archive/blender-archive
Archived
1
1
Fork 0
Code Pull Requests Packages Activity
This repository has been archived on 2023-10-09. You can view files and clone it, but cannot push or open issues or pull requests.
Files
197e889104697603f8da7793e8e55b6eebc45799
blender-archive/source/blender/compositor/intern/COM_NodeGraph.cpp

300 lines
9.5 KiB
C++
Raw Normal View History

Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
/*
* Copyright 2013, Blender Foundation.
*
* This program is free software; you can redistribute it and/or
* 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.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
* Contributor:
* Lukas Toenne
*/
#include <cstring>
extern "C" {
#include "BLI_listbase.h"
#include "BLI_utildefines.h"
#include "DNA_node_types.h"
#include "BKE_node.h"
}
#include "COM_CompositorContext.h"
#include "COM_Converter.h"
#include "COM_Debug.h"
#include "COM_Node.h"
#include "COM_SocketProxyNode.h"
#include "COM_NodeGraph.h" /* own include */
/*******************
**** NodeGraph ****
*******************/
NodeGraph::NodeGraph()
{
}
NodeGraph::~NodeGraph()
{
for (int index = 0; index < this->m_nodes.size(); index++) {
Node *node = this->m_nodes[index];
delete node;
}
}
void NodeGraph::from_bNodeTree(const CompositorContext &context, bNodeTree *tree)
{
add_bNodeTree(context, 0, tree, NODE_INSTANCE_KEY_BASE);
}
bNodeSocket *NodeGraph::find_b_node_input(bNode *b_group_node, const char *identifier)
{
for (bNodeSocket *b_sock = (bNodeSocket *)b_group_node->inputs.first; b_sock; b_sock = b_sock->next) {
if (STREQ(b_sock->identifier, identifier))
return b_sock;
}
return NULL;
}
bNodeSocket *NodeGraph::find_b_node_output(bNode *b_group_node, const char *identifier)
{
for (bNodeSocket *b_sock = (bNodeSocket *)b_group_node->outputs.first; b_sock; b_sock = b_sock->next) {
if (STREQ(b_sock->identifier, identifier))
return b_sock;
}
return NULL;
}
void NodeGraph::add_node(Node *node, bNodeTree *b_ntree, bNodeInstanceKey key, bool is_active_group)
{
node->setbNodeTree(b_ntree);
node->setInstanceKey(key);
node->setIsInActiveGroup(is_active_group);
m_nodes.push_back(node);
DebugInfo::node_added(node);
}
void NodeGraph::add_link(NodeOutput *fromSocket, NodeInput *toSocket)
{
m_links.push_back(Link(fromSocket, toSocket));
/* register with the input */
toSocket->setLink(fromSocket);
}
void NodeGraph::add_bNodeTree(const CompositorContext &context, int nodes_start, bNodeTree *tree, bNodeInstanceKey parent_key)
{
const bNodeTree *basetree = context.getbNodeTree();
/* update viewers in the active edittree as well the base tree (for backdrop) */
Fix T40986: crash on using the viewer node inside of group nodes. Viewers were activated both inside the active group as well as the top level tree (the latter being a quick fix for getting a fallback viewer). This caused a race condition on the shared viewer image. Now the active viewer is defined at node conversion time in the converter so that only one can be active at a time without each node having to follow complicated rules for exclusion.
2014-07-08 12:48:41 +02:00
bool is_active_group = (parent_key.value == basetree->active_viewer_key.value);
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
/* add all nodes of the tree to the node list */
for (bNode *node = (bNode *)tree->nodes.first; node; node = node->next) {
bNodeInstanceKey key = BKE_node_instance_key(parent_key, tree, node);
add_bNode(context, tree, node, key, is_active_group);
}
NodeRange node_range(m_nodes.begin() + nodes_start, m_nodes.end());
/* add all nodelinks of the tree to the link list */
for (bNodeLink *nodelink = (bNodeLink *)tree->links.first; nodelink; nodelink = nodelink->next) {
add_bNodeLink(node_range, nodelink);
}
}
void NodeGraph::add_bNode(const CompositorContext &context, bNodeTree *b_ntree, bNode *b_node, bNodeInstanceKey key, bool is_active_group)
{
/* replace muted nodes by proxies for internal links */
if (b_node->flag & NODE_MUTED) {
add_proxies_mute(b_ntree, b_node, key, is_active_group);
return;
}
/* replace slow nodes with proxies for fast execution */
if (context.isFastCalculation() && !Converter::is_fast_node(b_node)) {
add_proxies_skip(b_ntree, b_node, key, is_active_group);
return;
}
/* special node types */
if (b_node->type == NODE_GROUP) {
add_proxies_group(context, b_node, key);
}
Fix T39848: Reroute nodes were not handled properly after compositor cleanup patch. These simply need a proxy node, which will be replaced by direct operation links afterward.
2014-04-23 17:00:04 +02:00
else if (b_node->type == NODE_REROUTE) {
add_proxies_reroute(b_ntree, b_node, key, is_active_group);
}
else {
/* regular nodes, handled in Converter */
Node *node = Converter::convert(b_node);
if (node)
add_node(node, b_ntree, key, is_active_group);
}
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
}
Fix T40414: Multiple input nodes in a group not working. A node group can have multiple input nodes. In the compositor that means each of the input sockets has to be connected to the linked outputs, which is represented by a single link on the outside of the group.
2014-05-29 10:00:21 +02:00
NodeGraph::NodeInputs NodeGraph::find_inputs(const NodeRange &node_range, bNodeSocket *b_socket)
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
{
Fix T40414: Multiple input nodes in a group not working. A node group can have multiple input nodes. In the compositor that means each of the input sockets has to be connected to the linked outputs, which is represented by a single link on the outside of the group.
2014-05-29 10:00:21 +02:00
NodeInputs result;
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
for (NodeGraph::NodeIterator it = node_range.first; it != node_range.second; ++it) {
Node *node = *it;
for (int index = 0; index < node->getNumberOfInputSockets(); index++) {
NodeInput *input = node->getInputSocket(index);
Fix T40414: Multiple input nodes in a group not working. A node group can have multiple input nodes. In the compositor that means each of the input sockets has to be connected to the linked outputs, which is represented by a single link on the outside of the group.
2014-05-29 10:00:21 +02:00
if (input->getbNodeSocket() == b_socket) {
result.push_back(input);
}
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
}
}
Fix T40414: Multiple input nodes in a group not working. A node group can have multiple input nodes. In the compositor that means each of the input sockets has to be connected to the linked outputs, which is represented by a single link on the outside of the group.
2014-05-29 10:00:21 +02:00
return result;
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
}
NodeOutput *NodeGraph::find_output(const NodeRange &node_range, bNodeSocket *b_socket)
{
for (NodeGraph::NodeIterator it = node_range.first; it != node_range.second; ++it) {
Node *node = *it;
for (int index = 0; index < node->getNumberOfOutputSockets(); index++) {
NodeOutput *output = node->getOutputSocket(index);
Fix T40414: Multiple input nodes in a group not working. A node group can have multiple input nodes. In the compositor that means each of the input sockets has to be connected to the linked outputs, which is represented by a single link on the outside of the group.
2014-05-29 10:00:21 +02:00
if (output->getbNodeSocket() == b_socket) {
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
return output;
Fix T40414: Multiple input nodes in a group not working. A node group can have multiple input nodes. In the compositor that means each of the input sockets has to be connected to the linked outputs, which is represented by a single link on the outside of the group.
2014-05-29 10:00:21 +02:00
}
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
}
}
return NULL;
}
void NodeGraph::add_bNodeLink(const NodeRange &node_range, bNodeLink *b_nodelink)
{
/// @note: ignore invalid links
if (!(b_nodelink->flag & NODE_LINK_VALID))
return;
Fix T40414: Multiple input nodes in a group not working. A node group can have multiple input nodes. In the compositor that means each of the input sockets has to be connected to the linked outputs, which is represented by a single link on the outside of the group.
2014-05-29 10:00:21 +02:00
/* Note: a DNA input socket can have multiple NodeInput in the compositor tree! (proxies)
* The output then gets linked to each one of them.
*/
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
NodeOutput *output = find_output(node_range, b_nodelink->fromsock);
Fix T40414: Multiple input nodes in a group not working. A node group can have multiple input nodes. In the compositor that means each of the input sockets has to be connected to the linked outputs, which is represented by a single link on the outside of the group.
2014-05-29 10:00:21 +02:00
if (!output)
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
return;
Fix T40414: Multiple input nodes in a group not working. A node group can have multiple input nodes. In the compositor that means each of the input sockets has to be connected to the linked outputs, which is represented by a single link on the outside of the group.
2014-05-29 10:00:21 +02:00
NodeInputs inputs = find_inputs(node_range, b_nodelink->tosock);
for (NodeInputs::const_iterator it = inputs.begin(); it != inputs.end(); ++it) {
NodeInput *input = *it;
if (input->isLinked())
continue;
add_link(output, input);
}
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
}
/* **** Special proxy node type conversions **** */
void NodeGraph::add_proxies_mute(bNodeTree *b_ntree, bNode *b_node, bNodeInstanceKey key, bool is_active_group)
{
for (bNodeLink *b_link = (bNodeLink *)b_node->internal_links.first; b_link; b_link = b_link->next) {
Fix T41067: Muted nodes still do data type conversions. Proxy operations from muted nodes would still create conversion operations where the datatypes don't match, which creates unexpected behavior. Arguably datatype conversion could still happen even when the main operation is muted, but this would be a design change and so is disabled now.
2014-07-15 10:55:49 +02:00
SocketProxyNode *proxy = new SocketProxyNode(b_node, b_link->fromsock, b_link->tosock, false);
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
add_node(proxy, b_ntree, key, is_active_group);
}
}
void NodeGraph::add_proxies_skip(bNodeTree *b_ntree, bNode *b_node, bNodeInstanceKey key, bool is_active_group)
{
for (bNodeSocket *output = (bNodeSocket *)b_node->outputs.first; output; output = output->next) {
bNodeSocket *input;
/* look for first input with matching datatype for each output */
for (input = (bNodeSocket *)b_node->inputs.first; input; input = input->next) {
if (input->type == output->type)
break;
}
if (input) {
Fix T41067: Muted nodes still do data type conversions. Proxy operations from muted nodes would still create conversion operations where the datatypes don't match, which creates unexpected behavior. Arguably datatype conversion could still happen even when the main operation is muted, but this would be a design change and so is disabled now.
2014-07-15 10:55:49 +02:00
SocketProxyNode *proxy = new SocketProxyNode(b_node, input, output, true);
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
add_node(proxy, b_ntree, key, is_active_group);
}
}
}
void NodeGraph::add_proxies_group_inputs(bNode *b_node, bNode *b_node_io)
{
bNodeTree *b_group_tree = (bNodeTree *)b_node->id;
BLI_assert(b_group_tree); /* should have been checked in advance */
/* not important for proxies */
bNodeInstanceKey key = NODE_INSTANCE_KEY_BASE;
bool is_active_group = false;
for (bNodeSocket *b_sock_io = (bNodeSocket *)b_node_io->outputs.first; b_sock_io; b_sock_io = b_sock_io->next) {
bNodeSocket *b_sock_group = find_b_node_input(b_node, b_sock_io->identifier);
if (b_sock_group) {
Fix T41067: Muted nodes still do data type conversions. Proxy operations from muted nodes would still create conversion operations where the datatypes don't match, which creates unexpected behavior. Arguably datatype conversion could still happen even when the main operation is muted, but this would be a design change and so is disabled now.
2014-07-15 10:55:49 +02:00
SocketProxyNode *proxy = new SocketProxyNode(b_node_io, b_sock_group, b_sock_io, true);
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
add_node(proxy, b_group_tree, key, is_active_group);
}
}
}
void NodeGraph::add_proxies_group_outputs(bNode *b_node, bNode *b_node_io, bool use_buffer)
{
bNodeTree *b_group_tree = (bNodeTree *)b_node->id;
BLI_assert(b_group_tree); /* should have been checked in advance */
/* not important for proxies */
bNodeInstanceKey key = NODE_INSTANCE_KEY_BASE;
bool is_active_group = false;
for (bNodeSocket *b_sock_io = (bNodeSocket *)b_node_io->inputs.first; b_sock_io; b_sock_io = b_sock_io->next) {
bNodeSocket *b_sock_group = find_b_node_output(b_node, b_sock_io->identifier);
if (b_sock_group) {
if (use_buffer) {
SocketBufferNode *buffer = new SocketBufferNode(b_node_io, b_sock_io, b_sock_group);
add_node(buffer, b_group_tree, key, is_active_group);
}
else {
Fix T41067: Muted nodes still do data type conversions. Proxy operations from muted nodes would still create conversion operations where the datatypes don't match, which creates unexpected behavior. Arguably datatype conversion could still happen even when the main operation is muted, but this would be a design change and so is disabled now.
2014-07-15 10:55:49 +02:00
SocketProxyNode *proxy = new SocketProxyNode(b_node_io, b_sock_io, b_sock_group, true);
Structural cleanup and improvements for the compositor. Many parts of the compositor are unnecessarily complicated. This patch aims at reducing the complexity of writing nodes and making the code more transparent. == Separating Nodes and Operations == Currently these are both mixed in the same graph, even though they have very different purposes and are used at distinct stages in the compositing process. The patch introduces dedicated graph classes for nodes and for operations. This removes the need for a lot of special case checks (isOperation etc.) and explicit type casts. It simplifies the code since it becomes clear at every stage what type of node we are dealing with. The compiler can use static typing to avoid common bugs from mixing up these types and fewer runtime sanity checks are needed. == Simplified Node Conversion == Converting nodes to operations was previously based on "relinking", i.e. nodes would start with by mirroring links in the Blender DNA node trees, then add operations and redirect these links to them. This was very hard to follow in many cases and required a lot of attention to avoid invalid states. Now there is a helper class called the NodeConverter, which is passed to nodes and implements a much simpler API for this process. Nodes can add operations and explicit connections as before, but defining "external" links to the inputs/outputs of the original node now uses mapping instead of directly modifying link data. Input data (node graph) and result (operations graph) are cleanly separated. == Removed Redundant Data Structures == A few redundant data structures have been removed, notably the SocketConnection. These are only needed temporarily during graph construction. For executing the compositor operations it is perfectly sufficient to store only the direct input link pointers. A common pointer indirection is avoided this way (which might also give a little performance improvement). == Avoid virtual recursive functions == Recursive virtual functions are evil. They are very hard to follow during debugging. At least in the parts this patch is concerned with these functions have been replaced by a non-virtual recursive core function (which might then call virtual non-recursive functions if needed). See for example NodeOperationBuilder::group_operations.
2014-04-15 16:06:12 +02:00
add_node(proxy, b_group_tree, key, is_active_group);
}
}
}
}
void NodeGraph::add_proxies_group(const CompositorContext &context, bNode *b_node, bNodeInstanceKey key)
{
bNodeTree *b_group_tree = (bNodeTree *)b_node->id;
/* missing node group datablock can happen with library linking */
if (!b_group_tree) {
/* this error case its handled in convertToOperations() so we don't get un-convertred sockets */
return;
}
/* use node list size before adding proxies, so they can be connected in add_bNodeTree */
int nodes_start = m_nodes.size();
/* create proxy nodes for group input/output nodes */
for (bNode *b_node_io = (bNode *)b_group_tree->nodes.first; b_node_io; b_node_io = b_node_io->next) {
if (b_node_io->type == NODE_GROUP_INPUT)
add_proxies_group_inputs(b_node, b_node_io);
if (b_node_io->type == NODE_GROUP_OUTPUT && (b_node_io->flag & NODE_DO_OUTPUT))
add_proxies_group_outputs(b_node, b_node_io, context.isGroupnodeBufferEnabled());
}
add_bNodeTree(context, nodes_start, b_group_tree, key);
}
Fix T39848: Reroute nodes were not handled properly after compositor cleanup patch. These simply need a proxy node, which will be replaced by direct operation links afterward.
2014-04-23 17:00:04 +02:00
void NodeGraph::add_proxies_reroute(bNodeTree *b_ntree, bNode *b_node, bNodeInstanceKey key, bool is_active_group)
{
Fix T41067: Muted nodes still do data type conversions. Proxy operations from muted nodes would still create conversion operations where the datatypes don't match, which creates unexpected behavior. Arguably datatype conversion could still happen even when the main operation is muted, but this would be a design change and so is disabled now.
2014-07-15 10:55:49 +02:00
SocketProxyNode *proxy = new SocketProxyNode(b_node, (bNodeSocket *)b_node->inputs.first, (bNodeSocket *)b_node->outputs.first, false);
Fix T39848: Reroute nodes were not handled properly after compositor cleanup patch. These simply need a proxy node, which will be replaced by direct operation links afterward.
2014-04-23 17:00:04 +02:00
add_node(proxy, b_ntree, key, is_active_group);
}
Copy Permalink