[optimise] Namespace and indentation performance fixes

These are strictly incorrect, as they don’t do the XMLWriter substitution of special characters, BUT we could do this ONCE for every string; it does demonstrate that `XMLWriter.writeChars()` is a big part of the problem.
evolvedbinary · Dec 1, 2024 · 34e34dd · 34e34dd
1 parent 7fbb96b
commit 34e34dd
Show file tree

Hide file tree

Showing 2 changed files with 37 additions and 24 deletions.
diff --git a/exist-core/src/main/java/org/exist/util/serializer/IndentingXMLWriter.java b/exist-core/src/main/java/org/exist/util/serializer/IndentingXMLWriter.java
@@ -21,28 +21,28 @@
  */
 package org.exist.util.serializer;
 
-import java.io.IOException;
-import java.io.Writer;
-import java.util.ArrayDeque;
-import java.util.Deque;
-import java.util.Properties;
-
-import javax.xml.transform.OutputKeys;
-import javax.xml.transform.TransformerException;
-
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
 import org.exist.Namespaces;
 import org.exist.dom.QName;
 import org.exist.storage.serializers.EXistOutputKeys;
 
+import javax.xml.transform.OutputKeys;
+import javax.xml.transform.TransformerException;
+import java.io.IOException;
+import java.io.Writer;
+import java.util.ArrayDeque;
+import java.util.Arrays;
+import java.util.Deque;
+import java.util.Properties;
+
 public class IndentingXMLWriter extends XMLWriter {
 
     private final static Logger LOG = LogManager.getLogger(IndentingXMLWriter.class);
 
     private boolean indent = false;
     private int indentAmount = 4;
-    private String indentChars = "                                                                                           ";
+    private char[] indentChars = new char[0];
     private int level = 0;
     private boolean afterTag = false;
     private boolean sameline = false;
@@ -224,12 +224,19 @@ protected void indent() throws TransformerException {
         if (!indent || whitespacePreserve) {
             return;
         }
-        final int spaces = indentAmount * level;
-        while (spaces >= indentChars.length()) {
-            indentChars += indentChars;
+        final int padding = indentAmount * level + 1; /* 1 for leading \n */
+        if (indentChars.length < padding) {
+            indentChars = new char[padding + padding];
+            Arrays.fill(indentChars, ' ');
+            indentChars[0] = '\n';
+        }
+        try {
+            // aim is to use a single Writer.write to avoid efficiency overheads
+            // char[] is the most efficient form (least copies)
+            writer.write(indentChars, 0, padding);
+        } catch(final IOException ioe) {
+            throw new TransformerException(ioe.getMessage(), ioe);
         }
-        super.characters("\n");
-        super.characters(indentChars.subSequence(0, spaces));
         sameline = false;
     }
 

diff --git a/exist-core/src/main/java/org/exist/util/serializer/XMLWriter.java b/exist-core/src/main/java/org/exist/util/serializer/XMLWriter.java
@@ -44,6 +44,9 @@
 public class XMLWriter implements SerializerWriter {
 
     private final static IllegalStateException EX_CHARSET_NULL = new IllegalStateException("Charset should never be null!");
+
+    private static final char[] NAMESPACE_FN_XMLNS_PREFIX = " xmlns:".toCharArray();
+    private static final char[] NAMESPACE_FN_XMLNS_NO_PREFIX = " xmlns=\"".toCharArray();
 
     protected final static Properties defaultProperties = new Properties();
     static {
@@ -292,21 +295,22 @@ public void namespace(final String prefix, final String nsURI) throws Transforme
             }
 
             if(prefix != null && !prefix.isEmpty()) {
-                writer.write(' ');
-                writer.write("xmlns");
-                writer.write(':');
+                writer.write(NAMESPACE_FN_XMLNS_PREFIX);
                 writer.write(prefix);
                 writer.write("=\"");
-                writeChars(nsURI, true);
+                //TODO (AP) - test, just write as a string
+                //writeChars(nsURI, true);
+                writer.write(nsURI);
                 writer.write('"');
             } else {
                 if(defaultNamespace.equals(nsURI)) {
                     return;	
                 }
-                writer.write(' ');
-                writer.write("xmlns");
+                writer.write(NAMESPACE_FN_XMLNS_NO_PREFIX);
                 writer.write("=\"");
-                writeChars(nsURI, true);
+                //TODO (AP) - test, just write as a string
+                //writeChars(nsURI, true);
+                writer.write(nsURI);
                 writer.write('"');
                 defaultNamespace= nsURI;				
             }
@@ -348,7 +352,9 @@ public void attribute(final QName qname, final CharSequence value) throws Transf
             }
             writer.write(qname.getLocalPart());
             writer.write("=\"");
-            writeChars(value, true);
+            //TODO (AP) - perf hack - needs to respect specials (e.g. precompute)
+            //writeChars(value, true);
+            writer.append(value);
             writer.write('"');
         } catch(final IOException ioe) {
             throw new TransformerException(ioe.getMessage(), ioe);
@@ -364,7 +370,7 @@ public void characters(final CharSequence chars) throws TransformerException {
             if(tagIsOpen) {
                 closeStartTag(false);
             }
-            writeChars(chars, false);
+            writer.append(chars);
         } catch(final IOException ioe) {
             throw new TransformerException(ioe.getMessage(), ioe);
         }