Fix: Widget.ParseDa normalize TextDa to handle wrong formatting

MuPDF.NET.Test/WidgetTest.cs

@@ -98,5 +98,20 @@ public void Checkbox()
             //Assert.Pass();
         }
         */
+
+        [Test]
+        public void ShouldNotThrowOnGetWidgets()
+        {
+            Document doc = new Document("../../../resources/test_widget_parse.pdf");
+
+            var currentPage = 0;
+            while (currentPage < doc.PageCount)
+            {
+                var page = doc[currentPage];
+                var widgets = page.GetWidgets().ToList();
+                Assert.That(widgets.Count, Is.GreaterThanOrEqualTo(0));
+                currentPage++;
+            }
+        }
     }
 }

MuPDF.NET/Widget.cs

@@ -269,7 +269,7 @@ public void ParseDa()
             string font = "Helv";
             float fontSize = 0;
             float[] col = { 0, 0, 0 };
-            string[] dat = TextDa.Split(' ');    // split on any whitespace
+            string[] dat = TextDa.Split(new char[] { ' ' }, StringSplitOptions.RemoveEmptyEntries);    // split on any whitespace and remove empty entries
             for (int i = 0; i < dat.Length; i++)
             {
                 string item = dat[i];