blender::fn::lazy_function::NodeState Struct Reference

Public Attributes
InputState *	inputs
OutputState *	outputs
int	missing_required_inputs = 0
Mutex	mutex
bool	node_has_finished = false
bool	always_used_inputs_requested = false
bool	storage_and_defaults_initialized = false
bool	has_side_effects = false
bool	enabled_multi_threading = false
NodeScheduleState	schedule_state = NodeScheduleState::NotScheduled
void *	storage = nullptr

Detailed Description

Definition at line 132 of file lazy_function_graph_executor.cc.

Member Data Documentation

always_used_inputs_requested

bool blender::fn::lazy_function::NodeState::always_used_inputs_requested = false

Set to true once the always required inputs have been requested. This happens the first time the node is run.

Definition at line 161 of file lazy_function_graph_executor.cc.

enabled_multi_threading

bool blender::fn::lazy_function::NodeState::enabled_multi_threading = false

Whether this node has enabled multi-threading. If this is true, the node is allowed to call methods on Params from multiple threads.

Definition at line 176 of file lazy_function_graph_executor.cc.

has_side_effects

bool blender::fn::lazy_function::NodeState::has_side_effects = false

Nodes with side effects should always be executed when their required inputs have been computed.

Definition at line 171 of file lazy_function_graph_executor.cc.

Referenced by blender::fn::lazy_function::Executor::execute().

inputs

InputState* blender::fn::lazy_function::NodeState::inputs

States of the individual input and output sockets. One can index into these arrays without locking. However, to access data inside, a lock is needed unless noted otherwise. Those are not stored as Span to reduce memory usage. The number of inputs and outputs is stored on the node already.

Definition at line 139 of file lazy_function_graph_executor.cc.

missing_required_inputs

int blender::fn::lazy_function::NodeState::missing_required_inputs = 0

Counts the number of inputs that still have to be provided to this node, until it should run again. This is used as an optimization so that nodes are not scheduled unnecessarily in many cases.

Definition at line 146 of file lazy_function_graph_executor.cc.

mutex

Mutex blender::fn::lazy_function::NodeState::mutex

mutable

Needs to be locked when any data in this state is accessed that is not explicitly marked as not needing the lock.

Definition at line 151 of file lazy_function_graph_executor.cc.

node_has_finished

bool blender::fn::lazy_function::NodeState::node_has_finished = false

Is set to true once the node is done with its work, i.e. when all outputs that may be used have been computed.

Definition at line 156 of file lazy_function_graph_executor.cc.

outputs

OutputState* blender::fn::lazy_function::NodeState::outputs

Definition at line 140 of file lazy_function_graph_executor.cc.

schedule_state

NodeScheduleState blender::fn::lazy_function::NodeState::schedule_state = NodeScheduleState::NotScheduled

A node is always in one specific schedule state. This helps to ensure that the same node does not run twice at the same time accidentally.

Definition at line 181 of file lazy_function_graph_executor.cc.

storage

void* blender::fn::lazy_function::NodeState::storage = nullptr

Custom storage of the node.

Definition at line 185 of file lazy_function_graph_executor.cc.

storage_and_defaults_initialized

bool blender::fn::lazy_function::NodeState::storage_and_defaults_initialized = false

Set to true when the storage and defaults have been initialized. This happens the first time the node function is executed.

Definition at line 166 of file lazy_function_graph_executor.cc.

---

The documentation for this struct was generated from the following file:

• lazy_function_graph_executor.cc