Merge branch 'master' of 10.68.203.26:jhy/jsoup

CHANGES

@@ -1,14 +1,21 @@
 jsoup changelog
 
-*** Release 0.2.0 (pending)
+*** Release 0.2.2 (2010-Feb-07)
+ * jsoup packages are now available in the Maven central repository.
+
+ * New feature: supports Element#addClass, removeClass, toggleClass;
+    also collection class methods on Elements.
  * New feature: supports Element#wrap(html) and Elements#wrap(html).
  * New selector syntax: supports E + F adjacent sibling selector
  * New selector systax: supports E ~ F preceding sibling selector
  * New: supports Element#elementSiblingIndex()
+
+ * Improved document normalisation.
+ * Improved HTML string output format (pretty-print)
 
  * Fixed absolute URL resolution issue when a base tag has no href.
 
-*** Release 0.1.2 (2010-Jan-02)
+*** Release 0.1.2 (2010-Feb-02)
  * Fixed unrecognised tag handler to be more permissive
     <http://github.com/jhy/jsoup/issues/issue/1>
 

pom.xml

@@ -5,7 +5,7 @@
 
   <groupId>org.jsoup</groupId>
   <artifactId>jsoup</artifactId>
-  <version>0.1.3-SNAPSHOT</version>
+  <version>0.2.3-SNAPSHOT</version>
   <description>jsoup HTML parser</description>
   <url>http://jsoup.org/</url>
   <inceptionYear>2009</inceptionYear>
@@ -76,6 +76,48 @@
       </plugin>
     </plugins>
   </build>
+
+  <distributionManagement>
+    <snapshotRepository>
+      <id>sonatype-nexus-snapshots</id>
+      <name>Sonatype Nexus Snapshots</name>
+      <url>http://oss.sonatype.org/content/repositories/snapshots</url>
+    </snapshotRepository>
+    <repository>
+      <id>sonatype-nexus-staging</id>
+      <name>Nexus Release Repository</name>
+      <url>http://oss.sonatype.org/service/local/staging/deploy/maven2/</url>
+    </repository>
+  </distributionManagement>
+
+  <profiles>
+    <profile>
+      <id>release-sign-artifacts</id>
+      <activation>
+        <property>
+          <name>performRelease</name>
+          <value>true</value>
+        </property>
+      </activation>
+      <build>
+        <plugins>
+          <plugin>
+            <groupId>org.apache.maven.plugins</groupId>
+            <artifactId>maven-gpg-plugin</artifactId>
+            <executions>
+              <execution>
+                <id>sign-artifacts</id>
+                <phase>verify</phase>
+                <goals>
+                  <goal>sign</goal>
+                </goals>
+              </execution>
+            </executions>
+          </plugin>
+        </plugins>
+      </build>
+    </profile>
+  </profiles>
 
   <dependencies>
 
@@ -103,4 +145,16 @@
     <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
   </properties>
 
+  <developers>
+    <developer>
+      <id>jhy</id>
+      <name>Jonathan Hedley</name>
+      <email>jonathan@hedley.net</email>
+      <roles>
+        <role>Lead Developer</role>
+      </roles>
+      <timezone>+11</timezone>
+    </developer>
+  </developers>
+
 </project>

src/main/java/org/jsoup/nodes/Comment.java

@@ -30,6 +30,7 @@ public String getData() {
     }
 
     void outerHtml(StringBuilder accum) {
+        indent(accum);
         accum.append(String.format("<!--%s-->", getData()));
     }
 

src/main/java/org/jsoup/nodes/Document.java

@@ -3,6 +3,9 @@
 import org.apache.commons.lang.Validate;
 import org.jsoup.parser.Tag;
 
+import java.util.List;
+import java.util.ArrayList;
+
 /**
  A HTML Document.
 
@@ -84,6 +87,44 @@ public Element createElement(String tagName) {
         return new Element(Tag.valueOf(tagName), this.baseUri());
     }
 
+    /**
+     Normalise the document. This happens after the parse phase so generally does not need to be called.
+     Moves any text content that is not in the body element into the body.
+     @return this document after normalisation
+     */
+    public Document normalise() {
+        if (select("html").isEmpty())
+            appendElement("html");
+        if (head() == null)
+            select("html").first().appendElement("head");
+        if (body() == null)
+            select("html").first().appendElement("body");
+
+        normalise(this);
+        normalise(select("html").first());
+        normalise(head());
+
+        return this;
+    }
+
+    // does not recurse. the result order isn't great here (not intuitive); they are in the body though.
+    private void normalise(Element element) {
+        List<Node> toMove = new ArrayList<Node>();
+        for (Node node: element.childNodes) {
+            if (node instanceof TextNode) {
+                TextNode tn = (TextNode) node;
+                if (!tn.isBlank())
+                    toMove.add(tn);
+            }
+        }
+
+        for (Node node: toMove) {
+            element.removeChild(node);
+            body().appendChild(new TextNode(" ", ""));
+            body().appendChild(node);
+        }
+    }
+
     @Override
     public String outerHtml() {
         return super.html(); // no outer wrapper tag

src/main/java/org/jsoup/nodes/Element.java

@@ -1,6 +1,7 @@
 package org.jsoup.nodes;
 
 import org.apache.commons.lang.Validate;
+import org.apache.commons.lang.StringUtils;
 import org.jsoup.parser.Parser;
 import org.jsoup.parser.Tag;
 import org.jsoup.select.Collector;
@@ -571,6 +572,25 @@ public Element text(String text) {
         return this;
     }
 
+    /**
+     Test if this element has any text content (that is not just whitespace).
+     @return true if element has non-blank text content.
+     */
+    public boolean hasText() {
+        for (Node child: childNodes) {
+            if (child instanceof TextNode) {
+                TextNode textNode = (TextNode) child;
+                if (!textNode.isBlank())
+                    return true;
+            } else if (child instanceof Element) {
+                Element el = (Element) child;
+                if (el.hasText())
+                    return true;
+            }
+        }
+        return false;
+    }
+
     /**
      * Get the combined data of this element. Data is e.g. the inside of a {@code script} tag.
      * @return the data, or empty string if none
@@ -602,17 +622,29 @@ public String className() {
 
     /**
      * Get all of the element's class names. E.g. on element {@code <div class="header gray"}>},
-     * returns a set of two elements {@code "header", "gray"}.
+     * returns a set of two elements {@code "header", "gray"}. Note that modifications to this set are not pushed to
+     * the backing {@code class} attribute; use the {@link #classNames(java.util.Set)} method to persist them.
      * @return set of classnames, empty if no class attribute
      */
     public Set<String> classNames() {
         if (classNames == null) {
             String[] names = className().split("\\s+");
-            classNames = new HashSet<String>(Arrays.asList(names));
+            classNames = new LinkedHashSet<String>(Arrays.asList(names));
         }
         return classNames;
     }
 
+    /**
+     Set the element's {@code class} attribute to the supplied class names.
+     @param classNames set of classes
+     @return this element, for chaining
+     */
+    public Element classNames(Set<String> classNames) {
+        Validate.notNull(classNames);
+        attributes.put("class", StringUtils.join(classNames, " "));
+        return this;
+    }
+
     /**
      * Tests if this element has a class.
      * @param className name of class to check for
@@ -622,8 +654,57 @@ public boolean hasClass(String className) {
         return classNames().contains(className);
     }
 
+    /**
+     Add a class name to this element's {@code class} attribute.
+     @param className class name to add
+     @return this element
+     */
+    public Element addClass(String className) {
+        Validate.notNull(className);
+
+        Set<String> classes = classNames();
+        classes.add(className);
+        classNames(classes);
+
+        return this;
+    }
+
+    /**
+     Remove a class name from this element's {@code class} attribute.
+     @param className class name to remove
+     @return this element
+     */
+    public Element removeClass(String className) {
+        Validate.notNull(className);
+
+        Set<String> classes = classNames();
+        classes.remove(className);
+        classNames(classes);
+
+        return this;
+    }
+
+    /**
+     Toggle a class name on this element's {@code class} attribute: if present, remove it; otherwise add it.
+     @param className class name to toggle
+     @return this element
+     */
+    public Element toggleClass(String className) {
+        Validate.notNull(className);
+
+        Set<String> classes = classNames();
+        if (classes.contains(className))
+            classes.remove(className);
+        else
+            classes.add(className);
+        classNames(classes);
+
+        return this;
+    }
+
     void outerHtml(StringBuilder accum) {
-
+        if (isBlock() || (parent() != null && parent().tag().canContainBlock() && siblingIndex() == 0))
+            indent(accum);
         accum
                 .append("<")
                 .append(tagName())
@@ -633,12 +714,9 @@ void outerHtml(StringBuilder accum) {
             accum.append(" />");
         } else {
             accum.append(">");
-            if (tag.canContainBlock())
-                accum.append("\n");
             html(accum);
+            if (tag.canContainBlock()) indent(accum);
             accum.append("</").append(tagName()).append(">");
-            if (tag.isBlock())
-                accum.append("\n");
         }
     }
 

src/main/java/org/jsoup/nodes/Node.java

@@ -2,6 +2,7 @@
 
 import org.apache.commons.lang.NotImplementedException;
 import org.apache.commons.lang.Validate;
+import org.apache.commons.lang.StringUtils;
 
 import java.net.MalformedURLException;
 import java.net.URL;
@@ -222,6 +223,13 @@ protected void addChild(Node in) {
         childNodes.add(in);
         in.parentNode = this;
     }
+
+    protected int nodeDepth() {
+        if (parentNode == null)
+            return 0;
+        else
+            return parentNode.nodeDepth() + 1;
+    }
 
     /**
      Retrieves this node's sibling nodes. Effectively, {@link #childNodes()  node.parent.childNodes()}.
@@ -259,6 +267,16 @@ public Node previousSibling() {
             return null;
     }
 
+    /**
+     * Get the list index of this node in its node sibling list. I.e. if this is the first node
+     * sibling, returns 0.
+     * @return position in node sibling list
+     * @see org.jsoup.nodes.Element#elementSiblingIndex()
+     */
+    public Integer siblingIndex() {
+        return indexInList(this, parent().childNodes);
+    }
+
     protected static <N extends Node> Integer indexInList(N search, List<N> nodes) {
         Validate.notNull(search);
         Validate.notNull(nodes);
@@ -291,6 +309,10 @@ public String toString() {
         return outerHtml();
     }
 
+    protected void indent(StringBuilder accum) {
+        accum.append("\n").append(StringUtils.leftPad("", nodeDepth() -1 * 2));
+    }
+
     @Override
     public boolean equals(Object o) {
         if (this == o) return true;

src/main/java/org/jsoup/nodes/TextNode.java

@@ -2,6 +2,7 @@
 
 import org.apache.commons.lang.StringEscapeUtils;
 import org.apache.commons.lang.Validate;
+import org.apache.commons.lang.StringUtils;
 
 /**
  A text node.
@@ -34,12 +35,22 @@ public String getWholeText() {
         return attributes.get(TEXT_KEY);
     }
 
+    /**
+     Test if this text node is blank -- that is, empty or only whitespace (including newlines).
+     @return true if this document is empty or only whitespace, false if it contains any text content.
+     */
+    public boolean isBlank() {
+        return StringUtils.isBlank(normaliseWhitespace(getWholeText()));
+    }
+
     void outerHtml(StringBuilder accum) {
         String html = StringEscapeUtils.escapeHtml(getWholeText());
         if (parent() instanceof Element && !((Element) parent()).preserveWhitespace()) {
             html = normaliseWhitespace(html);
         }
 
+        if (!isBlank() && parentNode instanceof Element && ((Element) parentNode).tag().canContainBlock()  && siblingIndex() == 0)
+            indent(accum);
         accum.append(html);
     }
 

src/main/java/org/jsoup/parser/Parser.java

@@ -78,7 +78,7 @@ private Document parse() {
                 parseTextNode();
             }
         }
-        return doc;
+        return doc.normalise();
     }
 
     private void parseComment() {