feat: migrate to santuario c14n

instead of using a patched fork of apache c14n from 2013

This commit removes the block parameter from the JRuby API, because

a) it didn't previously work the same as CRuby/libxml2 (see earlier
commit that introduces block test coverage that failed)

b) with the mainline c14n library we can't hack it to support even the
broken callback semantics and would be forced to implement something
complicated ourselves

Given I'm not confident that C14n is widely-used anyway, removing a
misfeature like this doesn't seem like a controversial decision.

So, questions to the JRuby contributors:

Q1: Are we OK with all these new dependencies getting pulled in?
Q2: There's some Log4J warning messages being emitted at runtime that
    I'm not sure whether or how to quash. Any ideas?

Author: flavorjones

ext/java/nokogiri/NokogiriService.java

@@ -94,6 +94,8 @@ public class NokogiriService implements BasicLibraryService
     createSaxModule(ruby, xmlSaxModule, htmlSaxModule);
     createXsltModule(ruby, xsltModule);
     nokogiri.setInternalVariable("cache", populateNokogiriClassCahce(ruby));
+
+    org.apache.xml.security.Init.init();
   }
 
   private void

ext/java/nokogiri/XmlDocument.java

@@ -8,6 +8,7 @@
 import static nokogiri.internals.NokogiriHelpers.stringOrNil;
 
 import java.util.List;
+import java.io.ByteArrayOutputStream;
 
 import javax.xml.parsers.DocumentBuilderFactory;
 import javax.xml.parsers.ParserConfigurationException;
@@ -36,13 +37,13 @@
 import org.w3c.dom.Node;
 import org.w3c.dom.NodeList;
 
+import org.apache.xml.security.exceptions.XMLSecurityException;
+import org.apache.xml.security.c14n.Canonicalizer;
+
 import nokogiri.internals.NokogiriHelpers;
 import nokogiri.internals.NokogiriNamespaceCache;
 import nokogiri.internals.SaveContextVisitor;
 import nokogiri.internals.XmlDomParserContext;
-import nokogiri.internals.c14n.CanonicalFilter;
-import nokogiri.internals.c14n.CanonicalizationException;
-import nokogiri.internals.c14n.Canonicalizer;
 
 /**
  * Class for Nokogiri::XML::Document
@@ -673,15 +674,14 @@ private static class DocumentBuilderFactoryHolder
     try {
       Canonicalizer canonicalizer = Canonicalizer.getInstance(algorithmURI);
       XmlNode startingNode = getStartingNode(block);
-      byte[] result;
-      CanonicalFilter filter = new CanonicalFilter(context, block);
+      ByteArrayOutputStream writer = new ByteArrayOutputStream();
       if (inclusive_namespace == null) {
-        result = canonicalizer.canonicalizeSubtree(startingNode.getNode(), filter);
+        canonicalizer.canonicalizeSubtree(startingNode.getNode(), writer);
       } else {
-        result = canonicalizer.canonicalizeSubtree(startingNode.getNode(), inclusive_namespace, filter);
+        canonicalizer.canonicalizeSubtree(startingNode.getNode(), inclusive_namespace, writer);
       }
-      return RubyString.newString(context.runtime, new ByteList(result, UTF8Encoding.INSTANCE));
-    } catch (Exception e) {
+      return RubyString.newString(context.runtime, writer.toString());
+    } catch (XMLSecurityException e) {
       throw context.getRuntime().newRuntimeError(e.getMessage());
     }
   }