# PandaNode¶

class PandaNode

Bases: TypedWritableReferenceCount, Namable, LinkedListNode

A basic node of the scene graph or data graph. This is the base class of all specialized nodes, and also serves as a generic node with no special properties.

Inheritance diagram

class Children

This class is returned from get_children(). Use it to walk through the list of children. This is faster, and safer, than walking through the children one at a time via get_num_children()/get_child(), since the list of children is saved out ahead of time, rather than having to reacquire the lock with each iteration, or to keep the lock held for the entire pass.

std::size_t size(void) const
enum FancyBits
enumerator FB_transform = 1
enumerator FB_state = 2
enumerator FB_effects = 4
enumerator FB_tag = 16
enumerator FB_draw_mask = 32
enumerator FB_cull_callback = 64
class Parents

This class is returned from get_parents().

std::size_t size(void) const
class Stashed

Similarly for stashed children.

std::size_t size(void) const
enum UnexpectedChange
enumerator UC_parents = 1
enumerator UC_children = 2
enumerator UC_transform = 4
enumerator UC_state = 8
enumerator UC_draw_mask = 16
explicit PandaNode(std::string const &name)

Do not call the copy constructor directly; instead, use make_copy() or copy_subgraph() to make a copy of a node.

void add_child(PandaNode *child_node, int sort = 0, Thread *current_thread = Thread::get_current_thread())

Adds a new child to the node. The child is added in the relative position indicated by sort; if all children have the same sort index, the child is added at the end.

If the same child is added to a node more than once, the previous instance is first removed.

void add_stashed(PandaNode *child_node, int sort = 0, Thread *current_thread = Thread::get_current_thread())

Adds a new child to the node, directly as a stashed child. The child is not added in the normal sense, but will be revealed if unstash_child() is called on it later.

If the same child is added to a node more than once, the previous instance is first removed.

This can only be called from the top pipeline stage (i.e. from App).

void adjust_draw_mask(DrawMask show_mask, DrawMask hide_mask, DrawMask clear_mask)

Adjusts the hide/show bits of this particular node.

These three parameters can be used to adjust the _draw_control_mask and _draw_show_mask independently, which work together to provide per-camera visibility for the node and its descendents.

_draw_control_mask indicates the bits in _draw_show_mask that are significant. Each different bit corresponds to a different camera (and these bits are assigned via Camera::set_camera_mask()).

Where _draw_control_mask has a 1 bit, a 1 bit in _draw_show_mask indicates the node is visible to that camera, and a 0 bit indicates the node is hidden to that camera. Where _draw_control_mask is 0, the node is hidden only if a parent node is hidden.

The meaning of the three parameters is as follows:

set 1. Thus, show_mask indicates the set of cameras to which the node should be shown.

_draw_control_mask will be set 1. Thus, hide_mask indicates the set of cameras from which the node should be hidden.

clear_mask indicates the set of cameras from which the hidden state should be inherited from a parent.

virtual Light *as_light(void)

Cross-casts the node to a Light pointer, if it is one of the four kinds of Light nodes, or returns NULL if it is not.

void clear_attrib(TypeHandle type)
void clear_attrib(int slot)

Removes the render attribute of the given type from this node. This node, and the subgraph below, will now inherit the indicated render attribute from the nodes above this one.

void clear_bounds(void)

Reverses the effect of a previous call to set_bounds(), and allows the node’s bounding volume to be automatically computed once more based on the contents of the node.

void clear_effect(TypeHandle type)

Removes the render effect of the given type from this node.

void clear_effects(Thread *current_thread = Thread::get_current_thread())

Resets this node to have no render effects.

void clear_python_tag(PyObject *key)
void clear_state(Thread *current_thread = Thread::get_current_thread())

Resets this node to leave the render state alone. Nodes at this level and below will once again inherit their render state unchanged from the nodes above this level.

void clear_tag(std::string const &key, Thread *current_thread = Thread::get_current_thread())

Removes the value defined for this key on this particular node. After a call to clear_tag(), has_tag() will return false for the indicated key.

void clear_transform(Thread *current_thread = Thread::get_current_thread())

Resets the transform on this node to the identity transform.

void clear_unexpected_change(unsigned int flags)

Sets one or more of the PandaNode::UnexpectedChange bits off, indicating that the corresponding property may once again change on this node. See set_unexpected_change().

The input parameter is the union of bits that are to be cleared.

Since this is a developer debugging tool only, this function does nothing in a production (NDEBUG) build.

virtual PandaNode *combine_with(PandaNode *other)

published so that characters can be combined.

Collapses this PandaNode with the other PandaNode, if possible, and returns a pointer to the combined PandaNode, or NULL if the two PandaNodes cannot safely be combined.

The return value may be this, other, or a new PandaNode altogether.

This function is called from GraphReducer::flatten(), and need not deal with children; its job is just to decide whether to collapse the two PandaNodes and what the collapsed PandaNode should look like.

int compare_tags(PandaNode const *other) const

Returns a number less than 0, 0, or greater than 0, to indicate the similarity of tags between this node and the other one. If this returns 0, the tags are identical. If it returns other than 0, then the tags are different; and the nodes may be sorted into a consistent (but arbitrary) ordering based on this number.

void copy_all_properties(PandaNode *other)

Copies the TransformState, RenderState, RenderEffects, tags, Python tags, and the show/hide state from the other node onto this one. Typically this is used to prepare a node to replace another node in the scene graph (also see replace_node()).

void copy_children(PandaNode *other, Thread *current_thread = Thread::get_current_thread())

Makes another instance of all the children of the other node, copying them to this node.

PointerTo<PandaNode> copy_subgraph(Thread *current_thread = Thread::get_current_thread()) const

Allocates and returns a complete copy of this PandaNode and the entire scene graph rooted at this PandaNode. Some data may still be shared from the original (e.g. vertex index tables), but nothing that will impede normal use of the PandaNode.

void copy_tags(PandaNode *other)

Copies all of the tags stored on the other node onto this node. If a particular tag exists on both nodes, the contents of this node’s value is replaced by that of the other.

int count_num_descendants(void) const

Returns the number of nodes at and below this level.

static PointerTo<PandaNode> decode_from_bam_stream(vector_uchar data, BamReader *reader = nullptr)

Reads the bytes created by a previous call to encode_to_bam_stream(), and extracts and returns the single object on those bytes. Returns NULL on error.

This method is intended to replace decode_raw_from_bam_stream() when you know the stream in question returns an object of type PandaNode, allowing for easier reference count management. Note that the caller is still responsible for maintaining the reference count on the return value.

int find_child(PandaNode *node, Thread *current_thread = Thread::get_current_thread()) const

Returns the index of the indicated child node, if it is a child, or -1 if it is not.

int find_parent(PandaNode *node, Thread *current_thread = Thread::get_current_thread()) const

Returns the index of the indicated parent node, if it is a parent, or -1 if it is not.

int find_stashed(PandaNode *node, Thread *current_thread = Thread::get_current_thread()) const

Returns the index of the indicated stashed node, if it is a stashed child, or -1 if it is not.

DrawMask get_all_camera_mask(void)

Returns a DrawMask that is appropriate for rendering to all cameras.

ConstPointerTo<RenderAttrib> get_attrib(TypeHandle type) const
ConstPointerTo<RenderAttrib> get_attrib(int slot) const

Returns the render attribute of the indicated type, if it is defined on the node, or NULL if it is not. This checks only what is set on this particular node level, and has nothing to do with what render attributes may be inherited from parent nodes.

ConstPointerTo<BoundingVolume> get_bounds(Thread *current_thread = Thread::get_current_thread()) const
ConstPointerTo<BoundingVolume> get_bounds(UpdateSeq &seq, Thread *current_thread = Thread::get_current_thread()) const

Returns the external bounding volume of this node: a bounding volume that contains the user bounding volume, the internal bounding volume, and all of the children’s bounding volumes.

This flavor of get_bounds() return the external bounding volume, and also fills in seq with the bounding volume’s current sequence number. When this sequence number changes, it indicates that the bounding volume might have changed, e.g. because some nested child’s bounding volume has changed.

Although this might occasionally increment without changing the bounding volume, the bounding volume will never change without incrementing this counter, so as long as this counter remains unchanged you can be confident the bounding volume is also unchanged.

BoundingVolume::BoundsType get_bounds_type(void) const

Returns the bounding volume type set with set_bounds_type().

PandaNode *get_child(int n, Thread *current_thread = Thread::get_current_thread()) const

Returns the nth child node of this node. See get_num_children(). Also see get_children(), if your intention is to iterate through the complete list of children; get_children() is preferable in this case.

int get_child_sort(int n, Thread *current_thread = Thread::get_current_thread()) const

Returns the sort index of the nth child node of this node (that is, the number that was passed to add_child()). See get_num_children().

static TypeHandle get_class_type(void)
DrawMask get_draw_control_mask(void) const

Returns the set of bits in draw_show_mask that are considered meaningful. See adjust_draw_mask().

DrawMask get_draw_show_mask(void) const

Returns the hide/show bits of this particular node. See adjust_draw_mask().

ConstPointerTo<RenderEffect> get_effect(TypeHandle type) const

Returns the render effect of the indicated type, if it is defined on the node, or NULL if it is not.

ConstPointerTo<RenderEffects> get_effects(Thread *current_thread = Thread::get_current_thread()) const

Returns the complete RenderEffects that will be applied to this node.

int get_fancy_bits(Thread *current_thread = Thread::get_current_thread()) const

Returns the union of all of the enum FancyBits values corresponding to the various “fancy” attributes that are set on the node. If this returns 0, the node has nothing interesting about it. This is intended to speed traversal by quickly skipping past nodes that don’t particularly affect the render state.

ConstPointerTo<BoundingVolume> get_internal_bounds(Thread *current_thread = Thread::get_current_thread()) const

Returns the node’s internal bounding volume. This is the bounding volume around the node alone, without including children. If the user has called set_bounds(), it will be the specified bounding volume.

Returns the node’s internal bounding volume. This is the bounding volume around the node alone, without including children.

int get_internal_vertices(Thread *current_thread = Thread::get_current_thread()) const

Returns the total number of vertices that will be rendered by this particular node alone, not accounting for its children.

This may not include all vertices for certain dynamic effects.

CollideMask get_into_collide_mask(void) const

Returns the “into” collide mask for this node.

Returns the subset of CollideMask bits that may be set for this particular type of PandaNode. For most nodes, this is 0; it doesn’t make sense to set a CollideMask for most kinds of nodes.

For nodes that can be collided with, such as GeomNode and CollisionNode, this returns all bits on.

int get_nested_vertices(Thread *current_thread = Thread::get_current_thread()) const

Returns the total number of vertices that will be rendered by this node and all of its descendents.

This is not necessarily an accurate count of vertices that will actually be rendered, since this will include all vertices of all LOD’s, and it will also include hidden nodes. It may also omit or only approximate certain kinds of dynamic geometry. However, it will not include stashed nodes.

CollideMask get_net_collide_mask(Thread *current_thread = Thread::get_current_thread()) const

Returns the union of all into_collide_mask() values set at CollisionNodes at this level and below.

DrawMask get_net_draw_control_mask(void) const

Returns the set of bits in get_net_draw_show_mask() that have been explicitly set via adjust_draw_mask(), rather than implicitly inherited.

A 1 bit in any position of this mask indicates that (a) this node has renderable children, and (b) some child of this node has made an explicit hide() or show_through() call for the corresponding bit.

DrawMask get_net_draw_show_mask(void) const

Returns the union of all draw_show_mask values–of renderable nodes only– at this level and below. If any bit in this mask is 0, there is no reason to traverse below this node for a camera with the corresponding camera_mask.

The bits in this mask that do not correspond to a 1 bit in the net_draw_control_mask are meaningless (and will be set to 1). For bits that do correspond to a 1 bit in the net_draw_control_mask, a 1 bit indicates that at least one child should be visible, while a 0 bit indicates that all children are hidden.

int get_num_children(Thread *current_thread = Thread::get_current_thread()) const

Returns the number of child nodes this node has. The order of the child nodes is meaningful and is based on the sort number that was passed to add_child(), and also on the order in which the nodes were added.

int get_num_parents(Thread *current_thread = Thread::get_current_thread()) const

Returns the number of parent nodes this node has. If this number is greater than 1, the node has been multiply instanced. The order of the parent nodes is not meaningful and is not related to the order in which the node was instanced to them.

int get_num_stashed(Thread *current_thread = Thread::get_current_thread()) const

Returns the number of stashed nodes this node has. These are former children of the node that have been moved to the special stashed list via stash_child().

ConstPointerTo<RenderAttrib> get_off_clip_planes(Thread *current_thread = Thread::get_current_thread()) const

Returns a ClipPlaneAttrib which represents the union of all of the clip planes that have been turned off at this level and below.

DrawMask get_overall_bit(void)

Returns the special bit that, when specifically cleared in the node’s DrawMask, indicates that the node is hidden to all cameras, regardless of the remaining DrawMask bits.

PandaNode *get_parent(int n, Thread *current_thread = Thread::get_current_thread()) const

Returns the nth parent node of this node. See get_num_parents(). Also see get_parents(), if your intention is to iterate through the complete list of parents; get_parents() is preferable in this case.

ConstPointerTo<TransformState> get_prev_transform(Thread *current_thread = Thread::get_current_thread()) const

Returns the transform that has been set as this node’s “previous” position. See set_prev_transform().

PyObject *get_python_tag(PyObject *key) const
PyObject *get_python_tag_keys(void) const
PyObject *get_python_tags(void)
PandaNode *get_stashed(int n, Thread *current_thread = Thread::get_current_thread()) const
Stashed get_stashed(Thread *current_thread = Thread::get_current_thread()) const

Returns the nth stashed child of this node. See get_num_stashed(). Also see get_stashed(), if your intention is to iterate through the complete list of stashed children; get_stashed() is preferable in this case.

Returns an object that can be used to walk through the list of children of the node. When you intend to visit multiple children, using this is slightly faster than calling get_stashed() directly on the PandaNode, since this object avoids reopening the PipelineCycler each time.

This object also protects you from self-modifying loops (e.g. adding or removing children during traversal), since a virtual copy of the children is made ahead of time. The virtual copy is fast–it is a form of copy-on- write, so the list is not actually copied unless it is modified during the traversal.

int get_stashed_sort(int n, Thread *current_thread = Thread::get_current_thread()) const

Returns the sort index of the nth stashed node of this node (that is, the number that was passed to add_child()). See get_num_stashed().

ConstPointerTo<RenderState> get_state(Thread *current_thread = Thread::get_current_thread()) const

Returns the complete RenderState that will be applied to all nodes at this level and below, as set on this node. This returns only the RenderState set on this particular node, and has nothing to do with state that might be inherited from above.

std::string get_tag(std::string const &key, Thread *current_thread = Thread::get_current_thread()) const

Retrieves the user-defined value that was previously set on this node for the particular key, if any. If no value has been previously set, returns the empty string.

PyObject *get_tag_keys(void) const

Fills the given vector up with the list of tags on this PandaNode.

It is the user’s responsibility to ensure that the keys vector is empty before making this call; otherwise, the new keys will be appended to it.

ConstPointerTo<TransformState> get_transform(Thread *current_thread = Thread::get_current_thread()) const

Returns the transform that has been set on this particular node. This is not the net transform from the root, but simply the transform on this particular node.

unsigned int get_unexpected_change(unsigned int flags) const

Returns nonzero if any of the bits in the input parameter are set on this node, or zero if none of them are set. More specifically, this returns the particular set of bits (masked by the input parameter) that have been set on this node. See set_unexpected_change().

Since this is a developer debugging tool only, this function always returns zero in a production (NDEBUG) build.

bool has_attrib(TypeHandle type) const
bool has_attrib(int slot) const

Returns true if there is a render attribute of the indicated type defined on this node, or false if there is not.

bool has_dirty_prev_transform(void) const

Returns true if this node has the _dirty_prev_transform flag set, which indicates its _prev_transform is different from its _transform value (in pipeline stage 0). In this case, the node will be visited by reset_prev_transform().

bool has_effect(TypeHandle type) const

Returns true if there is a render effect of the indicated type defined on this node, or false if there is not.

bool has_python_tag(PyObject *key) const
bool has_tag(std::string const &key, Thread *current_thread = Thread::get_current_thread()) const

Returns true if a value has been defined on this node for the particular key (even if that value is the empty string), or false if no value has been set.

bool has_tags(void) const

Returns true if the node has any tags (or any Python tags) at all, false if it has none.

virtual bool is_ambient_light(void) const

Returns true if this is an AmbientLight, false if it is not a light, or it is some other kind of light.

bool is_bounds_stale(void) const

Returns true if the bounding volume of this node is stale and will be implicitly recomputed at the next call to get_bounds(), or false if it is fresh and need not be recomputed.

virtual bool is_collision_node(void) const

A simple downcast check. Returns true if this kind of node happens to inherit from CollisionNode, false otherwise.

This is provided as a a faster alternative to calling is_of_type(CollisionNode::get_class_type()).

bool is_final(Thread *current_thread = Thread::get_current_thread()) const

Returns the current state of the “final” flag. Initially, this flag is off (false), but it may be changed by an explicit call to set_final(). See set_final().

virtual bool is_geom_node(void) const

A simple downcast check. Returns true if this kind of node happens to inherit from GeomNode, false otherwise.

This is provided as a a faster alternative to calling is_of_type(GeomNode::get_class_type()), since this test is so important to rendering.

virtual bool is_lod_node(void) const

A simple downcast check. Returns true if this kind of node happens to inherit from LODNode, false otherwise.

This is provided as a a faster alternative to calling is_of_type(LODNode::get_class_type()).

bool is_overall_hidden(void) const

Returns true if the node has been hidden to all cameras by clearing its overall bit.

bool is_scene_root(void) const

Returns true if this particular node is known to be the render root of some active DisplayRegion associated with the global GraphicsEngine, false otherwise.

bool is_under_scene_root(void) const

Returns true if this particular node is in a live scene graph: that is, it is a child or descendent of a node that is itself a scene root. If this is true, this node may potentially be traversed by the render traverser. Stashed nodes don’t count for this purpose, but hidden nodes do.

void list_tags(std::ostream &out, std::string const &separator = "\n") const

Writes a list of all the tag keys assigned to the node to the indicated stream. Writes one instance of the separator following each key (but does not write a terminal separator). The value associated with each key is not written.

This is mainly for the benefit of the realtime user, to see the list of all of the associated tag keys.

void ls(std::ostream &out, int indent_level) const

Lists all the nodes at and below the current path hierarchically.

virtual PandaNode *make_copy(void) const

Returns a newly-allocated PandaNode that is a shallow copy of this one. It will be a different pointer, but its internal data may or may not be shared with that of the original PandaNode. No children will be copied.

void mark_bounds_stale(Thread *current_thread = Thread::get_current_thread()) const

Indicates that the bounding volume, or something that influences the bounding volume (or any of the other things stored in CData, like net_collide_mask), may have changed for this node, and that it must be recomputed.

Indicates that the bounding volume, or something that influences the bounding volume (or any of the other things stored in CData, like net_collide_mask), may have changed for this node, and that it must be recomputed.

With no parameters, this means to iterate through all stages including and upstream of the current pipeline stage.

This method is intended for internal use; usually it is not necessary for a user to call this directly. It will be called automatically by derived classes when appropriate.

void mark_internal_bounds_stale(Thread *current_thread = Thread::get_current_thread())

Should be called by a derived class to mark the internal bounding volume stale, so that recompute_internal_bounds() will be called when the bounding volume is next requested.

Should be called by a derived class to mark the internal bounding volume stale, so that compute_internal_bounds() will be called when the bounding volume is next requested.

With no parameters, this means to iterate through all stages including and upstream of the current pipeline stage.

It is normally not necessary to call this method directly; each node should be responsible for calling it when its internals have changed.

virtual void output(std::ostream &out) const
void prepare_scene(GraphicsStateGuardianBase *gsg, RenderState const *node_state)

Walks through the scene graph beginning at this node, and does whatever initialization is required to render the scene properly with the indicated GSG. It is not strictly necessary to call this, since the GSG will initialize itself when the scene is rendered, but this may take some of the overhead away from that process.

In particular, this will ensure that textures and vertex buffers within the scene are loaded into graphics memory.

void remove_all_children(Thread *current_thread = Thread::get_current_thread())

Removes all the children from the node at once, including stashed children.

This can only be called from the top pipeline stage (i.e. from App).

void remove_child(int child_index, Thread *current_thread = Thread::get_current_thread())
bool remove_child(PandaNode *child_node, Thread *current_thread = Thread::get_current_thread())

Removes the nth child from the node.

Removes the indicated child from the node. Returns true if the child was removed, false if it was not already a child of the node. This will also successfully remove the child if it had been stashed.

void remove_stashed(int child_index, Thread *current_thread = Thread::get_current_thread())

Removes the nth stashed child from the node.

bool replace_child(PandaNode *orig_child, PandaNode *new_child, Thread *current_thread = Thread::get_current_thread())

Searches for the orig_child node in the node’s list of children, and replaces it with the new_child instead. Returns true if the replacement is made, or false if the node is not a child or if there is some other problem.

void replace_node(PandaNode *other)

Inserts this node into the scene graph in place of the other one, and removes the other node. All scene graph attributes (TransformState, RenderState, etc.) are copied to this node.

All children are moved to this node, and removed from the old node. The new node is left in the same place in the old node’s parent’s list of children.

Even NodePaths that reference the old node are updated in-place to reference the new node instead.

This method is intended to be used to replace a node of a given type in the scene graph with a node of a different type.

static void reset_all_prev_transform(Thread *current_thread = Thread::get_current_thread())

Visits all nodes in the world with the _dirty_prev_transform flag–that is, all nodes whose _prev_transform is different from the _transform in pipeline stage 0–and resets the _prev_transform to be the same as _transform.

void reset_prev_transform(Thread *current_thread = Thread::get_current_thread())

Resets the transform that represents this node’s “previous” position to the same as the current transform. This is not the same thing as clearing it to identity.

void set_attrib(RenderAttrib const *attrib, int override = 0)

Adds the indicated render attribute to the scene graph on this node. This attribute will now apply to this node and everything below. If there was already an attribute of the same type, it is replaced.

void set_bound(BoundingVolume const *volume)

Deprecated. Use set_bounds() instead.

void set_bounds(BoundingVolume const *volume)

Resets the bounding volume so that it is the indicated volume. When it is explicitly set, the bounding volume will no longer be automatically computed according to the contents of the node itself, for nodes like GeomNodes and TextNodes that contain substance (but the bounding volume will still be automatically expanded to include its children).

Call clear_bounds() if you would like to return the bounding volume to its default behavior later.

void set_bounds_type(BoundingVolume::BoundsType bounds_type)

We define set_bounds() and get_bounds() functions so that set_bounds() sets the user bounding volume, while get_bounds() returns the external bounding volume. Although it might seem strange and confusing to do this, this is actually the natural way the user thinks about nodes and bounding volumes.

Specifies the desired type of bounding volume that will be created for this node. This is normally BoundingVolume::BT_default, which means to set the type according to the config variable “bounds-type”.

If this is BT_sphere or BT_box, a BoundingSphere or BoundingBox is explicitly created. If it is BT_best, the appropriate type to best enclose the node’s children is created.

This affects the bounding volume returned by get_bounds(), which is not exactly the same bounding volume modified by set_bounds(), because a new bounding volume has to be created that includes this node and all of its children.

void set_effect(RenderEffect const *effect)

Adds the indicated render effect to the scene graph on this node. If there was already an effect of the same type, it is replaced.

void set_effects(RenderEffects const *effects, Thread *current_thread = Thread::get_current_thread())

Sets the complete RenderEffects that will be applied this node. This completely replaces whatever has been set on this node via repeated calls to set_attrib().

void set_final(bool flag)

Sets the “final” flag on this PandaNode. If this is true, than no bounding volume need be tested below it; a positive intersection with this node’s bounding volume is deemed to be a positive intersection with all geometry inside.

This is useful to quickly force a larger bounding volume around a node when the GeomNodes themselves are inaccurate for some reason, without forcing a recompute of every nested bounding volume. It’s also helpful when the bounding volume is tricked by some special properties, like billboards, that may move geometry out of its bounding volume otherwise.

void set_into_collide_mask(CollideMask mask)

Sets the “into” CollideMask.

This specifies the set of bits that must be shared with a CollisionNode’s “from” CollideMask in order for the CollisionNode to detect a collision with this particular node.

The actual CollideMask that will be set is masked by the return value from get_legal_collide_mask(). Thus, the into_collide_mask cannot be set to anything other than nonzero except for those types of nodes that can be collided into, such as CollisionNodes and GeomNodes.

void set_overall_hidden(bool overall_hidden)

Sets or clears the hidden flag. When the hidden flag is true, the node and all of its children are invisible to all cameras, regardless of the setting of any draw masks. Setting the hidden flag to false restores the previous visibility as established by the draw masks.

This actually works by twiddling the reserved _overall_bit in the node’s draw mask, which has special meaning.

void set_prev_transform(TransformState const *transform, Thread *current_thread = Thread::get_current_thread())

Sets the transform that represents this node’s “previous” position, one frame ago, for the purposes of detecting motion for accurate collision calculations.

void set_python_tag(PyObject *key, PyObject *value)
void set_state(RenderState const *state, Thread *current_thread = Thread::get_current_thread())

Sets the complete RenderState that will be applied to all nodes at this level and below. (The actual state that will be applied to lower nodes is based on the composition of RenderStates from above this node as well). This completely replaces whatever has been set on this node via repeated calls to set_attrib().

void set_tag(std::string const &key, std::string const &value, Thread *current_thread = Thread::get_current_thread())

Associates a user-defined value with a user-defined key which is stored on the node. This value has no meaning to Panda; but it is stored indefinitely on the node until it is requested again.

Each unique key stores a different string value. There is no effective limit on the number of different keys that may be stored or on the length of any one key’s value.

void set_transform(TransformState const *transform, Thread *current_thread = Thread::get_current_thread())

Sets the transform that will be applied to this node and below. This defines a new coordinate space at this point in the scene graph and below.

void set_unexpected_change(unsigned int flags)

Sets one or more of the PandaNode::UnexpectedChange bits on, indicating that the corresponding property should not change again on this node. Once one of these bits has been set, if the property changes, an assertion failure will be raised, which is designed to assist the developer in identifying the troublesome code that modified the property unexpectedly.

The input parameter is the union of bits that are to be set. To clear these bits later, use clear_unexpected_change().

Since this is a developer debugging tool only, this function does nothing in a production (NDEBUG) build.

bool stash_child(PandaNode *child_node, Thread *current_thread = Thread::get_current_thread())
void stash_child(int child_index, Thread *current_thread = Thread::get_current_thread())

Stashes the indicated child node. This removes the child from the list of active children and puts it on a special list of stashed children. This child node no longer contributes to the bounding volume of the PandaNode, and is not visited in normal traversals. It is invisible and uncollidable. The child may later be restored by calling unstash_child().

This function returns true if the child node was successfully stashed, or false if it was not a child of the node in the first place (e.g. it was previously stashed).

Stashes the indicated child node. This removes the child from the list of active children and puts it on a special list of stashed children. This child node no longer contributes to the bounding volume of the PandaNode, and is not visited in normal traversals. It is invisible and uncollidable. The child may later be restored by calling unstash_child().

This can only be called from the top pipeline stage (i.e. from App).

void steal_children(PandaNode *other, Thread *current_thread = Thread::get_current_thread())

Moves all the children from the other node onto this node.

Any NodePaths to child nodes of the other node are truncated, rather than moved to the new parent.

bool unstash_child(PandaNode *child_node, Thread *current_thread = Thread::get_current_thread())
void unstash_child(int stashed_index, Thread *current_thread = Thread::get_current_thread())

Returns the indicated stashed node to normal child status. This removes the child from the list of stashed children and puts it on the normal list of active children. This child node once again contributes to the bounding volume of the PandaNode, and will be visited in normal traversals. It is visible and collidable.

This function returns true if the child node was successfully stashed, or false if it was not a child of the node in the first place (e.g. it was previously stashed).

Returns the indicated stashed node to normal child status. This removes the child from the list of stashed children and puts it on the normal list of active children. This child node once again contributes to the bounding volume of the PandaNode, and will be visited in normal traversals. It is visible and collidable.

This can only be called from the top pipeline stage (i.e. from App).

virtual void write(std::ostream &out, int indent_level) const