bgfx/examples/common/bgfx_utils.cpp
Branimir Karadžić b2405e5221 Cleanup.
2016-02-16 20:38:59 -08:00

696 lines
15 KiB
C++

/*
* Copyright 2011-2016 Branimir Karadzic. All rights reserved.
* License: https://github.com/bkaradzic/bgfx#license-bsd-2-clause
*/
#include <string.h> // strlen
#include "common.h"
#include <tinystl/allocator.h>
#include <tinystl/vector.h>
#include <tinystl/string.h>
namespace stl = tinystl;
#include <bgfx/bgfx.h>
#include <bx/commandline.h>
#include <bx/fpumath.h>
#include <bx/readerwriter.h>
#include <bx/string.h>
#include "entry/entry.h"
#include <ib-compress/indexbufferdecompression.h>
#include "bgfx_utils.h"
void* load(bx::FileReaderI* _reader, bx::AllocatorI* _allocator, const char* _filePath, uint32_t* _size)
{
if (bx::open(_reader, _filePath) )
{
uint32_t size = (uint32_t)bx::getSize(_reader);
void* data = BX_ALLOC(_allocator, size);
bx::read(_reader, data, size);
bx::close(_reader);
if (NULL != _size)
{
*_size = size;
}
return data;
}
else
{
DBG("Failed to open: %s.", _filePath);
}
if (NULL != _size)
{
*_size = 0;
}
return NULL;
}
void* load(const char* _filePath, uint32_t* _size)
{
return load(entry::getFileReader(), entry::getAllocator(), _filePath, _size);
}
void unload(void* _ptr)
{
BX_FREE(entry::getAllocator(), _ptr);
}
static const bgfx::Memory* loadMem(bx::FileReaderI* _reader, const char* _filePath)
{
if (bx::open(_reader, _filePath) )
{
uint32_t size = (uint32_t)bx::getSize(_reader);
const bgfx::Memory* mem = bgfx::alloc(size+1);
bx::read(_reader, mem->data, size);
bx::close(_reader);
mem->data[mem->size-1] = '\0';
return mem;
}
DBG("Failed to load %s.", _filePath);
return NULL;
}
static void* loadMem(bx::FileReaderI* _reader, bx::AllocatorI* _allocator, const char* _filePath, uint32_t* _size)
{
if (bx::open(_reader, _filePath) )
{
uint32_t size = (uint32_t)bx::getSize(_reader);
void* data = BX_ALLOC(_allocator, size);
bx::read(_reader, data, size);
bx::close(_reader);
if (NULL != _size)
{
*_size = size;
}
return data;
}
DBG("Failed to load %s.", _filePath);
return NULL;
}
static bgfx::ShaderHandle loadShader(bx::FileReaderI* _reader, const char* _name)
{
char filePath[512];
const char* shaderPath = "shaders/dx9/";
switch (bgfx::getRendererType() )
{
case bgfx::RendererType::Direct3D11:
case bgfx::RendererType::Direct3D12:
shaderPath = "shaders/dx11/";
break;
case bgfx::RendererType::OpenGL:
shaderPath = "shaders/glsl/";
break;
case bgfx::RendererType::Metal:
shaderPath = "shaders/metal/";
break;
case bgfx::RendererType::OpenGLES:
shaderPath = "shaders/gles/";
break;
default:
break;
}
strcpy(filePath, shaderPath);
strcat(filePath, _name);
strcat(filePath, ".bin");
return bgfx::createShader(loadMem(_reader, filePath) );
}
bgfx::ShaderHandle loadShader(const char* _name)
{
return loadShader(entry::getFileReader(), _name);
}
bgfx::ProgramHandle loadProgram(bx::FileReaderI* _reader, const char* _vsName, const char* _fsName)
{
bgfx::ShaderHandle vsh = loadShader(_reader, _vsName);
bgfx::ShaderHandle fsh = BGFX_INVALID_HANDLE;
if (NULL != _fsName)
{
fsh = loadShader(_reader, _fsName);
}
return bgfx::createProgram(vsh, fsh, true /* destroy shaders when program is destroyed */);
}
bgfx::ProgramHandle loadProgram(const char* _vsName, const char* _fsName)
{
return loadProgram(entry::getFileReader(), _vsName, _fsName);
}
typedef unsigned char stbi_uc;
extern "C" stbi_uc *stbi_load_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp, int req_comp);
bgfx::TextureHandle loadTexture(bx::FileReaderI* _reader, const char* _name, uint32_t _flags, uint8_t _skip, bgfx::TextureInfo* _info)
{
char filePath[512] = { '\0' };
if (NULL == strchr(_name, '/') )
{
strcpy(filePath, "textures/");
}
strcat(filePath, _name);
if (NULL != bx::stristr(_name, ".dds")
|| NULL != bx::stristr(_name, ".pvr")
|| NULL != bx::stristr(_name, ".ktx") )
{
const bgfx::Memory* mem = loadMem(_reader, filePath);
if (NULL != mem)
{
return bgfx::createTexture(mem, _flags, _skip, _info);
}
bgfx::TextureHandle handle = BGFX_INVALID_HANDLE;
DBG("Failed to load %s.", filePath);
return handle;
}
bgfx::TextureHandle handle = BGFX_INVALID_HANDLE;
bx::AllocatorI* allocator = entry::getAllocator();
uint32_t size = 0;
void* data = loadMem(_reader, allocator, filePath, &size);
if (NULL != data)
{
int width = 0;
int height = 0;
int comp = 0;
uint8_t* img = NULL;
img = stbi_load_from_memory( (uint8_t*)data, size, &width, &height, &comp, 4);
BX_FREE(allocator, data);
if (NULL != img)
{
handle = bgfx::createTexture2D(uint16_t(width), uint16_t(height), 1
, bgfx::TextureFormat::RGBA8
, _flags
, bgfx::copy(img, width*height*4)
);
free(img);
if (NULL != _info)
{
bgfx::calcTextureSize(*_info
, uint16_t(width)
, uint16_t(height)
, 0
, false
, 1
, bgfx::TextureFormat::RGBA8
);
}
}
}
else
{
DBG("Failed to load %s.", filePath);
}
return handle;
}
bgfx::TextureHandle loadTexture(const char* _name, uint32_t _flags, uint8_t _skip, bgfx::TextureInfo* _info)
{
return loadTexture(entry::getFileReader(), _name, _flags, _skip, _info);
}
void calcTangents(void* _vertices, uint16_t _numVertices, bgfx::VertexDecl _decl, const uint16_t* _indices, uint32_t _numIndices)
{
struct PosTexcoord
{
float m_x;
float m_y;
float m_z;
float m_pad0;
float m_u;
float m_v;
float m_pad1;
float m_pad2;
};
float* tangents = new float[6*_numVertices];
memset(tangents, 0, 6*_numVertices*sizeof(float) );
PosTexcoord v0;
PosTexcoord v1;
PosTexcoord v2;
for (uint32_t ii = 0, num = _numIndices/3; ii < num; ++ii)
{
const uint16_t* indices = &_indices[ii*3];
uint32_t i0 = indices[0];
uint32_t i1 = indices[1];
uint32_t i2 = indices[2];
bgfx::vertexUnpack(&v0.m_x, bgfx::Attrib::Position, _decl, _vertices, i0);
bgfx::vertexUnpack(&v0.m_u, bgfx::Attrib::TexCoord0, _decl, _vertices, i0);
bgfx::vertexUnpack(&v1.m_x, bgfx::Attrib::Position, _decl, _vertices, i1);
bgfx::vertexUnpack(&v1.m_u, bgfx::Attrib::TexCoord0, _decl, _vertices, i1);
bgfx::vertexUnpack(&v2.m_x, bgfx::Attrib::Position, _decl, _vertices, i2);
bgfx::vertexUnpack(&v2.m_u, bgfx::Attrib::TexCoord0, _decl, _vertices, i2);
const float bax = v1.m_x - v0.m_x;
const float bay = v1.m_y - v0.m_y;
const float baz = v1.m_z - v0.m_z;
const float bau = v1.m_u - v0.m_u;
const float bav = v1.m_v - v0.m_v;
const float cax = v2.m_x - v0.m_x;
const float cay = v2.m_y - v0.m_y;
const float caz = v2.m_z - v0.m_z;
const float cau = v2.m_u - v0.m_u;
const float cav = v2.m_v - v0.m_v;
const float det = (bau * cav - bav * cau);
const float invDet = 1.0f / det;
const float tx = (bax * cav - cax * bav) * invDet;
const float ty = (bay * cav - cay * bav) * invDet;
const float tz = (baz * cav - caz * bav) * invDet;
const float bx = (cax * bau - bax * cau) * invDet;
const float by = (cay * bau - bay * cau) * invDet;
const float bz = (caz * bau - baz * cau) * invDet;
for (uint32_t jj = 0; jj < 3; ++jj)
{
float* tanu = &tangents[indices[jj]*6];
float* tanv = &tanu[3];
tanu[0] += tx;
tanu[1] += ty;
tanu[2] += tz;
tanv[0] += bx;
tanv[1] += by;
tanv[2] += bz;
}
}
for (uint32_t ii = 0; ii < _numVertices; ++ii)
{
const float* tanu = &tangents[ii*6];
const float* tanv = &tangents[ii*6 + 3];
float normal[4];
bgfx::vertexUnpack(normal, bgfx::Attrib::Normal, _decl, _vertices, ii);
float ndt = bx::vec3Dot(normal, tanu);
float nxt[3];
bx::vec3Cross(nxt, normal, tanu);
float tmp[3];
tmp[0] = tanu[0] - normal[0] * ndt;
tmp[1] = tanu[1] - normal[1] * ndt;
tmp[2] = tanu[2] - normal[2] * ndt;
float tangent[4];
bx::vec3Norm(tangent, tmp);
tangent[3] = bx::vec3Dot(nxt, tanv) < 0.0f ? -1.0f : 1.0f;
bgfx::vertexPack(tangent, true, bgfx::Attrib::Tangent, _decl, _vertices, ii);
}
delete [] tangents;
}
struct Aabb
{
float m_min[3];
float m_max[3];
};
struct Obb
{
float m_mtx[16];
};
struct Sphere
{
float m_center[3];
float m_radius;
};
struct Primitive
{
uint32_t m_startIndex;
uint32_t m_numIndices;
uint32_t m_startVertex;
uint32_t m_numVertices;
Sphere m_sphere;
Aabb m_aabb;
Obb m_obb;
};
typedef stl::vector<Primitive> PrimitiveArray;
struct Group
{
Group()
{
reset();
}
void reset()
{
m_vbh.idx = bgfx::invalidHandle;
m_ibh.idx = bgfx::invalidHandle;
m_prims.clear();
}
bgfx::VertexBufferHandle m_vbh;
bgfx::IndexBufferHandle m_ibh;
Sphere m_sphere;
Aabb m_aabb;
Obb m_obb;
PrimitiveArray m_prims;
};
namespace bgfx
{
int32_t read(bx::ReaderI* _reader, bgfx::VertexDecl& _decl);
}
struct Mesh
{
void load(bx::ReaderSeekerI* _reader)
{
#define BGFX_CHUNK_MAGIC_VB BX_MAKEFOURCC('V', 'B', ' ', 0x1)
#define BGFX_CHUNK_MAGIC_IB BX_MAKEFOURCC('I', 'B', ' ', 0x0)
#define BGFX_CHUNK_MAGIC_IBC BX_MAKEFOURCC('I', 'B', 'C', 0x0)
#define BGFX_CHUNK_MAGIC_PRI BX_MAKEFOURCC('P', 'R', 'I', 0x0)
using namespace bx;
using namespace bgfx;
Group group;
bx::AllocatorI* allocator = entry::getAllocator();
uint32_t chunk;
bx::Error err;
while (4 == bx::read(_reader, chunk, &err)
&& err.isOk() )
{
switch (chunk)
{
case BGFX_CHUNK_MAGIC_VB:
{
read(_reader, group.m_sphere);
read(_reader, group.m_aabb);
read(_reader, group.m_obb);
read(_reader, m_decl);
uint16_t stride = m_decl.getStride();
uint16_t numVertices;
read(_reader, numVertices);
const bgfx::Memory* mem = bgfx::alloc(numVertices*stride);
read(_reader, mem->data, mem->size);
group.m_vbh = bgfx::createVertexBuffer(mem, m_decl);
}
break;
case BGFX_CHUNK_MAGIC_IB:
{
uint32_t numIndices;
read(_reader, numIndices);
const bgfx::Memory* mem = bgfx::alloc(numIndices*2);
read(_reader, mem->data, mem->size);
group.m_ibh = bgfx::createIndexBuffer(mem);
}
break;
case BGFX_CHUNK_MAGIC_IBC:
{
uint32_t numIndices;
bx::read(_reader, numIndices);
const bgfx::Memory* mem = bgfx::alloc(numIndices*2);
uint32_t compressedSize;
bx::read(_reader, compressedSize);
void* compressedIndices = BX_ALLOC(allocator, compressedSize);
bx::read(_reader, compressedIndices, compressedSize);
ReadBitstream rbs( (const uint8_t*)compressedIndices, compressedSize);
DecompressIndexBuffer( (uint16_t*)mem->data, numIndices / 3, rbs);
BX_FREE(allocator, compressedIndices);
group.m_ibh = bgfx::createIndexBuffer(mem);
}
break;
case BGFX_CHUNK_MAGIC_PRI:
{
uint16_t len;
read(_reader, len);
stl::string material;
material.resize(len);
read(_reader, const_cast<char*>(material.c_str() ), len);
uint16_t num;
read(_reader, num);
for (uint32_t ii = 0; ii < num; ++ii)
{
read(_reader, len);
stl::string name;
name.resize(len);
read(_reader, const_cast<char*>(name.c_str() ), len);
Primitive prim;
read(_reader, prim.m_startIndex);
read(_reader, prim.m_numIndices);
read(_reader, prim.m_startVertex);
read(_reader, prim.m_numVertices);
read(_reader, prim.m_sphere);
read(_reader, prim.m_aabb);
read(_reader, prim.m_obb);
group.m_prims.push_back(prim);
}
m_groups.push_back(group);
group.reset();
}
break;
default:
DBG("%08x at %d", chunk, bx::skip(_reader, 0) );
break;
}
}
}
void unload()
{
for (GroupArray::const_iterator it = m_groups.begin(), itEnd = m_groups.end(); it != itEnd; ++it)
{
const Group& group = *it;
bgfx::destroyVertexBuffer(group.m_vbh);
if (bgfx::isValid(group.m_ibh) )
{
bgfx::destroyIndexBuffer(group.m_ibh);
}
}
m_groups.clear();
}
void submit(uint8_t _id, bgfx::ProgramHandle _program, const float* _mtx, uint64_t _state) const
{
if (BGFX_STATE_MASK == _state)
{
_state = 0
| BGFX_STATE_RGB_WRITE
| BGFX_STATE_ALPHA_WRITE
| BGFX_STATE_DEPTH_WRITE
| BGFX_STATE_DEPTH_TEST_LESS
| BGFX_STATE_CULL_CCW
| BGFX_STATE_MSAA
;
}
uint32_t cached = bgfx::setTransform(_mtx);
for (GroupArray::const_iterator it = m_groups.begin(), itEnd = m_groups.end(); it != itEnd; ++it)
{
const Group& group = *it;
bgfx::setTransform(cached);
bgfx::setIndexBuffer(group.m_ibh);
bgfx::setVertexBuffer(group.m_vbh);
bgfx::setState(_state);
bgfx::submit(_id, _program);
}
}
void submit(const MeshState*const* _state, uint8_t _numPasses, const float* _mtx, uint16_t _numMatrices) const
{
uint32_t cached = bgfx::setTransform(_mtx, _numMatrices);
for (uint32_t pass = 0; pass < _numPasses; ++pass)
{
const MeshState& state = *_state[pass];
for (GroupArray::const_iterator it = m_groups.begin(), itEnd = m_groups.end(); it != itEnd; ++it)
{
const Group& group = *it;
bgfx::setTransform(cached, _numMatrices);
for (uint8_t tex = 0; tex < state.m_numTextures; ++tex)
{
const MeshState::Texture& texture = state.m_textures[tex];
bgfx::setTexture(texture.m_stage
, texture.m_sampler
, texture.m_texture
, texture.m_flags
);
}
bgfx::setIndexBuffer(group.m_ibh);
bgfx::setVertexBuffer(group.m_vbh);
bgfx::setState(state.m_state);
bgfx::submit(state.m_viewId, state.m_program);
}
}
}
bgfx::VertexDecl m_decl;
typedef stl::vector<Group> GroupArray;
GroupArray m_groups;
};
Mesh* meshLoad(bx::ReaderSeekerI* _reader)
{
Mesh* mesh = new Mesh;
mesh->load(_reader);
return mesh;
}
Mesh* meshLoad(const char* _filePath)
{
bx::FileReaderI* reader = entry::getFileReader();
if (bx::open(reader, _filePath) )
{
Mesh* mesh = meshLoad(reader);
bx::close(reader);
return mesh;
}
return NULL;
}
void meshUnload(Mesh* _mesh)
{
_mesh->unload();
delete _mesh;
}
MeshState* meshStateCreate()
{
MeshState* state = (MeshState*)BX_ALLOC(entry::getAllocator(), sizeof(MeshState) );
return state;
}
void meshStateDestroy(MeshState* _meshState)
{
BX_FREE(entry::getAllocator(), _meshState);
}
void meshSubmit(const Mesh* _mesh, uint8_t _id, bgfx::ProgramHandle _program, const float* _mtx, uint64_t _state)
{
_mesh->submit(_id, _program, _mtx, _state);
}
void meshSubmit(const Mesh* _mesh, const MeshState*const* _state, uint8_t _numPasses, const float* _mtx, uint16_t _numMatrices)
{
_mesh->submit(_state, _numPasses, _mtx, _numMatrices);
}
Args::Args(int _argc, char** _argv)
: m_type(bgfx::RendererType::Count)
, m_pciId(BGFX_PCI_ID_NONE)
{
bx::CommandLine cmdLine(_argc, (const char**)_argv);
if (cmdLine.hasArg("gl") )
{
m_type = bgfx::RendererType::OpenGL;
}
else if (cmdLine.hasArg("vk") )
{
m_type = bgfx::RendererType::Vulkan;
}
else if (cmdLine.hasArg("noop") )
{
m_type = bgfx::RendererType::Null;
}
else if (BX_ENABLED(BX_PLATFORM_WINDOWS) )
{
if (cmdLine.hasArg("d3d9") )
{
m_type = bgfx::RendererType::Direct3D9;
}
else if (cmdLine.hasArg("d3d11") )
{
m_type = bgfx::RendererType::Direct3D11;
}
else if (cmdLine.hasArg("d3d12") )
{
m_type = bgfx::RendererType::Direct3D12;
}
}
else if (BX_ENABLED(BX_PLATFORM_OSX) )
{
if (cmdLine.hasArg("mtl") )
{
m_type = bgfx::RendererType::Metal;
}
}
if (cmdLine.hasArg("amd") )
{
m_pciId = BGFX_PCI_ID_AMD;
}
else if (cmdLine.hasArg("nvidia") )
{
m_pciId = BGFX_PCI_ID_NVIDIA;
}
else if (cmdLine.hasArg("intel") )
{
m_pciId = BGFX_PCI_ID_INTEL;
}
else if (cmdLine.hasArg("sw") )
{
m_pciId = BGFX_PCI_ID_SOFTWARE_RASTERIZER;
}
}