/*

-----------------------------------------------------------------------------

This source file is part of OGRE

    (Object-oriented Graphics Rendering Engine)

For the latest info, see http://www.ogre3d.org/

Copyright (c) 2000-2014 Torus Knot Software Ltd

Permission is hereby granted, free of charge, to any person obtaining a copy

of this software and associated documentation files (the "Software"), to deal

in the Software without restriction, including without limitation the rights

to use, copy, modify, merge, publish, distribute, sublicense, and/or sell

copies of the Software, and to permit persons to whom the Software is

furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in

all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR

IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE

AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER

LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,

OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN

THE SOFTWARE.

-----------------------------------------------------------------------------

*/

#include "OgreStableHeaders.h"

#include "OgreVertexIndexData.h"

#include "OgreHardwareVertexBuffer.h"

#define INT10_MAX ((1 << 9) - 1)

namespace Ogre {

    static void swapPackedRB(uint32* ptr)

    {

        auto cptr = (uint8*)ptr;

        std::swap(cptr[0], cptr[2]);

    }

    struct int_10_10_10_2

    {

        int32_t x : 10;

        int32_t y : 10;

        int32_t z : 10;

        int32_t w :  2;

    };

    template<int INCLUDE_W>

    static void pack_10_10_10_2(uint8* pDst, uint8* pSrc, int elemOffset)

    {

        float* pFloat = (float*)(pSrc + elemOffset);

        int_10_10_10_2 packed = {int(INT10_MAX * pFloat[0]), int(INT10_MAX * pFloat[1]), int(INT10_MAX * pFloat[2]), 1};

        if(INCLUDE_W)

            packed.w = int(pFloat[3]);

        memcpy(pDst + elemOffset, &packed, sizeof(int_10_10_10_2));

    }

Unexecuted instantiation: OgreVertexIndexData.cpp:void Ogre::pack_10_10_10_2<0>(unsigned char*, unsigned char*, int)

Unexecuted instantiation: OgreVertexIndexData.cpp:void Ogre::pack_10_10_10_2<1>(unsigned char*, unsigned char*, int)

    template<int INCLUDE_W>

    static void unpack_10_10_10_2(uint8* pDst, uint8* pSrc, int elemOffset)

    {

        int_10_10_10_2* pPacked = (int_10_10_10_2*)(pSrc + elemOffset);

        float* pFloat = (float*)(pDst + elemOffset);

        pFloat[0] = float(pPacked->x) / INT10_MAX;

        pFloat[1] = float(pPacked->y) / INT10_MAX;

        pFloat[2] = float(pPacked->z) / INT10_MAX;

        if(INCLUDE_W)

            pFloat[3] = pPacked->w;

    }

Unexecuted instantiation: OgreVertexIndexData.cpp:void Ogre::unpack_10_10_10_2<0>(unsigned char*, unsigned char*, int)

Unexecuted instantiation: OgreVertexIndexData.cpp:void Ogre::unpack_10_10_10_2<1>(unsigned char*, unsigned char*, int)

    static void extract_float3(uint8* pDst, uint8* pSrc, int elemOffset)

    {

        memcpy(pDst, pSrc + elemOffset, sizeof(float) * 3);

    }

    static void pad_16x3(uint8* pDst, uint8* pSrc, int elemOffset)

    {

        // for half3, we want 1.0 in the 4th component

        // for others we dont care, so do it unconditionally

        static const uint16 one16f = Bitwise::floatToHalf(1.0f);

        memcpy(pDst + elemOffset, pSrc + elemOffset, sizeof(uint16) * 3);

        memcpy(pDst + elemOffset + sizeof(uint16) * 3, &one16f, sizeof(uint16));

    }

    static void float_to_half_3(uint8* pDst, uint8* pSrc, int elemOffset)

    {

        float* pFloat = (float*)(pSrc + elemOffset);

        uint16* pHalf = (uint16*)(pDst + elemOffset);

        pHalf[0] = Bitwise::floatToHalf(pFloat[0]);

        pHalf[1] = Bitwise::floatToHalf(pFloat[1]);

        pHalf[2] = Bitwise::floatToHalf(pFloat[2]);

    }

    /** Splice out an element from a vertex buffer

     * @param elem The element to splice out of the vertex

     * @param srcBuf Source buffer

     * @param pDst Destination buffer for the vertex without the element

     * @param pElemDst Destination buffer for the element (can be the same as pDst)

     */

    static void spliceElement(const VertexElement* elem, const HardwareVertexBufferPtr& srcBuf, uint8* pDst,

                              uint8* pElemDst, uint32 newElemSize, void (*elemConvert)(uint8*, uint8*, int))

    {

        auto vertexSize = srcBuf->getVertexSize();

        auto numVerts = srcBuf->getNumVertices();

        auto elemSize = elem->getSize();

        int elemOffset = elem->getOffset();

        auto postVertexOffset = elemOffset + elemSize;

        auto postVertexSize = vertexSize - postVertexOffset;

        auto elemDstSize = pDst == pElemDst ? newElemSize : 0;

        size_t newVertexSize = vertexSize - elemSize + elemDstSize;

        auto elemDstStep = pDst == pElemDst ? newVertexSize : newElemSize;

        HardwareBufferLockGuard srcLock(srcBuf, HardwareBuffer::HBL_READ_ONLY);

        uint8* pSrc = static_cast<uint8*>(srcLock.pData);

        for (uint32 v = 0; v < numVerts; ++v)

        {

            // copy and convert element from vertex

            elemConvert(pElemDst, pSrc, elemOffset);

            pElemDst += elemDstStep;

            // copy over other data

            if (elemOffset)

                memcpy(pDst, pSrc, elemOffset);

            if (postVertexSize)

                memcpy(pDst + elemOffset + elemDstSize, pSrc + postVertexOffset, postVertexSize);

            pSrc += vertexSize;

            pDst += newVertexSize;

        }

    }

    static void updateVertexDeclaration(VertexDeclaration* decl, const VertexElement* elem, VertexElementType newType, uint16 newSource)

    {

        auto elemSize = elem->getSize();

        auto oldElemOffset = elem->getOffset();

        auto newElemOffset = oldElemOffset;

        auto newElemSize = VertexElement::getTypeSize(newType);

        auto oldSource = elem->getSource();

        if(newSource != oldSource)

        {

            newElemOffset = 0;

            newElemSize = 0;

        }

        uint16 idx = 0;

        for (const auto& e : decl->getElements())

        {

            if (&e == elem)

            {

                // Modify element

                decl->modifyElement(idx, newSource, newElemOffset, newType, elem->getSemantic(), elem->getIndex());

            }

            else if (e.getSource() == oldSource && e.getOffset() > oldElemOffset)

            {

                // shift elements after this one

                decl->modifyElement(idx, e.getSource(), e.getOffset() - elemSize + newElemSize, e.getType(),

                                    e.getSemantic(), e.getIndex());

            }

            idx++;

        }

    }

    //-----------------------------------------------------------------------

    VertexData::VertexData(HardwareBufferManagerBase* mgr)

    {

        mMgr = mgr ? mgr : HardwareBufferManager::getSingletonPtr();

        vertexBufferBinding = mMgr->createVertexBufferBinding();

        vertexDeclaration = mMgr->createVertexDeclaration();

        mDeleteDclBinding = true;

        vertexCount = 0;

        vertexStart = 0;

        hwAnimDataItemsUsed = 0;

    }

    //---------------------------------------------------------------------

    VertexData::VertexData(VertexDeclaration* dcl, VertexBufferBinding* bind)

    {

        // this is a fallback rather than actively used

        mMgr = HardwareBufferManager::getSingletonPtr();

        vertexDeclaration = dcl;

        vertexBufferBinding = bind;

        mDeleteDclBinding = false;

        vertexCount = 0;

        vertexStart = 0;

        hwAnimDataItemsUsed = 0;

    }

    //-----------------------------------------------------------------------

    VertexData::~VertexData()

    {

        if (mDeleteDclBinding)

        {

            mMgr->destroyVertexBufferBinding(vertexBufferBinding);

            mMgr->destroyVertexDeclaration(vertexDeclaration);

        }

    }

    //-----------------------------------------------------------------------

    VertexData* VertexData::clone(bool copyData, HardwareBufferManagerBase* mgr) const

    {

        HardwareBufferManagerBase* pManager = mgr ? mgr : mMgr;

        VertexData* dest = OGRE_NEW VertexData(mgr);

        // Copy vertex buffers in turn

        const VertexBufferBinding::VertexBufferBindingMap& bindings = 

            this->vertexBufferBinding->getBindings();

        VertexBufferBinding::VertexBufferBindingMap::const_iterator vbi, vbend;

        vbend = bindings.end();

        for (vbi = bindings.begin(); vbi != vbend; ++vbi)

        {

            HardwareVertexBufferSharedPtr srcbuf = vbi->second;

            HardwareVertexBufferSharedPtr dstBuf;

            if (copyData)

            {

                // create new buffer with the same settings

                dstBuf = pManager->createVertexBuffer(

                        srcbuf->getVertexSize(), srcbuf->getNumVertices(), srcbuf->getUsage(),

                        srcbuf->hasShadowBuffer());

                // copy data

                dstBuf->copyData(*srcbuf, 0, 0, srcbuf->getSizeInBytes(), true);

            }

            else

            {

                // don't copy, point at existing buffer

                dstBuf = srcbuf;

            }

            // Copy binding

            dest->vertexBufferBinding->setBinding(vbi->first, dstBuf);

        }

        // Basic vertex info

        dest->vertexStart = this->vertexStart;

        dest->vertexCount = this->vertexCount;

        // Copy elements

        const VertexDeclaration::VertexElementList elems = 

            this->vertexDeclaration->getElements();

        VertexDeclaration::VertexElementList::const_iterator ei, eiend;

        eiend = elems.end();

        for (ei = elems.begin(); ei != eiend; ++ei)

        {

            dest->vertexDeclaration->addElement(

                ei->getSource(),

                ei->getOffset(),

                ei->getType(),

                ei->getSemantic(),

                ei->getIndex() );

        }

        // Copy reference to hardware shadow buffer, no matter whether copy data or not

        dest->hardwareShadowVolWBuffer = hardwareShadowVolWBuffer;

        // copy anim data

        dest->hwAnimationDataList = hwAnimationDataList;

        dest->hwAnimDataItemsUsed = hwAnimDataItemsUsed;

        return dest;

    }

    void VertexData::convertVertexElement(VertexElementSemantic semantic, VertexElementType dstType, uint16 index)

    {

        auto elem = vertexDeclaration->findElementBySemantic(semantic, index);

        if(!elem)

            return; // nothing to do

        auto srcBaseType = VertexElement::getBaseType(elem->getType());

        if (srcBaseType == VET_FLOAT1 && srcBaseType == VertexElement::getBaseType(dstType))

            return; // silently ignore floatX > floatY conversions

        auto srcType = elem->getType();

        auto vbuf = vertexBufferBinding->getBuffer(elem->getSource());

        uint32 newElemSize = VertexElement::getTypeSize(dstType);

        size_t newVertexSize = vbuf->getVertexSize() - elem->getSize() + newElemSize;

        auto newVBuf = vbuf->getManager()->createVertexBuffer(newVertexSize, vbuf->getNumVertices(), vbuf->getUsage(),

                                                              vbuf->hasShadowBuffer());

        {

            HardwareBufferLockGuard dst(newVBuf, HardwareBuffer::HBL_DISCARD);

            auto pDst = static_cast<uint8*>(dst.pData);

            if(dstType == VET_INT_10_10_10_2_NORM)

            {

                if(srcType == VET_FLOAT3)

                    spliceElement(elem, vbuf, pDst, pDst, newElemSize, pack_10_10_10_2<false>);

                else

                {

                    OgreAssert(srcType == VET_FLOAT4, "unsupported conversion");

                    spliceElement(elem, vbuf, pDst, pDst, newElemSize, pack_10_10_10_2<true>);

                }

            }

            else if(dstType == VET_FLOAT3)

            {

                OgreAssert(srcType == VET_INT_10_10_10_2_NORM, "unsupported conversion");

                spliceElement(elem, vbuf, pDst, pDst, newElemSize, unpack_10_10_10_2<false>);

            }

            else if(dstType == VET_FLOAT4)

            {

                OgreAssert(srcType == VET_INT_10_10_10_2_NORM, "unsupported conversion");

                spliceElement(elem, vbuf, pDst, pDst, newElemSize, unpack_10_10_10_2<true>);

            }

            else if(dstType == VET_HALF3)

            {

                OgreAssert(srcType == VET_FLOAT3, "unsupported conversion");

                spliceElement(elem, vbuf, pDst, pDst, newElemSize, float_to_half_3);

            }

            else if(dstType == VET_HALF4 || dstType == VET_SHORT4 || dstType == VET_USHORT4)

            {

                // pad 16x3 formats to 16x4

                OgreAssert(srcType == VET_HALF3 || srcType == VET_SHORT3 || srcType == VET_USHORT3, "unsupported conversion");

                spliceElement(elem, vbuf, pDst, pDst, newElemSize, pad_16x3);

            }

            else

            {

                OGRE_EXCEPT(Exception::ERR_INVALIDPARAMS, "unsupported dstType");

            }

        }

        // Bind the new buffer

        vertexBufferBinding->setBinding(elem->getSource(), newVBuf);

        updateVertexDeclaration(vertexDeclaration, elem, dstType, elem->getSource());

    }

    //-----------------------------------------------------------------------

    void VertexData::prepareForShadowVolume(void)

    {

        /* NOTE

        I would dearly, dearly love to just use a 4D position buffer in order to 

        store the extra 'w' value I need to differentiate between extruded and 

        non-extruded sections of the buffer, so that vertex programs could use that.

        Hey, it works fine for GL. However, D3D9 in it's infinite stupidity, does not

        support 4d position vertices in the fixed-function pipeline. If you use them, 

        you just see nothing. Since we can't know whether the application is going to use

        fixed function or vertex programs, we have to stick to 3d position vertices and

        store the 'w' in a separate 1D texture coordinate buffer, which is only used

        when rendering the shadow.

        */

        // Look for a position element

        const VertexElement* posElem = vertexDeclaration->findElementBySemantic(VES_POSITION);

        if (!posElem)

            return;

        // Upfront, lets check whether we have vertex program capability

        bool useVertexPrograms = Root::getSingleton().getRenderSystem() != 0;

        auto vbuf = vertexBufferBinding->getBuffer(posElem->getSource());

        // Are there other elements in the buffer except for the position?

        // We need to create another buffer to contain the remaining elements

        // Most drivers don't like gaps in the declaration, and in any case it's waste

        HardwareVertexBufferPtr newRemainderBuffer;

        if (vbuf->getVertexSize() > posElem->getSize())

        {

            newRemainderBuffer = vbuf->getManager()->createVertexBuffer(

                vbuf->getVertexSize() - posElem->getSize(), vbuf->getNumVertices(), vbuf->getUsage(),

                vbuf->hasShadowBuffer());

        }

        // Allocate new position buffer, will be FLOAT3 and 2x the size

        size_t oldVertexCount = vbuf->getNumVertices();

        size_t newVertexCount = oldVertexCount * 2;

        auto newPosBuffer = vbuf->getManager()->createVertexBuffer(

            VertexElement::getTypeSize(VET_FLOAT3), newVertexCount, vbuf->getUsage(), vbuf->hasShadowBuffer());

        // Point first destination pointer at the start of the new position buffer,

        // the other one half way along

        auto pDest = static_cast<float*>(newPosBuffer->lock(HardwareBuffer::HBL_DISCARD));

        auto pDest2 = pDest + oldVertexCount * 3;

        if (newRemainderBuffer)

        {

            // Basically we just memcpy the vertex excluding the position

            HardwareBufferLockGuard destRemLock(newRemainderBuffer, HardwareBuffer::HBL_DISCARD);

            spliceElement(posElem, vbuf, (uint8*)destRemLock.pData, (uint8*)pDest, posElem->getSize(), extract_float3);

        }

        else

        {

            // Unshared buffer, can block copy the whole thing

            vbuf->readData(0, vbuf->getSizeInBytes(), pDest);

        }

        memcpy(pDest2, pDest, oldVertexCount * 3 * sizeof(float));

        newPosBuffer->unlock();

        // At this stage, he original vertex buffer is going to be destroyed

        // So we should force the deallocation of any temporary copies

        vbuf->getManager()->_forceReleaseBufferCopies(vbuf);

        if (useVertexPrograms)

        {

            // Now it's time to set up the w buffer

            hardwareShadowVolWBuffer =

                vbuf->getManager()->createVertexBuffer(sizeof(float), newVertexCount, HBU_GPU_ONLY, false);

            // Fill the first half with 1.0, second half with 0.0

            pDest = static_cast<float*>(hardwareShadowVolWBuffer->lock(HardwareBuffer::HBL_DISCARD));

            for (size_t v = 0; v < oldVertexCount; ++v)

            {

                *pDest++ = 1.0f;

            }

            // Fill the second half with 0.0

            memset(pDest, 0, sizeof(float) * oldVertexCount);

            hardwareShadowVolWBuffer->unlock();

        }

        auto newPosBufferSource = posElem->getSource();

        if (newRemainderBuffer)

        {

            // Get the a new buffer binding index

            newPosBufferSource= vertexBufferBinding->getNextIndex();

            // Re-bind the old index to the remainder buffer

            vertexBufferBinding->setBinding(posElem->getSource(), newRemainderBuffer);

        }

        // Bind the new position buffer

        vertexBufferBinding->setBinding(newPosBufferSource, newPosBuffer);

        // Now, alter the vertex declaration to change the position source

        // and the offsets of elements using the same buffer

        // Note that we don't change vertexCount, because the other buffer(s) are still the same

        // size after all

        updateVertexDeclaration(vertexDeclaration, posElem, VET_FLOAT3, newPosBufferSource);

    }

    //-----------------------------------------------------------------------

    void VertexData::reorganiseBuffers(VertexDeclaration* newDeclaration, const BufferUsageList& bufferUsages,

                                       HardwareBufferManagerBase* mgr)

    {

        HardwareBufferManagerBase* pManager = mgr ? mgr : mMgr;

        // Firstly, close up any gaps in the buffer sources which might have arisen

        newDeclaration->closeGapsInSource();

        // Build up a list of both old and new elements in each buffer

        std::vector<uint8*> oldBufferLocks;

        std::vector<size_t> oldBufferVertexSizes;

        std::vector<uint8*> newBufferLocks;

        std::vector<size_t> newBufferVertexSizes;

        VertexBufferBinding* newBinding = pManager->createVertexBufferBinding();

        const auto& oldBindingMap = vertexBufferBinding->getBindings();

        VertexBufferBinding::VertexBufferBindingMap::const_iterator itBinding;

        // Pre-allocate old buffer locks

        if (!oldBindingMap.empty())

        {

            size_t count = oldBindingMap.rbegin()->first + 1;

            oldBufferLocks.resize(count);

            oldBufferVertexSizes.resize(count);

        }

        bool useShadowBuffer = false;

        // Lock all the old buffers for reading

        for (const auto& it : oldBindingMap)

        {

            assert(it.second->getNumVertices() >= vertexCount);

            oldBufferVertexSizes[it.first] = it.second->getVertexSize();

            oldBufferLocks[it.first] = (uint8*)it.second->lock(HardwareBuffer::HBL_READ_ONLY);

            useShadowBuffer |= it.second->hasShadowBuffer();

        }

        // Create new buffers and lock all for writing

        uint16 buf = 0;

        while (!newDeclaration->findElementsBySource(buf).empty())

        {

            size_t vertexSize = newDeclaration->getVertexSize(buf);

            auto vbuf = pManager->createVertexBuffer(vertexSize, vertexCount, bufferUsages[buf], useShadowBuffer);

            newBinding->setBinding(buf, vbuf);

            newBufferVertexSizes.push_back(vertexSize);

            newBufferLocks.push_back((uint8*)vbuf->lock(HardwareBuffer::HBL_DISCARD));

            buf++;

        }

        // Map from new to old elements

        std::map<const VertexElement*, const VertexElement*>  newToOldElementMap;

        const auto& newElemList = newDeclaration->getElements();

        for (const auto& newElem : newElemList)

        {

            // Find corresponding old element

            auto oldElem = vertexDeclaration->findElementBySemantic(newElem.getSemantic(), newElem.getIndex());

            if (!oldElem)

            {

                // Error, cannot create new elements with this method

                OGRE_EXCEPT(Exception::ERR_ITEM_NOT_FOUND, "Element not found in old vertex declaration");

            }

            newToOldElementMap[&newElem] = oldElem;

        }

        // Now iterate over the new buffers, pulling data out of the old ones

        // For each vertex

        for (size_t v = 0; v < vertexCount; ++v)

        {

            // For each (new) element

            for (const auto& newElem : newElemList)

            {

                const VertexElement* oldElem = newToOldElementMap[&newElem];

                auto oldBufferNo = oldElem->getSource();

                auto newBufferNo = newElem.getSource();

                auto pSrc = oldBufferLocks[oldBufferNo] + v * oldBufferVertexSizes[oldBufferNo];

                auto pDst = newBufferLocks[newBufferNo] + v * newBufferVertexSizes[newBufferNo];

                memcpy(pDst + newElem.getOffset(), pSrc + oldElem->getOffset(), newElem.getSize());

            }

        }

        // Unlock all buffers

        for (const auto& it : oldBindingMap)

        {

            it.second->unlock();

        }

        for (buf = 0; buf < newBinding->getBufferCount(); ++buf)

        {

            newBinding->getBuffer(buf)->unlock();

        }

        // Delete old binding & declaration

        if (mDeleteDclBinding)

        {

            pManager->destroyVertexBufferBinding(vertexBufferBinding);

            pManager->destroyVertexDeclaration(vertexDeclaration);

        }

        // Assign new binding and declaration

        vertexDeclaration = newDeclaration;

        vertexBufferBinding = newBinding;       

        // after this is complete, new manager should be used

        mMgr = pManager;

        mDeleteDclBinding = true; // because we created these through a manager

    }

    //-----------------------------------------------------------------------

    void VertexData::reorganiseBuffers(VertexDeclaration* newDeclaration, HardwareBufferManagerBase* mgr)

    {

        // Derive the buffer usages from looking at where the source has come

        // from

        BufferUsageList usages;

        for (unsigned short b = 0; b <= newDeclaration->getMaxSource(); ++b)

        {

            VertexDeclaration::VertexElementList destElems = newDeclaration->findElementsBySource(b);

            // Initialise with most restrictive version

            uint8 final = HBU_GPU_ONLY;

            for (VertexElement& destelem : destElems)

            {

                // get source

                auto srcelem = vertexDeclaration->findElementBySemantic(destelem.getSemantic(), destelem.getIndex());

                OgreAssert(srcelem, "Semantic not found in existing declaration");

                const auto& srcbuf = vertexBufferBinding->getBuffer(srcelem->getSource());

                // improve flexibility only

                if (srcbuf->getUsage() & HBU_CPU_ONLY)

                {

                    // remove static

                    final &= ~HBU_GPU_TO_CPU;

                    // add dynamic

                    final |= HBU_CPU_ONLY;

                }

                if (!(srcbuf->getUsage() & HBU_DETAIL_WRITE_ONLY))

                {

                    // remove write only

                    final &= ~HBU_DETAIL_WRITE_ONLY;

                }

            }

            usages.push_back(static_cast<HardwareBufferUsage>(final));

        }

        // Call specific method

        reorganiseBuffers(newDeclaration, usages, mgr);

    }

    //-----------------------------------------------------------------------

    void VertexData::closeGapsInBindings(void)

    {

        if (!vertexBufferBinding->hasGaps())

            return;

        // Check for error first

        const VertexDeclaration::VertexElementList& allelems = 

            vertexDeclaration->getElements();

        for (auto& e : allelems)

        {

            if (!vertexBufferBinding->isBufferBound(e.getSource()))

            {

                OGRE_EXCEPT(Exception::ERR_ITEM_NOT_FOUND,

                    "No buffer is bound to that element source.",

                    "VertexData::closeGapsInBindings");

            }

        }

        // Close gaps in the vertex buffer bindings

        VertexBufferBinding::BindingIndexMap bindingIndexMap;

        vertexBufferBinding->closeGaps(bindingIndexMap);

        // Modify vertex elements to reference to new buffer index

        unsigned short elemIndex = 0;

        for (auto ai = allelems.begin(); ai != allelems.end(); ++ai, ++elemIndex)

        {

            const VertexElement& elem = *ai;

            VertexBufferBinding::BindingIndexMap::const_iterator it =

                bindingIndexMap.find(elem.getSource());

            assert(it != bindingIndexMap.end());

            ushort targetSource = it->second;

            if (elem.getSource() != targetSource)

            {

                vertexDeclaration->modifyElement(elemIndex, 

                    targetSource, elem.getOffset(), elem.getType(), 

                    elem.getSemantic(), elem.getIndex());

            }

        }

    }

    //-----------------------------------------------------------------------

    void VertexData::removeUnusedBuffers(void)

    {

        std::set<ushort> usedBuffers;

        // Collect used buffers

        const VertexDeclaration::VertexElementList& allelems = 

            vertexDeclaration->getElements();

        for (auto& e : allelems)

        {

            usedBuffers.insert(e.getSource());

        }

        // Unset unused buffer bindings

        ushort count = vertexBufferBinding->getLastBoundIndex();

        for (ushort index = 0; index < count; ++index)

        {

            if (usedBuffers.find(index) == usedBuffers.end() &&

                vertexBufferBinding->isBufferBound(index))

            {

                vertexBufferBinding->unsetBinding(index);

            }

        }

        // Close gaps

        closeGapsInBindings();

    }

    //-----------------------------------------------------------------------

    void VertexData::convertPackedColour(VertexElementType, VertexElementType destType)

    {

        OgreAssert(destType == VET_UBYTE4_NORM, "Not supported");

        const VertexBufferBinding::VertexBufferBindingMap& bindMap = 

            vertexBufferBinding->getBindings();

        for (auto& m : bindMap)

        {

            const auto& elems =

                vertexDeclaration->findElementsBySource(m.first);

            bool conversionNeeded = false;

            for (auto& e : elems)

            {

                if (e.getType() == _DETAIL_SWAP_RB)

                {

                    conversionNeeded = true;

                }

            }

            if (conversionNeeded)

            {

                void* pBase = m.second->lock(HardwareBuffer::HBL_NORMAL);

                for (size_t v = 0; v < m.second->getNumVertices(); ++v)

                {

                    for (auto& e : elems)

                    {

                        if (e.getType() == _DETAIL_SWAP_RB)

                        {

                            uint32* pRGBA;

                            e.baseVertexPointerToElement(pBase, &pRGBA);

                            swapPackedRB(pRGBA);

                        }

                    }

                    pBase = static_cast<void*>(

                        static_cast<char*>(pBase) + m.second->getVertexSize());

                }

                m.second->unlock();

                // Modify the elements to reflect the changed type

                const VertexDeclaration::VertexElementList& allelems = 

                    vertexDeclaration->getElements();

                unsigned short elemIndex = 0;

                for (auto& e : allelems)

                {

                    if (e.getType() == _DETAIL_SWAP_RB)

                    {

                        vertexDeclaration->modifyElement(elemIndex,

                            e.getSource(), e.getOffset(), destType,

                            e.getSemantic(), e.getIndex());

                    }

                    ++elemIndex;

                }

            }

        } // each buffer

    }

    //-----------------------------------------------------------------------

    ushort VertexData::allocateHardwareAnimationElements(ushort count, bool animateNormals)

    {

        // Find first free texture coord set

        unsigned short texCoord = vertexDeclaration->getNextFreeTextureCoordinate();

        unsigned short freeCount = (ushort)(OGRE_MAX_TEXTURE_COORD_SETS - texCoord);

        if (animateNormals)

            // we need 2x the texture coords, round down

            freeCount /= 2;

        unsigned short supportedCount = std::min(freeCount, count);

        // Increase to correct size

        for (size_t c = hwAnimationDataList.size(); c < supportedCount; ++c)

        {

            // Create a new 3D texture coordinate set

            HardwareAnimationData data;

            data.targetBufferIndex = vertexBufferBinding->getNextIndex();

            vertexDeclaration->addElement(data.targetBufferIndex, 0, VET_FLOAT3, VES_TEXTURE_COORDINATES, texCoord++);

            if (animateNormals)

                    vertexDeclaration->addElement(data.targetBufferIndex, sizeof(float)*3, VET_FLOAT3, VES_TEXTURE_COORDINATES, texCoord++);

            hwAnimationDataList.push_back(data);

            // Vertex buffer will not be bound yet, we expect this to be done by the

            // caller when it becomes appropriate (e.g. through a VertexAnimationTrack)

        }

        return supportedCount;

    }

    VertexData* VertexData::_cloneRemovingBlendData() const

    {

        // Clone without copying data

        VertexData* ret = clone(false);

        bool removeIndices = Root::getSingleton().isBlendIndicesGpuRedundant();

        bool removeWeights = Root::getSingleton().isBlendWeightsGpuRedundant();

        unsigned short safeSource = 0xFFFF;

        auto blendIndexElem = vertexDeclaration->findElementBySemantic(VES_BLEND_INDICES);

        if (blendIndexElem && removeIndices)

        {

            //save the source in order to prevent the next stage from unbinding it.

            safeSource = blendIndexElem->getSource();

            // Remove buffer reference

            ret->vertexBufferBinding->unsetBinding(blendIndexElem->getSource());

        }

        // Remove blend weights

        const VertexElement* blendWeightElem = vertexDeclaration->findElementBySemantic(VES_BLEND_WEIGHTS);

        if (removeWeights && blendWeightElem && blendWeightElem->getSource() != safeSource)

        {

            // Remove buffer reference

            ret->vertexBufferBinding->unsetBinding(blendWeightElem->getSource());

        }

        // remove elements from declaration

        if (removeIndices)

            ret->vertexDeclaration->removeElement(VES_BLEND_INDICES);

        if (removeWeights)

            ret->vertexDeclaration->removeElement(VES_BLEND_WEIGHTS);

        // Close gaps in bindings for effective and safely

        if (removeWeights || removeIndices)

            ret->closeGapsInBindings();

        return ret;

    }

    //-----------------------------------------------------------------------

    //-----------------------------------------------------------------------

    IndexData::IndexData()

    {

        indexCount = 0;

        indexStart = 0;

    }

    //-----------------------------------------------------------------------

    IndexData::~IndexData()

    {

    }

    //-----------------------------------------------------------------------

    IndexData* IndexData::clone(bool copyData, HardwareBufferManagerBase* mgr) const

    {

        HardwareBufferManagerBase* pManager = mgr ? mgr : HardwareBufferManager::getSingletonPtr();

        IndexData* dest = OGRE_NEW IndexData();

        if (indexBuffer.get())

        {

            if (copyData)

            {

                dest->indexBuffer = pManager->createIndexBuffer(indexBuffer->getType(), indexBuffer->getNumIndexes(),

                    indexBuffer->getUsage(), indexBuffer->hasShadowBuffer());

                dest->indexBuffer->copyData(*indexBuffer, 0, 0, indexBuffer->getSizeInBytes(), true);

            }

            else

            {

                dest->indexBuffer = indexBuffer;

            }

        }

        dest->indexCount = indexCount;

        dest->indexStart = indexStart;

        return dest;

    }

    //-----------------------------------------------------------------------

    //-----------------------------------------------------------------------

    // Local Utility class for vertex cache optimizer

    class Triangle

    {

    public:

        enum EdgeMatchType {

            AB, BC, CA, ANY, NONE

        };

        uint32 a, b, c;     

        inline Triangle()

        {

        }

        inline Triangle( uint32 ta, uint32 tb, uint32 tc ) 

            : a( ta ), b( tb ), c( tc )

        {

        }

        inline Triangle( uint32 t[3] )

            : a( t[0] ), b( t[1] ), c( t[2] )

        {

        }

        inline Triangle( const Triangle& t )

            : a( t.a ), b( t.b ), c( t.c )

        {

        }

        inline Triangle& operator=(const Triangle& rhs) {

            a = rhs.a;

            b = rhs.b;

            c = rhs.c;

            return *this;

        }

        inline bool sharesEdge(const Triangle& t) const

        {

            return( (a == t.a && b == t.c) ||

                    (a == t.b && b == t.a) ||

                    (a == t.c && b == t.b) ||

                    (b == t.a && c == t.c) ||

                    (b == t.b && c == t.a) ||

                    (b == t.c && c == t.b) ||

                    (c == t.a && a == t.c) ||

                    (c == t.b && a == t.a) ||

                    (c == t.c && a == t.b) );

        }

        inline bool sharesEdge(const uint32 ea, const uint32 eb, const Triangle& t) const

        {

            return( (ea == t.a && eb == t.c) ||

                    (ea == t.b && eb == t.a) ||

                    (ea == t.c && eb == t.b) ); 

        }

        inline bool sharesEdge(const EdgeMatchType edge, const Triangle& t) const

        {

            if (edge == AB)

                return sharesEdge(a, b, t);

            else if (edge == BC)

                return sharesEdge(b, c, t);

            else if (edge == CA)

                return sharesEdge(c, a, t);

            else

                return (edge == ANY) == sharesEdge(t);

        }

        inline EdgeMatchType endoSharedEdge(const Triangle& t) const

        {

            if (sharesEdge(a, b, t)) return AB;

            if (sharesEdge(b, c, t)) return BC;

            if (sharesEdge(c, a, t)) return CA;

            return NONE;

        }

        inline EdgeMatchType exoSharedEdge(const Triangle& t) const

        {

            return t.endoSharedEdge(*this);

        }

        inline void shiftClockwise()

        {

            uint32 t = a;

            a = c;

            c = b;

            b = t;

        }

        inline void shiftCounterClockwise()

        {

            uint32 t = a;

            a = b;

            b = c;

            c = t;

        }

    };

    //-----------------------------------------------------------------------