Remove potential Unicode BOM after reading XML file

3 files changed, 11 insertions, 0 deletions

diff --git a/lib/docx/office-xml-reader.js b/lib/docx/office-xml-reader.js
index 39b700e..861b11e 100644
--- a/lib/docx/office-xml-reader.js
+++ b/lib/docx/office-xml-reader.js
@@ -22,6 +22,9 @@ function read(xmlString) {
 function readXmlFromZipFile(docxFile, path) {
     if (docxFile.exists(path)) {
         return docxFile.read(path, "utf-8")
+            .then(function (xmlString) {
+                return xmlString.replace(/^\uFEFF/g, '');
+            })
             .then(read);
     } else {
         return promises.resolve(null);
diff --git a/tests/mammoth.tests.js b/tests/mammoth.tests.js
index e0c1f6c..9be7006 100644
--- a/tests/mammoth.tests.js
+++ b/tests/mammoth.tests.js
@@ -32,6 +32,14 @@ describe('mammoth', function() {
             });
     });
     
+    test('should read docx xml files with unicode byte order mark', function() {
+        var docxPath = path.join(__dirname, "test-data/bom.docx");
+        return mammoth.convertToHtml({path: docxPath}).then(function (result) {
+            assert.equal(result.value, "<p>This XML has a byte order mark.</p>");
+            assert.deepEqual(result.messages, []);
+        });
+    });
+    
     test('style map can be expressed as string', function() {
         var docxFile = createFakeDocxFile({
             "word/document.xml": testData("simple/word/document.xml")
diff --git a/tests/test-data/bom.docx b/tests/test-data/bom.docx
new file mode 100644
index 0000000..3881238
--- /dev/null
+++ b/tests/test-data/bom.docx