fix: issue #373: handle bad encoding expcetion by start page renderin…

…g after page operator list is resolved, add i373_err_broken.pdf to verify (#378)

base/core/core.js

@@ -138,7 +138,6 @@ var Page = (function PageClosure() {
       return promise;
     },
     getOperatorList: function Page_getOperatorList(handler) {
-      var self = this;
       var promise = new Promise();
 
       function reject(e) {
@@ -148,8 +147,7 @@ var Page = (function PageClosure() {
       var pageListPromise = new Promise();
 
       var pdfManager = this.pdfManager;
-      var contentStreamPromise = pdfManager.ensure(this, 'getContentStream',
-                                                   []);
+      var contentStreamPromise = pdfManager.ensure(this, 'getContentStream', []);
       var resourcesPromise = this.loadResources([
         'ExtGState',
         'ColorSpace',
@@ -166,35 +164,31 @@ var Page = (function PageClosure() {
             this.pageIndex, 'p' + this.pageIndex + '_',
             this.idCounters, this.fontCache);
 
-      var dataPromises = Promise.all(
-          [contentStreamPromise, resourcesPromise], reject);
-      dataPromises.then(function(data) {
+      Promise.all([contentStreamPromise, resourcesPromise]).then(function(data) {
         var contentStream = data[0];
-
-
-        var opList = new OperatorList(handler, self.pageIndex);
-        try {
-            handler.send('StartRenderPage', {
-            transparency: partialEvaluator.hasBlendModes(self.resources),
-            pageIndex: self.pageIndex
-            });
-            partialEvaluator.getOperatorList(contentStream, self.resources, opList);
-            pageListPromise.resolve(opList);
-        }
-        catch(e) {
-            pageListPromise.reject(e);
-        }
-      });
+		try {
+			var opList = new OperatorList(handler, this.pageIndex);
+			partialEvaluator.getOperatorList(contentStream, this.resources, opList);
+			pageListPromise.resolve(opList);
+
+			handler.send('StartRenderPage', {
+				transparency: partialEvaluator.hasBlendModes(this.resources),
+				pageIndex: this.pageIndex
+			});
+		} catch (ex) {
+			reject(ex);
+		}
+      }.bind(this), reject);
 
       var annotationsPromise = pdfManager.ensure(this, 'annotations');
-      Promise.all([pageListPromise, annotationsPromise]).then(function(datas) {
-        var pageOpList = datas[0];
-        var annotations = datas[1];
+      Promise.all([pageListPromise, annotationsPromise]).then(function(data) {
+        var pageOpList = data[0];
+        var annotations = data[1];
 
         if (annotations.length === 0) {
           pageOpList.flush(true);
           promise.resolve(pageOpList);
-          return;
+          return promise;
         }
 
         var annotationsReadyPromise = Annotation.appendToOperatorList(

base/display/metadata.js

@@ -45,19 +45,18 @@ var Metadata = (PDFJS.Metadata = (function MetadataClosure() {
    }
 
    function Metadata(meta) {
-      if (typeof meta === 'string') {
+	this.metadata = {};
+	if (typeof meta === 'string') {
          // Ghostscript produces invalid metadata
          meta = fixMetadata(meta);
-
-         var parser = new DOMParser();
-         meta = parser.parseFromString(meta, 'application/xml');
-      } else if (!(meta instanceof Document)) {
-         error('Metadata: Invalid metadata object');
-      }
-
-      this.metaDocument = meta;
-      this.metadata = {};
-      this.parse();
+		 meta = meta.trim();
+		 if (meta.startsWith('<') && meta.endsWith('>')) {
+			var parser = new DOMParser();
+			meta = parser.parseFromString(meta, 'application/xml');
+			this.metaDocument = meta;
+			this.parse();
+		 }
+	}
    }
 
    Metadata.prototype = {

base/shared/util.js

@@ -197,8 +197,8 @@ function error(msg) {
 	} else {
 		//log('Error: ' + msg);
 	}
-	//log(backtrace());
-	//PDFJS.LogManager.notify('error', msg);
+	// log(backtrace());
+	PDFJS.LogManager.notify('error', msg);
 	throw new Error(msg);
 }
 

package.json

@@ -37,7 +37,7 @@
 		"test:jest": "jest --config ./jest.config.json",
 		"test": "jest --config ./jest.config.json && npm run parse-r && npm run parse-fd",
 		"test:forms": "cd ./test && sh p2j.forms.sh",
-		"test:misc": "cd ./test && sh p2j.one.sh misc . \"Expected: 13 success, 4 fail exception with stack trace\" ",
+		"test:misc": "cd ./test && sh p2j.one.sh misc . \"Expected: 13 success, 5 fail exception with stack trace\" ",
 		"parse": "./bin/pdf2json.js -f ./test/pdf/fd/form/F1040.pdf -o ./test/target/fd/form",
 		"parse-s": "./bin/pdf2json.js -f ./test/pdf/fd/form/F1040.pdf -o ./test/target/fd/form -s",
 		"parse-t": "./bin/pdf2json.js -f ./test/pdf/fd/form/F1040.pdf -o ./test/target/fd/form -s -t",
@@ -48,7 +48,7 @@
 		"parse-tb": "./bin/pdf2json.js -f ./test/pdf/misc/i242_testingWithTable.pdf -o ./test/target/misc",
 		"parse-tc": "./bin/pdf2json.js -f ./test/pdf/misc/i293_pdfpac.pdf -o ./test/target/misc",
 		"parse-rectFix": "./bin/pdf2json.js -f ./test/pdf/misc/pr298_rect_fix_from_upstream.pdf -o ./test/target/misc",
-		"parse-e": "./bin/pdf2json.js -f ./test/pdf/misc/i43_encrypted.pdf -o ./test/target/misc",
+		"parse-e": "./bin/pdf2json.js -f ./test/pdf/misc/i373_err_broken.pdf -o ./test/target/misc",
 		"build:rollup": "rollup -c ./rollup.config.js",
 		"build:bundle-pdfjs-base": "node rollup/bundle-pdfjs-base.js",
 		"build": "npm run build:bundle-pdfjs-base && npm run build:rollup",