From 29263836c24053180e7b38d4d2b2d885f6bda6de Mon Sep 17 00:00:00 2001
From: Brendan Dahl <brendan.dahl@gmail.com>
Date: Fri, 3 Aug 2012 16:11:43 -0700
Subject: [PATCH 1/2] Sanitize the document info.

---
 src/core.js                  | 43 +++++++++++++++++++++++++++++-------
 test/pdfs/issue1729.pdf.link |  2 ++
 test/test_manifest.json      |  8 +++++++
 3 files changed, 45 insertions(+), 8 deletions(-)
 create mode 100644 test/pdfs/issue1729.pdf.link

diff --git a/src/core.js b/src/core.js
index e8d008074..1ee01d7e5 100644
--- a/src/core.js
+++ b/src/core.js
@@ -403,6 +403,24 @@ var PDFDocument = (function PDFDocumentClosure() {
     return true; /* found */
   }
 
+  var DOCUMENT_INFO = {
+    get entries() {
+      // Lazily build this since all the validation functions below are not
+      // defined until after this file loads.
+      return shadow(this, 'entries', {
+        Title: isString,
+        Author: isString,
+        Subject: isString,
+        Keywords: isString,
+        Creator: isString,
+        Producer: isString,
+        CreationDate: isString,
+        ModDate: isString,
+        Trapped: isName
+      });
+    }
+  };
+
   PDFDocument.prototype = {
     get linearization() {
       var length = this.stream.length;
@@ -495,18 +513,27 @@ var PDFDocument = (function PDFDocumentClosure() {
       return shadow(this, 'numPages', num);
     },
     getDocumentInfo: function PDFDocument_getDocumentInfo() {
-      var info;
+      var docInfo;
       if (this.xref.trailer.has('Info')) {
         var infoDict = this.xref.trailer.get('Info');
 
-        info = {};
-        infoDict.forEach(function(key, value) {
-          info[key] = typeof value !== 'string' ? value :
-            stringToPDFString(value);
-        });
+        docInfo = {};
+        var validEntries = DOCUMENT_INFO.entries;
+        // Only fill the document info with valid entries from the spec.
+        for (var key in validEntries) {
+          if (infoDict.has(key)) {
+            var value = infoDict.get(key);
+            // Make sure the value conforms to the spec.
+            if (validEntries[key](value)) {
+              docInfo[key] = typeof value !== 'string' ? value :
+                             stringToPDFString(value);
+            } else {
+              info('Bad value in document info for "' + key + '"');
+            }
+          }
+        }
       }
-
-      return shadow(this, 'getDocumentInfo', info);
+      return shadow(this, 'getDocumentInfo', docInfo);
     },
     getFingerprint: function PDFDocument_getFingerprint() {
       var xref = this.xref, fileID;
diff --git a/test/pdfs/issue1729.pdf.link b/test/pdfs/issue1729.pdf.link
new file mode 100644
index 000000000..ca6fa690c
--- /dev/null
+++ b/test/pdfs/issue1729.pdf.link
@@ -0,0 +1,2 @@
+http://www.environmentallights.com/files/documents/ir_light_hazard.pdf
+
diff --git a/test/test_manifest.json b/test/test_manifest.json
index 0a22160b9..166b4fdc7 100644
--- a/test/test_manifest.json
+++ b/test/test_manifest.json
@@ -626,5 +626,13 @@
       "pageLimit": 1,
       "link": true,
       "type": "eq"
+    },
+    {  "id": "issue1729",
+      "file": "pdfs/issue1729.pdf",
+      "md5": "29b0eddc3e1dcb23a44384037032d470",
+      "rounds": 1,
+      "pageLimit": 1,
+      "link": true,
+      "type": "load"
     }
 ]

From 70e415900c0accbe1ef3bd84ec9c6867ef7f70f6 Mon Sep 17 00:00:00 2001
From: Brendan Dahl <brendan.dahl@gmail.com>
Date: Mon, 6 Aug 2012 14:32:54 -0700
Subject: [PATCH 2/2] Address review comments.

---
 src/core.js                  | 4 ++--
 test/pdfs/issue1729.pdf.link | 1 -
 2 files changed, 2 insertions(+), 3 deletions(-)

diff --git a/src/core.js b/src/core.js
index 1ee01d7e5..8519e2e68 100644
--- a/src/core.js
+++ b/src/core.js
@@ -403,7 +403,7 @@ var PDFDocument = (function PDFDocumentClosure() {
     return true; /* found */
   }
 
-  var DOCUMENT_INFO = {
+  var DocumentInfoValidators = {
     get entries() {
       // Lazily build this since all the validation functions below are not
       // defined until after this file loads.
@@ -518,7 +518,7 @@ var PDFDocument = (function PDFDocumentClosure() {
         var infoDict = this.xref.trailer.get('Info');
 
         docInfo = {};
-        var validEntries = DOCUMENT_INFO.entries;
+        var validEntries = DocumentInfoValidators.entries;
         // Only fill the document info with valid entries from the spec.
         for (var key in validEntries) {
           if (infoDict.has(key)) {
diff --git a/test/pdfs/issue1729.pdf.link b/test/pdfs/issue1729.pdf.link
index ca6fa690c..a6a73d5a7 100644
--- a/test/pdfs/issue1729.pdf.link
+++ b/test/pdfs/issue1729.pdf.link
@@ -1,2 +1 @@
 http://www.environmentallights.com/files/documents/ir_light_hazard.pdf
-