This repository has been archived on 2023-10-09.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
blender-archive/source/blender/modifiers/intern/MOD_nodes.cc

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

1613 lines
60 KiB
C++
Raw Normal View History

Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/*
* 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.
*
* The Original Code is Copyright (C) 2005 by the Blender Foundation.
* All rights reserved.
*/
/** \file
* \ingroup modifiers
*/
#include <cstring>
#include <iostream>
#include <string>
#include "MEM_guardedalloc.h"
#include "BLI_float3.hh"
#include "BLI_listbase.h"
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
#include "BLI_multi_value_map.hh"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
#include "BLI_set.hh"
#include "BLI_string.h"
#include "BLI_utildefines.h"
#include "DNA_collection_types.h"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
#include "DNA_defaults.h"
#include "DNA_mesh_types.h"
#include "DNA_meshdata_types.h"
#include "DNA_modifier_types.h"
#include "DNA_node_types.h"
#include "DNA_object_types.h"
#include "DNA_pointcloud_types.h"
#include "DNA_scene_types.h"
#include "DNA_screen_types.h"
#include "DNA_space_types.h"
#include "DNA_windowmanager_types.h"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
#include "BKE_customdata.h"
#include "BKE_geometry_set_instances.hh"
#include "BKE_global.h"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
#include "BKE_idprop.h"
#include "BKE_lib_query.h"
#include "BKE_main.h"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
#include "BKE_mesh.h"
#include "BKE_modifier.h"
#include "BKE_node_ui_storage.hh"
#include "BKE_object.h"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
#include "BKE_pointcloud.h"
#include "BKE_screen.h"
#include "BKE_simulation.h"
#include "BKE_workspace.h"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
#include "BLO_read_write.h"
#include "UI_interface.h"
#include "UI_resources.h"
#include "RNA_access.h"
#include "RNA_enum_types.h"
#include "DEG_depsgraph_build.h"
#include "DEG_depsgraph_query.h"
#include "MOD_modifiertypes.h"
#include "MOD_nodes.h"
#include "MOD_ui_common.h"
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
#include "ED_spreadsheet.h"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
#include "NOD_derived_node_tree.hh"
#include "NOD_geometry.h"
#include "NOD_geometry_exec.hh"
#include "NOD_node_tree_multi_function.hh"
#include "NOD_type_callbacks.hh"
#include "NOD_type_conversions.hh"
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
using blender::float3;
using blender::FunctionRef;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
using blender::IndexRange;
using blender::Map;
using blender::Set;
using blender::Span;
using blender::StringRef;
using blender::StringRefNull;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
using blender::Vector;
using blender::bke::PersistentCollectionHandle;
using blender::bke::PersistentDataHandleMap;
using blender::bke::PersistentObjectHandle;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
using blender::fn::GMutablePointer;
using blender::fn::GPointer;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
using blender::fn::GValueMap;
using blender::nodes::GeoNodeExecParams;
using namespace blender::fn::multi_function_types;
using namespace blender::nodes::derived_node_tree_types;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
static void initData(ModifierData *md)
{
NodesModifierData *nmd = (NodesModifierData *)md;
BLI_assert(MEMCMP_STRUCT_AFTER_IS_ZERO(nmd, modifier));
MEMCPY_STRUCT_AFTER(nmd, DNA_struct_default_get(NodesModifierData), modifier);
}
static void addIdsUsedBySocket(const ListBase *sockets, Set<ID *> &ids)
{
LISTBASE_FOREACH (const bNodeSocket *, socket, sockets) {
if (socket->type == SOCK_OBJECT) {
Object *object = ((bNodeSocketValueObject *)socket->default_value)->value;
if (object != nullptr) {
ids.add(&object->id);
}
}
else if (socket->type == SOCK_COLLECTION) {
Collection *collection = ((bNodeSocketValueCollection *)socket->default_value)->value;
if (collection != nullptr) {
ids.add(&collection->id);
}
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
static void find_used_ids_from_nodes(const bNodeTree &tree, Set<ID *> &ids)
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
Set<const bNodeTree *> handled_groups;
LISTBASE_FOREACH (const bNode *, node, &tree.nodes) {
addIdsUsedBySocket(&node->inputs, ids);
addIdsUsedBySocket(&node->outputs, ids);
if (ELEM(node->type, NODE_GROUP, NODE_CUSTOM_GROUP)) {
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const bNodeTree *group = (bNodeTree *)node->id;
if (group != nullptr && handled_groups.add(group)) {
find_used_ids_from_nodes(*group, ids);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
}
}
static void find_used_ids_from_settings(const NodesModifierSettings &settings, Set<ID *> &ids)
{
IDP_foreach_property(
settings.properties,
IDP_TYPE_FILTER_ID,
[](IDProperty *property, void *user_data) {
Set<ID *> *ids = (Set<ID *> *)user_data;
ID *id = IDP_Id(property);
if (id != nullptr) {
ids->add(id);
}
},
&ids);
}
static void add_collection_object_relations_recursive(const ModifierUpdateDepsgraphContext *ctx,
Collection &collection);
static void add_object_relation(const ModifierUpdateDepsgraphContext *ctx, Object &object)
{
DEG_add_object_relation(ctx->node, &object, DEG_OB_COMP_TRANSFORM, "Nodes Modifier");
if (&(ID &)object != &ctx->object->id) {
if (object.type == OB_EMPTY) {
Collection *collection_instance = object.instance_collection;
if (collection_instance != nullptr) {
add_collection_object_relations_recursive(ctx, *collection_instance);
}
}
else if (ELEM(object.type, OB_MESH, OB_POINTCLOUD, OB_VOLUME)) {
DEG_add_object_relation(ctx->node, &object, DEG_OB_COMP_GEOMETRY, "Nodes Modifier");
/* We don't know exactly what attributes from the other object we will need. */
CustomData_MeshMasks mask;
mask.vmask = CD_MASK_PROP_ALL | CD_MASK_MDEFORMVERT;
mask.pmask = CD_MASK_PROP_ALL;
mask.lmask = CD_MASK_PROP_ALL;
mask.fmask = CD_MASK_PROP_ALL;
mask.emask = CD_MASK_PROP_ALL;
DEG_add_customdata_mask(ctx->node, &object, &mask);
}
}
}
static void add_collection_object_relations_recursive(const ModifierUpdateDepsgraphContext *ctx,
Collection &collection)
{
LISTBASE_FOREACH (CollectionObject *, collection_object, &collection.gobject) {
BLI_assert(collection_object->ob != nullptr);
Object &object = *collection_object->ob;
add_object_relation(ctx, object);
}
LISTBASE_FOREACH (CollectionChild *, collection_child, &collection.children) {
BLI_assert(collection_child->collection != nullptr);
Collection &collection = *collection_child->collection;
add_collection_object_relations_recursive(ctx, collection);
}
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
static void updateDepsgraph(ModifierData *md, const ModifierUpdateDepsgraphContext *ctx)
{
NodesModifierData *nmd = reinterpret_cast<NodesModifierData *>(md);
DEG_add_modifier_to_transform_relation(ctx->node, "Nodes Modifier");
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (nmd->node_group != nullptr) {
DEG_add_node_tree_relation(ctx->node, nmd->node_group, "Nodes Modifier");
Set<ID *> used_ids;
find_used_ids_from_settings(nmd->settings, used_ids);
find_used_ids_from_nodes(*nmd->node_group, used_ids);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
for (ID *id : used_ids) {
if (GS(id->name) == ID_OB) {
Object *object = reinterpret_cast<Object *>(id);
add_object_relation(ctx, *object);
}
if (GS(id->name) == ID_GR) {
Collection *collection = reinterpret_cast<Collection *>(id);
add_collection_object_relations_recursive(ctx, *collection);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
}
/* TODO: Add dependency for adding and removing objects in collections. */
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
static void foreachIDLink(ModifierData *md, Object *ob, IDWalkFunc walk, void *userData)
{
NodesModifierData *nmd = reinterpret_cast<NodesModifierData *>(md);
walk(userData, ob, (ID **)&nmd->node_group, IDWALK_CB_USER);
struct ForeachSettingData {
IDWalkFunc walk;
void *userData;
Object *ob;
} settings = {walk, userData, ob};
IDP_foreach_property(
nmd->settings.properties,
IDP_TYPE_FILTER_ID,
[](IDProperty *id_prop, void *user_data) {
ForeachSettingData *settings = (ForeachSettingData *)user_data;
settings->walk(
settings->userData, settings->ob, (ID **)&id_prop->data.pointer, IDWALK_CB_USER);
},
&settings);
}
Geometry Nodes: show "Show Texture in texture tab" button This enables the quick access button [to show the relevant Texture in the Properties Editor] for textures used in geometry nodes. This goes in line to what we do for other textures: - modifier textures have this button - particle textures have this button - brush textures will soon have it, too (see D9813) When outside of the Properties Editor, the button will always show (if a texture is actually assigned), but will be inactive if no suiting Properties Editor to show the texture in can be found. Note this also changes the behavior to not show the button if _no_ texture is assigned (as in: we are still showing the "New" button). Previously it was always there (e.g. for modifier textures), even if it would take us to an empty texture tab. (Sure, we could add a texture there then, but imho it makes more sense to just start showing it once a texture is already there) For this to work with geometry nodes, the following chages were done: - implement foreachTexLink for geonode modifiers - new buttons_texture_user_node_property_add() that stores prop as well as node - also use NODE_ACTIVE_TEXTURE flag in geometry nodetrees notes: - this still uses the first suiting (as in: pinning does not interfere) Properties Editor it finds, this should (maybe?) find the _closest_ Property Editor instead (see related feedback in D9813). - this will already show the button for brush textures as well (disabled), but there is another mandatory change in an upcomming commit to make it work there as well (see D9813) ref. T85278 Maniphest Tasks: T85278 Differential Revision: https://developer.blender.org/D10293
2021-02-03 14:39:24 +01:00
static void foreachTexLink(ModifierData *md, Object *ob, TexWalkFunc walk, void *userData)
{
walk(userData, ob, md, "texture");
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
static bool isDisabled(const struct Scene *UNUSED(scene),
ModifierData *md,
bool UNUSED(useRenderParams))
{
NodesModifierData *nmd = reinterpret_cast<NodesModifierData *>(md);
if (nmd->node_group == nullptr) {
return true;
}
return false;
}
class GeometryNodesEvaluator {
public:
using LogSocketValueFn = std::function<void(DSocket, Span<GPointer>)>;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
private:
blender::LinearAllocator<> allocator_;
Map<std::pair<DInputSocket, DOutputSocket>, GMutablePointer> value_by_input_;
Vector<DInputSocket> group_outputs_;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
blender::nodes::MultiFunctionByNode &mf_by_node_;
const blender::nodes::DataTypeConversions &conversions_;
const PersistentDataHandleMap &handle_map_;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const Object *self_object_;
const ModifierData *modifier_;
Depsgraph *depsgraph_;
LogSocketValueFn log_socket_value_fn_;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
public:
GeometryNodesEvaluator(const Map<DOutputSocket, GMutablePointer> &group_input_data,
Vector<DInputSocket> group_outputs,
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
blender::nodes::MultiFunctionByNode &mf_by_node,
const PersistentDataHandleMap &handle_map,
const Object *self_object,
const ModifierData *modifier,
Depsgraph *depsgraph,
LogSocketValueFn log_socket_value_fn)
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
: group_outputs_(std::move(group_outputs)),
mf_by_node_(mf_by_node),
conversions_(blender::nodes::get_implicit_type_conversions()),
handle_map_(handle_map),
self_object_(self_object),
modifier_(modifier),
depsgraph_(depsgraph),
log_socket_value_fn_(std::move(log_socket_value_fn))
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
for (auto item : group_input_data.items()) {
this->log_socket_value(item.key, item.value);
this->forward_to_inputs(item.key, item.value);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
Vector<GMutablePointer> execute()
{
Vector<GMutablePointer> results;
for (const DInputSocket &group_output : group_outputs_) {
Vector<GMutablePointer> result = this->get_input_values(group_output);
this->log_socket_value(group_output, result);
results.append(result[0]);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
for (GMutablePointer value : value_by_input_.values()) {
value.destruct();
}
return results;
}
private:
Vector<GMutablePointer> get_input_values(const DInputSocket socket_to_compute)
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
Vector<DSocket> from_sockets;
socket_to_compute.foreach_origin_socket([&](DSocket socket) { from_sockets.append(socket); });
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (from_sockets.is_empty()) {
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/* The input is not connected, use the value from the socket itself. */
const CPPType &type = *blender::nodes::socket_cpp_type_get(*socket_to_compute->typeinfo());
return {get_unlinked_input_value(socket_to_compute, type)};
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
/* Multi-input sockets contain a vector of inputs. */
if (socket_to_compute->is_multi_input_socket()) {
return this->get_inputs_from_incoming_links(socket_to_compute, from_sockets);
}
const DSocket from_socket = from_sockets[0];
GMutablePointer value = this->get_input_from_incoming_link(socket_to_compute, from_socket);
return {value};
}
Vector<GMutablePointer> get_inputs_from_incoming_links(const DInputSocket socket_to_compute,
const Span<DSocket> from_sockets)
{
Vector<GMutablePointer> values;
for (const int i : from_sockets.index_range()) {
const DSocket from_socket = from_sockets[i];
const int first_occurence = from_sockets.take_front(i).first_index_try(from_socket);
if (first_occurence == -1) {
values.append(this->get_input_from_incoming_link(socket_to_compute, from_socket));
}
else {
2021-03-12 00:46:41 +11:00
/* If the same from-socket occurs more than once, we make a copy of the first value. This
* can happen when a node linked to a multi-input-socket is muted. */
GMutablePointer value = values[first_occurence];
const CPPType *type = value.type();
void *copy_buffer = allocator_.allocate(type->size(), type->alignment());
type->copy_to_uninitialized(value.get(), copy_buffer);
values.append({type, copy_buffer});
}
}
return values;
}
GMutablePointer get_input_from_incoming_link(const DInputSocket socket_to_compute,
const DSocket from_socket)
{
if (from_socket->is_output()) {
const DOutputSocket from_output_socket{from_socket};
const std::pair<DInputSocket, DOutputSocket> key = std::make_pair(socket_to_compute,
from_output_socket);
std::optional<GMutablePointer> value = value_by_input_.pop_try(key);
if (value.has_value()) {
/* This input has been computed before, return it directly. */
return {*value};
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/* Compute the socket now. */
this->compute_output_and_forward(from_output_socket);
return {value_by_input_.pop(key)};
}
/* Get value from an unlinked input socket. */
const CPPType &type = *blender::nodes::socket_cpp_type_get(*socket_to_compute->typeinfo());
const DInputSocket from_input_socket{from_socket};
return {get_unlinked_input_value(from_input_socket, type)};
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
void compute_output_and_forward(const DOutputSocket socket_to_compute)
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
const DNode node{socket_to_compute.context(), &socket_to_compute->node()};
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (!socket_to_compute->is_available()) {
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/* If the output is not available, use a default value. */
const CPPType &type = *blender::nodes::socket_cpp_type_get(*socket_to_compute->typeinfo());
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
void *buffer = allocator_.allocate(type.size(), type.alignment());
type.copy_to_uninitialized(type.default_value(), buffer);
this->forward_to_inputs(socket_to_compute, {type, buffer});
return;
}
/* Prepare inputs required to execute the node. */
GValueMap<StringRef> node_inputs_map{allocator_};
for (const InputSocketRef *input_socket : node->inputs()) {
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (input_socket->is_available()) {
DInputSocket dsocket{node.context(), input_socket};
Vector<GMutablePointer> values = this->get_input_values(dsocket);
this->log_socket_value(dsocket, values);
for (int i = 0; i < values.size(); ++i) {
/* Values from Multi Input Sockets are stored in input map with the format
* <identifier>[<index>]. */
blender::StringRefNull key = allocator_.copy_string(
input_socket->identifier() + (i > 0 ? ("[" + std::to_string(i)) + "]" : ""));
node_inputs_map.add_new_direct(key, std::move(values[i]));
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
/* Execute the node. */
GValueMap<StringRef> node_outputs_map{allocator_};
GeoNodeExecParams params{
node, node_inputs_map, node_outputs_map, handle_map_, self_object_, modifier_, depsgraph_};
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
this->execute_node(node, params);
/* Forward computed outputs to linked input sockets. */
for (const OutputSocketRef *output_socket : node->outputs()) {
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (output_socket->is_available()) {
const DOutputSocket dsocket{node.context(), output_socket};
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
GMutablePointer value = node_outputs_map.extract(output_socket->identifier());
this->log_socket_value(dsocket, value);
this->forward_to_inputs(dsocket, value);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
}
void log_socket_value(const DSocket socket, Span<GPointer> values)
{
if (log_socket_value_fn_) {
log_socket_value_fn_(socket, values);
}
}
void log_socket_value(const DSocket socket, Span<GMutablePointer> values)
{
this->log_socket_value(socket, values.cast<GPointer>());
}
void log_socket_value(const DSocket socket, GPointer value)
{
this->log_socket_value(socket, Span<GPointer>(&value, 1));
}
void execute_node(const DNode node, GeoNodeExecParams params)
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
const bNode &bnode = params.node();
/* Use the geometry-node-execute callback if it exists. */
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (bnode.typeinfo->geometry_node_execute != nullptr) {
bnode.typeinfo->geometry_node_execute(params);
return;
}
/* Use the multi-function implementation if it exists. */
const MultiFunction *multi_function = mf_by_node_.lookup_default(node, nullptr);
if (multi_function != nullptr) {
this->execute_multi_function_node(node, params, *multi_function);
return;
}
/* Just output default values if no implementation exists. */
this->execute_unknown_node(node, params);
}
void execute_multi_function_node(const DNode node,
GeoNodeExecParams params,
const MultiFunction &fn)
{
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
MFContextBuilder fn_context;
MFParamsBuilder fn_params{fn, 1};
Vector<GMutablePointer> input_data;
for (const InputSocketRef *socket_ref : node->inputs()) {
if (socket_ref->is_available()) {
GMutablePointer data = params.extract_input(socket_ref->identifier());
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
fn_params.add_readonly_single_input(GSpan(*data.type(), data.get(), 1));
input_data.append(data);
}
}
Vector<GMutablePointer> output_data;
for (const OutputSocketRef *socket_ref : node->outputs()) {
if (socket_ref->is_available()) {
const CPPType &type = *blender::nodes::socket_cpp_type_get(*socket_ref->typeinfo());
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
void *buffer = allocator_.allocate(type.size(), type.alignment());
fn_params.add_uninitialized_single_output(GMutableSpan(type, buffer, 1));
output_data.append(GMutablePointer(type, buffer));
}
}
fn.call(IndexRange(1), fn_params, fn_context);
for (GMutablePointer value : input_data) {
value.destruct();
}
int output_index = 0;
for (const int i : node->outputs().index_range()) {
if (node->output(i).is_available()) {
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
GMutablePointer value = output_data[output_index];
params.set_output_by_move(node->output(i).identifier(), value);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
value.destruct();
output_index++;
}
}
}
void execute_unknown_node(const DNode node, GeoNodeExecParams params)
{
for (const OutputSocketRef *socket : node->outputs()) {
if (socket->is_available()) {
const CPPType &type = *blender::nodes::socket_cpp_type_get(*socket->typeinfo());
params.set_output_by_copy(socket->identifier(), {type, type.default_value()});
}
}
}
void forward_to_inputs(const DOutputSocket from_socket, GMutablePointer value_to_forward)
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
/* For all sockets that are linked with the from_socket push the value to their node. */
Vector<DInputSocket> to_sockets_all;
auto handle_target_socket_fn = [&](DInputSocket to_socket) {
to_sockets_all.append_non_duplicates(to_socket);
};
auto handle_skipped_socket_fn = [&, this](DSocket socket) {
this->log_socket_value(socket, value_to_forward);
};
from_socket.foreach_target_socket(handle_target_socket_fn, handle_skipped_socket_fn);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const CPPType &from_type = *value_to_forward.type();
Vector<DInputSocket> to_sockets_same_type;
for (const DInputSocket &to_socket : to_sockets_all) {
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const CPPType &to_type = *blender::nodes::socket_cpp_type_get(*to_socket->typeinfo());
const std::pair<DInputSocket, DOutputSocket> key = std::make_pair(to_socket, from_socket);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (from_type == to_type) {
to_sockets_same_type.append(to_socket);
}
else {
void *buffer = allocator_.allocate(to_type.size(), to_type.alignment());
if (conversions_.is_convertible(from_type, to_type)) {
conversions_.convert_to_uninitialized(
from_type, to_type, value_to_forward.get(), buffer);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
else {
to_type.copy_to_uninitialized(to_type.default_value(), buffer);
}
add_value_to_input_socket(key, GMutablePointer{to_type, buffer});
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
if (to_sockets_same_type.size() == 0) {
/* This value is not further used, so destruct it. */
value_to_forward.destruct();
}
else if (to_sockets_same_type.size() == 1) {
/* This value is only used on one input socket, no need to copy it. */
const DInputSocket to_socket = to_sockets_same_type[0];
const std::pair<DInputSocket, DOutputSocket> key = std::make_pair(to_socket, from_socket);
add_value_to_input_socket(key, value_to_forward);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
else {
/* Multiple inputs use the value, make a copy for every input except for one. */
const DInputSocket first_to_socket = to_sockets_same_type[0];
Span<DInputSocket> other_to_sockets = to_sockets_same_type.as_span().drop_front(1);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const CPPType &type = *value_to_forward.type();
const std::pair<DInputSocket, DOutputSocket> first_key = std::make_pair(first_to_socket,
from_socket);
add_value_to_input_socket(first_key, value_to_forward);
for (const DInputSocket &to_socket : other_to_sockets) {
const std::pair<DInputSocket, DOutputSocket> key = std::make_pair(to_socket, from_socket);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
void *buffer = allocator_.allocate(type.size(), type.alignment());
type.copy_to_uninitialized(value_to_forward.get(), buffer);
add_value_to_input_socket(key, GMutablePointer{type, buffer});
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
}
void add_value_to_input_socket(const std::pair<DInputSocket, DOutputSocket> key,
GMutablePointer value)
{
value_by_input_.add_new(key, value);
}
GMutablePointer get_unlinked_input_value(const DInputSocket &socket,
const CPPType &required_type)
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
bNodeSocket *bsocket = socket->bsocket();
const CPPType &type = *blender::nodes::socket_cpp_type_get(*socket->typeinfo());
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
void *buffer = allocator_.allocate(type.size(), type.alignment());
if (bsocket->type == SOCK_OBJECT) {
Object *object = socket->default_value<bNodeSocketValueObject>()->value;
PersistentObjectHandle object_handle = handle_map_.lookup(object);
new (buffer) PersistentObjectHandle(object_handle);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
else if (bsocket->type == SOCK_COLLECTION) {
Collection *collection = socket->default_value<bNodeSocketValueCollection>()->value;
PersistentCollectionHandle collection_handle = handle_map_.lookup(collection);
new (buffer) PersistentCollectionHandle(collection_handle);
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
else {
blender::nodes::socket_cpp_value_get(*bsocket, buffer);
}
if (type == required_type) {
return {type, buffer};
}
if (conversions_.is_convertible(type, required_type)) {
void *converted_buffer = allocator_.allocate(required_type.size(),
required_type.alignment());
conversions_.convert_to_uninitialized(type, required_type, buffer, converted_buffer);
type.destruct(buffer);
return {required_type, converted_buffer};
}
void *default_buffer = allocator_.allocate(required_type.size(), required_type.alignment());
required_type.copy_to_uninitialized(required_type.default_value(), default_buffer);
return {required_type, default_buffer};
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
};
/**
* This code is responsible for creating the new property and also creating the group of
* properties in the prop_ui_container group for the UI info, the mapping for which is
* scattered about in RNA_access.c.
*
* TODO(Hans): Codify this with some sort of table or refactor IDProperty use in RNA_access.c.
*/
struct SocketPropertyType {
2021-02-05 16:23:34 +11:00
/* Create the actual property used to store the data for the modifier. */
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
IDProperty *(*create_prop)(const bNodeSocket &socket, const char *name);
/* Reused to build the "soft_min" property too. */
IDProperty *(*create_min_ui_prop)(const bNodeSocket &socket, const char *name);
/* Reused to build the "soft_max" property too. */
IDProperty *(*create_max_ui_prop)(const bNodeSocket &socket, const char *name);
/* This uses the same values as #create_prop, but sometimes the type is different, so it can't
* be the same function. */
IDProperty *(*create_default_ui_prop)(const bNodeSocket &socket, const char *name);
PropertyType (*rna_subtype_get)(const bNodeSocket &socket);
bool (*is_correct_type)(const IDProperty &property);
void (*init_cpp_value)(const IDProperty &property,
const PersistentDataHandleMap &handles,
void *r_value);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
};
static IDProperty *socket_add_property(IDProperty *settings_prop_group,
IDProperty *ui_container,
const SocketPropertyType &property_type,
const bNodeSocket &socket)
{
const char *new_prop_name = socket.identifier;
/* Add the property actually storing the data to the modifier's group. */
IDProperty *prop = property_type.create_prop(socket, new_prop_name);
IDP_AddToGroup(settings_prop_group, prop);
prop->flag |= IDP_FLAG_OVERRIDABLE_LIBRARY;
2021-03-12 00:46:41 +11:00
/* Make the group in the UI container group to hold the property's UI settings. */
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
IDProperty *prop_ui_group;
{
IDPropertyTemplate idprop = {0};
prop_ui_group = IDP_New(IDP_GROUP, &idprop, new_prop_name);
IDP_AddToGroup(ui_container, prop_ui_group);
}
/* Set property description (tooltip). */
IDPropertyTemplate property_description_template;
property_description_template.string.str = socket.description;
property_description_template.string.len = BLI_strnlen(socket.description, MAX_NAME) + 1;
property_description_template.string.subtype = IDP_STRING_SUB_UTF8;
IDProperty *description = IDP_New(IDP_STRING, &property_description_template, "description");
IDP_AddToGroup(prop_ui_group, description);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/* Create the properties for the socket's UI settings. */
if (property_type.create_min_ui_prop != nullptr) {
IDP_AddToGroup(prop_ui_group, property_type.create_min_ui_prop(socket, "min"));
IDP_AddToGroup(prop_ui_group, property_type.create_min_ui_prop(socket, "soft_min"));
}
if (property_type.create_max_ui_prop != nullptr) {
IDP_AddToGroup(prop_ui_group, property_type.create_max_ui_prop(socket, "max"));
IDP_AddToGroup(prop_ui_group, property_type.create_max_ui_prop(socket, "soft_max"));
}
if (property_type.create_default_ui_prop != nullptr) {
IDP_AddToGroup(prop_ui_group, property_type.create_default_ui_prop(socket, "default"));
}
if (property_type.rna_subtype_get != nullptr) {
const char *subtype_identifier = nullptr;
RNA_enum_identifier(rna_enum_property_subtype_items,
property_type.rna_subtype_get(socket),
&subtype_identifier);
if (subtype_identifier != nullptr) {
IDPropertyTemplate idprop = {0};
idprop.string.str = subtype_identifier;
idprop.string.len = BLI_strnlen(subtype_identifier, MAX_NAME) + 1;
IDP_AddToGroup(prop_ui_group, IDP_New(IDP_STRING, &idprop, "subtype"));
}
}
return prop;
}
static const SocketPropertyType *get_socket_property_type(const bNodeSocket &bsocket)
{
switch (bsocket.type) {
case SOCK_FLOAT: {
static const SocketPropertyType float_type = {
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueFloat *value = (bNodeSocketValueFloat *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.f = value->value;
return IDP_New(IDP_FLOAT, &idprop, name);
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueFloat *value = (bNodeSocketValueFloat *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.d = value->min;
return IDP_New(IDP_DOUBLE, &idprop, name);
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueFloat *value = (bNodeSocketValueFloat *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.d = value->max;
return IDP_New(IDP_DOUBLE, &idprop, name);
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueFloat *value = (bNodeSocketValueFloat *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.d = value->value;
return IDP_New(IDP_DOUBLE, &idprop, name);
},
[](const bNodeSocket &socket) {
return (PropertyType)((bNodeSocketValueFloat *)socket.default_value)->subtype;
},
[](const IDProperty &property) { return ELEM(property.type, IDP_FLOAT, IDP_DOUBLE); },
[](const IDProperty &property,
const PersistentDataHandleMap &UNUSED(handles),
void *r_value) {
if (property.type == IDP_FLOAT) {
*(float *)r_value = IDP_Float(&property);
}
else if (property.type == IDP_DOUBLE) {
*(float *)r_value = (float)IDP_Double(&property);
}
},
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
};
return &float_type;
}
case SOCK_INT: {
static const SocketPropertyType int_type = {
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueInt *value = (bNodeSocketValueInt *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.i = value->value;
return IDP_New(IDP_INT, &idprop, name);
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueInt *value = (bNodeSocketValueInt *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.i = value->min;
return IDP_New(IDP_INT, &idprop, name);
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueInt *value = (bNodeSocketValueInt *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.i = value->max;
return IDP_New(IDP_INT, &idprop, name);
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueInt *value = (bNodeSocketValueInt *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.i = value->value;
return IDP_New(IDP_INT, &idprop, name);
},
[](const bNodeSocket &socket) {
return (PropertyType)((bNodeSocketValueInt *)socket.default_value)->subtype;
},
[](const IDProperty &property) { return property.type == IDP_INT; },
[](const IDProperty &property,
const PersistentDataHandleMap &UNUSED(handles),
void *r_value) { *(int *)r_value = IDP_Int(&property); },
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
};
return &int_type;
}
case SOCK_VECTOR: {
static const SocketPropertyType vector_type = {
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueVector *value = (bNodeSocketValueVector *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.array.len = 3;
idprop.array.type = IDP_FLOAT;
IDProperty *property = IDP_New(IDP_ARRAY, &idprop, name);
copy_v3_v3((float *)IDP_Array(property), value->value);
return property;
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueVector *value = (bNodeSocketValueVector *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.d = value->min;
return IDP_New(IDP_DOUBLE, &idprop, name);
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueVector *value = (bNodeSocketValueVector *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.d = value->max;
return IDP_New(IDP_DOUBLE, &idprop, name);
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueVector *value = (bNodeSocketValueVector *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.array.len = 3;
idprop.array.type = IDP_FLOAT;
IDProperty *property = IDP_New(IDP_ARRAY, &idprop, name);
copy_v3_v3((float *)IDP_Array(property), value->value);
return property;
},
[](const bNodeSocket &socket) {
return (PropertyType)((bNodeSocketValueVector *)socket.default_value)->subtype;
},
[](const IDProperty &property) {
return property.type == IDP_ARRAY && property.subtype == IDP_FLOAT &&
property.len == 3;
},
[](const IDProperty &property,
const PersistentDataHandleMap &UNUSED(handles),
void *r_value) { copy_v3_v3((float *)r_value, (const float *)IDP_Array(&property)); },
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
};
return &vector_type;
}
case SOCK_BOOLEAN: {
static const SocketPropertyType boolean_type = {
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueBoolean *value = (bNodeSocketValueBoolean *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.i = value->value != 0;
return IDP_New(IDP_INT, &idprop, name);
},
[](const bNodeSocket &UNUSED(socket), const char *name) {
IDPropertyTemplate idprop = {0};
idprop.i = 0;
return IDP_New(IDP_INT, &idprop, name);
},
[](const bNodeSocket &UNUSED(socket), const char *name) {
IDPropertyTemplate idprop = {0};
idprop.i = 1;
return IDP_New(IDP_INT, &idprop, name);
},
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueBoolean *value = (bNodeSocketValueBoolean *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.i = value->value != 0;
return IDP_New(IDP_INT, &idprop, name);
},
nullptr,
[](const IDProperty &property) { return property.type == IDP_INT; },
[](const IDProperty &property,
const PersistentDataHandleMap &UNUSED(handles),
void *r_value) { *(bool *)r_value = IDP_Int(&property) != 0; },
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
};
return &boolean_type;
}
case SOCK_STRING: {
static const SocketPropertyType string_type = {
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueString *value = (bNodeSocketValueString *)socket.default_value;
return IDP_NewString(
value->value, name, BLI_strnlen(value->value, sizeof(value->value)) + 1);
},
nullptr,
nullptr,
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueString *value = (bNodeSocketValueString *)socket.default_value;
return IDP_NewString(
value->value, name, BLI_strnlen(value->value, sizeof(value->value)) + 1);
},
nullptr,
[](const IDProperty &property) { return property.type == IDP_STRING; },
[](const IDProperty &property,
const PersistentDataHandleMap &UNUSED(handles),
void *r_value) { new (r_value) std::string(IDP_String(&property)); },
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
};
return &string_type;
}
case SOCK_OBJECT: {
static const SocketPropertyType object_type = {
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueObject *value = (bNodeSocketValueObject *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.id = (ID *)value->value;
return IDP_New(IDP_ID, &idprop, name);
},
nullptr,
nullptr,
nullptr,
nullptr,
[](const IDProperty &property) { return property.type == IDP_ID; },
[](const IDProperty &property, const PersistentDataHandleMap &handles, void *r_value) {
ID *id = IDP_Id(&property);
Object *object = (id && GS(id->name) == ID_OB) ? (Object *)id : nullptr;
new (r_value) PersistentObjectHandle(handles.lookup(object));
},
};
return &object_type;
}
case SOCK_COLLECTION: {
static const SocketPropertyType collection_type = {
[](const bNodeSocket &socket, const char *name) {
bNodeSocketValueCollection *value = (bNodeSocketValueCollection *)socket.default_value;
IDPropertyTemplate idprop = {0};
idprop.id = (ID *)value->value;
return IDP_New(IDP_ID, &idprop, name);
},
nullptr,
nullptr,
nullptr,
nullptr,
[](const IDProperty &property) { return property.type == IDP_ID; },
[](const IDProperty &property, const PersistentDataHandleMap &handles, void *r_value) {
ID *id = IDP_Id(&property);
Collection *collection = (id && GS(id->name) == ID_GR) ? (Collection *)id : nullptr;
new (r_value) PersistentCollectionHandle(handles.lookup(collection));
},
};
return &collection_type;
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
default: {
return nullptr;
}
}
}
/**
* Rebuild the list of properties based on the sockets exposed as the modifier's node group
* inputs. If any properties correspond to the old properties by name and type, carry over
* the values.
*/
void MOD_nodes_update_interface(Object *object, NodesModifierData *nmd)
{
if (nmd->node_group == nullptr) {
return;
}
IDProperty *old_properties = nmd->settings.properties;
{
IDPropertyTemplate idprop = {0};
nmd->settings.properties = IDP_New(IDP_GROUP, &idprop, "Nodes Modifier Settings");
}
IDProperty *ui_container_group;
{
IDPropertyTemplate idprop = {0};
ui_container_group = IDP_New(IDP_GROUP, &idprop, "_RNA_UI");
IDP_AddToGroup(nmd->settings.properties, ui_container_group);
}
LISTBASE_FOREACH (bNodeSocket *, socket, &nmd->node_group->inputs) {
const SocketPropertyType *property_type = get_socket_property_type(*socket);
if (property_type == nullptr) {
continue;
}
IDProperty *new_prop = socket_add_property(
nmd->settings.properties, ui_container_group, *property_type, *socket);
if (old_properties != nullptr) {
IDProperty *old_prop = IDP_GetPropertyFromGroup(old_properties, socket->identifier);
if (old_prop != nullptr && property_type->is_correct_type(*old_prop)) {
IDP_CopyPropertyContent(new_prop, old_prop);
}
}
}
if (old_properties != nullptr) {
IDP_FreeProperty(old_properties);
}
DEG_id_tag_update(&object->id, ID_RECALC_GEOMETRY);
}
void MOD_nodes_init(Main *bmain, NodesModifierData *nmd)
{
bNodeTree *ntree = ntreeAddTree(bmain, "Geometry Nodes", ntreeType_Geometry->idname);
nmd->node_group = ntree;
ntreeAddSocketInterface(ntree, SOCK_IN, "NodeSocketGeometry", "Geometry");
ntreeAddSocketInterface(ntree, SOCK_OUT, "NodeSocketGeometry", "Geometry");
bNode *group_input_node = nodeAddStaticNode(nullptr, ntree, NODE_GROUP_INPUT);
bNode *group_output_node = nodeAddStaticNode(nullptr, ntree, NODE_GROUP_OUTPUT);
nodeSetSelected(group_input_node, false);
nodeSetSelected(group_output_node, false);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
group_input_node->locx = -200 - group_input_node->width;
group_output_node->locx = 200;
group_output_node->flag |= NODE_DO_OUTPUT;
nodeAddLink(ntree,
group_output_node,
(bNodeSocket *)group_output_node->inputs.first,
group_input_node,
(bNodeSocket *)group_input_node->outputs.first);
ntreeUpdateTree(bmain, ntree);
}
static void initialize_group_input(NodesModifierData &nmd,
const PersistentDataHandleMap &handle_map,
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const bNodeSocket &socket,
const CPPType &cpp_type,
void *r_value)
{
const SocketPropertyType *property_type = get_socket_property_type(socket);
if (property_type == nullptr) {
cpp_type.copy_to_uninitialized(cpp_type.default_value(), r_value);
return;
}
if (nmd.settings.properties == nullptr) {
blender::nodes::socket_cpp_value_get(socket, r_value);
return;
}
const IDProperty *property = IDP_GetPropertyFromGroup(nmd.settings.properties,
socket.identifier);
if (property == nullptr) {
blender::nodes::socket_cpp_value_get(socket, r_value);
return;
}
if (!property_type->is_correct_type(*property)) {
blender::nodes::socket_cpp_value_get(socket, r_value);
return;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
property_type->init_cpp_value(*property, handle_map, r_value);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
static void fill_data_handle_map(const NodesModifierSettings &settings,
const DerivedNodeTree &tree,
PersistentDataHandleMap &handle_map)
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
Set<ID *> used_ids;
find_used_ids_from_settings(settings, used_ids);
find_used_ids_from_nodes(*tree.root_context().tree().btree(), used_ids);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
int current_handle = 0;
for (ID *id : used_ids) {
handle_map.add(current_handle, *id);
current_handle++;
}
}
static void reset_tree_ui_storage(Span<const blender::nodes::NodeTreeRef *> trees,
const Object &object,
const ModifierData &modifier)
{
const NodeTreeEvaluationContext context = {object, modifier};
for (const blender::nodes::NodeTreeRef *tree : trees) {
bNodeTree *btree_cow = tree->btree();
bNodeTree *btree_original = (bNodeTree *)DEG_get_original_id((ID *)btree_cow);
BKE_nodetree_ui_storage_free_for_context(*btree_original, context);
}
}
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
static Vector<SpaceSpreadsheet *> find_spreadsheet_editors(Main *bmain)
{
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
Vector<SpaceSpreadsheet *> spreadsheets;
wmWindowManager *wm = (wmWindowManager *)bmain->wm.first;
LISTBASE_FOREACH (wmWindow *, window, &wm->windows) {
bScreen *screen = BKE_workspace_active_screen_get(window->workspace_hook);
LISTBASE_FOREACH (ScrArea *, area, &screen->areabase) {
SpaceLink *sl = (SpaceLink *)area->spacedata.first;
if (sl->spacetype == SPACE_SPREADSHEET) {
spreadsheets.append((SpaceSpreadsheet *)sl);
}
}
}
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
return spreadsheets;
}
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
using PreviewSocketMap = blender::MultiValueMap<DSocket, uint64_t>;
static DSocket try_find_preview_socket_in_node(const DNode node)
{
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
for (const SocketRef *socket : node->outputs()) {
if (socket->bsocket()->type == SOCK_GEOMETRY) {
return {node.context(), socket};
}
}
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
for (const SocketRef *socket : node->inputs()) {
if (socket->bsocket()->type == SOCK_GEOMETRY &&
(socket->bsocket()->flag & SOCK_MULTI_INPUT) == 0) {
return {node.context(), socket};
}
}
return {};
}
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
static DSocket try_get_socket_to_preview_for_spreadsheet(SpaceSpreadsheet *sspreadsheet,
NodesModifierData *nmd,
const ModifierEvalContext *ctx,
const DerivedNodeTree &tree)
{
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
Vector<SpreadsheetContext *> context_path = sspreadsheet->context_path;
if (context_path.size() < 3) {
return {};
}
if (context_path[0]->type != SPREADSHEET_CONTEXT_OBJECT) {
return {};
}
if (context_path[1]->type != SPREADSHEET_CONTEXT_MODIFIER) {
return {};
}
SpreadsheetContextObject *object_context = (SpreadsheetContextObject *)context_path[0];
if (object_context->object != DEG_get_original_object(ctx->object)) {
return {};
}
SpreadsheetContextModifier *modifier_context = (SpreadsheetContextModifier *)context_path[1];
if (StringRef(modifier_context->modifier_name) != nmd->modifier.name) {
return {};
}
for (SpreadsheetContext *context : context_path.as_span().drop_front(2)) {
if (context->type != SPREADSHEET_CONTEXT_NODE) {
return {};
}
}
Span<SpreadsheetContextNode *> nested_group_contexts =
context_path.as_span().drop_front(2).drop_back(1).cast<SpreadsheetContextNode *>();
SpreadsheetContextNode *last_context = (SpreadsheetContextNode *)context_path.last();
const DTreeContext *context = &tree.root_context();
for (SpreadsheetContextNode *node_context : nested_group_contexts) {
const NodeTreeRef &tree_ref = context->tree();
const NodeRef *found_node = nullptr;
for (const NodeRef *node_ref : tree_ref.nodes()) {
if (node_ref->name() == node_context->node_name) {
found_node = node_ref;
break;
}
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
}
if (found_node == nullptr) {
return {};
}
context = context->child_context(*found_node);
if (context == nullptr) {
return {};
}
}
const NodeTreeRef &tree_ref = context->tree();
for (const NodeRef *node_ref : tree_ref.nodes()) {
if (node_ref->name() == last_context->node_name) {
return try_find_preview_socket_in_node({context, node_ref});
}
}
return {};
}
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
static void find_sockets_to_preview(NodesModifierData *nmd,
const ModifierEvalContext *ctx,
const DerivedNodeTree &tree,
PreviewSocketMap &r_sockets_to_preview)
{
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
Main *bmain = DEG_get_bmain(ctx->depsgraph);
/* Based on every visible spreadsheet context path, get a list of sockets that need to have their
* intermediate geometries cached for display. */
Vector<SpaceSpreadsheet *> spreadsheets = find_spreadsheet_editors(bmain);
for (SpaceSpreadsheet *sspreadsheet : spreadsheets) {
const DSocket socket = try_get_socket_to_preview_for_spreadsheet(sspreadsheet, nmd, ctx, tree);
if (socket) {
const uint64_t key = ED_spreadsheet_context_path_hash(sspreadsheet);
r_sockets_to_preview.add_non_duplicates(socket, key);
}
}
}
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
static void log_preview_socket_value(const Span<GPointer> values,
Object *object,
Span<uint64_t> keys)
{
GeometrySet geometry_set = *(const GeometrySet *)values[0].get();
geometry_set.ensure_owns_direct_data();
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
for (uint64_t key : keys) {
BKE_object_preview_geometry_set_add(object, key, new GeometrySet(geometry_set));
}
}
static void log_ui_hints(const DSocket socket,
const Span<GPointer> values,
Object *self_object,
NodesModifierData *nmd)
{
const DNode node = socket.node();
if (node->is_reroute_node() || socket->typeinfo()->type != SOCK_GEOMETRY) {
return;
}
bNodeTree *btree_cow = node->btree();
bNodeTree *btree_original = (bNodeTree *)DEG_get_original_id((ID *)btree_cow);
const NodeTreeEvaluationContext context{*self_object, nmd->modifier};
for (const GPointer data : values) {
if (data.type() == &CPPType::get<GeometrySet>()) {
const GeometrySet &geometry_set = *(const GeometrySet *)data.get();
blender::bke::geometry_set_instances_attribute_foreach(
geometry_set,
[&](StringRefNull attribute_name, const AttributeMetaData &meta_data) {
BKE_nodetree_attribute_hint_add(*btree_original,
context,
*node->bnode(),
attribute_name,
meta_data.domain,
meta_data.data_type);
return true;
},
8);
}
}
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/**
* Evaluate a node group to compute the output geometry.
* Currently, this uses a fairly basic and inefficient algorithm that might compute things more
* often than necessary. It's going to be replaced soon.
*/
static GeometrySet compute_geometry(const DerivedNodeTree &tree,
Span<const NodeRef *> group_input_nodes,
const InputSocketRef &socket_to_compute,
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
GeometrySet input_geometry_set,
NodesModifierData *nmd,
const ModifierEvalContext *ctx)
{
blender::ResourceScope scope;
blender::LinearAllocator<> &allocator = scope.linear_allocator();
blender::nodes::MultiFunctionByNode mf_by_node = get_multi_function_per_node(tree, scope);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
PersistentDataHandleMap handle_map;
fill_data_handle_map(nmd->settings, tree, handle_map);
Map<DOutputSocket, GMutablePointer> group_inputs;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const DTreeContext *root_context = &tree.root_context();
for (const NodeRef *group_input_node : group_input_nodes) {
Span<const OutputSocketRef *> group_input_sockets = group_input_node->outputs().drop_back(1);
if (group_input_sockets.is_empty()) {
continue;
}
Span<const OutputSocketRef *> remaining_input_sockets = group_input_sockets;
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/* If the group expects a geometry as first input, use the geometry that has been passed to
* modifier. */
const OutputSocketRef *first_input_socket = group_input_sockets[0];
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (first_input_socket->bsocket()->type == SOCK_GEOMETRY) {
GeometrySet *geometry_set_in =
allocator.construct<GeometrySet>(input_geometry_set).release();
group_inputs.add_new({root_context, first_input_socket}, geometry_set_in);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
remaining_input_sockets = remaining_input_sockets.drop_front(1);
}
/* Initialize remaining group inputs. */
for (const OutputSocketRef *socket : remaining_input_sockets) {
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const CPPType &cpp_type = *blender::nodes::socket_cpp_type_get(*socket->typeinfo());
void *value_in = allocator.allocate(cpp_type.size(), cpp_type.alignment());
initialize_group_input(*nmd, handle_map, *socket->bsocket(), cpp_type, value_in);
group_inputs.add_new({root_context, socket}, {cpp_type, value_in});
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
/* Don't keep a reference to the input geometry components to avoid copies during evaluation. */
input_geometry_set.clear();
Vector<DInputSocket> group_outputs;
group_outputs.append({root_context, &socket_to_compute});
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
PreviewSocketMap preview_sockets;
find_sockets_to_preview(nmd, ctx, tree, preview_sockets);
auto log_socket_value = [&](const DSocket socket, const Span<GPointer> values) {
Spreadsheet: breadcrumbs and node pinning This introduces a context path to the spreadsheet editor, which contains information about what data is shown in the spreadsheet. The context path (breadcrumbs) can reference a specific node in a node group hierarchy. During object evaluation, the geometry nodes modifier checks what data is currently requested by visible spreadsheets and stores the corresponding geometry sets separately for later access. The context path can be updated by the user explicitely, by clicking on the new icon in the header of nodes. Under some circumstances, the context path is updated automatically based on Blender's context. This patch also consolidates the "Node" and "Final" object evaluation mode to just "Evaluated". Based on the current context path, either the final geometry set of an object will be displayed, or the data at a specific node. The new preview icon in geometry nodes now behaves more like a toggle. It can be clicked again to clear the context path in an open spreadsheet editor. Previously, only an object could be pinned in the spreadsheet editor. Now it is possible to pin the entire context path. That allows two different spreadsheets to display geometry data from two different nodes. The breadcrumbs in the spreadsheet header can be collapsed by clicking on the arrow icons. It's not ideal but works well for now. This might be changed again, if we get a data set region on the left. Differential Revision: https://developer.blender.org/D10931
2021-04-15 08:57:10 +02:00
if (!DEG_is_active(ctx->depsgraph)) {
return;
}
Span<uint64_t> keys = preview_sockets.lookup(socket);
if (!keys.is_empty()) {
log_preview_socket_value(values, ctx->object, keys);
}
log_ui_hints(socket, values, ctx->object, nmd);
};
GeometryNodesEvaluator evaluator{group_inputs,
group_outputs,
mf_by_node,
handle_map,
ctx->object,
(ModifierData *)nmd,
ctx->depsgraph,
log_socket_value};
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
Vector<GMutablePointer> results = evaluator.execute();
BLI_assert(results.size() == 1);
GMutablePointer result = results[0];
GeometrySet output_geometry = std::move(*(GeometrySet *)result.get());
return output_geometry;
}
/**
* \note This could be done in #initialize_group_input, though that would require adding the
* the object as a parameter, so it's likely better to this check as a separate step.
*/
static void check_property_socket_sync(const Object *ob, ModifierData *md)
{
NodesModifierData *nmd = reinterpret_cast<NodesModifierData *>(md);
int i = 0;
LISTBASE_FOREACH_INDEX (const bNodeSocket *, socket, &nmd->node_group->inputs, i) {
/* The first socket is the special geometry socket for the modifier object. */
if (i == 0 && socket->type == SOCK_GEOMETRY) {
continue;
}
IDProperty *property = IDP_GetPropertyFromGroup(nmd->settings.properties, socket->identifier);
if (property == nullptr) {
if (socket->type == SOCK_GEOMETRY) {
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
BKE_modifier_set_error(ob, md, "Node group can only have one geometry input");
}
else {
BKE_modifier_set_error(ob, md, "Missing property for input socket \"%s\"", socket->name);
}
continue;
}
const SocketPropertyType *property_type = get_socket_property_type(*socket);
if (!property_type->is_correct_type(*property)) {
BKE_modifier_set_error(
ob, md, "Property type does not match input socket \"(%s)\"", socket->name);
continue;
}
}
bool has_geometry_output = false;
LISTBASE_FOREACH (const bNodeSocket *, socket, &nmd->node_group->outputs) {
if (socket->type == SOCK_GEOMETRY) {
has_geometry_output = true;
}
}
if (!has_geometry_output) {
BKE_modifier_set_error(ob, md, "Node group must have a geometry output");
}
}
static void modifyGeometry(ModifierData *md,
const ModifierEvalContext *ctx,
GeometrySet &geometry_set)
{
NodesModifierData *nmd = reinterpret_cast<NodesModifierData *>(md);
if (nmd->node_group == nullptr) {
return;
}
check_property_socket_sync(ctx->object, md);
NodeTreeRefMap tree_refs;
DerivedNodeTree tree{*nmd->node_group, tree_refs};
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (tree.has_link_cycles()) {
BKE_modifier_set_error(ctx->object, md, "Node group has cycles");
return;
}
const NodeTreeRef &root_tree_ref = tree.root_context().tree();
Span<const NodeRef *> input_nodes = root_tree_ref.nodes_by_type("NodeGroupInput");
Span<const NodeRef *> output_nodes = root_tree_ref.nodes_by_type("NodeGroupOutput");
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (output_nodes.size() != 1) {
return;
}
Span<const InputSocketRef *> group_outputs = output_nodes[0]->inputs().drop_back(1);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (group_outputs.size() == 0) {
return;
}
const InputSocketRef *group_output = group_outputs[0];
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
if (group_output->idname() != "NodeSocketGeometry") {
return;
}
if (DEG_is_active(ctx->depsgraph)) {
reset_tree_ui_storage(tree.used_node_tree_refs(), *ctx->object, *md);
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
geometry_set = compute_geometry(
tree, input_nodes, *group_outputs[0], std::move(geometry_set), nmd, ctx);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
static Mesh *modifyMesh(ModifierData *md, const ModifierEvalContext *ctx, Mesh *mesh)
{
GeometrySet geometry_set = GeometrySet::create_with_mesh(mesh, GeometryOwnershipType::Editable);
geometry_set.get_component_for_write<MeshComponent>().copy_vertex_group_names_from_object(
*ctx->object);
modifyGeometry(md, ctx, geometry_set);
/* This function is only called when applying modifiers. In this case it makes sense to realize
* instances, otherwise in some cases there might be no results when applying the modifier. */
geometry_set = blender::bke::geometry_set_realize_mesh_for_modifier(geometry_set);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
Mesh *new_mesh = geometry_set.get_component_for_write<MeshComponent>().release();
if (new_mesh == nullptr) {
return BKE_mesh_new_nomain(0, 0, 0, 0, 0);
}
return new_mesh;
}
static void modifyGeometrySet(ModifierData *md,
const ModifierEvalContext *ctx,
GeometrySet *geometry_set)
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
{
modifyGeometry(md, ctx, *geometry_set);
}
/* Drawing the properties manually with #uiItemR instead of #uiDefAutoButsRNA allows using
* the node socket identifier for the property names, since they are unique, but also having
* the correct label displayed in the UI. */
static void draw_property_for_socket(uiLayout *layout,
PointerRNA *bmain_ptr,
PointerRNA *md_ptr,
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
const IDProperty *modifier_props,
const bNodeSocket &socket)
{
const SocketPropertyType *property_type = get_socket_property_type(socket);
if (property_type == nullptr) {
return;
}
/* The property should be created in #MOD_nodes_update_interface with the correct type. */
IDProperty *property = IDP_GetPropertyFromGroup(modifier_props, socket.identifier);
/* IDProperties can be removed with python, so there could be a situation where
* there isn't a property for a socket or it doesn't have the correct type. */
if (property != nullptr && property_type->is_correct_type(*property)) {
char socket_id_esc[sizeof(socket.identifier) * 2];
BLI_str_escape(socket_id_esc, socket.identifier, sizeof(socket_id_esc));
char rna_path[sizeof(socket_id_esc) + 4];
BLI_snprintf(rna_path, ARRAY_SIZE(rna_path), "[\"%s\"]", socket_id_esc);
/* Use #uiItemPointerR to draw pointer properties because #uiItemR would not have enough
* information about what type of ID to select for editing the values. This is because
* pointer IDProperties contain no information about their type. */
switch (socket.type) {
case SOCK_OBJECT: {
uiItemPointerR(
layout, md_ptr, rna_path, bmain_ptr, "objects", socket.name, ICON_OBJECT_DATA);
break;
}
case SOCK_COLLECTION: {
uiItemPointerR(layout,
md_ptr,
rna_path,
bmain_ptr,
"collections",
socket.name,
ICON_OUTLINER_COLLECTION);
break;
}
default:
uiItemR(layout, md_ptr, rna_path, 0, socket.name, ICON_NONE);
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
static void panel_draw(const bContext *C, Panel *panel)
{
uiLayout *layout = panel->layout;
Main *bmain = CTX_data_main(C);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
PointerRNA *ptr = modifier_panel_get_property_pointers(panel, nullptr);
NodesModifierData *nmd = static_cast<NodesModifierData *>(ptr->data);
uiLayoutSetPropSep(layout, true);
uiLayoutSetPropDecorate(layout, true);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
uiTemplateID(layout,
C,
ptr,
"node_group",
"node.new_geometry_node_group_assign",
nullptr,
nullptr,
0,
false,
nullptr);
if (nmd->node_group != nullptr && nmd->settings.properties != nullptr) {
PointerRNA bmain_ptr;
RNA_main_pointer_create(bmain, &bmain_ptr);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
LISTBASE_FOREACH (bNodeSocket *, socket, &nmd->node_group->inputs) {
draw_property_for_socket(layout, &bmain_ptr, ptr, nmd->settings.properties, *socket);
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
}
}
modifier_panel_end(layout, ptr);
}
static void panelRegister(ARegionType *region_type)
{
modifier_panel_register(region_type, eModifierType_Nodes, panel_draw);
}
static void blendWrite(BlendWriter *writer, const ModifierData *md)
{
const NodesModifierData *nmd = reinterpret_cast<const NodesModifierData *>(md);
if (nmd->settings.properties != nullptr) {
/* Note that the property settings are based on the socket type info
* and don't necessarily need to be written, but we can't just free them. */
IDP_BlendWrite(writer, nmd->settings.properties);
}
}
static void blendRead(BlendDataReader *reader, ModifierData *md)
{
NodesModifierData *nmd = reinterpret_cast<NodesModifierData *>(md);
BLO_read_data_address(reader, &nmd->settings.properties);
IDP_BlendDataRead(reader, &nmd->settings.properties);
}
static void copyData(const ModifierData *md, ModifierData *target, const int flag)
{
const NodesModifierData *nmd = reinterpret_cast<const NodesModifierData *>(md);
NodesModifierData *tnmd = reinterpret_cast<NodesModifierData *>(target);
BKE_modifier_copydata_generic(md, target, flag);
if (nmd->settings.properties != nullptr) {
tnmd->settings.properties = IDP_CopyProperty_ex(nmd->settings.properties, flag);
}
}
static void freeData(ModifierData *md)
{
NodesModifierData *nmd = reinterpret_cast<NodesModifierData *>(md);
if (nmd->settings.properties != nullptr) {
IDP_FreeProperty_ex(nmd->settings.properties, false);
nmd->settings.properties = nullptr;
}
}
static void requiredDataMask(Object *UNUSED(ob),
ModifierData *UNUSED(md),
CustomData_MeshMasks *r_cddata_masks)
{
/* We don't know what the node tree will need. If there are vertex groups, it is likely that the
* node tree wants to access them. */
r_cddata_masks->vmask |= CD_MASK_MDEFORMVERT;
r_cddata_masks->vmask |= CD_MASK_PROP_ALL;
}
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
ModifierTypeInfo modifierType_Nodes = {
/* name */ "GeometryNodes",
/* structName */ "NodesModifierData",
/* structSize */ sizeof(NodesModifierData),
/* srna */ &RNA_NodesModifier,
/* type */ eModifierTypeType_Constructive,
/* flags */
static_cast<ModifierTypeFlag>(
eModifierTypeFlag_AcceptsMesh | eModifierTypeFlag_SupportsEditmode |
eModifierTypeFlag_EnableInEditmode | eModifierTypeFlag_SupportsMapping),
/* icon */ ICON_NODETREE,
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/* copyData */ copyData,
/* deformVerts */ nullptr,
/* deformMatrices */ nullptr,
/* deformVertsEM */ nullptr,
/* deformMatricesEM */ nullptr,
/* modifyMesh */ modifyMesh,
/* modifyHair */ nullptr,
/* modifyGeometrySet */ modifyGeometrySet,
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/* modifyVolume */ nullptr,
/* initData */ initData,
/* requiredDataMask */ requiredDataMask,
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/* freeData */ freeData,
/* isDisabled */ isDisabled,
/* updateDepsgraph */ updateDepsgraph,
/* dependsOnTime */ nullptr,
/* dependsOnNormals */ nullptr,
/* foreachIDLink */ foreachIDLink,
Geometry Nodes: show "Show Texture in texture tab" button This enables the quick access button [to show the relevant Texture in the Properties Editor] for textures used in geometry nodes. This goes in line to what we do for other textures: - modifier textures have this button - particle textures have this button - brush textures will soon have it, too (see D9813) When outside of the Properties Editor, the button will always show (if a texture is actually assigned), but will be inactive if no suiting Properties Editor to show the texture in can be found. Note this also changes the behavior to not show the button if _no_ texture is assigned (as in: we are still showing the "New" button). Previously it was always there (e.g. for modifier textures), even if it would take us to an empty texture tab. (Sure, we could add a texture there then, but imho it makes more sense to just start showing it once a texture is already there) For this to work with geometry nodes, the following chages were done: - implement foreachTexLink for geonode modifiers - new buttons_texture_user_node_property_add() that stores prop as well as node - also use NODE_ACTIVE_TEXTURE flag in geometry nodetrees notes: - this still uses the first suiting (as in: pinning does not interfere) Properties Editor it finds, this should (maybe?) find the _closest_ Property Editor instead (see related feedback in D9813). - this will already show the button for brush textures as well (disabled), but there is another mandatory change in an upcomming commit to make it work there as well (see D9813) ref. T85278 Maniphest Tasks: T85278 Differential Revision: https://developer.blender.org/D10293
2021-02-03 14:39:24 +01:00
/* foreachTexLink */ foreachTexLink,
Geometry Nodes: initial scattering and geometry nodes This is the initial merge from the geometry-nodes branch. Nodes: * Attribute Math * Boolean * Edge Split * Float Compare * Object Info * Point Distribute * Point Instance * Random Attribute * Random Float * Subdivision Surface * Transform * Triangulate It includes the initial evaluation of geometry node groups in the Geometry Nodes modifier. Notes on the Generic attribute access API The API adds an indirection for attribute access. That has the following benefits: * Most code does not have to care about how an attribute is stored internally. This is mainly necessary, because we have to deal with "legacy" attributes such as vertex weights and attributes that are embedded into other structs such as vertex positions. * When reading from an attribute, we generally don't care what domain the attribute is stored on. So we want to abstract away the interpolation that that adapts attributes from one domain to another domain (this is not actually implemented yet). Other possible improvements for later iterations include: * Actually implement interpolation between domains. * Don't use inheritance for the different attribute types. A single class for read access and one for write access might be enough, because we know all the ways in which attributes are stored internally. We don't want more different internal structures in the future. On the contrary, ideally we can consolidate the different storage formats in the future to reduce the need for this indirection. * Remove the need for heap allocations when creating attribute accessors. It includes commits from: * Dalai Felinto * Hans Goudey * Jacques Lucke * Léo Depoix
2020-12-02 13:25:25 +01:00
/* freeRuntimeData */ nullptr,
/* panelRegister */ panelRegister,
/* blendWrite */ blendWrite,
/* blendRead */ blendRead,
};