feat: check "data-*" attributes name restrictions

rdeltour · rdeltour · commit 22f5994a90e3 · 2023-01-03T15:42:03.000+01:00
According to HTML: > A custom data attribute is an attribute in no namespace whose name > starts with the string "data-", has at least one character after the > hyphen, is XML-compatible, and contains no ASCII upper alphas. This commit reports invalid `data-*` attributes as errors with a new error code, `HTM-061`. This should be ideally delegated to validator.nu, but we add this check until we fully integrate the HTML checker. Close #1107
diff --git a/src/main/java/com/adobe/epubcheck/messages/DefaultSeverities.java b/src/main/java/com/adobe/epubcheck/messages/DefaultSeverities.java
@@ -145,6 +145,7 @@ private void initialize()
     severities.put(MessageId.HTM_059, Severity.ERROR);
     severities.put(MessageId.HTM_060a, Severity.USAGE);
     severities.put(MessageId.HTM_060b, Severity.USAGE);
+    severities.put(MessageId.HTM_061, Severity.ERROR);
 
     // Media
     severities.put(MessageId.MED_001, Severity.SUPPRESSED);
diff --git a/src/main/java/com/adobe/epubcheck/messages/MessageId.java b/src/main/java/com/adobe/epubcheck/messages/MessageId.java
@@ -139,6 +139,7 @@ public enum MessageId implements Comparable<MessageId>
   HTM_059("HTM_059"),
   HTM_060a("HTM_060a"),
   HTM_060b("HTM_060b"),
+  HTM_061("HTM_061"),
 
   // Messages associated with media (images, audio and video)
   MED_001("MED-001"),
diff --git a/src/main/java/com/adobe/epubcheck/xml/HTMLUtils.java b/src/main/java/com/adobe/epubcheck/xml/HTMLUtils.java
@@ -5,6 +5,8 @@
 import com.google.common.base.Preconditions;
 import com.google.common.collect.ImmutableSet;
 
+import net.sf.saxon.om.NameChecker;
+
 /**
  * Utilities for HTML-specific logic.
  * 
@@ -44,7 +46,7 @@ public static boolean isCustomNamespace(String namespace)
    * attributes.
    * 
    * @param name
-   *          the name of an attribute defined in the HTML specification
+   *        the name of an attribute defined in the HTML specification
    * @return <code>true</code> iff the attribute value is case-insensitive
    */
   public static boolean isCaseInsensitiveAttribute(String namespace, String name)
@@ -57,6 +59,25 @@ public static boolean isDataAttribute(String namespace, String name)
     return namespace.isEmpty() && name.startsWith("data-");
   }
 
+  /**
+   * Tells if a string is a valid <a href=
+   * "https://html.spec.whatwg.org/multipage/dom.html#custom-data-attribute">
+   * custom data attribute</a>, as
+   * defined in HTML.
+   *
+   * @param name
+   *        the data attribute to test
+   * @return true if {@code name} is a valid custom data attribute
+   */
+  public static boolean isValidDataAttribute(String name)
+  {
+    Preconditions.checkArgument(name != null && name.startsWith("data-"));
+    name = name.substring(5);
+    return !name.isEmpty()
+        && NameChecker.isValidNCName(name)
+        && !name.matches(".*[A-Z].*");
+  }
+
   private HTMLUtils()
   {
     // Not instanciable.
diff --git a/src/main/java/com/adobe/epubcheck/xml/handlers/PreprocessingDefaultHandler.java b/src/main/java/com/adobe/epubcheck/xml/handlers/PreprocessingDefaultHandler.java
@@ -86,6 +86,11 @@ private Attributes preprocessAttributes(String elemNamespace, Attributes atts)
           // Remove data-* attributes in both XHTML and SVG
           if (HTMLUtils.isDataAttribute(namespace, name))
           {
+            if (!HTMLUtils.isValidDataAttribute(name))
+            {
+              context.report.message(MessageId.HTM_061, LocationHandler.location(context, locator),
+                  name);
+            }
             attributes.removeAttribute(i);
           }
           // Remove custom namespace attributes in XHTML
diff --git a/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties b/src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties
@@ -64,6 +64,7 @@ HTM_058=HTML documents must be encoded in UTF-8, but UTF-16 was detected.
 HTM_059=Viewport "%1$s" property must not be defined more than once, but found values [%2$s].
 HTM_060a=EPUB reading systems must ignore secondary viewport meta elements in fixed-layout documents; viewport declaration "%1$s" will be ignored. 
 HTM_060b=EPUB reading systems must ignore viewport meta elements in reflowable documents; viewport declaration "%1$s" will be ignored.
+HTM_061="%1$s" is not a valid custom data attribute (it must have at least one character after the hyphen, be XML-compatible, and not contain ASCII uppercase letters).
 
 #media
 MED_003=Picture "img" elements must reference core media type resources, but found resource "%1$s" of type "%2$s".
diff --git a/src/test/resources/epub3/06-content-document/content-document-xhtml.feature b/src/test/resources/epub3/06-content-document/content-document-xhtml.feature
@@ -227,6 +227,11 @@ Feature: EPUB 3 — Content Documents — XHTML
     When checking document 'data-attr-valid.xhtml'
     Then no errors or warnings are reported
 
+  Scenario: Report invalid `data-*` attributes
+    When checking document 'data-attr-invalid-error.xhtml'
+    Then error HTM-061 is reported 3 times
+    And no other errors or warnings are reported
+
   Scenario: Report invalid elements after a `data-*` attribute
     See issue 189 - was allowed by stripping of `data-*` attributes
     When checking EPUB 'content-xhtml-data-attr-removal-markup-error'
diff --git a/src/test/resources/epub3/06-content-document/files/data-attr-invalid-error.xhtml b/src/test/resources/epub3/06-content-document/files/data-attr-invalid-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>data-* attributes</title>
+	</head>
+	<body>
+		<div data-="">invalid (no character after hte hyphen)</div>
+		<div data--test="">invalid (not an XML name)</div>
+		<div data-ERR="">invalid (contains upper alphas)</div>
+	</body>
+</html>

Original file line number	Diff line number	Diff line change
`@@ -86,6 +86,11 @@ private Attributes preprocessAttributes(String elemNamespace, Attributes atts)`
`86`	`86`	`// Remove data-* attributes in both XHTML and SVG`
`87`	`87`	`if (HTMLUtils.isDataAttribute(namespace, name))`
`88`	`88`	`{`
	`89`	`+ if (!HTMLUtils.isValidDataAttribute(name))`
	`90`	`+ {`
	`91`	`+ context.report.message(MessageId.HTM_061, LocationHandler.location(context, locator),`
	`92`	`+ name);`
	`93`	`+ }`
`89`	`94`	`attributes.removeAttribute(i);`
`90`	`95`	`}`
`91`	`96`	`// Remove custom namespace attributes in XHTML`