node.h revision d2e42b9425ccd4c2d4c20c2217455da177035e48
/** @file
* @brief Interface for XML nodes
*/
/* Authors:
* MenTaLguY <mental@rydia.net>
* Krzysztof Kosiński <tweenk.pl@gmail.com> (documentation)
*
* Copyright 2005-2008 Authors
*
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* See the file COPYING for details.
*/
#ifndef SEEN_INKSCAPE_XML_NODE_H
#define SEEN_INKSCAPE_XML_NODE_H
#include "gc-anchored.h"
struct AttributeRecord;
struct Document;
struct NodeEventVector;
/**
* @brief Enumeration containing all supported node types.
*/
enum NodeType {
DOCUMENT_NODE, ///< Top-level document node. Do not confuse with the root node.
ELEMENT_NODE, ///< Regular element node, e.g. <group />.
TEXT_NODE, ///< Text node, e.g. "Some text" in <group>Some text</group> is represented by a text node.
COMMENT_NODE, ///< Comment node, e.g. <!-- some comment -->
PI_NODE ///< Processing instruction node, e.g. <?xml version="1.0" encoding="utf-8" standalone="no"?>
};
// careful; GC::Anchored should only appear once in the inheritance
// hierarchy; otherwise there will be leaks
/**
* @brief Interface for refcounted XML nodes
*
* This class is an abstract base type for all nodes in an XML document - this includes
* everything except attributes. An XML document is also a node itself. This is the main
* class used for interfacing with Inkscape's documents. Everything that has to be stored
* in the SVG has to go through this class at some point.
*
* Each node unconditionally has to belong to a document. There no "documentless" nodes,
* and it's not possible to move nodes between documents - they have to be duplicated.
* Each node can only refer to the nodes in the same document. Name of the node is immutable,
* it cannot be changed after its creation. Same goes for the type of the node. To simplify
* the use of this class, you can perform all operations on all nodes, but only some of them
* make any sense. For example, only element nodes can have attributes, only element and
* document nodes can have children, and all nodes except element and document nodes can
* have content. Although you can set content for element nodes, it won't make any difference
* in the XML output.
*
* To create new nodes, use the methods of the Inkscape::XML::Document class. You can obtain
* the nodes' document using the document() method. To destroy a node, just unparent it
* by calling sp_repr_unparent() or node->parent->removeChild() and release any references
* to it. The garbage collector will reclaim the memory in the next pass. There are additional
*
* In addition to regular DOM manipulations, you can register observer objects that will
* receive notifications about changes made to the node. See the NodeObserver class.
*
* @see Inkscape::XML::Document
* @see Inkscape::XML::NodeObserver
*/
Node() {}
/**
* @name Retrieve information about the node
* @{
*/
/**
* @brief Get the type of the node
* @return NodeType enumeration member corresponding to the type of the node.
*/
/**
* @brief Get the name of the element node
*
* This method only makes sense for element nodes. Names are stored as
* GQuarks to accelerate conversions.
*
* @return Name for element nodes, NULL for others
*/
/**
* @brief Get the integer code corresponding to the node's name
* @return GQuark code corresponding to the name
*/
/**
* @brief Get the index of this node in parent's child order
*
* If this method is used on a node that doesn't have a parent, the method will return 0,
* and a warning will be printed on the console.
*
* @return The node's index, or 0 if the node does not have a parent
*/
/**
* @brief Get the number of children of this node
* @return The number of children
*/
virtual unsigned childCount() const=0;
/**
* @brief Get the content of a text or comment node
*
* This method makes no sense for element nodes. To retrieve the element node's name,
* use the name() method.
*
* @return The node's content
*/
/**
* @brief Get the string representation of a node's attribute
*
* If there is no attribute with the given name, the method will return NULL.
* All strings returned by this method are owned by the node and may not be freed.
* The returned pointer will become invalid when the attribute changes. If you need
* to store the return value, use g_strdup(). To parse the string, use methods
* in repr.h
*
* @param key The name of the node's attribute
*/
/**
* @brief Get a list of the node's attributes
*
* The returned list is a functional programming style list rather than a standard one.
*
* @return A list of AttributeRecord structures describing the attributes
* @todo This method should return std::map<Glib::Quark const, gchar const *>
* or something similar with a custom allocator
*/
/**
* @brief Check whether this node has any attribute that matches a string
*
* This method checks whether this node has any attributes whose names
* have @c partial_name as their substrings. The check is done using
* the strstr() function of the C library. I don't know what would require that
* functionality, because matchAttributeName("id") matches both "identity" and "hidden".
*
* @param partial_name The string to match against all attributes
* @return true if there is such an attribute, false otherwise
*/
/*@}*/
/**
* @name Modify the node
* @{
*/
/**
* @brief Set the position of this node in parent's child order
*
* To move the node to the end of the parent's child order, pass a negative argument.
*
* @param pos The new position in parent's child order
*/
/**
* @brief Set the content of a text or comment node
*
* This method doesn't make sense for element nodes.
*
* @param value The node's new content
*/
//@{
/**
* @brief Change an attribute of this node
*
* The strings passed to this method are copied, so you can free them after use.
*
* @param key Name of the attribute to change
* @param value The new value of the attribute
* @param is_interactive Ignored
*/
{
}
{
}
//@}
/**
* @brief Directly set the integer GQuark code for the name of the node
*
* This function is a hack to easily move elements with no namespace to the SVG namespace.
* Do not use this function unless you really have a good reason.
*
* @param code The integer value corresponding to the string to be set as the name of this node
*/
/*@}*/
/**
* @name Traverse the XML tree
* @{
*/
//@{
/**
* @brief Get the node's associated document
* @return The document to which the node belongs. Never NULL.
*/
//@}
//@{
/**
* @brief Get the root node of this node's document
*
* This method works on any node that is part of an XML document, and returns
* the root node of the document in which it resides. For detached node hierarchies
* (i.e. nodes that are not descendants of a document node) this method
* returns the highest-level element node. For detached non-element nodes this method
* returns NULL.
*
* @return A pointer to the root element node, or NULL if the node is detached
*/
//@}
//@{
/**
* @brief Get the parent of this node
*
* This method will return NULL for detached nodes.
*
* @return Pointer to the parent, or NULL
*/
//@}
//@{
/**
* @brief Get the next sibling of this node
*
* This method will return NULL if the node is the last sibling element of the parent.
* The nodes form a singly-linked list, so there is no "prev()" method. Use the provided
* external function for that.
*
* @return Pointer to the next sibling, or NULL
* @see Inkscape::XML::previous_node()
*/
//@}
//@{
/**
* @brief Get the first child of this node
*
* For nodes without any children, this method returns NULL.
*
* @return Pointer to the first child, or NULL
*/
//@}
//@{
/**
* @brief Get the last child of this node
*
* For nodes without any children, this method returns NULL.
*
* @return Pointer to the last child, or NULL
*/
//@}
//@{
/**
* @brief Get the child of this node with a given index
*
* If there is no child with the specified index number, this method will return NULL.
*
* @param index The zero-based index of the child to retrieve
* @return Pointer to the appropriate child, or NULL
*/
//@}
/*@}*/
/**
* @name Manipulate the XML tree
* @{
*/
/**
* @brief Create a duplicate of this node
*
* The newly created node has no parent, and a refcount equal 1.
* You need to manually insert it into the document, using e.g. appendChild().
* Afterwards, call Inkscape::GC::release on it, so that it will be
* automatically collected when the parent is collected.
*
* @param doc The document in which the duplicate should be created
* @return A pointer to the duplicated node
*/
/**
* @brief Insert another node as a child of this node
*
* When @c after is NULL, the inserted node will be placed as the first child
* of this node. @c after must be a child of this node.
*
* @param child The node to insert
* @param after The node after which the inserted node should be placed, or NULL
*/
/**
* @brief Append a node as the last child of this node
* @param child The node to append
*/
/**
* @brief Remove a child of this node
*
* Once the pointer to the removed node disappears from the stack, the removed node
* will be collected in the next GC pass, but only as long as its refcount is zero.
* You should keep a refcount of zero for all nodes in the document except for
* the document node itself, because they will be held in memory by the parent.
*
* @param child The child to remove
*/
/**
* @brief Move a given node in this node's child order
*
* Both @c child and @c after must be children of this node for the method to work.
*
* @param child The node to move in the order
* @param after The sibling node after which the moved node should be placed
*/
/**
* @brief Merge all children of another node with the current
*
* This method merges two node hierarchies, where @c src takes precedence.
* @c key is the name of the attribute that determines whether two nodes are
* corresponding (it must be the same for both, and all of their ancestors). If there is
* a corresponding node in @c src hierarchy, their attributes and content override the ones
* already present in this node's hierarchy. If there is no corresponding node,
* it is copied from @c src to this node. This method is used when merging the user's
* preferences file with the defaults, and has little use beyond that.
*
* @param src The node to merge into this node
* @param key The attribute to use as the identity attribute
*/
/*@}*/
/**
* @name Notify observers about operations on the node
* @{
*/
/**
* @brief Add an object that will be notified of the changes to this node
*
* @c observer must be an object deriving from the NodeObserver class.
* The virtual methods of this object will be called when a corresponding change
* happens to this node. You can also notify the observer of the node's current state
* using synthesizeEvents(NodeObserver &).
*
* @param observer The observer object
*/
/**
* @brief Remove an object from the list of observers
* @param observer The object to be removed
*/
/**
* @brief Generate a sequence of events corresponding to the state of this node
*
* This function notifies the specified observer of all the events that would
* recreate the current state of this node; e.g. the observer is notified of
* all the attributes, children and content like they were just created.
* This function can greatly simplify observer logic.
*
* @param observer The node observer to notify of the events
*/
/**
* @brief Add an object that will be notified of the changes to this node and its descendants
*
* The difference between adding a regular observer and a subtree observer is that
* the subtree observer will also be notified if a change occurs to any of the node's
* descendants, while a regular observer will only be notified of changes to the node
* it was assigned to.
*
* @param observer The observer object
*/
/**
* @brief Remove an object from the subtree observers list
* @param observer The object to be removed
*/
/**
* @brief Add a set node change callbacks with an associated data
* @deprecated Use addObserver(NodeObserver &) instead
*/
/**
* @brief Remove a set of node change callbacks by their associated data
* @deprecated Use removeObserver(NodeObserver &) instead
*/
/**
* @brief Generate a sequence of events corresponding to the state of this node
* @deprecated Use synthesizeEvents(NodeObserver &) instead
*/
/*@}*/
};
}
}
#endif
/*
Local Variables:
mode:c++
c-file-style:"stroustrup"
c-file-offsets:((innamespace . 0)(inline-open . 0)(case-label . +))
indent-tabs-mode:nil
fill-column:99
End:
*/
// vim: filetype=cpp:expandtab:shiftwidth=4:tabstop=8:softtabstop=4:fileencoding=utf-8:textwidth=99 :