Files
test/source/blender/modifiers/intern/MOD_explode.cc

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

1245 lines
36 KiB
C++
Raw Normal View History

/* SPDX-FileCopyrightText: 2005 Blender Authors
*
* SPDX-License-Identifier: GPL-2.0-or-later */
/** \file
* \ingroup modifiers
*/
2022-09-23 09:38:37 -05:00
#define DNA_DEPRECATED_ALLOW /* For #ME_FACE_SEL. */
#include "BLI_utildefines.h"
#include "BLI_kdtree.h"
#include "BLI_map.hh"
#include "BLI_math_matrix.h"
#include "BLI_math_rotation.h"
#include "BLI_math_vector.h"
#include "BLI_rand.h"
#include "BLT_translation.hh"
#include "DNA_defaults.h"
#include "DNA_mesh_types.h"
#include "DNA_meshdata_types.h"
#include "DNA_object_types.h"
#include "DNA_screen_types.h"
#include "BKE_customdata.hh"
2024-01-29 18:57:16 -05:00
#include "BKE_deform.hh"
2024-01-15 12:44:04 -05:00
#include "BKE_lib_id.hh"
#include "BKE_mesh.hh"
#include "BKE_mesh_legacy_convert.hh"
2023-11-14 09:30:40 +01:00
#include "BKE_modifier.hh"
#include "BKE_particle.h"
#include "BKE_scene.hh"
#include "UI_interface.hh"
#include "UI_resources.hh"
#include "RNA_access.hh"
#include "RNA_prototypes.hh"
#include "DEG_depsgraph_query.hh"
#include "MEM_guardedalloc.h"
#include "MOD_ui_common.hh"
static void init_data(ModifierData *md)
{
2012-05-06 13:38:33 +00:00
ExplodeModifierData *emd = (ExplodeModifierData *)md;
BLI_assert(MEMCMP_STRUCT_AFTER_IS_ZERO(emd, modifier));
MEMCPY_STRUCT_AFTER(emd, DNA_struct_default_get(ExplodeModifierData), modifier);
}
static void free_data(ModifierData *md)
{
2012-05-06 13:38:33 +00:00
ExplodeModifierData *emd = (ExplodeModifierData *)md;
2018-06-17 17:04:27 +02:00
MEM_SAFE_FREE(emd->facepa);
}
static void copy_data(const ModifierData *md, ModifierData *target, const int flag)
{
#if 0
const ExplodeModifierData *emd = (const ExplodeModifierData *)md;
#endif
2012-05-06 13:38:33 +00:00
ExplodeModifierData *temd = (ExplodeModifierData *)target;
BKE_modifier_copydata_generic(md, target, flag);
temd->facepa = nullptr;
}
static bool depends_on_time(Scene * /*scene*/, ModifierData * /*md*/)
{
return true;
}
static void required_data_mask(ModifierData *md, CustomData_MeshMasks *r_cddata_masks)
{
2012-05-06 13:38:33 +00:00
ExplodeModifierData *emd = (ExplodeModifierData *)md;
if (emd->vgroup) {
r_cddata_masks->vmask |= CD_MASK_MDEFORMVERT;
}
}
2018-05-12 08:04:56 +02:00
static void createFacepa(ExplodeModifierData *emd, ParticleSystemModifierData *psmd, Mesh *mesh)
{
ParticleSystem *psys = psmd->psys;
MFace *fa = nullptr, *mface = nullptr;
ParticleData *pa;
KDTree_3d *tree;
RNG *rng;
float center[3], co[3];
int *facepa = nullptr, *vertpa = nullptr, totvert = 0, totface = 0, totpart = 0;
int i, p, v1, v2, v3, v4 = 0;
const bool invert_vgroup = (emd->flag & eExplodeFlag_INVERT_VGROUP) != 0;
blender::MutableSpan<blender::float3> positions = mesh->vert_positions_for_write();
mface = (MFace *)CustomData_get_layer_for_write(
&mesh->fdata_legacy, CD_MFACE, mesh->totface_legacy);
totvert = mesh->verts_num;
totface = mesh->totface_legacy;
totpart = psmd->psys->totpart;
rng = BLI_rng_new_srandom(psys->seed);
if (emd->facepa) {
MEM_freeN(emd->facepa);
}
facepa = emd->facepa = MEM_calloc_arrayN<int>(totface, __func__);
vertpa = MEM_calloc_arrayN<int>(totvert, __func__);
/* initialize all faces & verts to no particle */
for (i = 0; i < totface; i++) {
facepa[i] = totpart;
}
for (i = 0; i < totvert; i++) {
vertpa[i] = totpart;
}
/* set protected verts */
if (emd->vgroup) {
const MDeformVert *dvert = mesh->deform_verts().data();
if (dvert) {
const int defgrp_index = emd->vgroup - 1;
for (i = 0; i < totvert; i++, dvert++) {
float val = BLI_rng_get_float(rng);
val = (1.0f - emd->protect) * val + emd->protect * 0.5f;
const float weight = invert_vgroup ? 1.0f - BKE_defvert_find_weight(dvert, defgrp_index) :
BKE_defvert_find_weight(dvert, defgrp_index);
if (val < weight) {
vertpa[i] = -1;
}
}
}
}
/* make tree of emitter locations */
tree = BLI_kdtree_3d_new(totpart);
for (p = 0, pa = psys->particles; p < totpart; p++, pa++) {
Remove Blender Internal and legacy viewport from Blender 2.8. Brecht authored this commit, but he gave me the honours to actually do it. Here it goes; Blender Internal. Bye bye, you did great! * Point density, voxel data, ocean, environment map textures were removed, as these only worked within BI rendering. Note that the ocean modifier and the Cycles point density shader node continue to work. * Dynamic paint using material shading was removed, as this only worked with BI. If we ever wanted to support this again probably it should go through the baking API. * GPU shader export through the Python API was removed. This only worked for the old BI GLSL shaders, which no longer exists. Doing something similar for Eevee would be significantly more complicated because it uses a lot of multiplass rendering and logic outside the shader, it's probably impractical. * Collada material import / export code is mostly gone, as it only worked for BI materials. We need to add Cycles / Eevee material support at some point. * The mesh noise operator was removed since it only worked with BI material texture slots. A displacement modifier can be used instead. * The delete texture paint slot operator was removed since it only worked for BI material texture slots. Could be added back with node support. * Not all legacy viewport features are supported in the new viewport, but their code was removed. If we need to bring anything back we can look at older git revisions. * There is some legacy viewport code that I could not remove yet, and some that I probably missed. * Shader node execution code was left mostly intact, even though it is not used anywhere now. We may eventually use this to replace the texture nodes with Cycles / Eevee shader nodes. * The Cycles Bake panel now includes settings for baking multires normal and displacement maps. The underlying code needs to be merged properly, and we plan to add back support for multires AO baking and add support to Cycles baking for features like vertex color, displacement, and other missing baking features. * This commit removes DNA and the Python API for BI material, lamp, world and scene settings. This breaks a lot of addons. * There is more DNA that can be removed or renamed, where Cycles or Eevee are reusing some old BI properties but the names are not really correct anymore. * Texture slots for materials, lamps and world were removed. They remain for brushes, particles and freestyle linestyles. * 'BLENDER_RENDER' remains in the COMPAT_ENGINES of UI panels. Cycles and other renderers use this to find all panels to show, minus a few panels that they have their own replacement for.
2018-04-19 17:34:44 +02:00
psys_particle_on_emitter(psmd,
psys->part->from,
pa->num,
pa->num_dmcache,
pa->fuv,
pa->foffset,
co,
nullptr,
nullptr,
nullptr,
nullptr);
BLI_kdtree_3d_insert(tree, p, co);
}
BLI_kdtree_3d_balance(tree);
/* set face-particle-indexes to nearest particle to face center */
for (i = 0, fa = mface; i < totface; i++, fa++) {
Mesh: Move positions to a generic attribute **Changes** As described in T93602, this patch removes all use of the `MVert` struct, replacing it with a generic named attribute with the name `"position"`, consistent with other geometry types. Variable names have been changed from `verts` to `positions`, to align with the attribute name and the more generic design (positions are not vertices, they are just an attribute stored on the point domain). This change is made possible by previous commits that moved all other data out of `MVert` to runtime data or other generic attributes. What remains is mostly a simple type change. Though, the type still shows up 859 times, so the patch is quite large. One compromise is that now `CD_MASK_BAREMESH` now contains `CD_PROP_FLOAT3`. With the general move towards generic attributes over custom data types, we are removing use of these type masks anyway. **Benefits** The most obvious benefit is reduced memory usage and the benefits that brings in memory-bound situations. `float3` is only 3 bytes, in comparison to `MVert` which was 4. When there are millions of vertices this starts to matter more. The other benefits come from using a more generic type. Instead of writing algorithms specifically for `MVert`, code can just use arrays of vectors. This will allow eliminating many temporary arrays or wrappers used to extract positions. Many possible improvements aren't implemented in this patch, though I did switch simplify or remove the process of creating temporary position arrays in a few places. The design clarity that "positions are just another attribute" brings allows removing explicit copying of vertices in some procedural operations-- they are just processed like most other attributes. **Performance** This touches so many areas that it's hard to benchmark exhaustively, but I observed some areas as examples. * The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster. * The Spring splash screen went from ~4.3 to ~4.5 fps. * The subdivision surface modifier/node was slightly faster RNA access through Python may be slightly slower, since now we need a name lookup instead of just a custom data type lookup for each index. **Future Improvements** * Remove uses of "vert_coords" functions: * `BKE_mesh_vert_coords_alloc` * `BKE_mesh_vert_coords_get` * `BKE_mesh_vert_coords_apply{_with_mat4}` * Remove more hidden copying of positions * General simplification now possible in many areas * Convert more code to C++ to use `float3` instead of `float[3]` * Currently `reinterpret_cast` is used for those C-API functions Differential Revision: https://developer.blender.org/D15982
2023-01-10 00:10:43 -05:00
add_v3_v3v3(center, positions[fa->v1], positions[fa->v2]);
add_v3_v3(center, positions[fa->v3]);
if (fa->v4) {
Mesh: Move positions to a generic attribute **Changes** As described in T93602, this patch removes all use of the `MVert` struct, replacing it with a generic named attribute with the name `"position"`, consistent with other geometry types. Variable names have been changed from `verts` to `positions`, to align with the attribute name and the more generic design (positions are not vertices, they are just an attribute stored on the point domain). This change is made possible by previous commits that moved all other data out of `MVert` to runtime data or other generic attributes. What remains is mostly a simple type change. Though, the type still shows up 859 times, so the patch is quite large. One compromise is that now `CD_MASK_BAREMESH` now contains `CD_PROP_FLOAT3`. With the general move towards generic attributes over custom data types, we are removing use of these type masks anyway. **Benefits** The most obvious benefit is reduced memory usage and the benefits that brings in memory-bound situations. `float3` is only 3 bytes, in comparison to `MVert` which was 4. When there are millions of vertices this starts to matter more. The other benefits come from using a more generic type. Instead of writing algorithms specifically for `MVert`, code can just use arrays of vectors. This will allow eliminating many temporary arrays or wrappers used to extract positions. Many possible improvements aren't implemented in this patch, though I did switch simplify or remove the process of creating temporary position arrays in a few places. The design clarity that "positions are just another attribute" brings allows removing explicit copying of vertices in some procedural operations-- they are just processed like most other attributes. **Performance** This touches so many areas that it's hard to benchmark exhaustively, but I observed some areas as examples. * The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster. * The Spring splash screen went from ~4.3 to ~4.5 fps. * The subdivision surface modifier/node was slightly faster RNA access through Python may be slightly slower, since now we need a name lookup instead of just a custom data type lookup for each index. **Future Improvements** * Remove uses of "vert_coords" functions: * `BKE_mesh_vert_coords_alloc` * `BKE_mesh_vert_coords_get` * `BKE_mesh_vert_coords_apply{_with_mat4}` * Remove more hidden copying of positions * General simplification now possible in many areas * Convert more code to C++ to use `float3` instead of `float[3]` * Currently `reinterpret_cast` is used for those C-API functions Differential Revision: https://developer.blender.org/D15982
2023-01-10 00:10:43 -05:00
add_v3_v3(center, positions[fa->v4]);
mul_v3_fl(center, 0.25);
}
else {
mul_v3_fl(center, 1.0f / 3.0f);
}
p = BLI_kdtree_3d_find_nearest(tree, center, nullptr);
v1 = vertpa[fa->v1];
v2 = vertpa[fa->v2];
v3 = vertpa[fa->v3];
if (fa->v4) {
v4 = vertpa[fa->v4];
}
if (v1 >= 0 && v2 >= 0 && v3 >= 0 && (fa->v4 == 0 || v4 >= 0)) {
facepa[i] = p;
}
if (v1 >= 0) {
vertpa[fa->v1] = p;
}
if (v2 >= 0) {
vertpa[fa->v2] = p;
}
if (v3 >= 0) {
vertpa[fa->v3] = p;
}
if (fa->v4 && v4 >= 0) {
vertpa[fa->v4] = p;
}
}
if (vertpa) {
MEM_freeN(vertpa);
}
BLI_kdtree_3d_free(tree);
BLI_rng_free(rng);
}
static int edgecut_get(const blender::Map<blender::OrderedEdge, int> &edgehash, uint v1, uint v2)
{
return edgehash.lookup({int(v1), int(v2)});
}
static const short add_faces[24] = {
0, 0, 0, 2, 0, 1, 2, 2, 0, 2, 1, 2, 2, 2, 2, 3, 0, 0, 0, 1, 0, 1, 1, 2,
};
static MFace *get_dface(Mesh *mesh, Mesh *split, int cur, int i, MFace *mf)
{
MFace *mfaces = static_cast<MFace *>(
CustomData_get_layer_for_write(&split->fdata_legacy, CD_MFACE, split->totface_legacy));
Mesh: Remove redundant custom data pointers For copy-on-write, we want to share attribute arrays between meshes where possible. Mutable pointers like `Mesh.mvert` make that difficult by making ownership vague. They also make code more complex by adding redundancy. The simplest solution is just removing them and retrieving layers from `CustomData` as needed. Similar changes have already been applied to curves and point clouds (e9f82d3dc7ee, 410a6efb747f). Removing use of the pointers generally makes code more obvious and more reusable. Mesh data is now accessed with a C++ API (`Mesh::edges()` or `Mesh::edges_for_write()`), and a C API (`BKE_mesh_edges(mesh)`). The CoW changes this commit makes possible are described in T95845 and T95842, and started in D14139 and D14140. The change also simplifies the ongoing mesh struct-of-array refactors from T95965. **RNA/Python Access Performance** Theoretically, accessing mesh elements with the RNA API may become slower, since the layer needs to be found on every random access. However, overhead is already high enough that this doesn't make a noticible differenc, and performance is actually improved in some cases. Random access can be up to 10% faster, but other situations might be a bit slower. Generally using `foreach_get/set` are the best way to improve performance. See the differential revision for more discussion about Python performance. Cycles has been updated to use raw pointers and the internal Blender mesh types, mostly because there is no sense in having this overhead when it's already compiled with Blender. In my tests this roughly halves the Cycles mesh creation time (0.19s to 0.10s for a 1 million face grid). Differential Revision: https://developer.blender.org/D15488
2022-09-05 11:56:34 -05:00
MFace *df = &mfaces[cur];
CustomData_copy_data(&mesh->fdata_legacy, &split->fdata_legacy, i, cur, 1);
*df = *mf;
return df;
}
#define SET_VERTS(a, b, c, d) \
{ \
v[0] = mf->v##a; \
uv[0] = a - 1; \
v[1] = mf->v##b; \
uv[1] = b - 1; \
v[2] = mf->v##c; \
uv[2] = c - 1; \
v[3] = mf->v##d; \
uv[3] = d - 1; \
} \
(void)0
#define GET_ES(v1, v2) edgecut_get(eh, v1, v2)
#define INT_UV(uvf, c0, c1) mid_v2_v2v2(uvf, mf->uv[c0], mf->uv[c1])
static void remap_faces_3_6_9_12(Mesh *mesh,
Mesh *split,
MFace *mf,
int *facepa,
const int *vertpa,
int i,
const blender::Map<blender::OrderedEdge, int> &eh,
int cur,
int v1,
int v2,
int v3,
int v4)
{
MFace *df1 = get_dface(mesh, split, cur, i, mf);
MFace *df2 = get_dface(mesh, split, cur + 1, i, mf);
MFace *df3 = get_dface(mesh, split, cur + 2, i, mf);
facepa[cur] = vertpa[v1];
df1->v1 = v1;
df1->v2 = GET_ES(v1, v2);
df1->v3 = GET_ES(v2, v3);
df1->v4 = v3;
df1->flag |= ME_FACE_SEL;
facepa[cur + 1] = vertpa[v2];
df2->v1 = GET_ES(v1, v2);
df2->v2 = v2;
df2->v3 = GET_ES(v2, v3);
df2->v4 = 0;
df2->flag &= ~ME_FACE_SEL;
facepa[cur + 2] = vertpa[v1];
df3->v1 = v1;
df3->v2 = v3;
df3->v3 = v4;
df3->v4 = 0;
df3->flag &= ~ME_FACE_SEL;
}
static void remap_uvs_3_6_9_12(
Mesh *mesh, Mesh *split, int layers_num, int i, int cur, int c0, int c1, int c2, int c3)
{
MTFace *mf, *df1, *df2, *df3;
int l;
for (l = 0; l < layers_num; l++) {
mf = static_cast<MTFace *>(CustomData_get_layer_n_for_write(
&split->fdata_legacy, CD_MTFACE, l, split->totface_legacy));
df1 = mf + cur;
df2 = df1 + 1;
df3 = df1 + 2;
mf = static_cast<MTFace *>(
CustomData_get_layer_n_for_write(&mesh->fdata_legacy, CD_MTFACE, l, mesh->totface_legacy));
mf += i;
copy_v2_v2(df1->uv[0], mf->uv[c0]);
INT_UV(df1->uv[1], c0, c1);
INT_UV(df1->uv[2], c1, c2);
copy_v2_v2(df1->uv[3], mf->uv[c2]);
INT_UV(df2->uv[0], c0, c1);
copy_v2_v2(df2->uv[1], mf->uv[c1]);
INT_UV(df2->uv[2], c1, c2);
copy_v2_v2(df3->uv[0], mf->uv[c0]);
copy_v2_v2(df3->uv[1], mf->uv[c2]);
copy_v2_v2(df3->uv[2], mf->uv[c3]);
}
}
static void remap_faces_5_10(Mesh *mesh,
Mesh *split,
MFace *mf,
int *facepa,
const int *vertpa,
int i,
const blender::Map<blender::OrderedEdge, int> &eh,
int cur,
int v1,
int v2,
int v3,
int v4)
{
MFace *df1 = get_dface(mesh, split, cur, i, mf);
MFace *df2 = get_dface(mesh, split, cur + 1, i, mf);
facepa[cur] = vertpa[v1];
df1->v1 = v1;
df1->v2 = v2;
df1->v3 = GET_ES(v2, v3);
df1->v4 = GET_ES(v1, v4);
df1->flag |= ME_FACE_SEL;
facepa[cur + 1] = vertpa[v3];
df2->v1 = GET_ES(v1, v4);
df2->v2 = GET_ES(v2, v3);
df2->v3 = v3;
df2->v4 = v4;
df2->flag |= ME_FACE_SEL;
}
static void remap_uvs_5_10(
Mesh *mesh, Mesh *split, int layers_num, int i, int cur, int c0, int c1, int c2, int c3)
{
MTFace *mf, *df1, *df2;
int l;
for (l = 0; l < layers_num; l++) {
mf = static_cast<MTFace *>(CustomData_get_layer_n_for_write(
&split->fdata_legacy, CD_MTFACE, l, split->totface_legacy));
df1 = mf + cur;
df2 = df1 + 1;
mf = static_cast<MTFace *>(
CustomData_get_layer_n_for_write(&mesh->fdata_legacy, CD_MTFACE, l, mesh->totface_legacy));
mf += i;
copy_v2_v2(df1->uv[0], mf->uv[c0]);
copy_v2_v2(df1->uv[1], mf->uv[c1]);
INT_UV(df1->uv[2], c1, c2);
INT_UV(df1->uv[3], c0, c3);
INT_UV(df2->uv[0], c0, c3);
INT_UV(df2->uv[1], c1, c2);
copy_v2_v2(df2->uv[2], mf->uv[c2]);
copy_v2_v2(df2->uv[3], mf->uv[c3]);
}
}
static void remap_faces_15(Mesh *mesh,
Mesh *split,
MFace *mf,
int *facepa,
const int *vertpa,
int i,
const blender::Map<blender::OrderedEdge, int> &eh,
int cur,
int v1,
int v2,
int v3,
int v4)
{
MFace *df1 = get_dface(mesh, split, cur, i, mf);
MFace *df2 = get_dface(mesh, split, cur + 1, i, mf);
MFace *df3 = get_dface(mesh, split, cur + 2, i, mf);
MFace *df4 = get_dface(mesh, split, cur + 3, i, mf);
facepa[cur] = vertpa[v1];
df1->v1 = v1;
df1->v2 = GET_ES(v1, v2);
df1->v3 = GET_ES(v1, v3);
df1->v4 = GET_ES(v1, v4);
df1->flag |= ME_FACE_SEL;
facepa[cur + 1] = vertpa[v2];
df2->v1 = GET_ES(v1, v2);
df2->v2 = v2;
df2->v3 = GET_ES(v2, v3);
df2->v4 = GET_ES(v1, v3);
df2->flag |= ME_FACE_SEL;
facepa[cur + 2] = vertpa[v3];
df3->v1 = GET_ES(v1, v3);
df3->v2 = GET_ES(v2, v3);
df3->v3 = v3;
df3->v4 = GET_ES(v3, v4);
df3->flag |= ME_FACE_SEL;
facepa[cur + 3] = vertpa[v4];
df4->v1 = GET_ES(v1, v4);
df4->v2 = GET_ES(v1, v3);
df4->v3 = GET_ES(v3, v4);
df4->v4 = v4;
df4->flag |= ME_FACE_SEL;
}
static void remap_uvs_15(
Mesh *mesh, Mesh *split, int layers_num, int i, int cur, int c0, int c1, int c2, int c3)
{
MTFace *mf, *df1, *df2, *df3, *df4;
int l;
for (l = 0; l < layers_num; l++) {
mf = static_cast<MTFace *>(CustomData_get_layer_n_for_write(
&split->fdata_legacy, CD_MTFACE, l, split->totface_legacy));
df1 = mf + cur;
df2 = df1 + 1;
df3 = df1 + 2;
df4 = df1 + 3;
mf = static_cast<MTFace *>(
CustomData_get_layer_n_for_write(&mesh->fdata_legacy, CD_MTFACE, l, mesh->totface_legacy));
mf += i;
copy_v2_v2(df1->uv[0], mf->uv[c0]);
INT_UV(df1->uv[1], c0, c1);
INT_UV(df1->uv[2], c0, c2);
INT_UV(df1->uv[3], c0, c3);
INT_UV(df2->uv[0], c0, c1);
copy_v2_v2(df2->uv[1], mf->uv[c1]);
INT_UV(df2->uv[2], c1, c2);
INT_UV(df2->uv[3], c0, c2);
INT_UV(df3->uv[0], c0, c2);
INT_UV(df3->uv[1], c1, c2);
copy_v2_v2(df3->uv[2], mf->uv[c2]);
INT_UV(df3->uv[3], c2, c3);
INT_UV(df4->uv[0], c0, c3);
INT_UV(df4->uv[1], c0, c2);
INT_UV(df4->uv[2], c2, c3);
copy_v2_v2(df4->uv[3], mf->uv[c3]);
}
}
static void remap_faces_7_11_13_14(Mesh *mesh,
Mesh *split,
MFace *mf,
int *facepa,
const int *vertpa,
int i,
const blender::Map<blender::OrderedEdge, int> &eh,
int cur,
int v1,
int v2,
int v3,
int v4)
{
MFace *df1 = get_dface(mesh, split, cur, i, mf);
MFace *df2 = get_dface(mesh, split, cur + 1, i, mf);
MFace *df3 = get_dface(mesh, split, cur + 2, i, mf);
facepa[cur] = vertpa[v1];
df1->v1 = v1;
df1->v2 = GET_ES(v1, v2);
df1->v3 = GET_ES(v2, v3);
df1->v4 = GET_ES(v1, v4);
df1->flag |= ME_FACE_SEL;
facepa[cur + 1] = vertpa[v2];
df2->v1 = GET_ES(v1, v2);
df2->v2 = v2;
df2->v3 = GET_ES(v2, v3);
df2->v4 = 0;
df2->flag &= ~ME_FACE_SEL;
facepa[cur + 2] = vertpa[v4];
df3->v1 = GET_ES(v1, v4);
df3->v2 = GET_ES(v2, v3);
df3->v3 = v3;
df3->v4 = v4;
df3->flag |= ME_FACE_SEL;
}
static void remap_uvs_7_11_13_14(
Mesh *mesh, Mesh *split, int layers_num, int i, int cur, int c0, int c1, int c2, int c3)
{
MTFace *mf, *df1, *df2, *df3;
int l;
for (l = 0; l < layers_num; l++) {
mf = static_cast<MTFace *>(CustomData_get_layer_n_for_write(
&split->fdata_legacy, CD_MTFACE, l, split->totface_legacy));
df1 = mf + cur;
df2 = df1 + 1;
df3 = df1 + 2;
mf = static_cast<MTFace *>(
CustomData_get_layer_n_for_write(&mesh->fdata_legacy, CD_MTFACE, l, mesh->totface_legacy));
mf += i;
copy_v2_v2(df1->uv[0], mf->uv[c0]);
INT_UV(df1->uv[1], c0, c1);
INT_UV(df1->uv[2], c1, c2);
INT_UV(df1->uv[3], c0, c3);
INT_UV(df2->uv[0], c0, c1);
copy_v2_v2(df2->uv[1], mf->uv[c1]);
INT_UV(df2->uv[2], c1, c2);
INT_UV(df3->uv[0], c0, c3);
INT_UV(df3->uv[1], c1, c2);
copy_v2_v2(df3->uv[2], mf->uv[c2]);
copy_v2_v2(df3->uv[3], mf->uv[c3]);
}
}
static void remap_faces_19_21_22(Mesh *mesh,
Mesh *split,
MFace *mf,
int *facepa,
const int *vertpa,
int i,
const blender::Map<blender::OrderedEdge, int> &eh,
int cur,
int v1,
int v2,
int v3)
{
MFace *df1 = get_dface(mesh, split, cur, i, mf);
MFace *df2 = get_dface(mesh, split, cur + 1, i, mf);
facepa[cur] = vertpa[v1];
df1->v1 = v1;
df1->v2 = GET_ES(v1, v2);
df1->v3 = GET_ES(v1, v3);
df1->v4 = 0;
df1->flag &= ~ME_FACE_SEL;
facepa[cur + 1] = vertpa[v2];
df2->v1 = GET_ES(v1, v2);
df2->v2 = v2;
df2->v3 = v3;
df2->v4 = GET_ES(v1, v3);
df2->flag |= ME_FACE_SEL;
}
static void remap_uvs_19_21_22(
Mesh *mesh, Mesh *split, int layers_num, int i, int cur, int c0, int c1, int c2)
{
MTFace *mf, *df1, *df2;
int l;
for (l = 0; l < layers_num; l++) {
mf = static_cast<MTFace *>(CustomData_get_layer_n_for_write(
&split->fdata_legacy, CD_MTFACE, l, split->totface_legacy));
df1 = mf + cur;
df2 = df1 + 1;
mf = static_cast<MTFace *>(
CustomData_get_layer_n_for_write(&mesh->fdata_legacy, CD_MTFACE, l, mesh->totface_legacy));
mf += i;
copy_v2_v2(df1->uv[0], mf->uv[c0]);
INT_UV(df1->uv[1], c0, c1);
INT_UV(df1->uv[2], c0, c2);
INT_UV(df2->uv[0], c0, c1);
copy_v2_v2(df2->uv[1], mf->uv[c1]);
copy_v2_v2(df2->uv[2], mf->uv[c2]);
INT_UV(df2->uv[3], c0, c2);
}
}
static void remap_faces_23(Mesh *mesh,
Mesh *split,
MFace *mf,
int *facepa,
const int *vertpa,
int i,
const blender::Map<blender::OrderedEdge, int> &eh,
int cur,
int v1,
int v2,
int v3)
{
MFace *df1 = get_dface(mesh, split, cur, i, mf);
MFace *df2 = get_dface(mesh, split, cur + 1, i, mf);
MFace *df3 = get_dface(mesh, split, cur + 2, i, mf);
facepa[cur] = vertpa[v1];
df1->v1 = v1;
df1->v2 = GET_ES(v1, v2);
df1->v3 = GET_ES(v2, v3);
df1->v4 = GET_ES(v1, v3);
df1->flag |= ME_FACE_SEL;
facepa[cur + 1] = vertpa[v2];
df2->v1 = GET_ES(v1, v2);
df2->v2 = v2;
df2->v3 = GET_ES(v2, v3);
df2->v4 = 0;
df2->flag &= ~ME_FACE_SEL;
facepa[cur + 2] = vertpa[v3];
df3->v1 = GET_ES(v1, v3);
df3->v2 = GET_ES(v2, v3);
df3->v3 = v3;
df3->v4 = 0;
df3->flag &= ~ME_FACE_SEL;
}
static void remap_uvs_23(
Mesh *mesh, Mesh *split, int layers_num, int i, int cur, int c0, int c1, int c2)
{
MTFace *mf, *df1, *df2;
int l;
for (l = 0; l < layers_num; l++) {
mf = static_cast<MTFace *>(CustomData_get_layer_n_for_write(
&split->fdata_legacy, CD_MTFACE, l, split->totface_legacy));
df1 = mf + cur;
df2 = df1 + 1;
mf = static_cast<MTFace *>(
CustomData_get_layer_n_for_write(&mesh->fdata_legacy, CD_MTFACE, l, mesh->totface_legacy));
mf += i;
copy_v2_v2(df1->uv[0], mf->uv[c0]);
INT_UV(df1->uv[1], c0, c1);
INT_UV(df1->uv[2], c1, c2);
INT_UV(df1->uv[3], c0, c2);
INT_UV(df2->uv[0], c0, c1);
copy_v2_v2(df2->uv[1], mf->uv[c1]);
INT_UV(df2->uv[2], c1, c2);
INT_UV(df2->uv[0], c0, c2);
INT_UV(df2->uv[1], c1, c2);
copy_v2_v2(df2->uv[2], mf->uv[c2]);
}
}
static Mesh *cutEdges(ExplodeModifierData *emd, Mesh *mesh)
{
Mesh *split_m;
MFace *mf = nullptr, *df1 = nullptr;
MFace *mface = static_cast<MFace *>(
CustomData_get_layer_for_write(&mesh->fdata_legacy, CD_MFACE, mesh->totface_legacy));
Mesh: Move positions to a generic attribute **Changes** As described in T93602, this patch removes all use of the `MVert` struct, replacing it with a generic named attribute with the name `"position"`, consistent with other geometry types. Variable names have been changed from `verts` to `positions`, to align with the attribute name and the more generic design (positions are not vertices, they are just an attribute stored on the point domain). This change is made possible by previous commits that moved all other data out of `MVert` to runtime data or other generic attributes. What remains is mostly a simple type change. Though, the type still shows up 859 times, so the patch is quite large. One compromise is that now `CD_MASK_BAREMESH` now contains `CD_PROP_FLOAT3`. With the general move towards generic attributes over custom data types, we are removing use of these type masks anyway. **Benefits** The most obvious benefit is reduced memory usage and the benefits that brings in memory-bound situations. `float3` is only 3 bytes, in comparison to `MVert` which was 4. When there are millions of vertices this starts to matter more. The other benefits come from using a more generic type. Instead of writing algorithms specifically for `MVert`, code can just use arrays of vectors. This will allow eliminating many temporary arrays or wrappers used to extract positions. Many possible improvements aren't implemented in this patch, though I did switch simplify or remove the process of creating temporary position arrays in a few places. The design clarity that "positions are just another attribute" brings allows removing explicit copying of vertices in some procedural operations-- they are just processed like most other attributes. **Performance** This touches so many areas that it's hard to benchmark exhaustively, but I observed some areas as examples. * The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster. * The Spring splash screen went from ~4.3 to ~4.5 fps. * The subdivision surface modifier/node was slightly faster RNA access through Python may be slightly slower, since now we need a name lookup instead of just a custom data type lookup for each index. **Future Improvements** * Remove uses of "vert_coords" functions: * `BKE_mesh_vert_coords_alloc` * `BKE_mesh_vert_coords_get` * `BKE_mesh_vert_coords_apply{_with_mat4}` * Remove more hidden copying of positions * General simplification now possible in many areas * Convert more code to C++ to use `float3` instead of `float[3]` * Currently `reinterpret_cast` is used for those C-API functions Differential Revision: https://developer.blender.org/D15982
2023-01-10 00:10:43 -05:00
float *dupve;
int totvert = mesh->verts_num;
int totface = mesh->totface_legacy;
int *facesplit = MEM_calloc_arrayN<int>(totface, __func__);
int *vertpa = MEM_calloc_arrayN<int>(totvert, __func__);
int *facepa = emd->facepa;
int *fs, totfsplit = 0, curdupface = 0;
int i, v1, v2, v3, v4, v[4] = {0, 0, 0, 0}, /* To quite gcc barking... */
uv[4] = {0, 0, 0, 0}; /* To quite gcc barking... */
int layers_num;
int totesplit = totvert;
blender::Map<blender::OrderedEdge, int> edgehash;
/* recreate vertpa from facepa calculation */
for (i = 0, mf = mface; i < totface; i++, mf++) {
vertpa[mf->v1] = facepa[i];
vertpa[mf->v2] = facepa[i];
vertpa[mf->v3] = facepa[i];
if (mf->v4) {
vertpa[mf->v4] = facepa[i];
}
}
/* mark edges for splitting and how to split faces */
for (i = 0, mf = mface, fs = facesplit; i < totface; i++, mf++, fs++) {
v1 = vertpa[mf->v1];
v2 = vertpa[mf->v2];
v3 = vertpa[mf->v3];
if (v1 != v2) {
edgehash.lookup_or_add_cb({mf->v1, mf->v2}, [&]() { return totesplit++; });
(*fs) |= 1;
}
if (v2 != v3) {
edgehash.lookup_or_add_cb({mf->v2, mf->v3}, [&]() { return totesplit++; });
(*fs) |= 2;
}
if (mf->v4) {
v4 = vertpa[mf->v4];
if (v3 != v4) {
edgehash.lookup_or_add_cb({mf->v3, mf->v4}, [&]() { return totesplit++; });
(*fs) |= 4;
}
if (v1 != v4) {
edgehash.lookup_or_add_cb({mf->v1, mf->v4}, [&]() { return totesplit++; });
(*fs) |= 8;
}
/* mark center vertex as a fake edge split */
if (*fs == 15) {
edgehash.lookup_or_add_cb({mf->v1, mf->v3}, [&]() { return totesplit++; });
}
}
else {
(*fs) |= 16; /* mark face as tri */
if (v1 != v3) {
edgehash.lookup_or_add_cb({mf->v1, mf->v3}, [&]() { return totesplit++; });
(*fs) |= 4;
}
}
}
/* count splits & create indexes for new verts */
/* count new faces due to splitting */
for (i = 0, fs = facesplit; i < totface; i++, fs++) {
totfsplit += add_faces[*fs];
}
split_m = BKE_mesh_new_nomain_from_template_ex(
mesh, totesplit, 0, totface + totfsplit, 0, 0, CD_MASK_EVERYTHING);
layers_num = CustomData_number_of_layers(&split_m->fdata_legacy, CD_MTFACE);
blender::MutableSpan<blender::float3> split_m_positions = split_m->vert_positions_for_write();
Mesh: Remove redundant custom data pointers For copy-on-write, we want to share attribute arrays between meshes where possible. Mutable pointers like `Mesh.mvert` make that difficult by making ownership vague. They also make code more complex by adding redundancy. The simplest solution is just removing them and retrieving layers from `CustomData` as needed. Similar changes have already been applied to curves and point clouds (e9f82d3dc7ee, 410a6efb747f). Removing use of the pointers generally makes code more obvious and more reusable. Mesh data is now accessed with a C++ API (`Mesh::edges()` or `Mesh::edges_for_write()`), and a C API (`BKE_mesh_edges(mesh)`). The CoW changes this commit makes possible are described in T95845 and T95842, and started in D14139 and D14140. The change also simplifies the ongoing mesh struct-of-array refactors from T95965. **RNA/Python Access Performance** Theoretically, accessing mesh elements with the RNA API may become slower, since the layer needs to be found on every random access. However, overhead is already high enough that this doesn't make a noticible differenc, and performance is actually improved in some cases. Random access can be up to 10% faster, but other situations might be a bit slower. Generally using `foreach_get/set` are the best way to improve performance. See the differential revision for more discussion about Python performance. Cycles has been updated to use raw pointers and the internal Blender mesh types, mostly because there is no sense in having this overhead when it's already compiled with Blender. In my tests this roughly halves the Cycles mesh creation time (0.19s to 0.10s for a 1 million face grid). Differential Revision: https://developer.blender.org/D15488
2022-09-05 11:56:34 -05:00
/* copy new faces & verts (is it really this painful with custom data??) */
for (i = 0; i < totvert; i++) {
CustomData_copy_data(&mesh->vert_data, &split_m->vert_data, i, i, 1);
}
/* override original facepa (original pointer is saved in caller function) */
/* TODO(@ideasman42): `(totfsplit * 2)` over allocation is used since the quads are
* later interpreted as triangles, for this to work right I think we probably
* have to stop using tessface. */
facepa = MEM_calloc_arrayN<int>(size_t(totface) + (size_t(totfsplit) * 2), __func__);
// memcpy(facepa, emd->facepa, totface*sizeof(int));
emd->facepa = facepa;
/* create new verts */
for (const auto [edge, esplit] : edgehash.items()) {
const int ed_v1 = edge.v_low;
const int ed_v2 = edge.v_high;
CustomData_free_elem(&split_m->vert_data, esplit, 1);
CustomData_copy_data(&split_m->vert_data, &split_m->vert_data, ed_v2, esplit, 1);
Mesh: Move positions to a generic attribute **Changes** As described in T93602, this patch removes all use of the `MVert` struct, replacing it with a generic named attribute with the name `"position"`, consistent with other geometry types. Variable names have been changed from `verts` to `positions`, to align with the attribute name and the more generic design (positions are not vertices, they are just an attribute stored on the point domain). This change is made possible by previous commits that moved all other data out of `MVert` to runtime data or other generic attributes. What remains is mostly a simple type change. Though, the type still shows up 859 times, so the patch is quite large. One compromise is that now `CD_MASK_BAREMESH` now contains `CD_PROP_FLOAT3`. With the general move towards generic attributes over custom data types, we are removing use of these type masks anyway. **Benefits** The most obvious benefit is reduced memory usage and the benefits that brings in memory-bound situations. `float3` is only 3 bytes, in comparison to `MVert` which was 4. When there are millions of vertices this starts to matter more. The other benefits come from using a more generic type. Instead of writing algorithms specifically for `MVert`, code can just use arrays of vectors. This will allow eliminating many temporary arrays or wrappers used to extract positions. Many possible improvements aren't implemented in this patch, though I did switch simplify or remove the process of creating temporary position arrays in a few places. The design clarity that "positions are just another attribute" brings allows removing explicit copying of vertices in some procedural operations-- they are just processed like most other attributes. **Performance** This touches so many areas that it's hard to benchmark exhaustively, but I observed some areas as examples. * The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster. * The Spring splash screen went from ~4.3 to ~4.5 fps. * The subdivision surface modifier/node was slightly faster RNA access through Python may be slightly slower, since now we need a name lookup instead of just a custom data type lookup for each index. **Future Improvements** * Remove uses of "vert_coords" functions: * `BKE_mesh_vert_coords_alloc` * `BKE_mesh_vert_coords_get` * `BKE_mesh_vert_coords_apply{_with_mat4}` * Remove more hidden copying of positions * General simplification now possible in many areas * Convert more code to C++ to use `float3` instead of `float[3]` * Currently `reinterpret_cast` is used for those C-API functions Differential Revision: https://developer.blender.org/D15982
2023-01-10 00:10:43 -05:00
dupve = split_m_positions[esplit];
copy_v3_v3(dupve, split_m_positions[ed_v2]);
Mesh: Move positions to a generic attribute **Changes** As described in T93602, this patch removes all use of the `MVert` struct, replacing it with a generic named attribute with the name `"position"`, consistent with other geometry types. Variable names have been changed from `verts` to `positions`, to align with the attribute name and the more generic design (positions are not vertices, they are just an attribute stored on the point domain). This change is made possible by previous commits that moved all other data out of `MVert` to runtime data or other generic attributes. What remains is mostly a simple type change. Though, the type still shows up 859 times, so the patch is quite large. One compromise is that now `CD_MASK_BAREMESH` now contains `CD_PROP_FLOAT3`. With the general move towards generic attributes over custom data types, we are removing use of these type masks anyway. **Benefits** The most obvious benefit is reduced memory usage and the benefits that brings in memory-bound situations. `float3` is only 3 bytes, in comparison to `MVert` which was 4. When there are millions of vertices this starts to matter more. The other benefits come from using a more generic type. Instead of writing algorithms specifically for `MVert`, code can just use arrays of vectors. This will allow eliminating many temporary arrays or wrappers used to extract positions. Many possible improvements aren't implemented in this patch, though I did switch simplify or remove the process of creating temporary position arrays in a few places. The design clarity that "positions are just another attribute" brings allows removing explicit copying of vertices in some procedural operations-- they are just processed like most other attributes. **Performance** This touches so many areas that it's hard to benchmark exhaustively, but I observed some areas as examples. * The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster. * The Spring splash screen went from ~4.3 to ~4.5 fps. * The subdivision surface modifier/node was slightly faster RNA access through Python may be slightly slower, since now we need a name lookup instead of just a custom data type lookup for each index. **Future Improvements** * Remove uses of "vert_coords" functions: * `BKE_mesh_vert_coords_alloc` * `BKE_mesh_vert_coords_get` * `BKE_mesh_vert_coords_apply{_with_mat4}` * Remove more hidden copying of positions * General simplification now possible in many areas * Convert more code to C++ to use `float3` instead of `float[3]` * Currently `reinterpret_cast` is used for those C-API functions Differential Revision: https://developer.blender.org/D15982
2023-01-10 00:10:43 -05:00
mid_v3_v3v3(dupve, dupve, split_m_positions[ed_v1]);
}
/* create new faces */
curdupface = 0; //=totface;
// curdupin=totesplit;
for (i = 0, fs = facesplit; i < totface; i++, fs++) {
Mesh: Remove redundant custom data pointers For copy-on-write, we want to share attribute arrays between meshes where possible. Mutable pointers like `Mesh.mvert` make that difficult by making ownership vague. They also make code more complex by adding redundancy. The simplest solution is just removing them and retrieving layers from `CustomData` as needed. Similar changes have already been applied to curves and point clouds (e9f82d3dc7ee, 410a6efb747f). Removing use of the pointers generally makes code more obvious and more reusable. Mesh data is now accessed with a C++ API (`Mesh::edges()` or `Mesh::edges_for_write()`), and a C API (`BKE_mesh_edges(mesh)`). The CoW changes this commit makes possible are described in T95845 and T95842, and started in D14139 and D14140. The change also simplifies the ongoing mesh struct-of-array refactors from T95965. **RNA/Python Access Performance** Theoretically, accessing mesh elements with the RNA API may become slower, since the layer needs to be found on every random access. However, overhead is already high enough that this doesn't make a noticible differenc, and performance is actually improved in some cases. Random access can be up to 10% faster, but other situations might be a bit slower. Generally using `foreach_get/set` are the best way to improve performance. See the differential revision for more discussion about Python performance. Cycles has been updated to use raw pointers and the internal Blender mesh types, mostly because there is no sense in having this overhead when it's already compiled with Blender. In my tests this roughly halves the Cycles mesh creation time (0.19s to 0.10s for a 1 million face grid). Differential Revision: https://developer.blender.org/D15488
2022-09-05 11:56:34 -05:00
mf = &mface[i];
switch (*fs) {
case 3:
case 10:
case 11:
case 15:
SET_VERTS(1, 2, 3, 4);
break;
case 5:
case 6:
case 7:
SET_VERTS(2, 3, 4, 1);
break;
case 9:
case 13:
SET_VERTS(4, 1, 2, 3);
break;
case 12:
case 14:
SET_VERTS(3, 4, 1, 2);
break;
case 21:
case 23:
SET_VERTS(1, 2, 3, 4);
break;
case 19:
SET_VERTS(2, 3, 1, 4);
break;
case 22:
SET_VERTS(3, 1, 2, 4);
break;
}
switch (*fs) {
case 3:
case 6:
case 9:
case 12:
remap_faces_3_6_9_12(
mesh, split_m, mf, facepa, vertpa, i, edgehash, curdupface, v[0], v[1], v[2], v[3]);
if (layers_num) {
remap_uvs_3_6_9_12(mesh, split_m, layers_num, i, curdupface, uv[0], uv[1], uv[2], uv[3]);
}
break;
case 5:
case 10:
remap_faces_5_10(
mesh, split_m, mf, facepa, vertpa, i, edgehash, curdupface, v[0], v[1], v[2], v[3]);
if (layers_num) {
remap_uvs_5_10(mesh, split_m, layers_num, i, curdupface, uv[0], uv[1], uv[2], uv[3]);
}
break;
case 15:
remap_faces_15(
mesh, split_m, mf, facepa, vertpa, i, edgehash, curdupface, v[0], v[1], v[2], v[3]);
if (layers_num) {
remap_uvs_15(mesh, split_m, layers_num, i, curdupface, uv[0], uv[1], uv[2], uv[3]);
}
break;
case 7:
case 11:
case 13:
case 14:
remap_faces_7_11_13_14(
mesh, split_m, mf, facepa, vertpa, i, edgehash, curdupface, v[0], v[1], v[2], v[3]);
if (layers_num) {
remap_uvs_7_11_13_14(
mesh, split_m, layers_num, i, curdupface, uv[0], uv[1], uv[2], uv[3]);
}
break;
case 19:
case 21:
case 22:
remap_faces_19_21_22(
mesh, split_m, mf, facepa, vertpa, i, edgehash, curdupface, v[0], v[1], v[2]);
if (layers_num) {
remap_uvs_19_21_22(mesh, split_m, layers_num, i, curdupface, uv[0], uv[1], uv[2]);
}
break;
case 23:
remap_faces_23(
mesh, split_m, mf, facepa, vertpa, i, edgehash, curdupface, v[0], v[1], v[2]);
if (layers_num) {
remap_uvs_23(mesh, split_m, layers_num, i, curdupface, uv[0], uv[1], uv[2]);
}
break;
case 0:
case 16:
df1 = get_dface(mesh, split_m, curdupface, i, mf);
facepa[curdupface] = vertpa[mf->v1];
if (df1->v4) {
df1->flag |= ME_FACE_SEL;
}
else {
df1->flag &= ~ME_FACE_SEL;
}
break;
}
curdupface += add_faces[*fs] + 1;
}
MFace *split_mface = static_cast<MFace *>(
CustomData_get_layer_for_write(&split_m->fdata_legacy, CD_MFACE, split_m->totface_legacy));
for (i = 0; i < curdupface; i++) {
Mesh: Remove redundant custom data pointers For copy-on-write, we want to share attribute arrays between meshes where possible. Mutable pointers like `Mesh.mvert` make that difficult by making ownership vague. They also make code more complex by adding redundancy. The simplest solution is just removing them and retrieving layers from `CustomData` as needed. Similar changes have already been applied to curves and point clouds (e9f82d3dc7ee, 410a6efb747f). Removing use of the pointers generally makes code more obvious and more reusable. Mesh data is now accessed with a C++ API (`Mesh::edges()` or `Mesh::edges_for_write()`), and a C API (`BKE_mesh_edges(mesh)`). The CoW changes this commit makes possible are described in T95845 and T95842, and started in D14139 and D14140. The change also simplifies the ongoing mesh struct-of-array refactors from T95965. **RNA/Python Access Performance** Theoretically, accessing mesh elements with the RNA API may become slower, since the layer needs to be found on every random access. However, overhead is already high enough that this doesn't make a noticible differenc, and performance is actually improved in some cases. Random access can be up to 10% faster, but other situations might be a bit slower. Generally using `foreach_get/set` are the best way to improve performance. See the differential revision for more discussion about Python performance. Cycles has been updated to use raw pointers and the internal Blender mesh types, mostly because there is no sense in having this overhead when it's already compiled with Blender. In my tests this roughly halves the Cycles mesh creation time (0.19s to 0.10s for a 1 million face grid). Differential Revision: https://developer.blender.org/D15488
2022-09-05 11:56:34 -05:00
mf = &split_mface[i];
BKE_mesh_mface_index_validate(
mf, &split_m->fdata_legacy, i, ((mf->flag & ME_FACE_SEL) ? 4 : 3));
}
MEM_freeN(facesplit);
MEM_freeN(vertpa);
BKE_mesh_calc_edges_tessface(split_m);
BKE_mesh_convert_mfaces_to_mpolys(split_m);
Mesh: Replace MPoly struct with offset indices Implements #95967. Currently the `MPoly` struct is 12 bytes, and stores the index of a face's first corner and the number of corners/verts/edges. Polygons and corners are always created in order by Blender, meaning each face's corners will be after the previous face's corners. We can take advantage of this fact and eliminate the redundancy in mesh face storage by only storing a single integer corner offset for each face. The size of the face is then encoded by the offset of the next face. The size of a single integer is 4 bytes, so this reduces memory usage by 3 times. The same method is used for `CurvesGeometry`, so Blender already has an abstraction to simplify using these offsets called `OffsetIndices`. This class is used to easily retrieve a range of corner indices for each face. This also gives the opportunity for sharing some logic with curves. Another benefit of the change is that the offsets and sizes stored in `MPoly` can no longer disagree with each other. Storing faces in the order of their corners can simplify some code too. Face/polygon variables now use the `IndexRange` type, which comes with quite a few utilities that can simplify code. Some: - The offset integer array has to be one longer than the face count to avoid a branch for every face, which means the data is no longer part of the mesh's `CustomData`. - We lose the ability to "reference" an original mesh's offset array until more reusable CoW from #104478 is committed. That will be added in a separate commit. - Since they aren't part of `CustomData`, poly offsets often have to be copied manually. - To simplify using `OffsetIndices` in many places, some functions and structs in headers were moved to only compile in C++. - All meshes created by Blender use the same order for faces and face corners, but just in case, meshes with mismatched order are fixed by versioning code. - `MeshPolygon.totloop` is no longer editable in RNA. This API break is necessary here unfortunately. It should be worth it in 3.6, since that's the best way to allow loading meshes from 4.0, which is important for an LTS version. Pull Request: https://projects.blender.org/blender/blender/pulls/105938
2023-04-04 20:39:28 +02:00
BKE_mesh_legacy_convert_polys_to_offsets(split_m);
return split_m;
}
static Mesh *explodeMesh(ExplodeModifierData *emd,
ParticleSystemModifierData *psmd,
const ModifierEvalContext *ctx,
Scene *scene,
Mesh *to_explode)
{
Mesh *explode, *mesh = to_explode;
MFace *mf = nullptr, *mface;
2022-06-03 13:39:37 +10:00
// ParticleSettings *part=psmd->psys->part; /* UNUSED */
ParticleSimulationData sim = {nullptr};
ParticleData *pa = nullptr, *pars = psmd->psys->particles;
ParticleKey state, birth;
float *vertco = nullptr, imat[4][4];
float rot[4];
float ctime;
2022-06-03 13:39:37 +10:00
// float timestep;
const int *facepa = emd->facepa;
int totdup = 0, totvert = 0, totface = 0, totpart = 0, delface = 0;
int i, u;
uint mindex = 0;
totface = mesh->totface_legacy;
totvert = mesh->verts_num;
mface = static_cast<MFace *>(
CustomData_get_layer_for_write(&mesh->fdata_legacy, CD_MFACE, mesh->totface_legacy));
totpart = psmd->psys->totpart;
sim.depsgraph = ctx->depsgraph;
sim.scene = scene;
sim.ob = ctx->object;
sim.psys = psmd->psys;
sim.psmd = psmd;
2022-06-03 13:39:37 +10:00
// timestep = psys_get_timestep(&sim);
ctime = BKE_scene_ctime_get(scene);
/* hash table for vertex <-> particle relations */
blender::Map<blender::OrderedEdge, int> vertpahash;
for (i = 0; i < totface; i++) {
if (facepa[i] != totpart) {
pa = pars + facepa[i];
if ((pa->alive == PARS_UNBORN && (emd->flag & eExplodeFlag_Unborn) == 0) ||
(pa->alive == PARS_ALIVE && (emd->flag & eExplodeFlag_Alive) == 0) ||
(pa->alive == PARS_DEAD && (emd->flag & eExplodeFlag_Dead) == 0))
{
delface++;
continue;
}
}
else {
pa = nullptr;
}
/* do mindex + totvert to ensure the vertex index to be the first. */
if (pa == nullptr || ctime < pa->time) {
mindex = totvert + totpart;
}
else {
mindex = totvert + facepa[i];
}
mf = &mface[i];
/* set face vertices to exist in particle group */
vertpahash.lookup_or_add_cb({mf->v1, mindex}, [&]() { return totdup++; });
vertpahash.lookup_or_add_cb({mf->v2, mindex}, [&]() { return totdup++; });
vertpahash.lookup_or_add_cb({mf->v3, mindex}, [&]() { return totdup++; });
if (mf->v4) {
vertpahash.lookup_or_add_cb({mf->v4, mindex}, [&]() { return totdup++; });
}
}
/* the final duplicated vertices */
explode = BKE_mesh_new_nomain_from_template_ex(
mesh, totdup, 0, totface - delface, 0, 0, CD_MASK_EVERYTHING);
MTFace *mtface = static_cast<MTFace *>(CustomData_get_layer_named_for_write(
&explode->fdata_legacy, CD_MTFACE, emd->uvname, explode->totface_legacy));
/* getting back to object space */
invert_m4_m4(imat, ctx->object->object_to_world().ptr());
psys_sim_data_init(&sim);
const blender::Span<blender::float3> positions = mesh->vert_positions();
blender::MutableSpan<blender::float3> explode_positions = explode->vert_positions_for_write();
Mesh: Remove redundant custom data pointers For copy-on-write, we want to share attribute arrays between meshes where possible. Mutable pointers like `Mesh.mvert` make that difficult by making ownership vague. They also make code more complex by adding redundancy. The simplest solution is just removing them and retrieving layers from `CustomData` as needed. Similar changes have already been applied to curves and point clouds (e9f82d3dc7ee, 410a6efb747f). Removing use of the pointers generally makes code more obvious and more reusable. Mesh data is now accessed with a C++ API (`Mesh::edges()` or `Mesh::edges_for_write()`), and a C API (`BKE_mesh_edges(mesh)`). The CoW changes this commit makes possible are described in T95845 and T95842, and started in D14139 and D14140. The change also simplifies the ongoing mesh struct-of-array refactors from T95965. **RNA/Python Access Performance** Theoretically, accessing mesh elements with the RNA API may become slower, since the layer needs to be found on every random access. However, overhead is already high enough that this doesn't make a noticible differenc, and performance is actually improved in some cases. Random access can be up to 10% faster, but other situations might be a bit slower. Generally using `foreach_get/set` are the best way to improve performance. See the differential revision for more discussion about Python performance. Cycles has been updated to use raw pointers and the internal Blender mesh types, mostly because there is no sense in having this overhead when it's already compiled with Blender. In my tests this roughly halves the Cycles mesh creation time (0.19s to 0.10s for a 1 million face grid). Differential Revision: https://developer.blender.org/D15488
2022-09-05 11:56:34 -05:00
for (const auto [edge, v] : vertpahash.items()) {
int ed_v1 = edge.v_low;
int ed_v2 = edge.v_high;
ed_v2 -= totvert;
Mesh: Move positions to a generic attribute **Changes** As described in T93602, this patch removes all use of the `MVert` struct, replacing it with a generic named attribute with the name `"position"`, consistent with other geometry types. Variable names have been changed from `verts` to `positions`, to align with the attribute name and the more generic design (positions are not vertices, they are just an attribute stored on the point domain). This change is made possible by previous commits that moved all other data out of `MVert` to runtime data or other generic attributes. What remains is mostly a simple type change. Though, the type still shows up 859 times, so the patch is quite large. One compromise is that now `CD_MASK_BAREMESH` now contains `CD_PROP_FLOAT3`. With the general move towards generic attributes over custom data types, we are removing use of these type masks anyway. **Benefits** The most obvious benefit is reduced memory usage and the benefits that brings in memory-bound situations. `float3` is only 3 bytes, in comparison to `MVert` which was 4. When there are millions of vertices this starts to matter more. The other benefits come from using a more generic type. Instead of writing algorithms specifically for `MVert`, code can just use arrays of vectors. This will allow eliminating many temporary arrays or wrappers used to extract positions. Many possible improvements aren't implemented in this patch, though I did switch simplify or remove the process of creating temporary position arrays in a few places. The design clarity that "positions are just another attribute" brings allows removing explicit copying of vertices in some procedural operations-- they are just processed like most other attributes. **Performance** This touches so many areas that it's hard to benchmark exhaustively, but I observed some areas as examples. * The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster. * The Spring splash screen went from ~4.3 to ~4.5 fps. * The subdivision surface modifier/node was slightly faster RNA access through Python may be slightly slower, since now we need a name lookup instead of just a custom data type lookup for each index. **Future Improvements** * Remove uses of "vert_coords" functions: * `BKE_mesh_vert_coords_alloc` * `BKE_mesh_vert_coords_get` * `BKE_mesh_vert_coords_apply{_with_mat4}` * Remove more hidden copying of positions * General simplification now possible in many areas * Convert more code to C++ to use `float3` instead of `float[3]` * Currently `reinterpret_cast` is used for those C-API functions Differential Revision: https://developer.blender.org/D15982
2023-01-10 00:10:43 -05:00
copy_v3_v3(explode_positions[v], positions[ed_v1]);
CustomData_copy_data(&mesh->vert_data, &explode->vert_data, ed_v1, v, 1);
Mesh: Move positions to a generic attribute **Changes** As described in T93602, this patch removes all use of the `MVert` struct, replacing it with a generic named attribute with the name `"position"`, consistent with other geometry types. Variable names have been changed from `verts` to `positions`, to align with the attribute name and the more generic design (positions are not vertices, they are just an attribute stored on the point domain). This change is made possible by previous commits that moved all other data out of `MVert` to runtime data or other generic attributes. What remains is mostly a simple type change. Though, the type still shows up 859 times, so the patch is quite large. One compromise is that now `CD_MASK_BAREMESH` now contains `CD_PROP_FLOAT3`. With the general move towards generic attributes over custom data types, we are removing use of these type masks anyway. **Benefits** The most obvious benefit is reduced memory usage and the benefits that brings in memory-bound situations. `float3` is only 3 bytes, in comparison to `MVert` which was 4. When there are millions of vertices this starts to matter more. The other benefits come from using a more generic type. Instead of writing algorithms specifically for `MVert`, code can just use arrays of vectors. This will allow eliminating many temporary arrays or wrappers used to extract positions. Many possible improvements aren't implemented in this patch, though I did switch simplify or remove the process of creating temporary position arrays in a few places. The design clarity that "positions are just another attribute" brings allows removing explicit copying of vertices in some procedural operations-- they are just processed like most other attributes. **Performance** This touches so many areas that it's hard to benchmark exhaustively, but I observed some areas as examples. * The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster. * The Spring splash screen went from ~4.3 to ~4.5 fps. * The subdivision surface modifier/node was slightly faster RNA access through Python may be slightly slower, since now we need a name lookup instead of just a custom data type lookup for each index. **Future Improvements** * Remove uses of "vert_coords" functions: * `BKE_mesh_vert_coords_alloc` * `BKE_mesh_vert_coords_get` * `BKE_mesh_vert_coords_apply{_with_mat4}` * Remove more hidden copying of positions * General simplification now possible in many areas * Convert more code to C++ to use `float3` instead of `float[3]` * Currently `reinterpret_cast` is used for those C-API functions Differential Revision: https://developer.blender.org/D15982
2023-01-10 00:10:43 -05:00
copy_v3_v3(explode_positions[v], positions[ed_v1]);
if (ed_v2 != totpart) {
/* get particle */
pa = pars + ed_v2;
psys_get_birth_coords(&sim, pa, &birth, 0, 0);
state.time = ctime;
psys_get_particle_state(&sim, ed_v2, &state, true);
Mesh: Move positions to a generic attribute **Changes** As described in T93602, this patch removes all use of the `MVert` struct, replacing it with a generic named attribute with the name `"position"`, consistent with other geometry types. Variable names have been changed from `verts` to `positions`, to align with the attribute name and the more generic design (positions are not vertices, they are just an attribute stored on the point domain). This change is made possible by previous commits that moved all other data out of `MVert` to runtime data or other generic attributes. What remains is mostly a simple type change. Though, the type still shows up 859 times, so the patch is quite large. One compromise is that now `CD_MASK_BAREMESH` now contains `CD_PROP_FLOAT3`. With the general move towards generic attributes over custom data types, we are removing use of these type masks anyway. **Benefits** The most obvious benefit is reduced memory usage and the benefits that brings in memory-bound situations. `float3` is only 3 bytes, in comparison to `MVert` which was 4. When there are millions of vertices this starts to matter more. The other benefits come from using a more generic type. Instead of writing algorithms specifically for `MVert`, code can just use arrays of vectors. This will allow eliminating many temporary arrays or wrappers used to extract positions. Many possible improvements aren't implemented in this patch, though I did switch simplify or remove the process of creating temporary position arrays in a few places. The design clarity that "positions are just another attribute" brings allows removing explicit copying of vertices in some procedural operations-- they are just processed like most other attributes. **Performance** This touches so many areas that it's hard to benchmark exhaustively, but I observed some areas as examples. * The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster. * The Spring splash screen went from ~4.3 to ~4.5 fps. * The subdivision surface modifier/node was slightly faster RNA access through Python may be slightly slower, since now we need a name lookup instead of just a custom data type lookup for each index. **Future Improvements** * Remove uses of "vert_coords" functions: * `BKE_mesh_vert_coords_alloc` * `BKE_mesh_vert_coords_get` * `BKE_mesh_vert_coords_apply{_with_mat4}` * Remove more hidden copying of positions * General simplification now possible in many areas * Convert more code to C++ to use `float3` instead of `float[3]` * Currently `reinterpret_cast` is used for those C-API functions Differential Revision: https://developer.blender.org/D15982
2023-01-10 00:10:43 -05:00
vertco = explode_positions[v];
mul_m4_v3(ctx->object->object_to_world().ptr(), vertco);
sub_v3_v3(vertco, birth.co);
/* apply rotation, size & location */
sub_qt_qtqt(rot, state.rot, birth.rot);
mul_qt_v3(rot, vertco);
if (emd->flag & eExplodeFlag_PaSize) {
mul_v3_fl(vertco, pa->size);
}
add_v3_v3(vertco, state.co);
mul_m4_v3(imat, vertco);
}
else {
pa = nullptr;
}
}
/* Map new vertices to faces. */
MFace *explode_mface = static_cast<MFace *>(
CustomData_get_layer_for_write(&explode->fdata_legacy, CD_MFACE, explode->totface_legacy));
for (i = 0, u = 0; i < totface; i++) {
MFace source;
int orig_v4;
if (facepa[i] != totpart) {
pa = pars + facepa[i];
if (pa->alive == PARS_UNBORN && (emd->flag & eExplodeFlag_Unborn) == 0) {
continue;
}
if (pa->alive == PARS_ALIVE && (emd->flag & eExplodeFlag_Alive) == 0) {
continue;
}
if (pa->alive == PARS_DEAD && (emd->flag & eExplodeFlag_Dead) == 0) {
continue;
}
}
else {
pa = nullptr;
}
Mesh: Remove redundant custom data pointers For copy-on-write, we want to share attribute arrays between meshes where possible. Mutable pointers like `Mesh.mvert` make that difficult by making ownership vague. They also make code more complex by adding redundancy. The simplest solution is just removing them and retrieving layers from `CustomData` as needed. Similar changes have already been applied to curves and point clouds (e9f82d3dc7ee, 410a6efb747f). Removing use of the pointers generally makes code more obvious and more reusable. Mesh data is now accessed with a C++ API (`Mesh::edges()` or `Mesh::edges_for_write()`), and a C API (`BKE_mesh_edges(mesh)`). The CoW changes this commit makes possible are described in T95845 and T95842, and started in D14139 and D14140. The change also simplifies the ongoing mesh struct-of-array refactors from T95965. **RNA/Python Access Performance** Theoretically, accessing mesh elements with the RNA API may become slower, since the layer needs to be found on every random access. However, overhead is already high enough that this doesn't make a noticible differenc, and performance is actually improved in some cases. Random access can be up to 10% faster, but other situations might be a bit slower. Generally using `foreach_get/set` are the best way to improve performance. See the differential revision for more discussion about Python performance. Cycles has been updated to use raw pointers and the internal Blender mesh types, mostly because there is no sense in having this overhead when it's already compiled with Blender. In my tests this roughly halves the Cycles mesh creation time (0.19s to 0.10s for a 1 million face grid). Differential Revision: https://developer.blender.org/D15488
2022-09-05 11:56:34 -05:00
source = mface[i];
mf = &explode_mface[u];
2018-06-17 17:04:27 +02:00
orig_v4 = source.v4;
/* Same as above in the first loop over mesh's faces. */
if (pa == nullptr || ctime < pa->time) {
mindex = totvert + totpart;
}
else {
mindex = totvert + facepa[i];
}
source.v1 = edgecut_get(vertpahash, source.v1, mindex);
source.v2 = edgecut_get(vertpahash, source.v2, mindex);
source.v3 = edgecut_get(vertpahash, source.v3, mindex);
if (source.v4) {
source.v4 = edgecut_get(vertpahash, source.v4, mindex);
}
CustomData_copy_data(&mesh->fdata_legacy, &explode->fdata_legacy, i, u, 1);
*mf = source;
/* override uv channel for particle age */
if (mtface) {
float age = (pa != nullptr) ? (ctime - pa->time) / pa->lifetime : 0.0f;
/* Clamp to this range to avoid flipping to the other side of the coordinates. */
CLAMP(age, 0.001f, 0.999f);
MTFace *mtf = mtface + u;
mtf->uv[0][0] = mtf->uv[1][0] = mtf->uv[2][0] = mtf->uv[3][0] = age;
mtf->uv[0][1] = mtf->uv[1][1] = mtf->uv[2][1] = mtf->uv[3][1] = 0.5f;
}
BKE_mesh_mface_index_validate(mf, &explode->fdata_legacy, u, (orig_v4 ? 4 : 3));
u++;
}
/* finalization */
BKE_mesh_calc_edges_tessface(explode);
BKE_mesh_convert_mfaces_to_mpolys(explode);
Mesh: Replace MPoly struct with offset indices Implements #95967. Currently the `MPoly` struct is 12 bytes, and stores the index of a face's first corner and the number of corners/verts/edges. Polygons and corners are always created in order by Blender, meaning each face's corners will be after the previous face's corners. We can take advantage of this fact and eliminate the redundancy in mesh face storage by only storing a single integer corner offset for each face. The size of the face is then encoded by the offset of the next face. The size of a single integer is 4 bytes, so this reduces memory usage by 3 times. The same method is used for `CurvesGeometry`, so Blender already has an abstraction to simplify using these offsets called `OffsetIndices`. This class is used to easily retrieve a range of corner indices for each face. This also gives the opportunity for sharing some logic with curves. Another benefit of the change is that the offsets and sizes stored in `MPoly` can no longer disagree with each other. Storing faces in the order of their corners can simplify some code too. Face/polygon variables now use the `IndexRange` type, which comes with quite a few utilities that can simplify code. Some: - The offset integer array has to be one longer than the face count to avoid a branch for every face, which means the data is no longer part of the mesh's `CustomData`. - We lose the ability to "reference" an original mesh's offset array until more reusable CoW from #104478 is committed. That will be added in a separate commit. - Since they aren't part of `CustomData`, poly offsets often have to be copied manually. - To simplify using `OffsetIndices` in many places, some functions and structs in headers were moved to only compile in C++. - All meshes created by Blender use the same order for faces and face corners, but just in case, meshes with mismatched order are fixed by versioning code. - `MeshPolygon.totloop` is no longer editable in RNA. This API break is necessary here unfortunately. It should be worth it in 3.6, since that's the best way to allow loading meshes from 4.0, which is important for an LTS version. Pull Request: https://projects.blender.org/blender/blender/pulls/105938
2023-04-04 20:39:28 +02:00
BKE_mesh_legacy_convert_polys_to_offsets(explode);
psys_sim_data_free(&sim);
return explode;
}
static ParticleSystemModifierData *findPrecedingParticlesystem(Object *ob, ModifierData *emd)
{
ModifierData *md;
2023-02-06 12:46:53 +11:00
ParticleSystemModifierData *psmd = nullptr;
for (md = static_cast<ModifierData *>(ob->modifiers.first); emd != md; md = md->next) {
if (md->type == eModifierType_ParticleSystem) {
psmd = (ParticleSystemModifierData *)md;
}
}
return psmd;
}
static Mesh *modify_mesh(ModifierData *md, const ModifierEvalContext *ctx, Mesh *mesh)
{
ExplodeModifierData *emd = (ExplodeModifierData *)md;
ParticleSystemModifierData *psmd = findPrecedingParticlesystem(ctx->object, md);
if (psmd) {
ParticleSystem *psys = psmd->psys;
if (psys == nullptr || psys->totpart == 0) {
return mesh;
}
if (psys->part == nullptr || psys->particles == nullptr) {
return mesh;
}
if (psmd->mesh_final == nullptr) {
return mesh;
}
Mesh: Replace MPoly struct with offset indices Implements #95967. Currently the `MPoly` struct is 12 bytes, and stores the index of a face's first corner and the number of corners/verts/edges. Polygons and corners are always created in order by Blender, meaning each face's corners will be after the previous face's corners. We can take advantage of this fact and eliminate the redundancy in mesh face storage by only storing a single integer corner offset for each face. The size of the face is then encoded by the offset of the next face. The size of a single integer is 4 bytes, so this reduces memory usage by 3 times. The same method is used for `CurvesGeometry`, so Blender already has an abstraction to simplify using these offsets called `OffsetIndices`. This class is used to easily retrieve a range of corner indices for each face. This also gives the opportunity for sharing some logic with curves. Another benefit of the change is that the offsets and sizes stored in `MPoly` can no longer disagree with each other. Storing faces in the order of their corners can simplify some code too. Face/polygon variables now use the `IndexRange` type, which comes with quite a few utilities that can simplify code. Some: - The offset integer array has to be one longer than the face count to avoid a branch for every face, which means the data is no longer part of the mesh's `CustomData`. - We lose the ability to "reference" an original mesh's offset array until more reusable CoW from #104478 is committed. That will be added in a separate commit. - Since they aren't part of `CustomData`, poly offsets often have to be copied manually. - To simplify using `OffsetIndices` in many places, some functions and structs in headers were moved to only compile in C++. - All meshes created by Blender use the same order for faces and face corners, but just in case, meshes with mismatched order are fixed by versioning code. - `MeshPolygon.totloop` is no longer editable in RNA. This API break is necessary here unfortunately. It should be worth it in 3.6, since that's the best way to allow loading meshes from 4.0, which is important for an LTS version. Pull Request: https://projects.blender.org/blender/blender/pulls/105938
2023-04-04 20:39:28 +02:00
BKE_mesh_tessface_ensure(mesh); /* BMESH - UNTIL MODIFIER IS UPDATED FOR POLYGONS */
/* 1. find faces to be exploded if needed */
if (emd->facepa == nullptr || psmd->flag & eParticleSystemFlag_Pars ||
emd->flag & eExplodeFlag_CalcFaces ||
MEM_allocN_len(emd->facepa) / sizeof(int) != mesh->totface_legacy)
{
if (psmd->flag & eParticleSystemFlag_Pars) {
psmd->flag &= ~eParticleSystemFlag_Pars;
}
if (emd->flag & eExplodeFlag_CalcFaces) {
emd->flag &= ~eExplodeFlag_CalcFaces;
}
createFacepa(emd, psmd, mesh);
}
/* 2. create new mesh */
Scene *scene = DEG_get_evaluated_scene(ctx->depsgraph);
if (emd->flag & eExplodeFlag_EdgeCut) {
int *facepa = emd->facepa;
Mesh *split_m = cutEdges(emd, mesh);
Mesh *explode = explodeMesh(emd, psmd, ctx, scene, split_m);
MEM_freeN(emd->facepa);
emd->facepa = facepa;
BKE_id_free(nullptr, split_m);
return explode;
}
return explodeMesh(emd, psmd, ctx, scene, mesh);
}
return mesh;
}
static void panel_draw(const bContext * /*C*/, Panel *panel)
{
uiLayout *row, *col;
uiLayout *layout = panel->layout;
const eUI_Item_Flag toggles_flag = UI_ITEM_R_TOGGLE | UI_ITEM_R_FORCE_BLANK_DECORATE;
PointerRNA ob_ptr;
PointerRNA *ptr = modifier_panel_get_property_pointers(panel, &ob_ptr);
PointerRNA obj_data_ptr = RNA_pointer_get(&ob_ptr, "data");
bool has_vertex_group = RNA_string_length(ptr, "vertex_group") != 0;
uiLayoutSetPropSep(layout, true);
uiItemPointerR(
layout, ptr, "particle_uv", &obj_data_ptr, "uv_layers", std::nullopt, ICON_GROUP_UVS);
row = &layout->row(true, IFACE_("Show"));
row->prop(ptr, "show_alive", toggles_flag, std::nullopt, ICON_NONE);
row->prop(ptr, "show_dead", toggles_flag, std::nullopt, ICON_NONE);
row->prop(ptr, "show_unborn", toggles_flag, std::nullopt, ICON_NONE);
uiLayoutSetPropSep(layout, true);
col = &layout->column(false);
col->prop(ptr, "use_edge_cut", UI_ITEM_NONE, std::nullopt, ICON_NONE);
col->prop(ptr, "use_size", UI_ITEM_NONE, std::nullopt, ICON_NONE);
modifier_vgroup_ui(layout, ptr, &ob_ptr, "vertex_group", "invert_vertex_group", std::nullopt);
row = &layout->row(false);
uiLayoutSetActive(row, has_vertex_group);
row->prop(ptr, "protect", UI_ITEM_NONE, std::nullopt, ICON_NONE);
layout->op("OBJECT_OT_explode_refresh", IFACE_("Refresh"), ICON_NONE);
modifier_error_message_draw(layout, ptr);
}
static void panel_register(ARegionType *region_type)
{
modifier_panel_register(region_type, eModifierType_Explode, panel_draw);
}
static void blend_read(BlendDataReader * /*reader*/, ModifierData *md)
{
ExplodeModifierData *psmd = (ExplodeModifierData *)md;
psmd->facepa = nullptr;
}
ModifierTypeInfo modifierType_Explode = {
/*idname*/ "Explode",
/*name*/ N_("Explode"),
/*struct_name*/ "ExplodeModifierData",
/*struct_size*/ sizeof(ExplodeModifierData),
/*srna*/ &RNA_ExplodeModifier,
/*type*/ ModifierTypeType::Constructive,
/*flags*/ eModifierTypeFlag_AcceptsMesh,
/*icon*/ ICON_MOD_EXPLODE,
/*copy_data*/ copy_data,
/*deform_verts*/ nullptr,
/*deform_matrices*/ nullptr,
/*deform_verts_EM*/ nullptr,
/*deform_matrices_EM*/ nullptr,
/*modify_mesh*/ modify_mesh,
/*modify_geometry_set*/ nullptr,
/*init_data*/ init_data,
/*required_data_mask*/ required_data_mask,
/*free_data*/ free_data,
/*is_disabled*/ nullptr,
/*update_depsgraph*/ nullptr,
/*depends_on_time*/ depends_on_time,
/*depends_on_normals*/ nullptr,
/*foreach_ID_link*/ nullptr,
/*foreach_tex_link*/ nullptr,
/*free_runtime_data*/ nullptr,
/*panel_register*/ panel_register,
/*blend_write*/ nullptr,
/*blend_read*/ blend_read,
/*foreach_cache*/ nullptr,
};