feat: make SVG content conformance checking informative-only

src/main/java/com/adobe/epubcheck/messages/DefaultSeverities.java

@@ -325,6 +325,8 @@ private void initialize()
     severities.put(MessageId.RSC_021, Severity.ERROR);
     severities.put(MessageId.RSC_022, Severity.INFO);
     severities.put(MessageId.RSC_023, Severity.WARNING);
+    severities.put(MessageId.RSC_024, Severity.USAGE);
+    severities.put(MessageId.RSC_025, Severity.USAGE);
 
     // Scripting
     severities.put(MessageId.SCP_001, Severity.USAGE);

src/main/java/com/adobe/epubcheck/messages/MessageId.java

@@ -319,6 +319,8 @@ public enum MessageId implements Comparable<MessageId>
   RSC_021("RSC-021"),
   RSC_022("RSC-022"),
   RSC_023("RSC-023"),
+  RSC_024("RSC-024"),
+  RSC_025("RSC-025"),
 
   // Messages relating to scripting
   SCP_001("SCP-001"),

src/main/java/com/adobe/epubcheck/ops/OPSChecker.java

@@ -56,11 +56,11 @@ public class OPSChecker extends PublicationResourceChecker
       .putAll(Predicates.and(mimetype("application/xhtml+xml"), version(EPUBVersion.VERSION_2)),
           XMLValidators.XHTML_20_NVDL, XMLValidators.XHTML_20_SCH, XMLValidators.IDUNIQUE_20_SCH)
       .putAll(Predicates.and(mimetype("application/xhtml+xml"), version(EPUBVersion.VERSION_3)),
-          XMLValidators.XHTML_30_RNC, XMLValidators.XHTML_30_SCH)
+          XMLValidators.XHTML_30_RNC, XMLValidators.XHTML_30_SCH, XMLValidators.SVG_INFORMATIVE_30_NVDL)
       .putAll(Predicates.and(mimetype("image/svg+xml"), version(EPUBVersion.VERSION_2)),
           XMLValidators.SVG_20_NVDL, XMLValidators.IDUNIQUE_20_SCH)
       .putAll(Predicates.and(mimetype("image/svg+xml"), version(EPUBVersion.VERSION_3)),
-          XMLValidators.SVG_30_NVDL)
+          XMLValidators.SVG_30_NVDL, XMLValidators.SVG_INFORMATIVE_30_NVDL)
       .putAll(
           and(or(profile(EPUBProfile.DICT), hasPubType(OPFData.DC_TYPE_DICT)),
               mimetype("application/xhtml+xml"), version(EPUBVersion.VERSION_3)),

src/main/java/com/adobe/epubcheck/xml/XMLParser.java

@@ -41,6 +41,7 @@
 import org.xml.sax.Attributes;
 import org.xml.sax.ContentHandler;
 import org.xml.sax.DTDHandler;
+import org.xml.sax.ErrorHandler;
 import org.xml.sax.InputSource;
 import org.xml.sax.Locator;
 import org.xml.sax.SAXException;
@@ -68,6 +69,46 @@
 
 public class XMLParser extends DefaultHandler implements LexicalHandler, DeclHandler
 {
+
+  private static final class UsageErrorHandler implements ErrorHandler
+  {
+    private final ValidationContext context;
+
+    public UsageErrorHandler(ValidationContext context)
+    {
+      this.context = context;
+    }
+
+    @Override
+    public void warning(SAXParseException exception)
+      throws SAXException
+    {
+      context.report.message(MessageId.RSC_024,
+          EPUBLocation.create(context.path, exception.getLineNumber(), exception.getColumnNumber()),
+          exception.getMessage());
+    }
+
+    @Override
+    public void error(SAXParseException exception)
+      throws SAXException
+    {
+      context.report.message(MessageId.RSC_025,
+          EPUBLocation.create(context.path, exception.getLineNumber(), exception.getColumnNumber()),
+          exception.getMessage());
+    }
+
+    @Override
+    public void fatalError(SAXParseException exception)
+      throws SAXException
+    {
+      context.report.message(MessageId.RSC_016,
+          EPUBLocation.create(context.path, exception.getLineNumber(), exception.getColumnNumber()),
+          exception.getMessage());
+
+    }
+
+  }
+
   private static final String SAXPROP_LEXICAL_HANDLER = "http://xml.org/sax/properties/lexical-handler";
   private static final String SAXPROP_DECL_HANDLER = "http://xml.org/sax/properties/declaration-handler";
   private SAXParser parser;
@@ -158,8 +199,9 @@ public void addXMLHandler(XMLHandler handler)
   public void addValidator(XMLValidator xv)
   {
     PropertyMapBuilder propertyMapBuilder = new PropertyMapBuilder();
-    propertyMapBuilder.put(ValidateProperty.ERROR_HANDLER, this);
-    Validator validator = xv.schema.createValidator(propertyMapBuilder.toPropertyMap());
+    ErrorHandler eh = xv.isNormative() ? this : new UsageErrorHandler(context);
+    propertyMapBuilder.put(ValidateProperty.ERROR_HANDLER, eh);
+    Validator validator = xv.getSchema().createValidator(propertyMapBuilder.toPropertyMap());
     ContentHandler contentHandler = validator.getContentHandler();
     if (contentHandler != null)
     {

src/main/java/com/adobe/epubcheck/xml/XMLValidator.java

@@ -28,7 +28,6 @@
 import java.net.URISyntaxException;
 import java.net.URL;
 
-import javax.xml.transform.TransformerException;
 import javax.xml.transform.TransformerFactory;
 
 import org.idpf.epubcheck.util.saxon.ColumnNumberFunction;
@@ -55,18 +54,16 @@
 
 import net.sf.saxon.Configuration;
 import net.sf.saxon.TransformerFactoryImpl;
-import net.sf.saxon.lib.FeatureKeys;
-import net.sf.saxon.lib.StandardErrorListener;
 import net.sf.saxon.sxpath.IndependentContext;
 import net.sf.saxon.sxpath.XPathStaticContext;
 import net.sf.saxon.trans.SymbolicName;
-import net.sf.saxon.trans.XPathException;
 
 
 public class XMLValidator
 {
 
-  Schema schema;
+  private final Schema schema;
+  private final boolean isNormative;
 
   /**
    * Basic Resolver from Jing modified to add support for resolving zip and
@@ -223,9 +220,10 @@ public void warning(SAXParseException exception) throws
     }
 
   }
-
-  public XMLValidator(String schemaName)
+  
+  public XMLValidator(String schemaName, boolean isNormative)
   {
+    this.isNormative = isNormative;
     try
     {
       String resourcePath = ResourceUtil.getResourcePath(schemaName);
@@ -271,4 +269,12 @@ public XMLValidator(String schemaName)
       throw new Error("Internal error: " + e + " " + schemaName);
     }
   }
+
+  public Schema getSchema() {
+    return schema;
+  }
+
+  public boolean isNormative() {
+    return isNormative;
+  }
 }

src/main/java/com/adobe/epubcheck/xml/XMLValidators.java

@@ -39,6 +39,7 @@ public enum XMLValidators
   SVG_20_NVDL("schema/20/rng/ops20-svg.nvdl"),
   SVG_30_RNC("schema/30/epub-svg-30.rnc"),
   SVG_30_NVDL("schema/30/epub-svg-30.nvdl"),
+  SVG_INFORMATIVE_30_NVDL("schema/30/epub-svg-informative.nvdl", false),
   SVG_30_SCH("schema/30/epub-svg-30.sch"),
   XHTML_20_NVDL("schema/20/rng/ops20.nvdl"),
   XHTML_20_SCH("schema/20/sch/xhtml.sch"),
@@ -55,7 +56,12 @@ public enum XMLValidators
 
   private XMLValidators(String schemaName)
   {
-    val = new XMLValidator(schemaName);
+    this(schemaName, true);
+  }
+
+  private XMLValidators(String schemaName, boolean isNormative)
+  {
+    this.val = new XMLValidator(schemaName, isNormative);
   }
 
   public XMLValidator get()

src/main/resources/com/adobe/epubcheck/messages/MessageBundle.properties

@@ -337,6 +337,8 @@ RSC_020="%1$s" is not a valid URI.
 RSC_021=A Search Key Map Document must point to Content Documents ("%1$s" was not found in the spine).
 RSC_022=Cannot check image details (requires Java version 7 or higher).
 RSC_023=Couldn’t parse host of URL "%1$s" (probably due to disallowed characters or missing slashes after the protocol)
+RSC_024=Informative parsing warning: %1$s
+RSC_025=Informative parsing error: %1$s
 
 #Scripting
 SCP_001=Use of Javascript eval() function in EPUB scripts is a security risk.

src/main/resources/com/adobe/epubcheck/schema/30/epub-svg-30.nvdl

@@ -4,7 +4,7 @@
 		<namespace ns="http://www.w3.org/2000/svg">
 			<validate schema="epub-svg-30.rnc" schemaType="application/relax-ng-compact-syntax"
 				useMode="allowForeignNS">
-				<context path="foreignObject" useMode="allowHTMLOnly"/>
+				<context path="foreignObject" useMode="attachAnyNS"/>
 			</validate>
 			<validate schema="epub-svg-30.sch" useMode="attachAnyNS"/>
 		</namespace>
@@ -29,23 +29,6 @@
 			<allow/>
 		</anyNamespace>
 	</mode>
-	<mode name="allowHTMLOnly">
-		<namespace ns="http://www.w3.org/1999/xhtml">
-			<attach/>
-		</namespace>
-		<namespace ns="http://www.idpf.org/2007/ops" match="attributes">
-			<attach/>
-		</namespace>
-		<namespace ns="http://www.w3.org/1999/xlink" match="attributes">
-			<attach/>
-		</namespace>
-		<namespace ns="http://www.w3.org/XML/1998/namespace" match="attributes">
-			<attach/>
-		</namespace>
-		<namespace ns="" match="attributes">
-			<attach/>
-		</namespace>
-	</mode>
 	<mode name="attachAnyNS">
 		<anyNamespace>
 			<attach/>

src/main/resources/com/adobe/epubcheck/schema/30/epub-svg-30.rnc

@@ -1 +1,17 @@
-include "./mod/epub-svg11.rnc"
+namespace svg = "http://www.w3.org/2000/svg"
+
+include "./mod/epub-xhtml.rnc" {
+  start = svg
+}
+include "./mod/epub-mathml3.rnc"
+include "./mod/epub-svg-inc.rnc"
+
+svg.title.content |= common.elem.phrasing
+
+svg.foreignObject.content |= 
+  ( body.elem
+  | common.inner.flow
+  | math 
+  )
+
+mtext.content |= common.elem.phrasing

src/main/resources/com/adobe/epubcheck/schema/30/epub-svg-30.sch

@@ -1,11 +1,7 @@
 <?xml version="1.0" encoding="UTF-8"?>
 <schema xmlns="http://purl.oclc.org/dsdl/schematron">    
 
-    <ns uri="http://www.w3.org/2000/svg" prefix="svg" />
-
-    <pattern></pattern>
     <include href="./mod/id-unique.sch"/>
-    <include href="./mod/epub-svg11-re.sch"/>
 
 </schema>
 

src/main/resources/com/adobe/epubcheck/schema/30/epub-svg-informative.nvdl

@@ -0,0 +1,12 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<rules xmlns="http://purl.oclc.org/dsdl/nvdl/ns/structure/1.0">
+	<namespace ns="http://www.w3.org/2000/svg">
+		<validate schema="epub-svg-informative.rnc" schemaType="application/relax-ng-compact-syntax"/>
+	</namespace>
+	<namespace ns="" match="attributes">
+		<attach/>
+	</namespace>
+	<anyNamespace match="elements attributes">
+		<allow/>
+	</anyNamespace>
+</rules>

src/main/resources/com/adobe/epubcheck/schema/30/epub-svg-informative.rnc

@@ -0,0 +1,30 @@
+namespace svg = "http://www.w3.org/2000/svg"
+
+include "./mod/epub-xhtml.rnc" {
+  # This schema applies both to XHTML and SVG documents
+  start = svg | html.elem
+}
+include "./mod/svg11/svg11-inc.rnc" {
+  # ID datatype is set here to the lower HTML requirements,
+  # proper checking is done by the normative schema
+  SVG.id.attrib = attribute id { datatype.html5.token }?
+}
+include "./mod/svg11/inkscape.rnc"
+
+## EPUB-specific additions:
+
+SVG.Core.attrib &= aria.global?
+SVG.Core.extra.attrib &= epub.type.attr?
+
+## Do not check restricted elements, they are checked in normative schemas
+SVG.foreignObject.content |= any.content
+SVG.title.content |= any.content
+
+## Anything
+
+## Any attribute from any namespace, other than the special cases
+any.attr = attribute * { text }*
+## Any element from any namespace, other than the special cases
+any.elem = element * { any.content & any.attr }
+## Any content from any namespace
+any.content = text & any.elem*

src/main/resources/com/adobe/epubcheck/schema/30/epub-xhtml-30.rnc

@@ -1,7 +1,17 @@
-## XHTML5
+namespace svg = "http://www.w3.org/2000/svg"
+
 include "./mod/epub-xhtml.rnc"
+include "./mod/epub-mathml3.rnc"
+include "./mod/epub-svg-inc.rnc" {
+  svg.attr.id = attribute id { datatype.html5.token }?
+}
+
+common.elem.phrasing |= svg
+common.elem.phrasing |= math
 
-## SVG and MathML
+svg.title.content |= common.elem.phrasing
 
-include "./mod/epub-xhtml-mathml3.rnc"
-include "./mod/epub-xhtml-svg11.rnc"
+svg.foreignObject.content |= 
+  ( common.inner.flow
+  | math 
+  )

src/main/resources/com/adobe/epubcheck/schema/30/epub-xhtml-30.sch

@@ -402,6 +402,4 @@
         </rule>
     </pattern>
 
-    <include href="./mod/epub-svg11-re.sch"/>
-
 </schema>

src/main/resources/com/adobe/epubcheck/schema/30/epub30-catalog.xml

@@ -18,6 +18,12 @@
    <system systemId="http://www.idpf.org/epub/30/schema/epub-svg-30.sch"
            uri="epub-svg-30.sch"/>
 
+   <system systemId="http://www.idpf.org/epub/30/schema/epub-svg-informative.nvdl"
+           uri="epub-svg-informative.nvdl"/>
+
+   <system systemId="http://www.idpf.org/epub/30/schema/epub-svg-informative.rnc"
+           uri="epub-svg-informative.rnc"/>
+
    <system systemId="http://www.idpf.org/epub/30/schema/epub-xhtml-30.nvdl"
            uri="epub-xhtml-30.nvdl"/>
 
@@ -52,13 +58,9 @@
            uri="mod/epub-ssml-attrs.rnc"/>
    <system systemId="epub-ssml-attrs.rnc" uri="mod/epub-ssml-attrs.rnc"/>
 
-   <system systemId="http://www.idpf.org/epub/30/schema/mod/epub-svg11.rnc"
-           uri="mod/epub-svg11.rnc"/>
-   <system systemId="epub-svg11.rnc" uri="mod/epub-svg11.rnc"/>
-
-   <system systemId="http://www.idpf.org/epub/30/schema/mod/epub-svg11-re.sch"
-           uri="mod/epub-svg11-re.sch"/>
-   <system systemId="epub-svg11-re.sch" uri="mod/epub-svg11-re.sch"/>
+   <system systemId="http://www.idpf.org/epub/30/schema/mod/epub-svg-inc.rnc"
+           uri="mod/epub-svg-inc.rnc"/>
+   <system systemId="epub-svg-inc.rnc" uri="mod/epub-svg-inc.rnc"/>
 
    <system systemId="http://www.idpf.org/epub/30/schema/mod/epub-switch.rnc"
            uri="mod/epub-switch.rnc"/>

src/main/resources/com/adobe/epubcheck/schema/30/mod/epub-svg-inc.rnc

@@ -0,0 +1,39 @@
+default namespace = "http://www.w3.org/2000/svg"
+
+# #####################################################################
+##  Anything-goes schema for svg                                      #
+# #####################################################################
+
+## SVG root
+svg = svg.elem
+svg.elem =
+  element svg { any.content & any.attr }
+
+## Special cases
+specific.elems =
+  ( svg.foreignObject
+  | svg.title
+  )
+specific.attrs = svg.attr.id
+
+### SVG ID datatype
+svg.attr.id = attribute id { xsd:ID }?
+
+### SVG foreignObject element restrictions
+svg.foreignObject = 
+  element foreignObject { any.attr & svg.foreignObject.content }
+svg.foreignObject.content = notAllowed
+
+### SVG title element restrictions
+svg.title =
+  element title { any.attr & svg.title.content }
+svg.title.content = text
+
+## Anything
+
+## Any attribute from any namespace, other than the special cases
+any.attr = attribute * -id { text }* & svg.attr.id
+## Any element from any namespace, other than the special cases
+any.elem = element * - (foreignObject|title) { any.content & any.attr }
+## Any content from any namespace
+any.content = text & any.elem* & specific.elems*