Handle missing complex field start characters

3 files changed, 47 insertions, 0 deletions

diff --git a/NEWS b/NEWS
index fa7fdb7..250c1e1 100644
--- a/NEWS
+++ b/NEWS
@@ -10,6 +10,9 @@
   Note that it is still strongly recommended to process untrusted documents in
   a separate thread with a timeout to avoid potential similar issues.
 
+* Handle complex field separator and end characters without corresponding start
+  characters.
+
 # 1.12.1
 
 * Detect and ignore numbering levels that use numStyleLink to refer to
diff --git a/src/main/java/org/zwobble/mammoth/internal/docx/StatefulBodyXmlReader.java b/src/main/java/org/zwobble/mammoth/internal/docx/StatefulBodyXmlReader.java
index 2cc7e4b..06326d0 100644
--- a/src/main/java/org/zwobble/mammoth/internal/docx/StatefulBodyXmlReader.java
+++ b/src/main/java/org/zwobble/mammoth/internal/docx/StatefulBodyXmlReader.java
@@ -321,6 +321,12 @@ class StatefulBodyXmlReader {
             complexFieldStack.add(ComplexField.begin(element));
             currentInstrText.setLength(0);
         } else if (type.equals("end")) {
+            if (complexFieldStack.isEmpty()) {
+                return emptyWithWarning(
+                    "Ignoring complex field end character without corresponding start character"
+                );
+            }
+
             ComplexField complexField = complexFieldStack.remove();
             if (complexField instanceof BeginComplexField) {
                 complexField = parseCurrentInstrText(complexField);
@@ -329,6 +335,12 @@ class StatefulBodyXmlReader {
                 return success(new Checkbox(((CheckboxComplexField) complexField).checked));
             }
         } else if (type.equals("separate")) {
+            if (complexFieldStack.isEmpty()) {
+                return emptyWithWarning(
+                    "Ignoring complex field separator character without corresponding start character"
+                );
+            }
+
             ComplexField complexFieldSeparate = complexFieldStack.remove();
             ComplexField complexField = parseCurrentInstrText(complexFieldSeparate);
             complexFieldStack.add(complexField);
diff --git a/src/test/java/org/zwobble/mammoth/tests/docx/BodyXmlTests.java b/src/test/java/org/zwobble/mammoth/tests/docx/BodyXmlTests.java
index 135ad54..d3fc3b1 100644
--- a/src/test/java/org/zwobble/mammoth/tests/docx/BodyXmlTests.java
+++ b/src/test/java/org/zwobble/mammoth/tests/docx/BodyXmlTests.java
@@ -626,6 +626,38 @@ public class BodyXmlTests {
                 isEmptyRun()
             )));
         }
+
+        @Test
+        public void separatorCharacterWithoutCorrespondingStartCharacterIsIgnored() {
+            XmlElement element = paragraphXml(list(
+                SEPARATE_COMPLEX_FIELD
+            ));
+
+            InternalResult<DocumentElement> result = read(bodyReader(), element);
+
+            assertThat(result, isInternalResult(
+                isParagraph(hasChildren(
+                    isEmptyRun()
+                )),
+                list("Ignoring complex field separator character without corresponding start character")
+            ));
+        }
+
+        @Test
+        public void endCharacterWithoutCorrespondingStartCharacterIsIgnored() {
+            XmlElement element = paragraphXml(list(
+                END_COMPLEX_FIELD
+            ));
+
+            InternalResult<DocumentElement> result = read(bodyReader(), element);
+
+            assertThat(result, isInternalResult(
+                isParagraph(hasChildren(
+                    isEmptyRun()
+                )),
+                list("Ignoring complex field end character without corresponding start character")
+            ));
+        }
     }
 
     @Nested