diff -r 000000000000 -r 5d03bc08d59c graphicsdeviceinterface/gdi/inc/blendingalgorithms.h
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/graphicsdeviceinterface/gdi/inc/blendingalgorithms.h	Tue Feb 02 01:47:50 2010 +0200
@@ -0,0 +1,129 @@
+// Copyright (c) 2006-2009 Nokia Corporation and/or its subsidiary(-ies).
+// All rights reserved.
+// This component and the accompanying materials are made available
+// under the terms of "Eclipse Public License v1.0"
+// which accompanies this distribution, and is available
+// at the URL "http://www.eclipse.org/legal/epl-v10.html".
+//
+// Initial Contributors:
+// Nokia Corporation - initial contribution.
+//
+// Contributors:
+//
+// Description:
+//
+
+#ifndef __BLENDINGALGORITHMS_H__
+#define __BLENDINGALGORITHMS_H__
+
+#include <e32def.h>
+/**
+@file
+@internalTechnology
+*/
+
+/** Premultiplied Alpha (PMA) Blending algorithm.
+Assumes src and target pixels are in PMA format (i.e., the color channels are already multiplied
+with the alpha channel.
+Blending Equations:
+Cd' = Cs + ((255 - As) * Cd)/255
+Ad' = As + ((255 - As) * Ad)/255
+where Cd' and Ad' are destination values for color and alpha channels respectively, after blending,
+Cd and Ad are destination values before blending, and Cs and As are source values (color and
+alpha values respectively).
+Optimisations:
+Use of 256 instead of 255 converts the division to shift (faster).
+Red and Blue channels are calculated together in one 32 bit operation.
+Alpha and Green channels are calculated in the same way. However, due the
+position of these channels within the pixel format, an extra shift is required.
+The equations used are:
+Cd' = Cs + (((0x100 - As) * Cd)>>8)
+Ad' = As + (((0x100 - As) * Ad)>>8)
+For the results to not overflow, it is required that the value of any of the color channels
+is no greater than the alpha channel, which is what is exepected of pma format.
+It is also assumed that the aMask is equal to source alpha.
+@param 	aDestPixel	The destination pixel value.
+@param	aSrcPixel	The source pixel value.
+@param	aMask		The source alpha. It is assumed to be same as source alpha.
+@return the alpha-blended pixel (within the PMA space).
+@internalTechnology
+@released
+*/
+inline TUint32 PMAPixelBlend(TUint32 aDestPixel, TUint32 aSrcPixel, TUint8 aMask);
+
+/** Same as inline TUint32 PMAPixelBlend(TUint32 aDestPixel, TUint32 aSrcPixel, TUint8 aMask)
+Calls inline TUint32 PMAPixelBlend(TUint32 aDestPixel, TUint32 aSrcPixel, TUint8 aMask), with aMask
+extracted from the source pixel.
+@see PMAPixelBlend(TUint32 aDestPixel, TUint32 aSrcPixel, TUint8 aMask)
+@internalTechnology
+@released
+*/
+inline TUint32 PMAPixelBlend(TUint32 aDestPixel, TUint32 aSrcPixel);
+
+/**
+Does the premultiplied alpha blend, but does not check for any pre/post conditions.
+Please DO NOT Add any optimisations to check for these conditions in this code!
+@internalTechnology
+@released
+*/
+inline TUint32 PMABlend_noChecks(TUint32 aDestPixel, TUint32 aSrcPixel, TUint8 aMaskingFactor);
+
+/**
+The version of PMABlend_noChecks which does the blending in place.
+@internalTechnology
+@released
+*/
+inline void PMABlend_noChecksInplace(TUint32& aDest_io, const TUint32& aSrcPixel, TUint8 aMaskingFactor);
+
+/**
+The in-place version of the PMAPixelBlend algorithm
+@internalTechnology
+@released
+*/
+inline void PMAInplaceBlend(TUint32& aDest_io, TUint32& aSrc_in);
+/**
+Premultiplies the color channel values with the Alpha channel value.
+Alpha value remains unchanged. An approximation is used in the operation where the division
+by 255 is approximated by a shift-by-8-bits operation (i.e. division by 256).
+@param	aPixel	The 32 bit pixel value to be pre-multiplied.
+@return	The PMA value.
+@internalTechnology
+@released
+*/
+inline TUint32 NonPMA2PMAPixel(TUint32 aPixel);
+
+/**
+Divives the PMA pixel color channels with the Alpha value, to convert them to non-PMA format.
+Alpha value remains unchanged.
+@param	aPixel	the premultiplied 32 bit pixel value.
+@param	aNormTable	The lookup table used to do the normalisation (the table converts the division
+					to multiplication operation).
+					The table is usually obtainable by a call to the method:
+					PtrTo16BitNormalisationTable, which is defined in lookuptable.dll(.lib).
+					The lookup table for normalised alpha is compluted using this equation: 
+					Table[index] = (255*256) / index (where index is an 8 bit value).
+@return The NON-PMA 32 bit pixel value.
+@internalTechnology
+@released
+*/
+inline TUint32 PMA2NonPMAPixel(TUint32 aPixel, const TUint16* aNormTable);
+
+/**
+In-place version of NonPMA2PMAPixel.
+@see NonPMA2PMAPixel
+@internalTechnology
+@released
+*/
+inline void Convert2PMA(TUint32& aInOutValue);
+
+/**
+In-place version of PMA2NonPMAPixel
+@see PMA2NonPMAPixel
+@internalTechnology
+@released
+*/
+inline void Convert2NonPMA(TUint32& aInOutValue, const TUint16* aNormTable);
+
+#include <graphics/blendingalgorithms.inl>
+
+#endif //__BLENDINGALGORITHMS_H__