString` to sanitizing bad "Prefix" entries in Page Label dictionaries, and add more validation

src/core/obj.js

@@ -283,7 +283,6 @@ var Catalog = (function CatalogClosure() {
       var pageLabels = new Array(this.numPages);
       var style = null;
       var prefix = '';
-      var start = 1;
 
       var numberTree = new NumberTree(obj, this.xref);
       var nums = numberTree.getAll();
@@ -300,14 +299,16 @@ var Catalog = (function CatalogClosure() {
 
           var s = labelDict.get('S');
           assert(!s || isName(s), 'Invalid style in PageLabel dictionary.');
-          style = (s ? s.name : null);
+          style = s ? s.name : null;
 
-          prefix = labelDict.get('P') || '';
+          var p = labelDict.get('P');
-          assert(isString(prefix), 'Invalid prefix in PageLabel dictionary.');
+          assert(!p || isString(p), 'Invalid prefix in PageLabel dictionary.');
+          prefix = p ? stringToPDFString(p) : '';
 
-          start = labelDict.get('St') || 1;
+          var st = labelDict.get('St');
-          assert(isInt(start), 'Invalid start in PageLabel dictionary.');
+          assert(!st || (isInt(st) && st >= 1),
-          currentIndex = start;
+                 'Invalid start in PageLabel dictionary.');
+          currentIndex = st || 1;
         }
 
         switch (style) {

test/pdfs/.gitignore

@@ -40,6 +40,7 @@
 !issue7544.pdf
 !issue7598.pdf
 !issue7665.pdf
+!bad-PageLabels.pdf
 !filled-background.pdf
 !ArabicCIDTrueType.pdf
 !ThuluthFeatures.pdf

test/unit/api_spec.js

@@ -503,14 +503,24 @@ describe('api', function() {
         return pdfDoc.getPageLabels();
       });
 
-      Promise.all([promise0, promise1, promise2]).then(function (pageLabels) {
+      // PageLabels with bad "Prefix" entries.
+      var url3 = new URL('../pdfs/bad-PageLabels.pdf', window.location).href;
+      var loadingTask3 = new PDFJS.getDocument(url3);
+      var promise3 = loadingTask3.promise.then(function (pdfDoc) {
+        return pdfDoc.getPageLabels();
+      });
+
+      Promise.all([promise0, promise1, promise2, promise3]).then(
+          function (pageLabels) {
         expect(pageLabels[0]).toEqual(['i', 'ii', 'iii', '1']);
         expect(pageLabels[1]).toEqual(['Front Page1']);
         expect(pageLabels[2]).toEqual(['1', '2']);
+        expect(pageLabels[3]).toEqual(['X1']);
 
         loadingTask0.destroy();
         loadingTask1.destroy();
         loadingTask2.destroy();
+        loadingTask3.destroy();
         done();
       }).catch(function (reason) {
         done.fail(reason);