xml improvements for 2.2.1 (see changes)

Author: Burak Emir

src/library/scala/xml/Group.scala

@@ -25,6 +25,7 @@ case class Group(val nodes: Seq[Node]) extends Node {
 
   /** structural equality */
   override def equals(x: Any) = x match {
+    case z:Group     => (length == z.length) && sameElements(z)
     case z:Node      => (length == 1) && z == apply(0)
     case z:Seq[Node] => sameElements(z)
     case z:String    => text == z

src/library/scala/xml/PrettyPrinter.scala

@@ -196,6 +196,7 @@ class PrettyPrinter( width:Int, step:Int ) {
             makeBreak()
             traverse(node.child.elements, node.scope, ind + step)
             makeBox(cur, etg)
+            makeBreak()
           } else { // give up
             makeBox(ind, test)
             makeBreak()

src/library/scala/xml/UnprefixedAttribute.scala

@@ -22,7 +22,7 @@ class UnprefixedAttribute(val key: String, val value: Seq[Node], next1: MetaData
 
   /** same as this(key, Utility.parseAttributeValue(value), next) */
   def this(key: String, value: String, next: MetaData) =
-    this(key, if(value!=null) Utility.parseAttributeValue(value) else {val z:NodeSeq=null;z}, next)
+    this(key, if(value!=null) Text(value) else {val z:NodeSeq=null;z}, next)
 
   /** returns a copy of this unprefixed attribute with the given next field*/
   def copy(next: MetaData) =

src/library/scala/xml/Utility.scala

@@ -33,11 +33,25 @@ object Utility extends AnyRef with parsing.TokenTests {
       case '>' => s.append(">")
       case '&' => s.append("&")
       case '"' => s.append(""")
+      case '\'' => s.append("'")
       case _   => s.append(c)
     }
     s
   }
 
+  /* appends unescaped string to s, amp becomes & lt becomes <
+   * @returns null if ref was not a predefined entity
+   */
+  final def unescape(ref: String, s: StringBuilder): StringBuilder =
+    ref match {
+      case "lt"   => s.append('<')
+      case "gt"   => s.append('>')
+      case "amp"  => s.append('&')
+      case "quot" => s.append('"')
+      case "apos" => s.append(''')
+      case _   => null
+    }
+
   /**
    * Returns a set of all namespaces used in a sequence of nodes
    * and all their descendants, including the empty namespaces.
@@ -265,38 +279,45 @@ object Utility extends AnyRef with parsing.TokenTests {
    */
   def parseAttributeValue(value:String):Seq[Node] = {
     val zs:Seq[Char] = value
-    val sb = new StringBuilder()
+    val sb  = new StringBuilder
+    var rfb: StringBuilder = null
     val nb = new NodeBuffer()
     val it = zs.elements
     while(it.hasNext) {
       var c = it.next
       c match {
-        case '&' =>
-          if(sb.length() > 0) {
-            nb += Text(sb.toString())
-            sb.setLength(0)
-          }
+        case '&' => // entity! flush buffer into text node
           it.next match {
             case '#' =>
               c = it.next
               val theChar = parseCharRef ({ ()=> c },{ () => c = it.next },{s => throw new RuntimeException(s)})
               sb.append(theChar)
 
             case x =>
-              sb.append(x)
+              if(rfb==null) rfb = new StringBuilder()
+              rfb.append(x)
               c = it.next
               while(c != ';') {
-                sb.append(c)
+                rfb.append(c)
                 c = it.next
               }
-              nb += EntityRef(sb.toString())
-              sb.setLength(0)
+              val ref = rfb.toString()
+              rfb.setLength(0)
+              unescape(ref,sb) match {
+                case null =>
+                  if(sb.length() > 0) {          // flush buffer
+                    nb += Text(sb.toString())
+                    sb.setLength(0)
+                  }
+                  nb += EntityRef(sb.toString()) // add entityref
+                case _ =>
+              }
           }
         case x   =>
           sb.append(x)
       }
     }
-    if(sb.length() > 0) {
+    if(sb.length() > 0) { // flush buffer
       val x = Text(sb.toString())
       if(nb.length == 0)
         return x

test/files/jvm/xml01.check

@@ -22,6 +22,9 @@ passed ok
 -- group nodes
 <f><a></a><b></b><c></c></f>
 <a></a><f><a></a><b></b><c></c></f><a></a><b></b><c></c>
+passed ok
+passed ok
+passed ok
 attribute value normalization
 passed ok
 passed ok

test/files/jvm/xml01.scala

@@ -189,6 +189,15 @@ object Test {
     val zx2: Node = Group { List(<a/>,zy1,zx1) }
     Console println zx2.toString()
 
+    val zz1 = <xml:group><a/><b/><c/></xml:group>
+
+    assertTrue(zx1 == zz1)
+    assertTrue(zz1.length == 3)
+
+    // unparsed
+
+    val uup = <xml:unparsed>&<<>""^%@$!#</xml:unparsed>
+    assertTrue(uup == "&<<>\"\"^%@$!#")
     // test unicode escapes backslash u
 
   Console println ("attribute value normalization");

test/files/jvm/xml02.scala

@@ -21,7 +21,8 @@ def main(args:Array[String]) = {
   Console.println("three");
   assertEquals(bx \ "@foo", "bar&x")
   Console.println("four");
-  assertSameElements(bx \ "@foo", List(xml.Text("bar"),xml.EntityRef("amp"),xml.Text("x")))
+  assertSameElements(bx \ "@foo", List(xml.Text("bar&x")))
+  //assertSameElements(bx \ "@foo", List(xml.Text("bar"),xml.EntityRef("amp"),xml.Text("x")))
 
   Console.println("five");
   assertEquals(bx.toString, "<hello foo=\"bar&amp;x\"></hello>")

test/files/jvm/xmlattr.check

@@ -8,3 +8,5 @@ true
 
 None
 null
+<b x="&amp;"></b>
+<b x="&amp;"></b>

test/files/jvm/xmlattr.scala

@@ -20,4 +20,10 @@ object Test extends Application {
   Console.println(x.get("foo")) // None
   Console.println(x("foo")) // null
   }
+
+  val x1 = <b x="&amp;"/>
+  val x2 = <b x={"&"}/>
+  Console.println(x1)
+  Console.println(x2)
+
 }