Add method SafeSetAttribute to safely write attributes

Strumenta · Feb 4, 2023 · 1ea4acd · 1ea4acd
1 parent 3ea8a93
commit 1ea4acd
Show file tree

Hide file tree

Showing 3 changed files with 16 additions and 10 deletions.
diff --git a/src/SmartReader/NodeUtility.cs b/src/SmartReader/NodeUtility.cs
@@ -6,6 +6,7 @@
 
 using AngleSharp.Dom;
 using AngleSharp.Html.Dom;
+using AngleSharp.Text;
 
 namespace SmartReader
 {
@@ -59,11 +60,7 @@ internal static IElement SetNodeTag(IElement node, string tag)
             {
                 var attr = node.Attributes[i]!;
 
-                // the possible result of malformed HTML
-                if (!attr.Name.Contains("<") && !attr.Name.Contains(">"))
-                {
-                    replacement.SetAttribute(attr.Name, attr.Value);
-                }
+                NodeUtility.SafeSetAttribute(replacement, attr);                
             }
             return replacement;
         }
@@ -622,5 +619,17 @@ internal static List<INode> GetNodeAncestors(INode node, int maxDepth = 0)
             }
             return next as IElement;
         }
+
+        /// <summary>
+        /// This is a safe method to write attributes
+        /// because AngleSharp cannot handle attribute names that would be invalid in XML
+        /// </summary>  
+        internal static void SafeSetAttribute(this IElement child, IAttr attribute)
+        {
+            if (attribute.Name.IsXmlName())
+                child.SetAttribute(attribute.Name, attribute.Value);
+            else
+                child.SetAttribute(attribute.Name.CleanXmlName(), attribute.Value);
+        }
     }
 }
diff --git a/src/SmartReader/Readability.cs b/src/SmartReader/Readability.cs
@@ -188,10 +188,7 @@ internal static void SimplifyNestedElements(IElement articleContent)
                         var child = node.Children[0];
                         for (var i = 0; i < node.Attributes.Length; i++)
                         {
-                            if (node.Attributes[i]!.Name.IsXmlName())
-                                child.SetAttribute(node.Attributes[i]!.Name, node.Attributes[i]!.Value);
-                            else
-                                child.SetAttribute(node.Attributes[i]!.Name.CleanXmlName(), node.Attributes[i]!.Value);
+                            NodeUtility.SafeSetAttribute(child, node.Attributes[i]);                            
                         }
                         node.Parent.ReplaceChild(child, node);
                         node = child;

diff --git a/src/SmartReader/TextUtility.cs b/src/SmartReader/TextUtility.cs
@@ -42,6 +42,6 @@ internal static string CleanXmlName(this string str)
             }
             else
                 return str;
-        }        
+        }
     }
 }