feat: report reserved string in XHTML custom attribute namespaces

rdeltour · rdeltour · commit bc86db861b27 · 2022-01-24T00:30:49.000+01:00
EPUB 3.3 disallows "w3.org" and "idpf.org" in domains of XHTML custom attribute namespaces. Fix #1190
diff --git a/src/main/java/com/adobe/epubcheck/messages/DefaultSeverities.java b/src/main/java/com/adobe/epubcheck/messages/DefaultSeverities.java
@@ -135,6 +135,7 @@ private void initialize()
     severities.put(MessageId.HTM_051, Severity.WARNING);
     severities.put(MessageId.HTM_052, Severity.ERROR);
     severities.put(MessageId.HTM_053, Severity.INFO);
+    severities.put(MessageId.HTM_054, Severity.ERROR);
 
     // Media
     severities.put(MessageId.MED_001, Severity.ERROR);
diff --git a/src/main/java/com/adobe/epubcheck/messages/MessageId.java b/src/main/java/com/adobe/epubcheck/messages/MessageId.java
@@ -129,6 +129,7 @@ public enum MessageId implements Comparable<MessageId>
   HTM_051("HTM-051"),
   HTM_052("HTM-052"),
   HTM_053("HTM_053"),
+  HTM_054("HTM_054"),
 
   // Messages associated with media (images, audio and video)
   MED_001("MED-001"),
diff --git a/src/main/java/com/adobe/epubcheck/xml/XMLParser.java b/src/main/java/com/adobe/epubcheck/xml/XMLParser.java
@@ -28,6 +28,8 @@
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.StringReader;
+import java.net.URI;
+import java.net.URISyntaxException;
 import java.util.HashSet;
 import java.util.Hashtable;
 import java.util.Locale;
@@ -581,15 +583,20 @@ private Attributes preprocessAttributes(String elemNamespace, String elemName, S
       {
         if (context.version == EPUBVersion.VERSION_3)
         {
+          String attrNamespace = attributes.getURI(i);
           // Remove data-* attributes in both XHTML and SVG
           if (isDataAttribute(attributes, i))
           {
             attributes.removeAttribute(i);
           }
           // Remove custom namespace attributes in XHTML
           else if ("application/xhtml+xml".equals(context.mimeType)
-              && isHTMLCustomNamespace(attributes.getURI(i)))
+              && isHTMLCustomNamespace(attrNamespace))
           {
+            String reserved = findReservedStringInHTMLCustomNamespace(attrNamespace);
+            if (reserved != null) {
+              report.message(MessageId.HTM_054, getLocation(), attrNamespace, reserved);
+            }
             attributes.removeAttribute(i);
           }
           // Normalize case of case-insensitive attributes in XHTML
@@ -624,6 +631,22 @@ private static boolean isHTMLCustomNamespace(String namespace)
     return !KNOWN_XHTML_NAMESPACES.contains(namespace.trim());
   }
 
+  private static String findReservedStringInHTMLCustomNamespace(String namespace)
+  {
+    if (namespace != null) {
+      try
+      {
+        URI uri = new URI(namespace);
+        if (uri.getHost().contains("w3.org")) return "w3.org";
+        if (uri.getHost().contains("idpf.org")) return "idpf.org";
+      } catch (URISyntaxException e)
+      {
+        // ignore
+      }
+    }
+    return null;
+  }
+
   private static boolean isCaseInsensitiveAttribute(Attributes attributes, int index)
   {
     return (attributes.getURI(index).isEmpty()
diff --git a/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties b/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties
@@ -122,6 +122,7 @@ HTM_050=Found epub:type="pagebreak" attribute in content document.
 HTM_051=Found Microdata semantic enrichments but no RDFa. EDUPUB recommends using RDFa Lite.
 HTM_052=The property "region-based" is only allowed on nav elements in Data Navigation Documents.
 HTM_053=Found an external file link (file://) in file: "%1$s".
+HTM_054=Custom attribute namespace ("%1$s") must not include the string "%2$s" in its domain.
 
 #media
 MED_001=Video poster must have core media image type.
diff --git a/src/test/resources/epub3/content-document-xhtml.feature b/src/test/resources/epub3/content-document-xhtml.feature
@@ -615,7 +615,11 @@ Feature: EPUB 3 ▸ Content Documents ▸ XHTML Document Checks
   Scenario: Verify attributes in custom namespaces are ignored
     When checking document 'attrs-custom-ns-valid.xhtml'
     Then no errors or warnings are reported
-
+    
+  Scenario: Report custom attributes using reserved strings in their namespace
+    When checking document 'attrs-custom-ns-reserved-error.xhtml'
+    Then error HTM-054 is reported 2 times
+    And no other errors or warnings are reported
 
   ## 2.5 HTML Deviations and Constraints
 
diff --git a/src/test/resources/epub3/files/content-document-xhtml/attrs-custom-ns-reserved-error.xhtml b/src/test/resources/epub3/files/content-document-xhtml/attrs-custom-ns-reserved-error.xhtml
@@ -0,0 +1,12 @@
+<!DOCTYPE html>
+<html xmlns="https://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
+	<head>
+		<meta charset="utf-8" />
+		<title>Test</title>
+	</head>
+	<body w3:attr="disallowed" idpf:attr="disallowed" ok:attr="allowed"
+		xmlns:w3="https://example.w3.org" xmlns:idpf="https://example.idpf.org"
+		xmlns:ok="https://example.org/w3.org">
+		<h1>Test</h1>
+	</body>
+</html>
diff --git a/src/test/resources/epub3/files/content-document-xhtml/attrs-custom-ns-valid.xhtml b/src/test/resources/epub3/files/content-document-xhtml/attrs-custom-ns-valid.xhtml
@@ -1,11 +1,10 @@
-<!DOCTYPE html>	 
-<html xmlns="https://www.w3.org/1999/xhtml" xmlns:foo="https://example.org" xml:lang="en" lang="en">
+<!DOCTYPE html>
+<html xmlns="https://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
 	<head>
-		<meta charset="utf-8"/>
-		<title>Attributes in custom namespace</title>
+		<meta charset="utf-8" />
+		<title>Test</title>
 	</head>
-	<body>
+	<body custom:attribute="allowed" xmlns:custom="https://example.org">
 		<h1>Test</h1>
-		<p foo:bar1="baz" foo:bar2="baz" foo:bar3="baz">custom attribute!</p>
 	</body>
 </html>
