<?xml version="1.0" encoding="UTF-8"?> <!-- $Id: traversal.xml 225913 2001-06-01 11:15:37Z dims $ --> <!-- ************************************************************************* * BEGINNING OF TRAVERSAL * ************************************************************************* --> <div1 id="Traversal"> <head>Document Object Model Traversal</head> <orglist role="editors"> <member> <name>Joe Kesselman</name> <affiliation>IBM</affiliation> </member> <member> <name>Jonathan Robie</name> <affiliation>Software AG</affiliation> </member> <member> <name>Mike Champion</name> <affiliation>Software AG</affiliation> </member> </orglist><?GENERATE-MINI-TOC?> <!-- ****************************************************** | INTRODUCTION | ****************************************************** --> <div2 id="Traversal-overview"> <head>Overview</head> <p>This chapter describes the optional DOM Level 2 <emph>Traversal</emph> feature. Its <code>TreeWalker</code>, <code>NodeIterator</code>, and <code>NodeFilter</code> interfaces provide easy-to-use, robust, selective traversal of a document's contents.</p> <p> The interfaces found within this section are not mandatory. A DOM application may use the <code>hasFeature(feature, version)</code> method of the <code>DOMImplementation</code> interface with parameter values "Traversal" and "2.0" (respectively) to determine whether or not this module is supported by the implementation. In order to fully support this module, an implementation must also support the "Core" feature defined defined in the DOM Level 2 Core specification <bibref ref="DOMCore"/>. Please refer to additional information about <xspecref href='&core.latest.url;/introduction.html#ID-Conformance'>conformance</xspecref> in the DOM Level 2 Core specification <bibref ref="DOMCore"/>. </p> <p> <code>NodeIterators</code> and <code>TreeWalkers</code> are two different ways of representing the nodes of a document subtree and a position within the nodes they present. A <code>NodeIterator</code> presents a flattened view of the subtree as an ordered sequence of nodes, presented in document order. Because this view is presented without respect to hierarchy, iterators have methods to move forward and backward, but not to move up and down. Conversely, a <code>TreeWalker</code> maintains the hierarchical relationships of the subtree, allowing navigation of this hierarchy. In general, <code>TreeWalkers</code> are better for tasks in which the structure of the document around selected nodes will be manipulated, while <code>NodeIterators</code> are better for tasks that focus on the content of each selected node. </p> <p><code>NodeIterators</code> and <code>TreeWalkers</code> each present a view of a document subtree that may not contain all nodes found in the subtree. In this specification, we refer to this as the <term id="Traversal-Logical-View">logical view</term> to distinguish it from the <term>physical view</term>, which corresponds to the document subtree per se. When an iterator or <code>TreeWalker</code> is created, it may be associated with a <code>NodeFilter</code>, which examines each node and determines whether it should appear in the logical view. In addition, flags may be used to specify which node types should occur in the logical view.</p> <p><code>NodeIterators</code> and <code>TreeWalkers</code> are dynamic - the logical view changes to reflect changes made to the underlying document. However, they differ in how they respond to those changes. <code>NodeIterators</code>, which present the nodes sequentially, attempt to maintain their location relative to a position in that sequence when the sequence's contents change. <code>TreeWalkers</code>, which present the nodes as a filtered tree, maintain their location relative to their current node and remain attached to that node if it is moved to a new context. We will discuss these behaviors in greater detail below.</p> <div3 id="Iterator-overview"> <head><code>NodeIterators</code></head> <p>A <code>NodeIterator</code> allows the members of a list of nodes to be returned sequentially. In the current DOM interfaces, this list will always consist of the nodes of a subtree, presented in <termref def="dt-documentorder">document order</termref>. When an iterator is first created, calling its <code>nextNode() </code>method returns the first node in the logical view of the subtree; in most cases, this is the root of the subtree. Each successive call advances the <code></code><code>NodeIterator</code> through the list, returning the next node available in the logical view. When no more nodes are visible, <code>nextNode()</code> returns <code>null</code>.</p> <p><code>NodeIterators</code> are created using the <code>createNodeIterator</code> method found in the <code>DocumentTraversal</code> interface. When a <code>NodeIterator</code> is created, flags can be used to determine which node types will be "visible" and which nodes will be "invisible" while traversing the tree; these flags can be combined using the <code>OR</code> operator. Nodes that are "invisible" are skipped over by the iterator as though they did not exist. </p> <p>The following code creates an iterator, then calls a function to print the name of each element:</p> <eg xml:space="preserve"> NodeIterator iter= ((DocumentTraversal)document).createNodeIterator( root, NodeFilter.SHOW_ELEMENT, null); while (Node n = iter.nextNode()) printMe(n); </eg> <div4 id="Iterator-Moving"> <head>Moving Forward and Backward</head> <p><code>NodeIterators</code> present nodes as an ordered list, and move forward and backward within this list. The iterator's position is always either between two nodes, before the first node, or after the last node. When an iterator is first created, the position is set before the first item. The following diagram shows the list view that an iterator might provide for a particular subtree, with the position indicated by an asterisk '*' :</p> <eg xml:space="preserve"> * A B C D E F G H I</eg> <p>Each call to <code>nextNode()</code> returns the next node and advances the position. For instance, if we start with the above position, the first call to <code>nextNode()</code> returns "A" and advances the iterator:</p> <eg xml:space="preserve"> [A] * B C D E F G H I</eg> <p>The position of a <code>NodeIterator</code> can best be described with respect to the last node returned, which we will call the <term>reference node</term>. When an iterator is created, the first node is the reference node, and the iterator is positioned before the reference node. In these diagrams, we use square brackets to indicate the reference node.</p> <p>A call to <code>previousNode()</code> returns the previous node and moves the position backward. For instance, if we start with the <code>NodeIterator</code> between "A" and "B", it would return "A" and move to the position shown below:</p> <eg xml:space="preserve"> * [A] B C D E F G H I</eg> <p>If <code>nextNode()</code> is called at the end of a list, or <code>previousNode()</code> is called at the beginning of a list, it returns <code>null</code> and does not change the position of the iterator. When a <code>NodeIterator</code> is first created, the reference node is the first node:</p> <eg xml:space="preserve"> * [A] B C D E F G H I</eg> </div4> <div4 id="Iterator-Robustness"> <head>Robustness</head> <p>A <code>NodeIterator</code> may be active while the data structure it navigates is being edited, so an iterator must behave gracefully in the face of change. Additions and removals in the underlying data structure do not invalidate a <code>NodeIterator</code>; in fact, a <code>NodeIterator</code> is never invalidated unless its <code>detach()</code> method is invoked. To make this possible, the iterator uses the reference node to maintain its position. The state of an iterator also depends on whether the iterator is positioned before or after the reference node.</p> <p>If changes to the iterated list do not remove the reference node, they do not affect the state of the <code>NodeIterator</code>. For instance, the iterator's state is not affected by inserting new nodes in the vicinity of the iterator or removing nodes other than the reference node. Suppose we start from the following position:</p> <eg xml:space="preserve">A B C [D] * E F G H I</eg> <p>Now let's remove "E". The resulting state is:</p> <eg xml:space="preserve">A B C [D] * F G H I</eg> <p>If a new node is inserted, the <code>NodeIterator</code> stays close to the reference node, so if a node is inserted between "D" and "F", it will occur between the iterator and "F": </p> <eg xml:space="preserve">A B C [D] * X F G H I</eg> <p>Moving a node is equivalent to a removal followed by an insertion. If we move "I" to the position before "X" the result is:</p> <eg xml:space="preserve">A B C [D] * I X F G H</eg> <p>If the reference node is removed from the list being iterated over, a different node is selected as the reference node. If the reference node's position is before that of the <code>NodeIterator</code>, which is usually the case after <code>nextNode()</code> has been called, the nearest node before the iterator is chosen as the new reference node. Suppose we remove the "D" node, starting from the following state:</p> <eg xml:space="preserve">A B C [D] * F G H I</eg> <p>The "C" node becomes the new reference node, since it is the nearest node to the <code>NodeIterator</code> that is before the iterator:</p> <eg xml:space="preserve">A B [C] * F G H I</eg> <p>If the reference node is after the <code>NodeIterator</code>, which is usually the case after <code>previousNode()</code> has been called, the nearest node after the iterator is chosen as the new reference node. Suppose we remove "E", starting from the following state:</p> <eg xml:space="preserve">A B C D * [E] F G H I</eg> <p>The "F" node becomes the new reference node, since it is the nearest node to the <code>NodeIterator</code> that is after the iterator:</p> <eg xml:space="preserve">A B C D * [F] G H I</eg> <p>As noted above, moving a node is equivalent to a removal followed by an insertion. Suppose we wish to move the "D" node to the end of the list, starting from the following state:</p> <eg xml:space="preserve">A B C [D] * F G H I C</eg> <p>The resulting state is as follows:</p> <eg xml:space="preserve">A B [C] * F G H I D</eg> <p>One special case arises when the reference node is the last node in the list and the reference node is removed. Suppose we remove node "C", starting from the following state:</p> <eg xml:space="preserve">A B * [C]</eg> <p>According to the rules we have given, the new reference node should be the nearest node after the <code>NodeIterator</code>, but there are no further nodes after "C". The same situation can arise when <code>previousNode()</code> has just returned the first node in the list, which is then removed. Hence: If there is no node in the original direction of the reference node, the nearest node in the opposite direction is selected as the reference node:</p> <eg xml:space="preserve">A [B] *</eg> <p>If the <code>NodeIterator</code> is positioned within a block of nodes that is removed, the above rules clearly indicate what is to be done. For instance, suppose "C" is the <termref def="dt-parent">parent</termref> node of "D", "E", and "F", and we remove "C", starting with the following state:</p> <eg xml:space="preserve">A B C [D] * E F G H I D</eg> <p>The resulting state is as follows:</p> <eg xml:space="preserve">A [B] * G H I D</eg> <p>Finally, note that removing a <code>NodeIterator</code>'s <code>root</code> node from its <termref def="dt-parent">parent</termref> does not alter the list being iterated over, and thus does not change the iterator's state.</p> </div4> <div4 id="Iterator-Visibility"> <head>Visibility of Nodes</head> <p>The underlying data structure that is being iterated may contain nodes that are not part of the logical view, and therefore will not be returned by the <code>NodeIterator</code>. If nodes that are to be excluded because of the value of the <code>whatToShow</code> flag, <code>nextNode()</code> returns the next visible node, skipping over the excluded "invisible" nodes. If a <code>NodeFilter</code> is present, it is applied before returning a node; if the filter does not accept the node, the process is repeated until a node is accepted by the filter and is returned. If no visible nodes are encountered, a <code>null</code> is returned and the iterator is positioned at the end of the list. In this case, the reference node is the last node in the list, whether or not it is visible. The same approach is taken, in the opposite direction, for <code>previousNode()</code>.</p> <p> In the following examples, we will use lowercase letters to represent nodes that are in the data structure, but which are not in the logical view. For instance, consider the following list:</p> <eg xml:space="preserve">A [B] * c d E F G</eg> <p>A call to <code>nextNode()</code> returns E and advances to the following position:</p> <eg xml:space="preserve">A B c d [E] * F G</eg> <p> Nodes that are not visible may nevertheless be used as reference nodes if a reference node is removed. Suppose node "E" is removed, started from the state given above. The resulting state is:</p> <eg xml:space="preserve">A B c [d] * F G</eg> <p>Suppose a new node "X", which is visible, is inserted before "d". The resulting state is:</p> <eg xml:space="preserve">A B c X [d] * F G</eg> <p>Note that a call to <code>previousNode()</code> now returns node X. It is important not to skip over invisible nodes when the reference node is removed, because there are cases, like the one just given above, where the wrong results will be returned. When "E" was removed, if the new reference node had been "B" rather than "d", calling <code>previousNode()</code> would not return "X".</p> </div4> </div3> <div3 id="Traversal-Filters"> <head><code>NodeFilters</code></head> <p><code>NodeFilters</code> allow the user to create objects that "filter out" nodes. Each filter contains a user-written function that looks at a node and determines whether or not it should be presented as part of the traversal's logical view of the document. To use a <code>NodeFilter</code>, you create a <code>NodeIterator</code> or a <code>TreeWalker</code> that uses the filter. The traversal engine applies the filter to each node, and if the filter does not accept the node, traversal skips over the node as though it were not present in the document. <code>NodeFilters</code> need not know how to navigate the structure that contains the nodes on which they operate. </p> <!--JKESS 20000305: Traversal has no memory, and no anticipation, of filters--> <p>Filters will be consulted when a traversal operation is performed, or when a <code>NodeIterator</code>'s reference node is removed from the subtree being iterated over and it must select a new one. However, the exact timing of these filter calls may vary from one DOM implementation to another. For that reason, <code>NodeFilters</code> should not attempt to maintain state based on the history of past invocations; the resulting behavior may not be portable.</p> <p>Similarly, <code>TreeWalkers</code> and <code>NodeIterators</code> should behave as if they have no memory of past filter results, and no anticipation of future results. If the conditions a <code>NodeFilter</code> is examining have changed (e.g., an attribute which it tests has been added or removed) since the last time the traversal logic examined this node, this change in visibility will be discovered only when the next traversal operation is performed. For example: if the filtering for the current node changes from <code>FILTER_SHOW</code> to <code>FILTER_SKIP</code>, a <code>TreeWalker</code> will be able to navigate off that node in any direction, but not back to it unless the filtering conditions change again. <code>NodeFilters</code> which change during a traversal can be written, but their behavior may be confusing and they should be avoided when possible.</p> <!-- End JKESS 20000305 --> <div4 id="Traversal-Filters-Usage"> <head>Using <code>NodeFilters</code></head> <p>A <code>NodeFilter</code> contains one method named <code>acceptNode()</code>, which allows a <code>NodeIterator</code> or <code>TreeWalker</code> to pass a <code>Node</code> to a filter and ask whether it should be present in the logical view. The <code>acceptNode()</code> function returns one of three values to state how the <code>Node</code> should be treated. If <code>acceptNode()</code> returns <code>FILTER_ACCEPT</code>, the <code>Node</code> will be present in the logical view; if it returns <code>FILTER_SKIP</code>, the <code>Node</code> will not be present in the logical view, but the children of the <code>Node</code> may; if it returns <code>FILTER_REJECT</code>, neither the <code>Node</code> nor its <termref def="dt-descendant">descendants</termref> will be present in the logical view. Since iterators present nodes as an ordered list, without hierarchy, <code>FILTER_REJECT</code> and <code>FILTER_SKIP</code> are synonyms for <code>NodeIterators</code>, skipping only the single current node.</p> <p>Consider a filter that accepts the named anchors in an HTML document. In HTML, an HREF can refer to any A element that has a NAME attribute. Here is a <code>NodeFilter</code> in Java that looks at a node and determines whether it is a named anchor:</p> <eg xml:space="preserve"> class NamedAnchorFilter implements NodeFilter { short acceptNode(Node n) { if (n.getNodeType()==Node.ELEMENT_NODE) { Element e = (Element)n; if (! e.getNodeName().equals("A")) return FILTER_SKIP; if (e.getAttributeNode("NAME") != null) return FILTER_ACCEPT; } return FILTER_SKIP; } }</eg> <p>If the above <code>NodeFilter</code> were to be used only with <code>NodeIterators</code>, it could have used <code>FILTER_REJECT</code> wherever <code>FILTER_SKIP</code> is used, and the behavior would not change. For <code>TreeWalker</code>, though, <code>FILTER_REJECT</code> would reject the children of any element that is not a named anchor, and since named anchors are always contained within other elements, this would have meant that no named anchors would be found. <code>FILTER_SKIP</code> rejects the given node, but continues to examine the children; therefore, the above filter will work with either a <code>NodeIterator</code> or a <code>TreeWalker</code>.</p> <p>To use this filter, the user would create an instance of the <code>NodeFilter</code> and create a <code>NodeIterator</code> using it:</p> <eg xml:space="preserve"> NamedAnchorFilter myFilter = new NamedAnchorFilter(); NodeIterator iter= ((DocumentTraversal)document).createNodeIterator( node, NodeFilter.SHOW_ELEMENT, myFilter); </eg> <p>Note that the use of the <code>SHOW_ELEMENT</code> flag is not strictly necessary in this example, since our sample <code>NodeFilter</code> tests the <code>nodeType</code>. However, some implementations of the Traversal interfaces may be able to improve <code>whatToShow</code> performance by taking advantage of knowledge of the document's structure, which makes the use of <code>SHOW_ELEMENT</code> worthwhile. Conversely, while we could remove the <code>nodeType</code> test from our filter, that would make it dependent upon <code>whatToShow</code> to distinguish between <code>Elements</code>, <code>Attr</code>'s, and <code>ProcessingInstructions</code>. </p> </div4> <div4 id="Traversal-Filters-Exceptions"> <head><code>NodeFilters</code> and Exceptions</head> <p>When writing a <code>NodeFilter</code>, users should avoid writing code that can throw an exception. However, because a DOM implementation can not prevent exceptions from being thrown, it is important that the behavior of filters that throw an exception be well-defined. A <code>TreeWalker</code> or <code>NodeIterator</code> does not catch or alter an exception thrown by a filter, but lets it propagate up to the user's code. The following functions may invoke a <code>NodeFilter</code>, and may therefore propagate an exception if one is thrown by a filter: <olist> <item> <p><code>NodeIterator</code><code>.nextNode()</code> </p> </item> <item> <p><code>NodeIterator</code><code>.previousNode()</code> </p> </item> <item> <p><code>TreeWalker</code><code>.firstChild()</code> </p> </item> <item> <p><code>TreeWalker</code><code>.lastChild()</code> </p> </item> <item> <p><code>TreeWalker</code><code>.nextSibling()</code> </p> </item> <item> <p><code>TreeWalker</code><code>.previousSibling()</code> </p> </item> <item> <p><code>TreeWalker</code><code>.nextNode()</code> </p> </item> <item> <p><code>TreeWalker</code><code>.previousNode()</code> </p> </item> <item> <p><code>TreeWalker</code><code>.parentNode()</code> </p> </item> </olist> </p> </div4> <div4 id="Traversal-Filters-Mutation"> <head><code>NodeFilters</code> and Document Mutation</head> <p>Well-designed <code>NodeFilters</code> should not have to modify the underlying structure of the document. But a DOM implementation can not prevent a user from writing filter code that does alter the document structure. Traversal does not provide any special processing to handle this case. For instance, if a <code>NodeFilter</code> removes a node from a document, it can still accept the node, which means that the node may be returned by the <code>NodeIterator</code> or <code>TreeWalker</code> even though it is no longer in the subtree being traversed. In general, this may lead to inconsistent, confusing results, so we encourage users to write <code>NodeFilters</code> that make no changes to document structures. Instead, do your editing in the loop controlled by the traversal object.</p> <!--A second case that was given by Andy was modifications that change the node so that the filter's tests would no longer accept the node. However, since the Filter makes the changes and applies the test, it is the Filter function itself that defines whether a node is accepted or rejected, so if the Filter modifies a node and then accepts it, this is simply the definition of the Filter. We decided to omit the case.--> </div4> <div4 id="Traversal-Filters-Flags"> <head><code>NodeFilters</code> and <code>whatToShow</code> flags </head> <p><code>NodeIterator</code> and <code>TreeWalker</code> apply their <code>whatToShow</code> flags before applying filters. If a node is <!-- JKESS 20000217: Clarify whatToShow behavior, potential interaction with FILTER_REJECT. --> skipped by the active <code>whatToShow</code> flags, a <code>NodeFilter</code> will not be called to evaluate that node. Please note that this behavior is similar to that of <code>FILTER_SKIP</code>; children of that node will be considered, and filters may be called to evaluate them. Also note that it will in fact be a "skip" even if the <code>NodeFilter</code> would have preferred to reject the entire subtree; if this would cause a problem in your application, consider setting <code>whatToShow</code> to <code>SHOW_ALL</code> and performing the <code>nodeType</code> test inside your filter. <!-- End JKESS 20000217 --> </p> </div4> </div3> <div3 id="TreeWalker"> <head><code>TreeWalker</code></head> <p>The <code>TreeWalker</code> interface provides many of the same benefits as the <code>NodeIterator</code> interface. The main difference between these two interfaces is that the <code>TreeWalker</code> presents a tree-oriented view of the nodes in a subtree, rather than the iterator's list-oriented view. In other words, an iterator allows you to move forward or back, but a <code>TreeWalker</code> allows you to also move to the <termref def="dt-parent">parent</termref> of a node, to one of its children, or to a <termref def="dt-sibling">sibling</termref>.</p> <p>Using a <code>TreeWalker</code> is quite similar to navigation using the Node directly, and the navigation methods for the two interfaces are analogous. For instance, here is a function that recursively walks over a tree of nodes in document order, taking separate actions when first entering a node and after processing any children:</p> <eg xml:space="preserve"> processMe(Node n) { nodeStartActions(n); for (Node child=n.firstChild(); child != null; child=child.nextSibling()) { processMe(child); } nodeEndActions(n); }</eg> <p>Doing the same thing using a <code>TreeWalker</code> is quite similar. There is one difference: since navigation on the <code>TreeWalker</code> changes the current position, the position at the end of the function has changed. A read/write attribute named <code>currentNode</code> allows the current node for a <code>TreeWalker</code> to be both queried and set. We will use this to ensure that the position of the <code>TreeWalker</code> is restored when this function is completed:</p> <eg xml:space="preserve"> processMe(TreeWalker tw) { Node n = tw.getCurrentNode(); nodeStartActions(tw); for (Node child=tw.firstChild(); child!=null; child=tw.nextSibling()) { processMe(tw); } tw.setCurrentNode(n); nodeEndActions(tw); }</eg> <p>The advantage of using a <code>TreeWalker</code> instead of direct <code>Node</code> navigation is that the <code>TreeWalker</code> allows the user to choose an appropriate view of the tree. Flags may be used to show or hide <code>Comments</code> or <code>ProcessingInstructions</code>; entities may be expanded or shown as <code>EntityReference</code> nodes. In addition, <code>NodeFilters</code> may be used to present a custom view of the tree. Suppose a program needs a view of a document that shows which tables occur in each chapter, listed by chapter. In this view, only the chapter elements and the tables that they contain are seen. The first step is to write an appropriate filter:</p> <eg xml:space="preserve"> class TablesInChapters implements NodeFilter { short acceptNode(Node n) { if (n.getNodeType()==Node.ELEMENT_NODE) { if (n.getNodeName().equals("CHAPTER")) return FILTER_ACCEPT; if (n.getNodeName().equals("TABLE")) return FILTER_ACCEPT; if (n.getNodeName().equals("SECT1") || n.getNodeName().equals("SECT2") || n.getNodeName().equals("SECT3") || n.getNodeName().equals("SECT4") || n.getNodeName().equals("SECT5") || n.getNodeName().equals("SECT6") || n.getNodeName().equals("SECT7")) return FILTER_SKIP; } return FILTER_REJECT; } }</eg> <p> </p> <p>This filter assumes that TABLE elements are contained directly in CHAPTER or SECTn elements. If another kind of element is encountered, it and its children are rejected. If a SECTn element is encountered, it is skipped, but its children are explored to see if they contain any TABLE elements.</p> <p>Now the program can create an instance of this <code>NodeFilter</code>, create a <code>TreeWalker</code> that uses it, and pass this <code>TreeWalker</code> to our ProcessMe() function:</p> <eg>TablesInChapters tablesInChapters = new TablesInChapters(); TreeWalker tw = ((DocumentTraversal)document).createTreeWalker( root, NodeFilter.SHOW_ELEMENT, tablesInChapters); processMe(tw);</eg> <p>(Again, we've chosen to both test the <code>nodeType</code> in the filter's logic and use <code>SHOW_ELEMENT</code>, for the reasons discussed in the earlier <code>NodeIterator</code> example.)</p> <p>Without making any changes to the above <code>ProcessMe()</code> function, it now processes only the CHAPTER and TABLE elements. The programmer can write other filters or set other flags to choose different sets of nodes; if functions use <code>TreeWalker</code> to navigate, they will support any view of the document defined with a <code>TreeWalker</code>.</p> <p>Note that the structure of a <code>TreeWalker</code>'s filtered view of a document may differ significantly from that of the document itself. For example, a <code>TreeWalker</code> with only <code>SHOW_TEXT</code> specified in its <code>whatToShow</code> parameter would present all the <code>Text</code> nodes as if they were <termref def="dt-sibling">siblings</termref> of each other yet had no <termref def="dt-parent">parent</termref>.</p> <div4 id="TreeWalker-Robustness"> <head>Robustness</head> <p>As with <code>NodeIterators</code>, a <code>TreeWalker</code> may be active while the data structure it navigates is being edited, and must behave gracefully in the face of change. Additions and removals in the underlying data structure do not invalidate a <code>TreeWalker</code>; in fact, a <code>TreeWalker</code> is never invalidated.</p> <p>But a <code>TreeWalker</code>'s response to these changes is quite different from that of a <code>NodeIterator</code>. While <code>NodeIterators</code> respond to editing by maintaining their position within the list that they are iterating over, <code>TreeWalkers</code> will instead remain attached to their <code>currentNode</code>. All the <code>TreeWalker</code>'s navigation methods operate in terms of the context of the <code>currentNode</code> at the time they are invoked, no matter what has happened to, or around, that node since the last time the <code>TreeWalker</code> was accessed. This remains true even if the <code>currentNode</code> is moved out of its original subtree.</p> <p>As an example, consider the following document fragment:</p> <eg xml:space="preserve"> ... <subtree> <twRoot> <currentNode/> <anotherNode/> </twRoot> </subtree> ... </eg> <p>Let's say we have created a <code>TreeWalker</code> whose <code>root</code> node is the <twRoot/> element and whose <code>currentNode</code> is the <currentNode/> element. For this illustration, we will assume that all the nodes shown above are accepted by the <code>TreeWalker</code>'s <code>whatToShow</code> and filter settings.</p> <p>If we use <code>removeChild()</code> to remove the <currentNode/> element from its <termref def="dt-parent">parent</termref>, that element remains the <code>TreeWalker</code>'s <code>currentNode</code>, even though it is no longer within the <code>root</code> node's subtree. We can still use the <code>TreeWalker</code> to navigate through any children that the orphaned <code>currentNode</code> may have, but are no longer able to navigate outward from the <code>currentNode</code> since there is no <termref def="dt-parent">parent</termref> available.</p> <p>If we use <code>insertBefore()</code> or <code>appendChild()</code> to give the <currentNode/> a new <termref def="dt-parent">parent</termref>, then <code>TreeWalker</code> navigation will operate from the <code>currentNode</code>'s new location. For example, if we inserted the <currentNode/> immediately after the <anotherNode/> element, the <code>TreeWalker</code>'s <code>previousSibling()</code> operation would move it back to the <anotherNode/>, and calling <code>parentNode()</code> would move it up to the <twRoot/>.</p> <p>If we instead insert the <code>currentNode</code> into the <subtree/> element, like so:</p> <eg xml:space="preserve"> ... <subtree> <currentNode/> <twRoot> <anotherNode/> </twRoot> </subtree> ...</eg> <p>we have moved the <code>currentNode</code> out from under the <code>TreeWalker</code>'s <code>root</code> node. This does not invalidate the <code>TreeWalker</code>; it may still be used to navigate relative to the <code>currentNode</code>. Calling its <code>parentNode()</code> operation, for example, would move it to the <subtree/> element, even though that too is outside the original <code>root</code> node. However, if the <code>TreeWalker</code>'s navigation should take it back into the original <code>root</code> node's subtree -- for example, if rather than calling <code>parentNode()</code> we called <code>nextNode()</code>, moving the <code>TreeWalker</code> to the <twRoot/> element -- the <code>root</code> node will "recapture" the <code>TreeWalker</code>, and prevent it from traversing back out.</p> <!--JKESS 20000301: Miles' "transient sibling" case--> <p>This becomes a bit more complicated when filters are in use. Relocation of the <code>currentNode</code> -- or explicit selection of a new <code>currentNode</code>, or changes in the conditions that the <code>NodeFilter</code> is basing its decisions on -- can result in a <code>TreeWalker</code> having a <code>currentNode</code> which would not otherwise be visible in the filtered (logical) view of the document. This node can be thought of as a "transient member" of that view. When you ask the <code>TreeWalker</code> to navigate off this node the result will be just as if it had been visible, but you may be unable to navigate back to it unless conditions change to make it visible again.</p> <p>In particular: If the <code>currentNode</code> becomes part of a subtree that would otherwise have been Rejected by the filter, that entire subtree may be added as transient members of the logical view. You will be able to navigate within that subtree (subject to all the usual filtering) until you move upward past the Rejected <termref def="dt-ancestor">ancestor</termref>. The behavior is as if the Rejected node had only been Skipped (since we somehow wound up inside its subtree) until we leave it; thereafter, standard filtering applies.</p> <!--End JKESS 20000301--> </div4> </div3> </div2> <div2 id="Traversal-IDLDefinition"> <head>Formal Interface Definition</head> <definitions> <interface name="NodeIterator" id="Traversal-NodeIterator" since="DOM Level 2"> <descr> <p><code>Iterators</code> are used to step through a set of nodes, e.g. the set of nodes in a <code>NodeList</code>, the document subtree governed by a particular <code>Node</code>, the results of a query, or any other set of nodes. The set of nodes to be iterated is determined by the implementation of the <code>NodeIterator</code>. DOM Level 2 specifies a single <code>NodeIterator</code> implementation for document-order traversal of a document subtree. Instances of these iterators are created by calling <code>DocumentTraversal</code><code>.createNodeIterator()</code>.</p> </descr> <!-- JKESS: 200000217: New attribute. Approved in 2/16 telecon --> <attribute id="Traversal-NodeIterator-root" name="root" type="Node" readonly="yes"> <descr> <p>The root node of the <code>NodeIterator</code>, as specified when it was created.</p> </descr> </attribute> <!-- End JKESS: 200000217:--> <!-- JKESS: 2/2/2000: whatToShow changed from long to unsigned long --> <attribute id="Traversal-NodeIterator-whatToShow" name="whatToShow" type="unsigned long" readonly="yes"> <descr> <p>This attribute determines which node types are presented via the iterator. The available set of constants is defined in the <code>NodeFilter</code> interface. <!--JKESS 20000228: Document interaction --> Nodes not accepted by <code>whatToShow</code> will be skipped, but their children may still be considered. Note that this skip takes precedence over the filter, if any. <!--End JKESS 20000228--></p> </descr> </attribute> <attribute id="Traversal-NodeIterator-filter" name="filter" type="NodeFilter" readonly="yes"> <descr> <p>The <code>NodeFilter</code> used to screen nodes.</p> </descr> </attribute> <attribute id="Traversal-NodeIterator-expandEntityReferences" name="expandEntityReferences" type="boolean" readonly="yes"> <descr> <p> The value of this flag determines whether the children of entity reference nodes are visible to the iterator. If false, they <!--JKESS 20000228: "will be skipped over" should be "rejected"--> and their <termref def="dt-descendant">descendants</termref> will be rejected. Note that this rejection takes precedence over <code>whatToShow</code> and the filter. Also note that this is currently the only situation where <code>NodeIterators</code> may reject a complete subtree rather than skipping individual nodes. <!--end JKESS 20000228--></p> <p></p> <p> To produce a view of the document that has entity references expanded and does not expose the entity reference node itself, use the <code>whatToShow</code> flags to hide the entity reference node and set <code>expandEntityReferences</code> to true when creating the iterator. To produce a view of the document that has entity reference nodes but no entity expansion, use the <code>whatToShow</code> flags to show the entity reference node and set <code>expandEntityReferences</code> to false.</p> </descr> </attribute> <method name="nextNode" id="Traversal-NodeIterator-nextNode"> <descr> <p>Returns the next node in the set and advances the position of the iterator in the set. After a <code>NodeIterator</code> is created, the first call to <code>nextNode()</code> returns the first node in the set.</p> </descr> <parameters> </parameters> <returns type="Node"> <descr> <p>The next <code>Node</code> in the set being iterated over, or <code>null</code> if there are no more members in that set.</p> </descr> </returns> <raises> <exception name="DOMException"> <descr> <p>INVALID_STATE_ERR: Raised if this method is called after the <code>detach</code> method was invoked.</p> </descr> </exception> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> <method name="previousNode" id="Traversal-NodeIterator-previousNode"> <descr> <p>Returns the previous node in the set and moves the position of the <code>NodeIterator</code> backwards in the set.</p> </descr> <parameters> </parameters> <returns type="Node"> <descr> <p>The previous <code>Node</code> in the set being iterated over, or <code>null</code> if there are no more members in that set. </p> </descr> </returns> <raises> <exception name="DOMException"> <descr> <p>INVALID_STATE_ERR: Raised if this method is called after the <code>detach</code> method was invoked.</p> </descr> </exception> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> <method name="detach" id="Traversal-NodeIterator-detach"> <descr> <p>Detaches the <code>NodeIterator</code> from the set which it iterated over, releasing any computational resources and placing the iterator in the INVALID state. After <code>detach</code> has been invoked, calls to <code>nextNode</code> or <code>previousNode</code> will raise the exception INVALID_STATE_ERR.</p> </descr> <parameters> </parameters> <returns type="void"> <descr> <p></p> </descr> </returns> <raises> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> </interface> <interface name="NodeFilter" id="Traversal-NodeFilter" since="DOM Level 2" role="special"> <descr> <p>Filters are objects that know how to "filter out" nodes. If a <code>NodeIterator</code> or <code>TreeWalker</code> is given a <code>NodeFilter</code>, it applies the filter before it returns the next node. If the filter says to accept the node, the traversal logic returns it; otherwise, traversal looks for the next node and pretends that the node that was rejected was not there.</p> <p>The DOM does not provide any filters. <code>NodeFilter</code> is just an interface that users can implement to provide their own filters. </p> <p><code>NodeFilters</code> do not need to know how to traverse from node to node, nor do they need to know anything about the data structure that is being traversed. This makes it very easy to write filters, since the only thing they have to know how to do is evaluate a single node. One filter may be used with a number of different kinds of traversals, encouraging code reuse.</p> </descr> <group name="Constants returned by acceptNode" id="Traversal-NodeFilter-acceptNode-constants"> <descr> <p>The following constants are returned by the acceptNode() method:</p> </descr> <constant name="FILTER_ACCEPT" type="short" value="1"> <descr> <p>Accept the node. Navigation methods defined for <code>NodeIterator</code> or <code>TreeWalker</code> will return this node.</p> </descr> </constant> <constant name="FILTER_REJECT" type="short" value="2"> <descr> <p>Reject the node. Navigation methods defined for <code>NodeIterator</code> or <code>TreeWalker</code> will not return this node. For <code>TreeWalker</code>, the children of this node will also be rejected. <code>NodeIterators</code> treat this as a synonym for <code>FILTER_SKIP</code>.</p> </descr> </constant> <constant name="FILTER_SKIP" type="short" value="3"> <descr> <p>Skip this single node. Navigation methods defined for <code>NodeIterator</code> or <code>TreeWalker</code> will not return this node. For both <code>NodeIterator</code> and <code>TreeWalker</code>, the children of this node will still be considered. </p> </descr> </constant> </group> <group id="Traversal-NodeFilter-whatToShow-constants" name="Constants for whatToShow"> <descr> <p>These are the available values for the <code>whatToShow</code> parameter used in <code>TreeWalkers</code> and <code>NodeIterators</code>. They are the same as the set of possible types for <code>Node</code>, and their values are derived by using a bit position corresponding to the value of <code>nodeType</code> for the equivalent node type. <!-- JKESS 20000217: Added behavioral clarifications --> If a bit in <code>whatToShow</code> is set false, that will be taken as a request to skip over this type of node; the behavior in that case is similar to that of <code>FILTER_SKIP</code>. </p> <p> Note that if node types greater than 32 are ever introduced, they may not be individually testable via <code>whatToShow</code>. If that need should arise, it can be handled by selecting <code>SHOW_ALL</code> together with an appropriate <code>NodeFilter</code>.</p> <!-- End JKESS 20000217 --> </descr> <!-- JKESS 20000217: Corrected from erroneous value 0x0000FFFF. Approved in 0216 telecon. --> <constant name="SHOW_ALL" type="unsigned long" value="0xFFFFFFFF"> <!-- End JKESS 20000217 --> <descr> <p>Show all <code>Nodes</code>.</p> </descr> </constant> <constant name="SHOW_ELEMENT" type="unsigned long" value="0x00000001"> <descr> <p>Show <code>Element</code> nodes.</p> </descr> </constant> <constant name="SHOW_ATTRIBUTE" type="unsigned long" value="0x00000002"> <descr> <p>Show <code>Attr</code> nodes. This is meaningful only when creating an iterator or tree-walker with an attribute node as its <code>root</code>; in this case, it means that the attribute node will appear in the first position of the iteration or traversal. Since attributes are never children of other nodes, they do not appear when traversing over the document tree.</p> </descr> </constant> <constant name="SHOW_TEXT" type="unsigned long" value="0x00000004"> <descr> <p>Show <code>Text</code> nodes.</p> </descr> </constant> <constant name="SHOW_CDATA_SECTION" type="unsigned long" value="0x00000008"> <descr> <p>Show <code>CDATASection</code> nodes.</p> </descr> </constant> <constant name="SHOW_ENTITY_REFERENCE" type="unsigned long" value="0x00000010"> <descr> <p>Show <code>EntityReference</code> nodes.</p> </descr> </constant> <constant name="SHOW_ENTITY" type="unsigned long" value="0x00000020"> <descr> <p>Show <code>Entity</code> nodes. This is meaningful only when creating an iterator or tree-walker with an<code> Entity</code> node as its <code>root</code>; in this case, it means that the <code>Entity</code> node will appear in the first position of the traversal. Since entities are not part of the document tree, they do not appear when traversing over the document tree.</p> </descr> </constant> <constant name="SHOW_PROCESSING_INSTRUCTION" type="unsigned long" value="0x00000040"> <descr> <p>Show <code>ProcessingInstruction</code> nodes.</p> </descr> </constant> <constant name="SHOW_COMMENT" type="unsigned long" value="0x00000080"> <descr> <p>Show <code>Comment</code> nodes.</p> </descr> </constant> <constant name="SHOW_DOCUMENT" type="unsigned long" value="0x00000100"> <descr> <p>Show <code>Document</code> nodes.</p> </descr> </constant> <constant name="SHOW_DOCUMENT_TYPE" type="unsigned long" value="0x00000200"> <descr> <p>Show <code>DocumentType</code> nodes.</p> </descr> </constant> <constant name="SHOW_DOCUMENT_FRAGMENT" type="unsigned long" value="0x00000400"> <descr> <p>Show <code>DocumentFragment</code> nodes.</p> </descr> </constant> <constant name="SHOW_NOTATION" type="unsigned long" value="0x00000800"> <descr> <p>Show <code>Notation</code> nodes. This is meaningful only when creating an iterator or tree-walker with a <code>Notation</code> node as its <code>root</code>; in this case, it means that the <code>Notation</code> node will appear in the first position of the traversal. Since notations are not part of the document tree, they do not appear when traversing over the document tree.</p> </descr> </constant> </group> <method name="acceptNode" id="Traversal-NodeFilter-acceptNode"> <descr> <p>Test whether a specified node is visible in the logical view of a <code>TreeWalker</code> or <code>NodeIterator</code>. This function will be called by the implementation of <code>TreeWalker</code> and <code>NodeIterator</code>; it is not normally called directly from user code. (Though you could do so if you wanted to use the same filter to guide your own application logic.)</p> </descr> <parameters> <param name="n" type="Node" attr="in"> <descr> <p>The node to check to see if it passes the filter or not.</p> </descr> </param> </parameters> <returns type="short"> <descr> <p>a constant to determine whether the node is accepted, rejected, or skipped, as defined <loc href="#Traversal-NodeFilter-acceptNode-constants">above</loc>.</p> </descr> </returns> <raises> <!-- No exceptions --> </raises> </method> </interface> <interface name="TreeWalker" id="Traversal-TreeWalker" since="DOM Level 2"> <descr> <p><code>TreeWalker</code> objects are used to navigate a document tree or subtree using the view of the document defined by their <code>whatToShow</code> flags and filter (if any). Any function which performs navigation using a <code>TreeWalker</code> will automatically support any view defined by a <code>TreeWalker</code>.</p> <p>Omitting nodes from the logical view of a subtree can result in a structure that is substantially different from the same subtree in the complete, unfiltered document. Nodes that are <termref def="dt-sibling">siblings</termref> in the <code>TreeWalker</code> view may be children of different, widely separated nodes in the original view. For instance, consider a <code>NodeFilter</code> that skips all nodes except for Text nodes and the root node of a document. In the logical view that results, all text nodes will be <termref def="dt-sibling">siblings</termref> and appear as direct children of the root node, no matter how deeply nested the structure of the original document.</p> </descr> <!-- JKESS: 200000217: New attribute. Approved in 2/16 telecon --> <attribute id="Traversal-TreeWalker-root" name="root" type="Node" readonly="yes"> <descr> <p>The <code>root</code> node of the <code>TreeWalker</code>, as specified when it was created.</p> </descr> </attribute> <!-- End JKESS: 200000217:--> <!-- JKESS: 2/2/2000: whatToShow changed from long to unsigned long --> <attribute id="Traversal-TreeWalker-whatToShow" name="whatToShow" type="unsigned long" readonly="yes"> <descr> <p>This attribute determines which node types are presented via the <code>TreeWalker</code>. The available set of constants is defined in the <code>NodeFilter</code> interface. <!--JKESS 20000228: Document interaction --> Nodes not accepted by <code>whatToShow</code> will be skipped, but their children may still be considered. Note that this skip takes precedence over the filter, if any. <!--End JKESS 20000228--></p> </descr> </attribute> <attribute id="Traversal-TreeWalker-filter" name="filter" type="NodeFilter" readonly="yes"> <descr> <p>The filter used to screen nodes.</p> </descr> </attribute> <attribute id="Traversal-TreeWalker-expandEntityReferences" name="expandEntityReferences" type="boolean" readonly="yes"> <descr> <p>The value of this flag determines whether the children of entity reference nodes are visible to the <code>TreeWalker</code>. If false, they <!--JKESS 20000228: "will be skipped over" should be "rejected"--> and their <termref def="dt-descendant">descendants</termref> will be rejected. Note that this rejection takes precedence over <code>whatToShow</code> and the filter, if any. <!--end JKESS 20000228--></p> <p> To produce a view of the document that has entity references expanded and does not expose the entity reference node itself, use the <code>whatToShow</code> flags to hide the entity reference node and set <code>expandEntityReferences</code> to true when creating the <code>TreeWalker</code>. To produce a view of the document that has entity reference nodes but no entity expansion, use the <code>whatToShow</code> flags to show the entity reference node and set <code>expandEntityReferences</code> to false.</p> </descr> </attribute> <attribute id="Traversal-TreeWalker-currentNode" name="currentNode" type="Node" readonly="no"> <descr> <p>The node at which the <code>TreeWalker</code> is currently positioned.</p> <p>Alterations to the DOM tree may cause the current node to no longer be accepted by the <code>TreeWalker</code>'s associated filter. <code>currentNode</code> may also be explicitly set to any node, whether or not it is within the subtree specified by the <code>root</code> node or would be accepted by the filter and <code>whatToShow</code> flags. Further traversal occurs relative to <code>currentNode</code> even if it is not part of the current view, by applying the filters in the requested direction; if no traversal is possible, <code>currentNode</code> is not changed. </p> </descr> <setraises> <exception name="DOMException"> <descr> <p>NOT_SUPPORTED_ERR: Raised if an attempt is made to set <code>currentNode</code> to <code>null</code>.</p> </descr> </exception> </setraises> </attribute> <method name="parentNode" id="Traversal-TreeWalker-parentNode"> <descr> <p>Moves to and returns the closest visible <termref def="dt-ancestor">ancestor</termref> node of the current node. If the search for <code>parentNode</code> attempts to step upward from the <code>TreeWalker</code>'s <code>root</code> node, or if it fails to find a visible <termref def="dt-ancestor">ancestor</termref> node, this method retains the current position and returns <code>null</code>.</p> </descr> <parameters> </parameters> <returns type="Node"> <descr> <p>The new <termref def="dt-parent">parent</termref> node, or <code>null</code> if the current node has no parent <!-- JKESS20000217: Added phrase --> in the <code>TreeWalker</code>'s logical view. <!-- End JKESS20000217 --> </p> </descr> </returns> <raises> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> <method name="firstChild" id="Traversal-TreeWalker-firstChild"> <descr> <p>Moves the <code>TreeWalker</code> to the first visible <termref def="dt-child">child</termref> of the current node, and returns the new node. If the current node has no visible children, returns <code>null</code>, and retains the current node.</p> </descr> <parameters> </parameters> <returns type="Node"> <descr> <p>The new node, or <code>null</code> if the current node has no visible children <!-- JKESS20000217: Added phrase --> in the <code>TreeWalker</code>'s logical view. <!-- End JKESS20000217 --> </p> </descr> </returns> <raises> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> <method name="lastChild" id="Traversal-TreeWalker-lastChild"> <descr> <p>Moves the <code>TreeWalker</code> to the last visible <termref def="dt-child">child</termref> of the current node, and returns the new node. If the current node has no visible children, returns <code>null</code>, and retains the current node.</p> </descr> <parameters> </parameters> <returns type="Node"> <descr> <p>The new node, or <code>null</code> if the current node has no children <!-- JKESS20000217: Added phrase --> in the <code>TreeWalker</code>'s logical view. <!-- End JKESS20000217 --> </p> </descr> </returns> <raises> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> <method name="previousSibling" id="Traversal-TreeWalker-previousSibling"> <descr> <p>Moves the <code>TreeWalker</code> to the previous <termref def="dt-sibling">sibling</termref> of the current node, and returns the new node. If the current node has no visible previous <termref def="dt-sibling">sibling</termref>, returns <code>null</code>, and retains the current node.</p> </descr> <parameters> </parameters> <returns type="Node"> <descr> <p>The new node, or <code>null</code> if the current node has no previous <termref def="dt-sibling">sibling</termref>. <!-- JKESS20000217: Added phrase --> in the <code>TreeWalker</code>'s logical view. <!-- End JKESS20000217 --> </p> </descr> </returns> <raises> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> <method name="nextSibling" id="Traversal-TreeWalker-nextSibling"> <descr> <p>Moves the <code>TreeWalker</code> to the next <termref def="dt-sibling">sibling</termref> of the current node, and returns the new node. If the current node has no visible next <termref def="dt-sibling">sibling</termref>, returns <code>null</code>, and retains the current node.</p> </descr> <parameters> </parameters> <returns type="Node"> <descr> <p>The new node, or <code>null</code> if the current node has no next <termref def="dt-sibling">sibling</termref>. <!-- JKESS20000217: Added phrase --> in the <code>TreeWalker</code>'s logical view. <!-- End JKESS20000217 --> </p> </descr> </returns> <raises> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> <method name="previousNode" id="Traversal-TreeWalker-previousNode"> <descr> <p>Moves the <code>TreeWalker</code> to the previous visible node in document order relative to the current node, and returns the new node. If the current node has no previous node, <!-- JKESS: Since this may involve an upward search... --> or if the search for <code>previousNode</code> attempts to step upward from the <code>TreeWalker</code>'s <code>root</code> node, <!-- END JKESS --> returns <code>null</code>, and retains the current node. </p> </descr> <parameters> </parameters> <returns type="Node"> <descr> <p>The new node, or <code>null</code> if the current node has no previous node <!-- JKESS20000217: Added phrase --> in the <code>TreeWalker</code>'s logical view. <!-- End JKESS20000217 --> </p> </descr> </returns> <raises> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> <method name="nextNode" id="Traversal-TreeWalker-nextNode"> <descr> <p>Moves the <code>TreeWalker</code> to the next visible node in document order relative to the current node, and returns the new node. If the current node has no next node, or if the search for nextNode attempts to step upward from the <code>TreeWalker</code>'s <code>root</code> node, returns <code>null</code>, and retains the current node.</p> </descr> <parameters> </parameters> <returns type="Node"> <descr> <p>The new node, or <code>null</code> if the current node has no next node <!-- JKESS20000217: Added phrase --> in the <code>TreeWalker</code>'s logical view. <!-- End JKESS20000217 --> </p> </descr> </returns> <raises> <!-- <exception name="Exceptions from user code"> <descr><p>Any exceptions raised by a user-written Filter will propagate through.</p></descr> </exception> --> </raises> </method> </interface> <interface name="DocumentTraversal" id="Traversal-Document" since="DOM Level 2"> <descr> <p><code>DocumentTraversal</code> contains methods that create iterators and tree-walkers to traverse a node and its children in document order (depth first, pre-order traversal, which is equivalent to the order in which the start tags occur in the text representation of the document). In DOMs which support the Traversal feature, <code>DocumentTraversal</code> will be implemented by the same objects that implement the Document interface.</p> </descr> <method name="createNodeIterator" id="Traversal-NodeIteratorFactory-createNodeIterator"> <descr> <p>Create a new <code>NodeIterator</code> over the subtree rooted at the specified node.</p> </descr> <parameters> <param name="root" type="Node" attr="in"> <descr> <p>The node which will be iterated together with its children. The iterator is initially positioned just before this node. The <code>whatToShow</code> flags and the filter, if any, are not considered when setting this position. The root must not be <code>null</code>.</p> </descr> </param> <!-- JKESS: 2/2/2000: whatToShow changed from long to unsigned long --> <param name="whatToShow" type="unsigned long" attr="in"> <descr> <p>This flag specifies which node types may appear in the logical view of the tree presented by the iterator. See the description of <code>NodeFilter</code> for the set of possible <code>SHOW_</code> values.</p> <p>These flags can be combined using <code>OR</code>.</p> </descr> </param> <param name="filter" type="NodeFilter" attr="in"> <descr> <p>The <code>NodeFilter</code> to be used with this <code>TreeWalker</code>, or <code>null</code> to indicate no filter.</p> </descr> </param> <param name="entityReferenceExpansion" type="boolean" attr="in"> <descr> <p>The value of this flag determines whether entity reference nodes are expanded.</p> </descr> </param> </parameters> <returns type="NodeIterator"> <descr> <p>The newly created <code>NodeIterator</code>.</p> </descr> </returns> <raises> <exception name="DOMException"> <descr> <p>NOT_SUPPORTED_ERR: Raised if the specified <code>root</code> is <code>null</code>.</p> </descr> </exception> </raises> </method> <method name="createTreeWalker" id="NodeIteratorFactory-createTreeWalker"> <descr> <p>Create a new <code>TreeWalker</code> over the subtree rooted at the specified node.</p> </descr> <parameters> <param name="root" type="Node" attr="in"> <descr> <p>The node which will serve as the <code>root</code> for the <code>TreeWalker</code>. The <code>whatToShow</code> flags and the <code>NodeFilter</code> are not considered when setting this value; any node type will be accepted as the <code>root</code>. The <code>currentNode</code> of the <code>TreeWalker</code> is initialized to this node, whether or not it is visible. The <code>root</code> functions as a stopping point for traversal methods that look upward in the document structure, such as <code>parentNode</code> and nextNode. The <code>root</code> must not be <code>null</code>.</p> </descr> </param> <!-- JKESS: 2/2/2000: whatToShow changed from long to unsigned long --> <param name="whatToShow" type="unsigned long" attr="in"> <descr> <p>This flag specifies which node types may appear in the logical view of the tree presented by the tree-walker. See the description of <code>NodeFilter</code> for the set of possible SHOW_ values.</p> <p>These flags can be combined using <code>OR</code>.</p> </descr> </param> <param name="filter" type="NodeFilter" attr="in"> <descr> <p>The <code>NodeFilter</code> to be used with this <code>TreeWalker</code>, or <code>null</code> to indicate no filter.</p> </descr> </param> <param name="entityReferenceExpansion" type="boolean" attr="in"> <descr> <p>If this flag is false, the contents of <code>EntityReference</code> nodes are not presented in the logical view.</p> </descr> </param> </parameters> <returns type="TreeWalker"> <descr> <p>The newly created <code>TreeWalker</code>.</p> </descr> </returns> <raises> <exception name="DOMException"> <descr> <p> NOT_SUPPORTED_ERR: Raised if the specified <code>root</code> is <code>null</code>.</p> </descr> </exception> </raises> </method> </interface> </definitions> </div2> </div1>