HTMLCollectionImplpublic class HTMLCollectionImpl extends Object implements HTMLCollectionImplements {@link org.w3c.dom.html.HTMLCollection} to traverse any named
elements on a {@link org.w3c.dom.html.HTMLDocument}. The elements type to
look for is identified in the constructor by code. This collection is not
optimized for traversing large trees.
The collection has to meet two requirements: it has to be live, and it has
to traverse depth first and always return results in that order. As such,
using an object container (such as {@link java.util.Vector}) is expensive on
insert/remove operations. Instead, the collection has been implemented using
three traversing functions. As a result, operations on large documents will
result in traversal of the entire document tree and consume a considerable
amount of time.
Note that synchronization on the traversed document cannot be achieved.
The document itself cannot be locked, and locking each traversed node is
likely to lead to a dead lock condition. Therefore, there is a chance of the
document being changed as results are fetched; in all likelihood, the results
might be out dated, but not erroneous. |
Fields Summary |
---|
static final short | ANCHORRequest collection of all anchors in document: <A> elements that
have a name attribute. | static final short | FORMRequest collection of all forms in document: <FORM> elements. | static final short | IMAGERequest collection of all images in document: <IMAGE> elements. | static final short | APPLETRequest collection of all Applets in document: <APPLET> and
<OBJECT> elements (<OBJECT> must contain an Applet). | static final short | LINKRequest collection of all links in document: <A> and <AREA>
elements (must have a href attribute). | static final short | OPTIONRequest collection of all options in selection: <OPTION> elments in
<SELECT> or <OPTGROUP>. | static final short | ROWRequest collection of all rows in table: <TR> elements in table or
table section. | static final short | ELEMENTRequest collection of all form elements: <INPUT>, <BUTTON>,
<SELECT>, <TEXT> and <TEXTAREA> elements inside form
<FORM>. | static final short | AREARequest collection of all areas in map: <AREA> element in <MAP>
(non recursive). | static final short | TBODYRequest collection of all table bodies in table: <TBODY> element in
table <TABLE> (non recursive). | static final short | CELLRequest collection of all cells in row: <TD> elements in <TR>
(non recursive). | private short | _lookingForIndicates what this collection is looking for. Holds one of the enumerated
values and used by {@link #collectionMatch}. Set by the constructor and
determine the collection's use for its life time. | private Element | _topLevelThis is the top level element underneath which the collection exists. |
Constructors Summary |
---|
HTMLCollectionImpl(HTMLElement topLevel, short lookingFor)Construct a new collection that retrieves element of the specific type
(lookingFor ) from the specific document portion
(topLevel ).
if ( topLevel == null )
throw new NullPointerException( "HTM011 Argument 'topLevel' is null." );
_topLevel = topLevel;
_lookingFor = lookingFor;
|
Methods Summary |
---|
protected boolean | collectionMatch(org.w3c.dom.Element elem, java.lang.String name)Determines if current element matches based on what we're looking for.
The element is passed along with an optional identifier name. If the
element is the one we're looking for, return true. If the name is also
specified, the name must match the id attribute
(match name first for anchors).
boolean match;
synchronized ( elem )
{
// Begin with no matching. Depending on what we're looking for,
// attempt to match based on the element type. This is the quickest
// way to match involving only a cast. Do the expensive string
// comparison later on.
match = false;
switch ( _lookingFor )
{
case ANCHOR:
// Anchor is an <A> element with a 'name' attribute. Otherwise, it's
// just a link.
match = ( elem instanceof HTMLAnchorElement ) &&
elem.getAttribute( "name" ).length() > 0;
break;
case FORM:
// Any <FORM> element.
match = ( elem instanceof HTMLFormElement );
break;
case IMAGE:
// Any <IMG> element. <OBJECT> elements with images are not returned.
match = ( elem instanceof HTMLImageElement );
break;
case APPLET:
// Any <APPLET> element, and any <OBJECT> element which represents an
// Applet. This is determined by 'codetype' attribute being
// 'application/java' or 'classid' attribute starting with 'java:'.
match = ( elem instanceof HTMLAppletElement ) ||
( elem instanceof HTMLObjectElement &&
( "application/java".equals( elem.getAttribute( "codetype" ) ) ||
elem.getAttribute( "classid" ).startsWith( "java:" ) ) );
break;
case ELEMENT:
// All form elements implement HTMLFormControl for easy identification.
match = ( elem instanceof HTMLFormControl );
break;
case LINK:
// Any <A> element, and any <AREA> elements with an 'href' attribute.
match = ( ( elem instanceof HTMLAnchorElement ||
elem instanceof HTMLAreaElement ) &&
elem.getAttribute( "href" ).length() > 0 );
break;
case AREA:
// Any <AREA> element.
match = ( elem instanceof HTMLAreaElement );
break;
case OPTION:
// Any <OPTION> element.
match = ( elem instanceof HTMLOptionElement );
break;
case ROW:
// Any <TR> element.
match = ( elem instanceof HTMLTableRowElement );
break;
case TBODY:
// Any <TBODY> element (one of three table section types).
match = ( elem instanceof HTMLTableSectionElement &&
elem.getTagName().equals( "tbody" ) );
break;
case CELL:
// Any <TD> element.
match = ( elem instanceof HTMLTableCellElement );
break;
}
// If element type was matched and a name was specified, must also match
// the name against either the 'id' or the 'name' attribute. The 'name'
// attribute is relevant only for <A> elements for backward compatibility.
if ( match && name != null )
{
// If an anchor and 'name' attribute matches, return true. Otherwise,
// try 'id' attribute.
if ( elem instanceof HTMLAnchorElement &&
name.equals( elem.getAttribute( "name" ) ) )
return true;
match = name.equals( elem.getAttribute( "id" ) );
}
}
return match;
| public final int | getLength()Returns the length of the collection. This method might traverse the
entire document tree.
// Call recursive function on top-level element.
return getLength( _topLevel );
| private int | getLength(org.w3c.dom.Element topLevel)Recursive function returns the number of elements of a particular type
that exist under the top level element. This is a recursive function
and the top level element is passed along.
int length;
Node node;
synchronized ( topLevel )
{
// Always count from zero and traverse all the childs of the
// current element in the order they appear.
length = 0;
node = topLevel.getFirstChild();
while ( node != null )
{
// If a particular node is an element (could be HTML or XML),
// do two things: if it's the one we're looking for, count
// another matched element; at any rate, traverse it's
// children as well.
if ( node instanceof Element )
{
if ( collectionMatch( (Element) node, null ) )
++ length;
else if ( recurse() )
length += getLength( (Element) node );
}
node = node.getNextSibling();
}
}
return length;
| public final org.w3c.dom.Node | item(int index)Retrieves the indexed node from the collection. Nodes are numbered in
tree order - depth-first traversal order. This method might traverse
the entire document tree.
if ( index < 0 )
throw new IllegalArgumentException( "HTM012 Argument 'index' is negative." );
// Call recursive function on top-level element.
return item( _topLevel, new CollectionIndex( index ) );
| private org.w3c.dom.Node | item(org.w3c.dom.Element topLevel, com.sun.org.apache.html.internal.dom.CollectionIndex index)Recursive function returns the numbered element of a particular type
that exist under the top level element. This is a recursive function
and the top level element is passed along.
Note that this function must call itself with an index and get back both
the element (if one was found) and the new index which is decremeneted
for any like element found. Since integers are only passed by value,
this function makes use of a separate class ({@link CollectionIndex})
to hold that index.
Node node;
Node result;
synchronized ( topLevel )
{
// Traverse all the childs of the current element in the order
// they appear. Count from the index backwards until you reach
// matching element with an index of zero. Return that element.
node = topLevel.getFirstChild();
while ( node != null )
{
// If a particular node is an element (could be HTML or XML),
// do two things: if it's the one we're looking for, decrease
// the index and if zero, return this node; at any rate,
// traverse it's children as well.
if ( node instanceof Element )
{
if ( collectionMatch( (Element) node, null ) )
{
if ( index.isZero() )
return node;
index.decrement();
} else if ( recurse() )
{
result = item( (Element) node, index );
if ( result != null )
return result;
}
}
node = node.getNextSibling();
}
}
return null;
| public final org.w3c.dom.Node | namedItem(java.lang.String name)Retrieves the named node from the collection. The name is matched case
sensitive against the id attribute of each element in the
collection, returning the first match. The tree is traversed in
depth-first order. This method might traverse the entire document tree.
if ( name == null )
throw new NullPointerException( "HTM013 Argument 'name' is null." );
// Call recursive function on top-level element.
return namedItem( _topLevel, name );
| private org.w3c.dom.Node | namedItem(org.w3c.dom.Element topLevel, java.lang.String name)Recursive function returns an element of a particular type with the
specified name (id attribute).
Node node;
Node result;
synchronized ( topLevel )
{
// Traverse all the childs of the current element in the order
// they appear.
node = topLevel.getFirstChild();
while ( node != null )
{
// If a particular node is an element (could be HTML or XML),
// do two things: if it's the one we're looking for, and the
// name (id attribute) attribute is the one we're looking for,
// return this element; otherwise, traverse it's children.
if ( node instanceof Element )
{
if ( collectionMatch( (Element) node, name ) )
return node;
else if ( recurse() )
{
result = namedItem( (Element) node, name );
if ( result != null )
return result;
}
}
node = node.getNextSibling();
}
return node;
}
| protected boolean | recurse()Returns true if scanning methods should iterate through the collection.
When looking for elements in the document, recursing is needed to traverse
the full document tree. When looking inside a specific element (e.g. for a
cell inside a row), recursing can lead to erroneous results.
return _lookingFor > 0;
|
|