|
|
/*++
Copyright (c) 1990-1991 Microsoft Corporation
Module Name:
htalias.c
Abstract:
This module contains all low levels halftone rendering functions.
Author:
22-Jan-1991 Tue 12:49:03 created -by- Daniel Chou (danielc)
[Environment:]
GDI Device Driver - Halftone.
[Notes:]
Revision History:
--*/
#define DBGP_VARNAME dbgpHTAlias
#include "htp.h"
#include "htmapclr.h"
#include "limits.h"
#include "htpat.h"
#include "htalias.h"
#include "htstret.h"
#include "htrender.h"
#include "htgetbmp.h"
#include "htsetbmp.h"
#define DBGP_BUILD 0x00000001
#define DBGP_BUILD2 0x00000002
#define DBGP_BUILD3 0x00000004
#define DBGP_EXPMATRIX 0x00000008
#define DBGP_EXP 0x00000010
#define DBGP_AAHEADER 0x00000020
#define DBGP_OUTAA 0x00000040
#define DBGP_FUNC 0x00000080
#define DBGP_AAHT 0x00000100
#define DBGP_AAHT_MEM 0x00000200
#define DBGP_AAHT_TIME 0x00000400
#define DBGP_AAHTPAT 0x00000800
#define DBGP_FIXUPDIB 0x00001000
#define DBGP_INPUT 0x00002000
#define DBGP_VGA256XLATE 0x00004000
#define DBGP_EXPAND 0x00008000
#define DBGP_GETFIXUP 0x00010000
#define DBGP_PSD 0x00020000
#define DBGP_MASK 0x00040000
#define DBGP_PALETTE 0x00080000
#define DBGP_PAL_CHKSUM 0x00100000
#define DBGP_LUT_MAP 0x00200000
DEF_DBGPVAR(BIT_IF(DBGP_BUILD, 0) | BIT_IF(DBGP_BUILD2, 0) | BIT_IF(DBGP_BUILD3, 0) | BIT_IF(DBGP_EXPMATRIX, 0) | BIT_IF(DBGP_EXP, 0) | BIT_IF(DBGP_AAHEADER, 0) | BIT_IF(DBGP_OUTAA, 0) | BIT_IF(DBGP_FUNC, 0) | BIT_IF(DBGP_AAHT, 0) | BIT_IF(DBGP_AAHT_MEM, 0) | BIT_IF(DBGP_AAHT_TIME, 0) | BIT_IF(DBGP_AAHTPAT, 0) | BIT_IF(DBGP_FIXUPDIB, 0) | BIT_IF(DBGP_INPUT, 0) | BIT_IF(DBGP_VGA256XLATE, 0) | BIT_IF(DBGP_EXPAND, 0) | BIT_IF(DBGP_GETFIXUP, 0) | BIT_IF(DBGP_PSD, 0) | BIT_IF(DBGP_MASK, 0) | BIT_IF(DBGP_PALETTE, 0) | BIT_IF(DBGP_PAL_CHKSUM, 0) | BIT_IF(DBGP_LUT_MAP, 0))
extern CONST RGBORDER SrcOrderTable[];extern DWORD dwABPreMul[256];
#define SIZE_AAINFO _ALIGN_MEM(sizeof(AAINFO))
//
// Following computation is based on
//
// NTSC_R_INT = 299000
// NTSC_G_INT = 587000
// NTSC_B_INT = 114000
// GRAY_MAX_IDX = 0xFFFF
//
// NTSC_R_GRAY_MAX = (((NTSC_R_INT * GRAY_MAX_IDX) + 500000) / 1000000)
// NTSC_B_GRAY_MAX = (((NTSC_B_INT * GRAY_MAX_IDX) + 500000) / 1000000)
// NTSC_G_GRAY_MAX = (GRAY_MAX_IDX - NTSC_R_GRAY_MAX - NTSC_B_GRAY_MAX)
//
#define NTSC_R_GRAY_MAX (DWORD)0x4c8b
#define NTSC_B_GRAY_MAX (DWORD)0x1d2f
#define NTSC_G_GRAY_MAX (DWORD)(0xFFFF - NTSC_R_GRAY_MAX - NTSC_B_GRAY_MAX)
VOIDHTENTRYSetGrayColorTable( PLONG pIdxBGR, PAASURFINFO pAASI )
/*++
Routine Description:
Arguments:
pIdxBGR - Pointer to how to translate from RGB to gray scale, typically this pointer is computed with NTSC gray standard plus any device transform or color adjustment, but if this pointer is NULL then we are reading from the device (1bpp, 8bpp) so we will only do NTSC standard mapping.
Return Value:
Author:
19-Feb-1999 Fri 13:14:01 created -by- Daniel Chou (danielc)
Revision History:
08-Aug-2000 Tue 18:34:22 updated -by- Daniel Chou (danielc) Fixing bug for alpha blending, in gray scale mode, the destination can only be 1bpp or 8bpp mask mono, so when we read back from the destination to do alpha blending, it will double color mapping pixels. In gray scale mode, the input function will map the source RGB value to gray value with the current device transform, color adjustment and so on, so if we read back from destination then this transform is not desired.
--*/
{ PRGB4B prgb4b; LONG cSrcTable;
ASSERT(NTSC_R_INT == 299000); ASSERT(NTSC_G_INT == 587000); ASSERT(NTSC_B_INT == 114000); ASSERT(GRAY_MAX_IDX == 0xFFFF);
if (cSrcTable = (LONG)pAASI->cClrTable) {
prgb4b = pAASI->pClrTable;
if (pIdxBGR) {
//
// This is a reading from the source
//
ASSERT(pAASI->Flags & AASIF_GRAY); ASSERT(pAASI->pIdxBGR == pIdxBGR);
while (cSrcTable--) {
prgb4b++->a = IDXBGR_2_GRAY_BYTE(pIdxBGR, prgb4b->b, prgb4b->g, prgb4b->r); }
} else {
//
// We are reading from the destination so only do NTSC standard,
// since the destination surface alreay halftoned.
//
while (cSrcTable--) {
prgb4b++->a = (BYTE)((((DWORD)prgb4b->r * NTSC_R_GRAY_MAX) + ((DWORD)prgb4b->g * NTSC_G_GRAY_MAX) + ((DWORD)prgb4b->b * NTSC_B_GRAY_MAX) + (0x7FFF)) / 0xFFFF); } }
} else if (pIdxBGR != pAASI->pIdxBGR) {
//
// This is the source surface info, we will see if this is the
// gray color table
//
ASSERT(pAASI->Flags & AASIF_GRAY); ASSERT(pIdxBGR); ASSERT(pAASI->pIdxBGR);
DBGP_IF(DBGP_INPUT, DBGP("Copy pIdxBGR [BGR] order from 012 to %ld%ld%ld" ARGDW(pAASI->AABFData.MaskRGB[2]) ARGDW(pAASI->AABFData.MaskRGB[1]) ARGDW(pAASI->AABFData.MaskRGB[0])));
CopyMemory(&pAASI->pIdxBGR[pAASI->AABFData.MaskRGB[2] * 256], &pIdxBGR[0 * 256], sizeof(LONG) * 256); CopyMemory(&pAASI->pIdxBGR[pAASI->AABFData.MaskRGB[1] * 256], &pIdxBGR[1 * 256], sizeof(LONG) * 256); CopyMemory(&pAASI->pIdxBGR[pAASI->AABFData.MaskRGB[0] * 256], &pIdxBGR[2 * 256], sizeof(LONG) * 256); }}
VOIDHTENTRYComputeInputColorInfo( LPBYTE pSrcTable, UINT cPerTable, UINT PrimaryOrder, PBFINFO pBFInfo, PAASURFINFO pAASI )
/*++
Routine Description:
Arguments:
Return Value:
Author:
19-Feb-1999 Fri 13:14:01 created -by- Daniel Chou (danielc)
Revision History:
--*/
{ PRGB4B prgb4b; PAABFDATA pAABFData; LONG cSrcTable; LONG Count; LONG LS; LONG RS; BYTE Mask;
pAABFData = &pAASI->AABFData;
if ((pSrcTable) && (Count = cSrcTable = (LONG)pAASI->cClrTable)) {
UINT iR; UINT iG; UINT iB;
prgb4b = pAASI->pClrTable; iR = SrcOrderTable[PrimaryOrder].Order[0]; iG = SrcOrderTable[PrimaryOrder].Order[1]; iB = SrcOrderTable[PrimaryOrder].Order[2];
switch (pAABFData->Format) {
case BMF_1BPP:
pAASI->InputFunc = (AAINPUTFUNC)Input1BPPToAA24; break;
case BMF_4BPP_VGA16: case BMF_4BPP:
pAASI->InputFunc = (AAINPUTFUNC)Input4BPPToAA24; break;
case BMF_8BPP_VGA256: case BMF_8BPP:
pAASI->InputFunc = (AAINPUTFUNC)Input8BPPToAA24; break;
default:
DBGP("ComputeInputColorInfo() Invalid Bitmap Format=%ld" ARGDW(pAABFData->Format));
break; }
while (Count--) {
prgb4b->r = *(pSrcTable + iR); prgb4b->g = *(pSrcTable + iG); prgb4b->b = *(pSrcTable + iB);
DBGP_IF(DBGP_PALETTE, DBGP("Source Pal [%3ld] = %3ld:%3ld:%3ld" ARGDW((LONG)cSrcTable - Count - 1) ARGDW(prgb4b->r) ARGDW(prgb4b->g) ARGDW(prgb4b->b)));
++prgb4b; pSrcTable += cPerTable; }
} else {
pAASI->InputFunc = InputAABFDATAToAA24;
if (pBFInfo->Flags & BFIF_RGB_888) {
pAABFData->Flags |= AABF_MASK_IS_ORDER; pAABFData->MaskRGB[0] = (BYTE)pBFInfo->RGBOrder.Order[0]; pAABFData->MaskRGB[1] = (BYTE)pBFInfo->RGBOrder.Order[1]; pAABFData->MaskRGB[2] = (BYTE)pBFInfo->RGBOrder.Order[2];
} else {
//
// This is bitfield, figure out how to do it, we want to lshift to
// to edge then right shift to 8bpp then mask off the unwanted bit
//
//
cSrcTable = 3;
while (cSrcTable--) {
LS = 0; RS = (LONG)pBFInfo->BitStart[cSrcTable]; Count = (LONG)pBFInfo->BitCount[cSrcTable];
if (Count >= 8) {
RS += (Count - 8); Mask = 0xFF;
} else {
LS = 8 - Count; Mask = (0xFF << LS);
if ((RS -= LS) < 0) {
LS = -RS; RS = 0;
} else {
LS = 0; } }
pAABFData->MaskRGB[cSrcTable] = (BYTE)Mask; pAABFData->LShiftRGB[cSrcTable] = (BYTE)LS; pAABFData->RShiftRGB[cSrcTable] = (BYTE)RS;
DBGP_IF(DBGP_FUNC | DBGP_PALETTE, DBGP("BFData[%ld]: Bits=%08lx, LS=%2ld, RS=%2ld, Mask=%02lx -->%02lx" ARGDW(cSrcTable) ARGDW(pBFInfo->BitsRGB[cSrcTable]) ARGDW(LS) ARGDW(RS) ARGDW(Mask) ARGDW(((pBFInfo->BitsRGB[cSrcTable] >> RS) << LS) & Mask))); } }
switch (pBFInfo->BitmapFormat) {
case BMF_16BPP: case BMF_16BPP_555: case BMF_16BPP_565:
pAABFData->cbSrcInc = 2; break;
case BMF_24BPP:
ASSERT(pAABFData->Flags & AABF_MASK_IS_ORDER);
if (pBFInfo->RGBOrder.Index == PRIMARY_ORDER_BGR) {
ASSERT((pAABFData->MaskRGB[0] == 2) && (pAABFData->MaskRGB[1] == 1) && (pAABFData->MaskRGB[2] == 0));
pAABFData->Flags |= AABF_SRC_IS_BGR8; }
pAABFData->cbSrcInc = 3;
break;
case BMF_32BPP:
if (pAASI->Flags & AASIF_AB_PREMUL_SRC) {
ASSERT(pAABFData->Flags & AABF_MASK_IS_ORDER); ASSERT(dwABPreMul[0] == 0);
switch (pBFInfo->RGBOrder.Index) {
case PRIMARY_ORDER_BGR:
ASSERT((pAABFData->MaskRGB[0] == 2) && (pAABFData->MaskRGB[1] == 1) && (pAABFData->MaskRGB[2] == 0));
pAABFData->Flags |= AABF_SRC_IS_BGR_ALPHA; break;
case PRIMARY_ORDER_RGB:
ASSERT((pAABFData->MaskRGB[0] == 0) && (pAABFData->MaskRGB[1] == 1) && (pAABFData->MaskRGB[2] == 2));
pAABFData->Flags |= AABF_SRC_IS_RGB_ALPHA; break;
default:
break; }
if (dwABPreMul[0] == 0) {
pAASI->InputFunc = InputPreMul32BPPToAA24; } }
pAABFData->cbSrcInc = 4; break;
default:
DBGP("ERROR: Invalid BFInfo Format=%ld" ARGDW(pBFInfo->BitmapFormat)); break; }
DBGP_IF(DBGP_FUNC, DBGP("Flags=%02lx. cbSrcInc=%ld, Mask=%02lx:%02lx:%02lx, LS=%2ld:%2ld:%2ld, RS=%2ld:%2ld:%2ld" ARGDW(pAABFData->Flags) ARGDW(pAABFData->cbSrcInc) ARGDW(pAABFData->MaskRGB[0]) ARGDW(pAABFData->MaskRGB[1]) ARGDW(pAABFData->MaskRGB[2]) ARGDW(pAABFData->LShiftRGB[0]) ARGDW(pAABFData->LShiftRGB[1]) ARGDW(pAABFData->LShiftRGB[2]) ARGDW(pAABFData->RShiftRGB[0]) ARGDW(pAABFData->RShiftRGB[1]) ARGDW(pAABFData->RShiftRGB[2]))); }
DBGP_IF(DBGP_FUNC, DBGP("+++++ InputFunc = %hs(SrcFmt=%ld), cClrTable=%ld\n" ARGPTR(GetAAInputFuncName(pAASI->InputFunc)) ARGDW(pBFInfo->BitmapFormat) ARGDW(pAASI->cClrTable)));}
PAAINFOHTENTRYBuildTileAAInfo( PDEVICECOLORINFO pDCI, DWORD AAHFlags, PLONG piSrcBeg, PLONG piSrcEnd, LONG SrcSize, LONG IdxDst, LONG IdxDstEnd, PLONG piDstBeg, PLONG piDstEnd, LONG cbExtra )
/*++
Routine Description:
Arguments:
piSrcBeg - Passed in as begining source index, When return this is the real source begining index. This always well ordered
piSrcEnd - Passed in as ending source index, When return this is the real source ending index. This always well ordered
SrcSize - The real size of source in pixel
IdxDst - Starting index of destination pixel
IdxDstEnd - Ending index of destination pixel, the iDxdst and IdxDstEnd must be well ordered.
piDstBeg - Clipped destination start index when passed in, at return it adjusted to the real destination starting index.
piDstEnd - Clipped destination end index when passed in, at return it adjusted to the real destination ending index.
cbExtra - Extra byte count to be allocated
NOTE: 1) piDstBeg/piDstEnd When passed in this must be well ordered, and when it returned this is well ordered.
Return Value:
At enter of this function, the *piSrcEnd, *piDstEnd is exclusive but when return from this function the *piSrcEnd and *piDstEnd is inclusive
*piSrcBeg, *piSrcEnd, *piDstBeg, *piDstEnd are updated if return value is not NULL
Author:
22-Mar-1998 Sun 18:36:28 created -by- Daniel Chou (danielc)
Revision History:
--*/
{ PAAINFO pAAInfo; PEXPDATA pED; LONG cMatrix; LONG cM; LONG cIn; LONG cOut; LONG IdxSrc; LONG iSrcBeg; LONG iSrcEnd; LONG iDstBeg; LONG iDstEnd; LONG jSrcBeg; LONG jSrcEnd; LONG jDstBeg; LONG jDstEnd; LONG cLoop; LONG cAAData;
iSrcBeg = *piSrcBeg; iSrcEnd = *piSrcEnd;
//
// The source always clipped to the visiable surface area
//
if (iSrcBeg < 0) {
iSrcBeg = 0; }
if (iSrcEnd > SrcSize) {
iSrcEnd = SrcSize; }
cIn = iSrcEnd - (IdxSrc = iSrcBeg); cOut = IdxDstEnd - IdxDst;
if (cIn <= 0) {
return(NULL); }
ASSERT(cOut > 0);
iDstBeg = *piDstBeg; iDstEnd = *piDstEnd; jSrcBeg = -1;
ASSERT(iDstBeg < iDstEnd);
DBGP_IF(DBGP_BUILD, DBGP("\nTile(%ld-%ld): iSrc=%ld-%ld, cSrc=%ld, iDst=%ld-%ld, Idx=%ld:%ld" ARGDW(cIn) ARGDW(cOut) ARGDW(iSrcBeg) ARGDW(iSrcEnd) ARGDW(SrcSize) ARGDW(iDstBeg) ARGDW(iDstEnd) ARGDW(IdxSrc) ARGDW(IdxDst)));
ALIGN_MEM(cbExtra, cbExtra);
if (pAAInfo = (PAAINFO)HTAllocMem((LPVOID)pDCI, HTMEM_BLTAA, LPTR, SIZE_AAINFO + cbExtra)) {
SETDBGVAR(pAAInfo->cbAlloc, SIZE_AAINFO + cbExtra);
pAAInfo->pbExtra = (LPBYTE)pAAInfo + SIZE_AAINFO; cLoop = cOut;
while (cLoop--) {
if ((IdxSrc >= iSrcBeg) && (IdxSrc < iSrcEnd) && (IdxDst >= iDstBeg) && (IdxDst < iDstEnd)) {
if (jSrcBeg == -1) {
jSrcBeg = IdxSrc; jDstBeg = IdxDst; }
jSrcEnd = IdxSrc; jDstEnd = IdxDst;
} else if (jSrcBeg != -1) {
break; }
if (++IdxSrc >= iSrcEnd) {
IdxSrc = iSrcBeg; }
++IdxDst; }
if (jSrcBeg == -1) {
HTFreeMem(pAAInfo); return(NULL); }
pAAInfo->iSrcBeg = jSrcBeg - iSrcBeg; jSrcBeg = iSrcBeg; jSrcEnd = iSrcEnd - 1; pAAInfo->Mask.iBeg = *piSrcBeg = jSrcBeg; *piSrcEnd = jSrcEnd; *piDstBeg = jDstBeg; *piDstEnd = jDstEnd; pAAInfo->Mask.iSize = pAAInfo->cIn = jSrcEnd - jSrcBeg + 1; pAAInfo->cOut = jDstEnd - jDstBeg + 1; pAAInfo->cAAData = pAAInfo->cAALoad = (DWORD)pAAInfo->cOut; pAAInfo->Mask.cIn = cIn; pAAInfo->Mask.cOut = cOut;
DBGP_IF(DBGP_BUILD, DBGP("TILE(%ld->%ld): iSrc=%ld:%ld (%ld), iDst=%ld:%ld, cAAData=%ld, cbExtra=%ld, Flags=%04lx" ARGDW(pAAInfo->cIn) ARGDW(pAAInfo->cOut) ARGDW(*piSrcBeg) ARGDW(*piSrcEnd) ARGDW(pAAInfo->iSrcBeg) ARGDW(*piDstBeg) ARGDW(*piDstEnd) ARGDW(pAAInfo->cAAData) ARGDW(cbExtra) ARGDW(pAAInfo->Flags))); }
return(pAAInfo);}
PAAINFOHTENTRYBuildBltAAInfo( PDEVICECOLORINFO pDCI, DWORD AAHFlags, PLONG piSrcBeg, PLONG piSrcEnd, LONG SrcSize, LONG IdxDst, LONG IdxDstEnd, PLONG piDstBeg, PLONG piDstEnd, LONG cbExtra )
/*++
Routine Description:
Arguments:
piSrcBeg - Passed in as begining source index, When return this is the real source begining index. This always well ordered
piSrcEnd - Passed in as ending source index, When return this is the real source ending index. This always well ordered
SrcSize - The real size of source in pixel
IdxDst - Starting index of destination pixel
IdxDstEnd - Ending index of destination pixel, the iDxdst and IdxDstEnd must be well ordered.
piDstBeg - Clipped destination start index when passed in, at return it adjusted to the real destination starting index.
piDstEnd - Clipped destination end index when passed in, at return it adjusted to the real destination ending index.
cbExtra - Extra byte count to be allocated
NOTE: 1) piDstBeg/piDstEnd When passed in this must be well ordered, and when it returned this is well ordered.
Return Value:
At enter of this function, the *piSrcEnd, *piDstEnd is exclusive but when return from this function the *piSrcEnd and *piDstEnd is inclusive
*piSrcBeg, *piSrcEnd, *piDstBeg, *piDstEnd are updated if return value is not NULL
Author:
22-Mar-1998 Sun 18:36:28 created -by- Daniel Chou (danielc)
Revision History:
--*/
{ PAAINFO pAAInfo; PEXPDATA pED; LONG cMatrix; LONG cM; LONG cIn; LONG cOut; LONG IdxSrc; LONG iSrcBeg; LONG iSrcEnd; LONG iDstBeg; LONG iDstEnd; LONG jSrcBeg; LONG jSrcEnd; LONG jDstBeg; LONG jDstEnd; LONG cLoop; LONG cAAData;
cIn = (LONG)((iSrcEnd = *piSrcEnd) - (iSrcBeg = IdxSrc = *piSrcBeg)); cOut = IdxDstEnd - IdxDst;
ASSERT(cOut > 0);
if (iSrcBeg < 0) {
iSrcBeg = 0; }
if (iSrcEnd > SrcSize) {
iSrcEnd = SrcSize; }
iDstBeg = *piDstBeg; iDstEnd = *piDstEnd; jSrcBeg = -1;
ASSERT(iDstBeg < iDstEnd); ASSERT(cIn == cOut);
DBGP_IF(DBGP_BUILD, DBGP("\nBlt(%ld-%ld): iSrc=%ld-%ld, cSrc=%ld, iDst=%ld-%ld, Idx=%ld:%ld" ARGDW(cIn) ARGDW(cOut) ARGDW(iSrcBeg) ARGDW(iSrcEnd) ARGDW(SrcSize) ARGDW(iDstBeg) ARGDW(iDstEnd) ARGDW(IdxSrc) ARGDW(IdxDst)));
ALIGN_MEM(cbExtra, cbExtra);
if (pAAInfo = (PAAINFO)HTAllocMem((LPVOID)pDCI, HTMEM_BLTAA, LPTR, SIZE_AAINFO + cbExtra)) {
SETDBGVAR(pAAInfo->cbAlloc, SIZE_AAINFO + cbExtra);
pAAInfo->pbExtra = (LPBYTE)pAAInfo + SIZE_AAINFO; cLoop = cOut;
while (cLoop--) {
if ((IdxSrc >= iSrcBeg) && (IdxSrc < iSrcEnd) && (IdxDst >= iDstBeg) && (IdxDst < iDstEnd)) {
if (jSrcBeg == -1) {
jSrcBeg = IdxSrc; jDstBeg = IdxDst; }
jSrcEnd = IdxSrc; jDstEnd = IdxDst;
} else if (jSrcBeg != -1) {
break; }
++IdxSrc; ++IdxDst; }
if (jSrcBeg == -1) {
HTFreeMem(pAAInfo); return(NULL); }
pAAInfo->Mask.iBeg = *piSrcBeg = jSrcBeg; *piSrcEnd = jSrcEnd; *piDstBeg = jDstBeg; *piDstEnd = jDstEnd; pAAInfo->Mask.iSize = pAAInfo->cIn = jSrcEnd - jSrcBeg + 1; pAAInfo->cOut = jDstEnd - jDstBeg + 1; pAAInfo->cAAData = pAAInfo->cAALoad = (DWORD)pAAInfo->cOut; pAAInfo->Mask.cIn = cIn; pAAInfo->Mask.cOut = cOut;
DBGP_IF(DBGP_BUILD, DBGP("BLT(%ld->%ld): iSrc=%ld:%ld, iDst=%ld:%ld, cAAData=%ld, cbExtra=%ld, Flags=%4lx" ARGDW(pAAInfo->cIn) ARGDW(pAAInfo->cOut) ARGDW(*piSrcBeg) ARGDW(*piSrcEnd) ARGDW(*piDstBeg) ARGDW(*piDstEnd) ARGDW(pAAInfo->cAAData) ARGDW(cbExtra) ARGDW(pAAInfo->Flags))); }
return(pAAInfo);}
#define _MATRIX_POW (FD6)1414214
#if DBG
FD6 MATRIX_POWER = _MATRIX_POW;#else
#define MATRIX_POWER _MATRIX_POW
#endif
BOOLHTENTRYBuildRepData( PSRCBLTINFO pSBInfo, LONG IdxSrc, LONG IdxDst )
/*++
Routine Description:
Arguments:
piSrcBeg - Passed in as begining source index, When return this is the real source begining index. This always well ordered
piSrcEnd - Passed in as ending source index, When return this is the real source ending index. This always well ordered
SrcSize - The real size of source in pixel
IdxDst - Starting index of destination pixel
IdxDstEnd - Ending index of destination pixel, the iDxdst and IdxDstEnd must be well ordered. (exclusive)
piDstBeg - Clipped destination start index when passed in, at return it adjusted to the real destination starting index.
piDstEnd - Clipped destination end index when passed in, at return it adjusted to the real destination ending index.
cbExtra - Extra byte count to be allocated
NOTE: 1) piDstBeg/piDstEnd When passed in this must be well ordered, and when it returned this is well ordered.
Return Value:
At enter of this function, the *piSrcEnd, *piDstEnd is exclusive but when return from this function the *piSrcEnd and *piDstEnd is inclusive
*piSrcBeg, *piSrcEnd, *piDstBeg, *piDstEnd are updated if return value is not NULL
Author:
22-Mar-1998 Sun 18:36:28 created -by- Daniel Chou (danielc)
Revision History:
--*/
{ PREPDATA pRep; PREPDATA pRepEnd; SRCBLTINFO SBInfo; PLONG pSrcInc; PLONG pDstInc; LONG MinRep; LONG MaxRep; LONG cRem; LONG iRep; LONG cRep; LONG cTot; LONG cIn; LONG cOut; LONG jSrcBeg; LONG jSrcEnd; LONG jDstBeg; LONG jDstEnd; UINT cPrevSrc; UINT cNextSrc;
SBInfo = *pSBInfo; pRep = SBInfo.pRep; pRepEnd = SBInfo.pRepEnd; jSrcBeg = -1;
if (SBInfo.cIn < SBInfo.cOut) {
//
// Expanding
//
cIn = SBInfo.cIn; cOut = SBInfo.cOut; pSrcInc = &IdxSrc; pDstInc = &IdxDst;
} else if (SBInfo.cIn > SBInfo.cOut) {
//
// Shrinking
//
cIn = SBInfo.cOut; cOut = SBInfo.cIn; pSrcInc = &IdxDst; pDstInc = &IdxSrc;
} else {
ASSERT(SBInfo.cIn != SBInfo.cOut); return(FALSE); }
MinRep = cOut / cIn; MaxRep = MinRep + 1;
DBGP_IF(DBGP_BUILD, DBGP("\nEXPREP(%ld-%ld): iSrc=%ld-%ld, iDst=%ld-%ld, Idx=%ld:%ld" ARGDW(SBInfo.cIn) ARGDW(SBInfo.cOut) ARGDW(SBInfo.iSrcBeg) ARGDW(SBInfo.iSrcEnd) ARGDW(SBInfo.iDstBeg) ARGDW(SBInfo.iDstEnd) ARGDW(IdxSrc) ARGDW(IdxDst)));
SBInfo.cFirstSkip = SBInfo.cLastSkip = 0; iRep = cRep = cTot = 0;
//
// Multiply In/Out by 2, so we do not have run down/up problem, the
// cRem will be initialized with extra 0.5 for rounding.
//
cRem = (cOut <<= 1) + cIn; cIn <<= 1;
while (IdxDst < SBInfo.iDstEnd) {
if ((cRem -= cIn) < 0) {
++*pSrcInc;
if (jSrcBeg != -1) {
ASSERT(pRep < pRepEnd); ASSERT(cRep > 0); ASSERT(cRep <= MaxRep);
pRep++->c = (WORD)cRep; cTot += cRep;
DBGP_IF(DBGP_BUILD2, DBGP(" Src=%4ld [%4ld-%4ld], Dst=%4ld [%4ld-%4ld], Rep %4ld=%3ld [%4ld -> %4ld, Rem=%4ld, cTot=%4ld]" ARGDW(IdxSrc) ARGDW(SBInfo.iSrcBeg) ARGDW(SBInfo.iSrcEnd) ARGDW(IdxDst) ARGDW(SBInfo.iDstBeg) ARGDW(SBInfo.iDstEnd) ARGDW(pRep - SBInfo.pRep) ARGDW(cRep) ARGDW(cIn) ARGDW(cOut) ARGDW(cRem) ARGDW(cTot))); }
DBGP_IF(DBGP_BUILD3, DBGP("IdxSrc=%4ld, IdxDst=%4ld, Rep=%3ld / %3ld, Rem=%5ld, cIn=%5ld" ARGDW(IdxSrc) ARGDW(IdxDst) ARGDW(cRep) ARGDW(1) ARGDW(cRem) ARGDW(cIn)));
cRem += cOut; iRep = cRep = 0;
} else {
DBGP_IF(DBGP_BUILD3, DBGP("IdxSrc=%4ld, IdxDst=%4ld, Rep=%3ld / %3ld, Rem=%5ld, cIn=%5ld" ARGDW(IdxSrc) ARGDW(IdxDst) ARGDW(cRep) ARGDW(iRep + 1) ARGDW(cRem) ARGDW(cIn)));
}
++iRep;
if ((IdxSrc >= SBInfo.iSrcBeg) && (IdxSrc < SBInfo.iSrcEnd) && (IdxDst >= SBInfo.iDstBeg) && (IdxDst < SBInfo.iDstEnd)) {
++cRep;
if (jSrcBeg == -1) {
//
// Any iRep will be the first pixel on the destination that
// corresponding to the current source, so minus 1 is the
// total count to be skip for this group of destination
//
jSrcBeg = IdxSrc; jDstBeg = IdxDst; SBInfo.cFirstSkip = (BYTE)(iRep - 1);
DBGP_IF(DBGP_BUILD3, DBGP(" @@@ Set cFirstSkip=%ld at IdxDst=%ld @@@" ARGDW(SBInfo.cFirstSkip) ARGDW(IdxDst))); }
jSrcEnd = IdxSrc; jDstEnd = IdxDst;
} else if (jSrcBeg != -1) {
break; }
++*pDstInc; }
if (jSrcBeg == -1) {
DBGP_IF(DBGP_BUILD3, DBGP(" Nothing in the source is on the destination"));
return(FALSE); }
if (cRep) {
ASSERT(pRep < pRepEnd);
pRep++->c = (WORD)cRep; cTot += cRep;
DBGP_IF(DBGP_BUILD2, DBGP(" ****** Total pRep=%ld, Last cRep=%ld, cTot=%ld" ARGDW(pRep - SBInfo.pRep) ARGDW(cRep) ARGDW(cTot))); }
while ((cRem -= cIn) >= 0) {
++SBInfo.cLastSkip;
DBGP_IF(DBGP_BUILD3, DBGP(" @@@ Set cLastSkip=%4ld, cRem=%5ld @@@" ARGDW(SBInfo.cLastSkip) ARGDW(cRem)));
}
if (SBInfo.cIn < SBInfo.cOut) {
//
// Expanding, checking for maximum 2 sourcs pixels on each side
//
cPrevSrc = cNextSrc = 2;
} else {
//
// Shrinking, checking only cFirstSkip and cLastSkip
//
cPrevSrc = (UINT)SBInfo.cFirstSkip; cNextSrc = (UINT)SBInfo.cLastSkip; }
//
// Check src begin
//
IdxSrc = jSrcBeg;
while ((cPrevSrc) && (IdxSrc > SBInfo.iSrcBeg)) {
--IdxSrc; --cPrevSrc; }
SBInfo.cPrevSrc = (BYTE)(jSrcBeg - IdxSrc); IdxSrc = jSrcEnd;
while ((cNextSrc) && (IdxSrc < (SBInfo.iSrcEnd - 1))) {
++IdxSrc; --cNextSrc; }
SBInfo.cNextSrc = (BYTE)(IdxSrc - jSrcEnd);
DBGP_IF(DBGP_BUILD3, DBGP("cFirstSkip=%ld (%ld), cLastSkip=%ld (%ld), cPrevSrc=%ld, cNextSrc=%ld" ARGDW(SBInfo.cFirstSkip) ARGDW(SBInfo.pRep->c) ARGDW(SBInfo.cLastSkip) ARGDW((pRep - 1)->c) ARGDW(SBInfo.cPrevSrc) ARGDW(SBInfo.cNextSrc)));
// Bug 27036: ensure SBInfo.iSrcEnd is always exclusive
SBInfo.iBeg = SBInfo.iSrcBeg = jSrcBeg; SBInfo.iSrcEnd = jSrcEnd + 1; SBInfo.iDstBeg = jDstBeg; SBInfo.iDstEnd = jDstEnd + 1; SBInfo.iSize = jSrcEnd - jSrcBeg + 1; SBInfo.pRepEnd = pRep; SBInfo.cRep = 1;
DBGP_IF((DBGP_BUILD | DBGP_BUILD2 | DBGP_BUILD3), DBGP("EXPREP(%ld->%ld): iSrc=%ld-%ld, iDst=%ld-%ld, iRepSize=%ld" ARGDW(SBInfo.cIn) ARGDW(SBInfo.cOut) ARGDW(SBInfo.iSrcBeg) ARGDW(SBInfo.iSrcEnd) ARGDW(SBInfo.iDstBeg) ARGDW(SBInfo.iDstEnd) ARGDW(SBInfo.pRepEnd - SBInfo.pRep)));
*pSBInfo = SBInfo;
return(TRUE);}
PAAINFOHTENTRYBuildExpandAAInfo( PDEVICECOLORINFO pDCI, DWORD AAHFlags, PLONG piSrcBeg, PLONG piSrcEnd, LONG SrcSize, LONG IdxDst, LONG IdxDstEnd, PLONG piDstBeg, PLONG piDstEnd, LONG cbExtra )
/*++
Routine Description:
Arguments:
piSrcBeg - Passed in as begining source index, When return this is the real source begining index. This always well ordered
piSrcEnd - Passed in as ending source index, When return this is the real source ending index. This always well ordered
SrcSize - The real size of source in pixel
IdxDst - Starting index of destination pixel
IdxDstEnd - Ending index of destination pixel, the iDxdst and IdxDstEnd must be well ordered. (exclusive)
piDstBeg - Clipped destination start index when passed in, at return it adjusted to the real destination starting index.
piDstEnd - Clipped destination end index when passed in, at return it adjusted to the real destination ending index.
cbExtra - Extra byte count to be allocated
NOTE: 1) piDstBeg/piDstEnd When passed in this must be well ordered, and when it returned this is well ordered.
Return Value:
At enter of this function, the *piSrcEnd, *piDstEnd is exclusive but when return from this function the *piSrcEnd and *piDstEnd is inclusive
*piSrcBeg, *piSrcEnd, *piDstBeg, *piDstEnd are updated if return value is not NULL
Author:
22-Mar-1998 Sun 18:36:28 created -by- Daniel Chou (danielc)
Revision History:
--*/
{ PAAINFO pAAInfo; PEXPDATA pED; LONG cMatrix; LONG cM; LONG cIn; LONG cOut; LONG IdxSrc; LONG iSrcBeg; LONG iSrcEnd; LONG iDstBeg; LONG iDstEnd; LONG jSrcBeg; LONG jSrcEnd; LONG jDstBeg; LONG jDstEnd; LONGLONG cTot; LONG MulAdd; LONG Mul; DWORD cAALoad; LONG cbRep; LONG cRem; DWORD cbED;
cIn = (LONG)((iSrcEnd = *piSrcEnd) - (iSrcBeg = IdxSrc = *piSrcBeg)); cOut = IdxDstEnd - IdxDst;
ASSERT(cOut > 0);
if (iSrcBeg < 0) {
iSrcBeg = 0; }
if (iSrcEnd > SrcSize) {
iSrcEnd = SrcSize; }
iDstBeg = *piDstBeg; iDstEnd = *piDstEnd; jSrcBeg = -1; cAALoad = 0;
ASSERT(iDstBeg < iDstEnd); ASSERT(cIn < cOut);
DBGP_IF(DBGP_BUILD, DBGP("\nEXP(%ld-%ld): iSrc=%ld-%ld, cSrc=%ld, iDst=%ld-%ld, Idx=%ld:%ld" ARGDW(cIn) ARGDW(cOut) ARGDW(iSrcBeg) ARGDW(iSrcEnd) ARGDW(SrcSize) ARGDW(iDstBeg) ARGDW(iDstEnd) ARGDW(IdxSrc) ARGDW(IdxDst)));
if (AAHFlags & (AAHF_HAS_MASK | AAHF_ALPHA_BLEND | AAHF_FAST_EXP_AA | AAHF_BBPF_AA_OFF)) {
ALIGN_MEM(cbRep, (iSrcEnd - iSrcBeg + 3) * sizeof(REPDATA));
} else {
cbRep = 0; }
if (AAHFlags & (AAHF_BBPF_AA_OFF | AAHF_FAST_EXP_AA)) {
cMatrix = cbED = cM = 0;
} else {
cMatrix = (LONG)((((cOut + (cIn - 1)) / cIn) << 1) - 1); cM = sizeof(DWORD) * cMatrix;
ALIGN_MEM(cbED, (iDstEnd - iDstBeg) * sizeof(EXPDATA)); }
DBGP_IF(DBGP_BUILD, DBGP("BuildEXP(%ld, %ld), cMatrix=%ld, cb=%ld+%ld+%ld=%ld" ARGDW(cIn) ARGDW(cOut) ARGDW(cMatrix) ARGDW(cbRep) ARGDW(cbED) ARGDW(cM) ARGDW(cbRep + cbED + cM)));
ALIGN_MEM(cbExtra, cbExtra);
if (pAAInfo = (PAAINFO)HTAllocMem((LPVOID)pDCI, HTMEM_EXPAA, LPTR, cbRep + cbED + cbExtra + cM + SIZE_AAINFO)) {
LPBYTE pbExtra;
SETDBGVAR(pAAInfo->cbAlloc, cbRep + cbED + cbExtra + cM + SIZE_AAINFO);
pbExtra = (LPBYTE)pAAInfo + SIZE_AAINFO;
if (cbExtra) {
pAAInfo->pbExtra = (LPBYTE)pbExtra; pbExtra += cbExtra; }
if (cbRep) {
pAAInfo->Src.cIn = cIn; pAAInfo->Src.cOut = cOut; pAAInfo->Src.iSrcBeg = iSrcBeg; pAAInfo->Src.iSrcEnd = iSrcEnd; pAAInfo->Src.iDstBeg = iDstBeg; pAAInfo->Src.iDstEnd = iDstEnd; pAAInfo->Src.pRep = (PREPDATA)pbExtra; pAAInfo->Src.pRepEnd = pAAInfo->Src.pRep + (iSrcEnd - iSrcBeg); pbExtra += cbRep;
if (!BuildRepData(&(pAAInfo->Src), IdxSrc, IdxDst)) {
HTFreeMem(pAAInfo); return(NULL); }
pAAInfo->AB = pAAInfo->Mask = pAAInfo->Src;
if (AAHFlags & AAHF_FAST_EXP_AA) {
pAAInfo->Src.iSrcBeg -= pAAInfo->Src.cPrevSrc; pAAInfo->Src.iSrcEnd += pAAInfo->Src.cNextSrc; } }
if (cbED) {
PEXPDATA pED; PEXPDATA pEDEnd; LPDWORD pM; LPDWORD pM1; LPDWORD pM2; LONG cRem2; LONG MincM; LONGLONG ExpMul[4]; EXPDATA ed; WORD EDFlags; LONGLONG cNum; LONG cLoop; LONG cMul0; LONG cMul1; LONG cMaskRem; LONG iMaskBeg; LONG iMaskEnd;
pED = (PEXPDATA)pbExtra; pAAInfo->pAAData = (LPVOID)pED; pM = (LPDWORD)((LPBYTE)pED + cbED); pM1 = pM; pM2 = pM1 + cMatrix - 1; cTot = 0;
DBGP_IF(DBGP_BUILD, DBGP("Allocate cbExtra=%ld, pbExtra=%p:%p" ARGDW(cbExtra) ARGPTR(pAAInfo->pbExtra) ARGPTR((LPBYTE)pAAInfo->pbExtra + cbExtra)));
pM2 = (pM1 += (cMatrix >> 1)); cTot = (LONGLONG)(*pM1 = FD6_1);
if (AAHFlags & AAHF_BBPF_AA_OFF) {
pAAInfo->Flags |= AAIF_EXP_NO_SHARPEN;
} else {
MulAdd = cOut;
while (((MulAdd -= cIn) > 0) && (--pM1 >= pM)) {
Mul = (LONG)DivFD6(MulAdd, cOut);
if (Mul != 500000) {
Mul = (LONG)RaisePower((FD6)Mul, MATRIX_POWER, (WORD)((Mul <= 500000) ? 0 : RPF_RADICAL)); }
DBGP_IF(DBGP_EXPMATRIX, DBGP("(%4ld, %4ld) = %s ^ %s = %s" ARGDW(MulAdd) ARGDW(cOut) ARGFD6((DivFD6(MulAdd, cOut)), 1, 6) ARGFD6(MATRIX_POWER, 1, 6) ARGFD6(Mul, 1, 6)));
*pM1 = *(++pM2) = Mul; cTot += (Mul << 1); } }
#if DBG
{ FD6 PMPrev = FD6_0; FD6 PMCur;
pM1 = pM; cLoop = (LONG)cMatrix;
while (cLoop--) {
PMCur = DivFD6(*pM1, (FD6)cTot);
DBGP_IF(DBGP_EXPMATRIX, DBGP("%3ld: %7ld [%s], Dif=%s, cTot=%ld" ARGDW((pM1 - pM) + 1) ARGDW(*pM1) ARGFD6(PMCur, 1, 6) ARGFD6((PMCur - PMPrev), 1, 6) ARGDW(cTot)));
PMPrev = PMCur; ++pM1; } } #endif
cTot *= (LONGLONG)cIn; cRem = cOut + (cIn * (LONG)(cMatrix >> 1)); cLoop = cOut; cMul0 = cMul1 = 0;
while (cLoop--) {
cRem2 = cRem;
if ((cRem -= cIn) <= 0) {
cRem += cOut; }
pM1 = pM; cM = cMatrix; MincM = (cMatrix >> 1) - cLoop; EDFlags = 0;
ZeroMemory(ExpMul, sizeof(ExpMul));
while (cM--) {
LONG cMul;
Mul = *pM1++;
if ((cRem2 < cIn) && (cM >= MincM)) {
if (cMul = cRem2) {
ExpMul[3] += (LONGLONG)cRem2 * (LONGLONG)Mul; }
cRem2 -= cIn;
ASSERTMSG("BuildEXP: Shift more than 3 times", !ExpMul[0]);
CopyMemory(&ExpMul[0], &ExpMul[1], sizeof(ExpMul[0]) * 3);
ExpMul[3] = (LONGLONG)-cRem2 * (LONGLONG)Mul; cRem2 += cOut;
if (!cM) {
if ((++IdxSrc >= iSrcBeg) && (IdxSrc < iSrcEnd)) {
EDFlags |= EDF_LOAD_PIXEL;
++cAALoad; ++IdxSrc;
if ((IdxSrc < iSrcBeg) || (IdxSrc >= iSrcEnd)) {
EDFlags |= EDF_NO_NEWSRC; }
--IdxSrc; } }
} else {
ExpMul[3] += (LONGLONG)(cMul = cIn) * (LONGLONG)Mul; cRem2 -= cIn; }
DBGP_IF(DBGP_BUILD3, DBGP("%5ld-%7ld:%7ld:%7ld:%7ld, %4ld+, cRem2=%4ld, cM=%5ld:%5ld, cMul=%5ldx%5ld%hs%hs" ARGDW(cOut - cLoop - 1) ARGDW(ExpMul[0]) ARGDW(ExpMul[1]) ARGDW(ExpMul[2]) ARGDW(ExpMul[3]) ARGDW(cMul * Mul) ARGDW(cRem2) ARGDW(cM) ARGDW(MincM) ARGDW(cMul) ARGDW(Mul) ARGPTR((EDFlags & EDF_LOAD_PIXEL) ? ", Load Pixel" : "") ARGPTR((EDFlags & EDF_NO_NEWSRC) ? ", NO New Src" : ""))); }
if ((IdxSrc >= iSrcBeg) && (IdxSrc < iSrcEnd) && (IdxDst >= iDstBeg) && (IdxDst < iDstEnd)) {
GET_FIRST_EDMUL(ed.Mul[3], ExpMul[3], cNum, cTot); GET_NEXT_EDMUL( ed.Mul[2], ExpMul[2], cNum, cTot);
if (ExpMul[1]) {
++cMul1;
GET_NEXT_EDMUL(ed.Mul[1], ExpMul[1], cNum, cTot);
if (ExpMul[0]) {
++cMul0;
GET_NEXT_EDMUL(ed.Mul[0], ExpMul[0], cNum, cTot);
} else {
ed.Mul[0] = 0; }
} else {
ed.Mul[1] = ed.Mul[0] = 0; }
ASSERTMSG("ed.Mul[0] > DI_MAX_NUM", ed.Mul[0] <= DI_MAX_NUM); ASSERTMSG("ed.Mul[1] > DI_MAX_NUM", ed.Mul[1] <= DI_MAX_NUM); ASSERTMSG("ed.Mul[2] > DI_MAX_NUM", ed.Mul[2] <= DI_MAX_NUM); ASSERTMSG("ed.Mul[3] > DI_MAX_NUM", ed.Mul[3] <= DI_MAX_NUM);
DBGP_IF(DBGP_BUILD2, DBGP("--%5ld=%7ld:%7ld:%7ld:%7ld, IdxSrc=%5ld --%hs%hs--" ARGDW(IdxDst) ARGDW(ed.Mul[0]) ARGDW(ed.Mul[1]) ARGDW(ed.Mul[2]) ARGDW(ed.Mul[3]) ARGDW(IdxSrc) ARGPTR((EDFlags & EDF_LOAD_PIXEL) ? ", Load Pixel" : "") ARGPTR((EDFlags & EDF_NO_NEWSRC) ? ", NO New Src" : "")));
ed.Mul[0] |= EDFlags; *pED++ = ed;
if (jSrcBeg == -1) {
iMaskBeg = (cOut - cLoop - 1); jSrcBeg = IdxSrc; jDstBeg = IdxDst; }
jSrcEnd = IdxSrc; jDstEnd = IdxDst;
} else if (jSrcBeg != -1) {
break; }
++IdxDst; }
if (jSrcBeg == -1) {
HTFreeMem(pAAInfo); return(NULL); }
//
// 10-Jun-1998 Wed 08:41:16 updated -by- Daniel Chou (danielc)
// Fixed for the problem that we need to read extra source for the
// last loaded pixel because before expand we will sharpen the pixel
// by its 4 neighbors, this will make sure Last+1 pixel/scan line
// got read it for sharpen purpose
//
// BEGIN FIX
//
++jSrcEnd;
if ((jSrcEnd < iSrcBeg) || (jSrcEnd >= iSrcEnd)) {
--jSrcEnd; }
//
// END FIX
//
IdxSrc = *piSrcBeg = jSrcBeg; *piSrcEnd = jSrcEnd; *piDstBeg = jDstBeg; *piDstEnd = jDstEnd; pEDEnd = pED; pED = (PEXPDATA)(pAAInfo->pAAData); pAAInfo->cAAData = (DWORD)(pEDEnd - pED); pAAInfo->cAALoad = cAALoad; pAAInfo->cMaxMul = (DWORD)((cMul1) ? ((cMul0) ? 4 : 3) : 2); ed = *pED; cLoop = 4;
if (ed.Mul[0] & EDF_LOAD_PIXEL) {
--IdxSrc; --cLoop;
} else {
++IdxSrc;
if ((IdxSrc < iSrcBeg) || (IdxSrc >= iSrcEnd)) {
pAAInfo->Flags |= AAIF_EXP_NO_LAST_RIGHT; }
--IdxSrc; }
cMul0 = 0;
while ((cMul0 < cLoop) && (!(ed.Mul[cMul0] & ~(EDF_LOAD_PIXEL | EDF_NO_NEWSRC)))) {
++cMul0; }
DBGP_IF(DBGP_BUILD, DBGP("cMul0=%ld, cLoop=%ld, IdxSrc=%ld, iSrcBeg=%ld, iSrcEnd=%ld" ARGDW(cMul0) ARGDW(cLoop) ARGDW(IdxSrc) ARGDW(iSrcBeg) ARGDW(iSrcEnd)));
while (cLoop-- > cMul0) {
if ((IdxSrc >= iSrcBeg) && (IdxSrc < iSrcEnd)) {
*piSrcBeg = IdxSrc; pAAInfo->cPreLoad += 0x01;
} else {
pAAInfo->cPreLoad += 0x10; }
--IdxSrc; }
if (pAAInfo->cPreLoad) {
if ((IdxSrc >= iSrcBeg) && (IdxSrc < iSrcEnd)) {
*piSrcBeg = IdxSrc; pAAInfo->Flags |= AAIF_EXP_HAS_1ST_LEFT; } }
if (!(pAAInfo->cPreLoad)) {
DBGP("*** cPreLOAD=0, BuildExpandAAInfo [%04lx:%04lx:%04lx:%04lx], cPreload=0x%02lx, Flags=0x%04lx" ARGDW(ed.Mul[0]) ARGDW(ed.Mul[1]) ARGDW(ed.Mul[2]) ARGDW(ed.Mul[3]) ARGDW(pAAInfo->cPreLoad) ARGDW(pAAInfo->Flags)); }
} else {
// Bug 27036: ensure jSrcEnd is less than iSrcEnd
*piSrcBeg = pAAInfo->Src.iSrcBeg; *piSrcEnd = pAAInfo->Src.iSrcEnd - 1; *piDstBeg = pAAInfo->Src.iDstBeg; *piDstEnd = pAAInfo->Src.iDstEnd - 1; }
pAAInfo->cIn = *piSrcEnd - *piSrcBeg + 1; pAAInfo->cOut = *piDstEnd - *piDstBeg + 1;
DBGP_IF((DBGP_BUILD | DBGP_BUILD2 | DBGP_BUILD3), DBGP("EXP(%ld->%ld): iSrc=%ld-%ld, iDst=%ld-%ld, cAAData=%ld, cPreLoad=0x%02lx, Flags=0x%04lx" ARGDW(pAAInfo->cIn) ARGDW(pAAInfo->cOut) ARGDW(*piSrcBeg) ARGDW(*piSrcEnd) ARGDW(*piDstBeg) ARGDW(*piDstEnd) ARGDW(pAAInfo->cAAData) ARGDW(pAAInfo->cPreLoad) ARGDW(pAAInfo->Flags))); }
return(pAAInfo);}
PAAINFOHTENTRYBuildShrinkAAInfo( PDEVICECOLORINFO pDCI, DWORD AAHFlags, PLONG piSrcBeg, PLONG piSrcEnd, LONG SrcSize, LONG IdxDst, LONG IdxDstEnd, PLONG piDstBeg, PLONG piDstEnd, LONG cbExtra )
/*++
Routine Description:
Arguments:
piSrcBeg - Passed in as begining source index, When return this is the real source begining index. This always well ordered
piSrcEnd - Passed in as ending source index, When return this is the real source ending index. This always well ordered
SrcSize - The real size of source in pixel
IdxDst - Starting index of destination pixel
IdxDstEnd - Ending index of destination pixel, the iDxdst and IdxDstEnd must be well ordered.
piDstBeg - Clipped destination start index when passed in, at return it adjusted to the real destination starting index.
piDstEnd - Clipped destination end index when passed in, at return it adjusted to the real destination ending index.
cbExtra - Extra byte count to be allocated
NOTE: 1) piDstBeg/piDstEnd When passed in this must be well ordered, and when it returned this is well ordered.
Return Value:
PSHRINKINFO, if NULL then memory allocation failed
Author:
20-Mar-1998 Fri 12:29:17 created -by- Daniel Chou (danielc)
Revision History:
--*/
{ PAAINFO pAAInfo; PLONG pMap; PLONG pMapEnd; PSHRINKDATA pSD; PSHRINKDATA pSDEnd; LONG cIn; LONG cOut; LONG IdxSrc; LONG IdxDstOrg; LONG iSrcBeg; LONG iSrcEnd; LONG iDstBeg; LONG iDstEnd; LONG jSrcBeg; LONG jSrcEnd; LONG jDstBeg; LONG jDstEnd; LONG cLoop; LONG cbRep; LONG Mul; LONG NextMul; LONG CurMul; LONG cCur; LONG MinPixel; DWORD cAAData; DWORD cAADone; LONGLONG cNum;
cIn = (LONG)((iSrcEnd = *piSrcEnd) - (iSrcBeg = IdxSrc = *piSrcBeg)); cOut = IdxDstEnd - (IdxDstOrg = IdxDst);
ASSERT(cOut > 0);
if (iSrcBeg < 0) {
iSrcBeg = 0; }
if (iSrcEnd > SrcSize) {
iSrcEnd = SrcSize; }
//
// For shrinking we will enlarge the destination by 1 on both side to
// obtained the source pixel that for sharpening purpose
//
iDstBeg = *piDstBeg - 1; iDstEnd = *piDstEnd; jSrcBeg = -1; cAADone = 0;
ASSERT(iDstBeg < iDstEnd); ASSERT(cIn > cOut);
DBGP_IF(DBGP_BUILD, DBGP("\nSRK(%ld-%ld): iSrc=%ld-%ld, cSrc=%ld, iDst=%ld-%ld (%ld-%ld), Idx=%ld:%ld" ARGDW(cIn) ARGDW(cOut) ARGDW(iSrcBeg) ARGDW(iSrcEnd) ARGDW(SrcSize) ARGDW(iDstBeg) ARGDW(iDstEnd) ARGDW(*piDstBeg) ARGDW(*piDstEnd) ARGDW(IdxSrc) ARGDW(IdxDst)));
//
// Firstable figure out how may SHRINKDATA needed
//
cAAData = (DWORD)((((iDstEnd - iDstBeg + 1) * cIn) + (cOut-1)) / cOut) + 4;
if ((LONG)cAAData > cIn) {
(LONG)cAAData = cIn; }
DBGP_IF(DBGP_BUILD, DBGP("BuildShrink(%ld-%ld): cSD estimated=%ld (%ld), iDst=%ld-%ld=%ld" ARGDW(cIn) ARGDW(cOut) ARGDW(cAAData) ARGDW(cIn) ARGDW(iDstBeg) ARGDW(iDstEnd) ARGDW(iDstEnd - iDstBeg)));
ALIGN_MEM(CurMul, 256 * 2 * sizeof(LONG)); ALIGN_MEM(cCur, (cAAData + 1) * sizeof(SHRINKDATA));
CurMul += cCur; MinPixel = (LONG)(((LONGLONG)cOut * (LONGLONG)DI_MAX_NUM) / (LONGLONG)cIn);
if (AAHFlags & (AAHF_HAS_MASK | AAHF_ALPHA_BLEND | AAHF_FAST_EXP_AA | AAHF_BBPF_AA_OFF)) {
ALIGN_MEM(cbRep, (iDstEnd - iDstBeg + 4) * sizeof(REPDATA));
if (AAHFlags & AAHF_BBPF_AA_OFF) {
CurMul = 0; }
} else {
cbRep = 0; }
ALIGN_MEM(cbExtra, cbExtra);
if (pAAInfo = (PAAINFO)HTAllocMem((LPVOID)pDCI, HTMEM_SRKAA, LPTR, SIZE_AAINFO + CurMul + cbRep + cbExtra)) {
LPBYTE pbExtra;
SETDBGVAR(pAAInfo->cbAlloc, SIZE_AAINFO + CurMul + cbRep + cbExtra);
pbExtra = (LPBYTE)pAAInfo + SIZE_AAINFO;
if (cbExtra) {
pAAInfo->pbExtra = (LPBYTE)pbExtra; pbExtra += cbExtra; }
if (cbRep) {
pAAInfo->Src.cIn = cIn; pAAInfo->Src.cOut = cOut; pAAInfo->Src.iSrcBeg = iSrcBeg; pAAInfo->Src.iSrcEnd = iSrcEnd; pAAInfo->Src.iDstBeg = iDstBeg + 1; pAAInfo->Src.iDstEnd = iDstEnd; pAAInfo->Src.pRep = (PREPDATA)pbExtra; pAAInfo->Src.pRepEnd = pAAInfo->Src.pRep + (iDstEnd-iDstBeg+1); pbExtra += cbRep;
if (!BuildRepData(&(pAAInfo->Src), IdxSrc, IdxDst)) {
HTFreeMem(pAAInfo); return(NULL); }
pAAInfo->AB = pAAInfo->Mask = pAAInfo->Src; }
if (CurMul) {
pAAInfo->cPreLoad = 1; pMap = pAAInfo->pMapMul = (PLONG)pbExtra; pMapEnd = pMap + 256; pSD = (PSHRINKDATA)(pMap + (256 * 2)); pSDEnd = pSD + cAAData; pAAInfo->pAAData = (LPVOID)pSD; Mul = -MinPixel; CurMul = MinPixel + 1; NextMul = -CurMul;
ASSERT_MEM_ALIGN(pAAInfo->pMapMul, sizeof(LONG));
//
// Build InMax 256 multiplication table
//
do {
pMap[ 0] = (Mul += MinPixel); pMap[256] = (NextMul += CurMul);
} while (++pMap < pMapEnd);
//
// Build the SHRINKINFO table
//
CurMul = 0; cCur = cLoop = cIn; cNum = (LONGLONG)0; cAAData = 0; --pSD;
while (cLoop--) {
WORD SDFlags;
if ((cCur -= cOut) <= 0) {
Mul = cCur + cOut; NextMul = -cCur; cCur += cIn; SDFlags = SDF_DONE;
++IdxDst;
} else {
Mul = cOut; SDFlags = 0; }
if ((IdxDst >= (iDstBeg - 1)) && (IdxDst <= iDstEnd)) {
cNum += ((LONGLONG)Mul * (LONGLONG)DI_MAX_NUM);
if ((Mul = (LONG)(cNum / cIn)) > MinPixel) {
SDFlags |= SDF_LARGE_MUL; }
CurMul += Mul; cNum %= cIn;
if (SDFlags & SDF_DONE) {
cNum = (LONGLONG)NextMul * (LONGLONG)DI_MAX_NUM; NextMul = (LONG)(cNum / cIn);
if ((Mul + NextMul) > MinPixel) {
SDFlags |= SDF_LARGE_MUL;
} else {
SDFlags &= ~SDF_LARGE_MUL; }
cNum %= cIn; CurMul = NextMul; NextMul = 0; } }
if ((IdxDst >= iDstBeg) && (IdxDst <= iDstEnd)) {
if ((IdxSrc >= iSrcBeg) && (IdxSrc < iSrcEnd)) {
//
// Save it first
//
cAADone += (SDFlags & SDF_DONE) ? 1 : 0;
if (++pSD >= pSDEnd) {
DBGP("Error(1): cAAData Overrun of %ld, Fixed it" ARGDW(++cAAData)); ASSERT(pSD < pSDEnd);
--pSD; }
pSD->Mul = (WORD)Mul | SDFlags;
ASSERTMSG("sd.Mul > DI_MAX_NUM", Mul <= DI_MAX_NUM);
if (jSrcBeg == -1) {
jSrcBeg = jSrcEnd = IdxSrc; jDstBeg = jDstEnd = IdxDst;
if (SDFlags & SDF_DONE) {
//
// If we just finished a pixel then we need to see
// if it is a source index or a detination index
// cause the output become valid
//
if (IdxDst == iDstBeg) {
//
// The destination just become valid now,
// PreMul is the start of current detination
// until next done pixel
//
DBGP_IF(DBGP_BUILD, DBGP("@@ FIRST DEST: PreMul=CurMul=%ld, No PSD, IncSrc" ARGDW(CurMul)));
pAAInfo->PreMul = (WORD)CurMul; pAAInfo->PreSrcInc = 1; --cAADone; --pSD;
} else {
//
// The source just become valid now, need to
// PreMul all prev sources for this detination
// and save this pSD for done pixel
//
DBGP_IF(DBGP_BUILD, DBGP("@@ FIRST SRC: PreMul=%ld - Mul (%ld)=%ld" ARGDW(DI_MAX_NUM) ARGDW(Mul) ARGDW(DI_MAX_NUM - Mul)));
pAAInfo->PreMul = (WORD)(DI_MAX_NUM - Mul);
--jDstBeg; --jDstEnd;
ASSERTMSG("!!! Error: jDstBeg is WRONG", (jDstBeg >= iDstBeg) && (jDstBeg <= iDstEnd)); }
} else {
//
// We are in the middle of compositions, so the
// source just become valid, notice that PreMul
// could be zero
//
DBGP_IF(DBGP_BUILD, DBGP("@@ FIRST MIDDLE: PreMul=CurMul (%ld) - Mul (%ld)=%ld" ARGDW(CurMul) ARGDW(Mul) ARGDW(CurMul - Mul)));
pAAInfo->PreMul = (WORD)(CurMul - Mul); }
} else {
jSrcEnd = IdxSrc; jDstEnd = IdxDst; }
} else if (jSrcBeg != -1) {
//
// Source got cut off early, so wrap it up now
//
DBGP_IF(DBGP_BUILD, DBGP("@@ END SRC: Mul=%ld, CurMul=%ld" ARGDW(Mul) ARGDW(CurMul)));
if (++pSD >= pSDEnd) {
DBGP("Error(2): cAAData Overrun of %ld, Fixed it" ARGDW(++cAAData)); ASSERT(pSD < pSDEnd);
--pSD; }
if (!(SDFlags & SDF_DONE)) {
Mul += (DI_MAX_NUM - CurMul); }
pSD->Mul = (WORD)Mul | (SDFlags |= SDF_DONE); cLoop = 0; ++cAADone;
ASSERTMSG("sd.Mul > DI_MAX_NUM", Mul <= DI_MAX_NUM); }
} else if (jSrcBeg != -1) {
//
// we just pass the iDstEnd so this one MUST have SDF_DONE
// bit set and we need to save this one, if this one is not
// SDF_DONE then something is wrong
//
ASSERTMSG("End Dest but not SDF_DONE", SDFlags & SDF_DONE);
DBGP_IF(DBGP_BUILD, DBGP("@@ PASS IdxDst: Mul=%ld, CurMul=%ld" ARGDW(Mul) ARGDW(CurMul)));
if (++pSD >= pSDEnd) {
DBGP("Error(3): cAAData Overrun of %ld, Fixed it" ARGDW(++cAAData)); ASSERT(pSD < pSDEnd);
--pSD; }
jSrcEnd = IdxSrc; Mul = DI_MAX_NUM - CurMul; pSD->Mul = (WORD)Mul | (SDFlags = SDF_DONE); cLoop = 0; ++cAADone;
ASSERTMSG("sd.Mul > DI_MAX_NUM", Mul <= DI_MAX_NUM); }#if DBG
if ((pSD >= (PSHRINKDATA)(pAAInfo->pAAData)) || (pAAInfo->PreSrcInc)) {
BOOL HasSD;
HasSD = (BOOL)(pSD >= (PSHRINKDATA)(pAAInfo->pAAData));
if (SDFlags & SDF_DONE) {
DBGP_IF(DBGP_BUILD2, DBGP("%hscLoop=%5ld (%5ld/%5ld), iSrc=%5ld, Mul=%5ld [%5ld], Flags=0x%04lx%hs, Done Pixel" ARGPTR((HasSD) ? "" : " >>") ARGDW(cIn - cLoop - 1) ARGDW(IdxDst) ARGDW(IdxDst - iDstBeg) ARGDW(IdxSrc) ARGDW((Mul) ? Mul : ((SDFlags & SDF_LARGE_MUL) ? MinPixel + 1 : MinPixel)) ARGDW(CurMul) ARGDW(SDFlags) ARGPTR((SDFlags & SDF_LARGE_MUL) ? ", Large Mul" : "")));
} else {
DBGP_IF(DBGP_BUILD2, DBGP("%hscLoop=%5ld iSrc=%5ld, Mul=%5ld [%5ld], Flags=0x%04lx%hs" ARGPTR((HasSD) ? "" : " >>") ARGDW(cIn - cLoop - 1) ARGDW(IdxSrc) ARGDW((Mul) ? Mul : ((SDFlags & SDF_LARGE_MUL) ? MinPixel + 1 : MinPixel)) ARGDW(CurMul) ARGDW(SDFlags) ARGPTR((SDFlags & SDF_LARGE_MUL) ? ", Large Mul" : ""))); } }#endif
++IdxSrc; }
//
// For the last one ZERO
//
++pSD;
if ((jSrcBeg == -1) || (pSD == (PSHRINKDATA)(pAAInfo->pAAData))) {
HTFreeMem(pAAInfo); return(NULL); }
++iDstBeg;
DBGP_IF(DBGP_BUILD, DBGP("*** Final jDstBeg/End=%ld:%ld, REAL=(%ld:%ld)" ARGDW(jDstBeg) ARGDW(jDstEnd) ARGDW(iDstBeg) ARGDW(iDstEnd)));
// Bug 27036: ensure jSrcEnd is less than iSrcEnd
if (jSrcEnd >= iSrcEnd) { jSrcEnd = iSrcEnd - 1; }
if (jDstBeg < iDstBeg) {
++(pAAInfo->cPreLoad);
jDstBeg = iDstBeg; }
if (jDstEnd >= iDstEnd) {
jDstEnd = iDstEnd - 1; }
if ((pAAInfo->PreSrcInc) && (pAAInfo->PreMul == 0)) {
pAAInfo->PreSrcInc = 0; ++jSrcBeg; }#if 0
//
// 04-Aug-2000 Fri 15:31:03 updated -by- Daniel Chou (danielc)
// This assert does not applyed here when anti-aliasing will use
// surounding 3 pixels (L/T/R/B) if source available (when clipped
// source) but the Rep does not use suround pixels.
//
if (cbRep) {
ASSERT(jSrcBeg == pAAInfo->Mask.iBeg);
if (jSrcEnd != (pAAInfo->Mask.iBeg + pAAInfo->Mask.iSize - 1)) {
DBGP("jSrcEnd=%ld, Mask: iBeg=%ld, iSize=%ld" ARGDW(jSrcEnd) ARGDW(pAAInfo->Mask.iBeg) ARGDW(pAAInfo->Mask.iSize));
ASSERT(jSrcEnd == pAAInfo->Mask.iBeg + pAAInfo->Mask.iSize - 1); } }#endif
pAAInfo->cAAData = (DWORD)(pSD - (PSHRINKDATA)(pAAInfo->pAAData)); pAAInfo->cAADone = cAADone; pSD->Mul = 0;
} else {
ASSERT(cbRep);
// Bug 27036: ensure jSrcEnd is less than iSrcEnd
jSrcBeg = pAAInfo->Src.iSrcBeg; jSrcEnd = pAAInfo->Src.iSrcEnd - 1; jDstBeg = pAAInfo->Src.iDstBeg; jDstEnd = pAAInfo->Src.iDstEnd - 1; }
*piSrcBeg = jSrcBeg; *piSrcEnd = jSrcEnd; *piDstBeg = jDstBeg; *piDstEnd = jDstEnd; pAAInfo->cIn = jSrcEnd - jSrcBeg + 1; pAAInfo->cOut = jDstEnd - jDstBeg + 1;
DBGP_IF(DBGP_BUILD, DBGP("SRK(%ld->%ld): iSrc=%ld-%ld, iDst=%ld-%ld, cAAData=%ld, cAADone=%ld, PreMul=%4ld, PresrcInc=%ld, cPreLoad=%ld" ARGDW(pAAInfo->cIn) ARGDW(pAAInfo->cOut) ARGDW(jSrcBeg) ARGDW(jSrcEnd) ARGDW(jDstBeg) ARGDW(jDstEnd) ARGDW(pAAInfo->cAAData) ARGDW(cAADone) ARGDW(pAAInfo->PreMul) ARGDW(pAAInfo->PreSrcInc) ARGDW(pAAInfo->cPreLoad))); }
return(pAAInfo);}
#if DBG
BOOL ExpExp = TRUE;BOOL SrkSrk = TRUE;#endif
LONGHTENTRYComputeAABBP( PBITBLTPARAMS pBBP, PHTSURFACEINFO pDstSI, PAABBP pAABBP, BOOL GrayFunc )
/*++
Routine Description:
Arguments:
Return Value:
Author:
01-Apr-1998 Wed 20:32:36 created -by- Daniel Chou (danielc)
Revision History:
05-Aug-1998 Wed 19:38:56 updated -by- Daniel Chou (danielc) Fix banding problem
10-Aug-1998 Mon 16:05:32 updated -by- Daniel Chou (danielc) Fix rectangle banding with flip (X or Y) computation, the computation is done first by flip the Destination rectangles (Orginal and final) first by compute from right to left for flipping X, and bottom to top for flipping Y, after computation for stretch, we flip all the rectangles back
--*/
{ BITBLTPARAMS BBP; RECTL rclSurf; RECTL rclPhyDst; LONG cxIn; LONG cyIn; LONG cxOut; LONG cyOut; LONG Tmp; DWORD AAHFlags;
BBP = *pBBP; AAHFlags = pAABBP->AAHFlags;
DBGP_IF(DBGP_AAHEADER, DBGP(" Input: rclSrc=(%6ld, %6ld)-(%6ld, %6ld)=%6ld x %6ld" ARGDW(pBBP->rclSrc.left) ARGDW(pBBP->rclSrc.top) ARGDW(pBBP->rclSrc.right) ARGDW(pBBP->rclSrc.bottom) ARGDW(pBBP->rclSrc.right - pBBP->rclSrc.left) ARGDW(pBBP->rclSrc.bottom - pBBP->rclSrc.top)));
DBGP_IF(DBGP_AAHEADER, DBGP(" Input: rclDst=(%6ld, %6ld)-(%6ld, %6ld)=%6ld x %6ld" ARGDW(pBBP->rclDest.left) ARGDW(pBBP->rclDest.top) ARGDW(pBBP->rclDest.right) ARGDW(pBBP->rclDest.bottom) ARGDW(pBBP->rclDest.right - pBBP->rclDest.left) ARGDW(pBBP->rclDest.bottom - pBBP->rclDest.top)));
if (BBP.Flags & BBPF_HAS_DEST_CLIPRECT) {
DBGP_IF(DBGP_AAHEADER, DBGP(" Input: rclClip=(%6ld, %6ld)-(%6ld, %6ld)=%6ld x %6ld" ARGDW(pBBP->rclClip.left) ARGDW(pBBP->rclClip.top) ARGDW(pBBP->rclClip.right) ARGDW(pBBP->rclClip.bottom) ARGDW(pBBP->rclClip.right - pBBP->rclClip.left) ARGDW(pBBP->rclClip.bottom - pBBP->rclClip.top))); }
if (BBP.Flags & BBPF_HAS_BANDRECT) {
DBGP_IF(DBGP_AAHEADER, DBGP(" Input: rclBand=(%6ld, %6ld)-(%6ld, %6ld)=%6ld x %6ld" ARGDW(pBBP->rclBand.left) ARGDW(pBBP->rclBand.top) ARGDW(pBBP->rclBand.right) ARGDW(pBBP->rclBand.bottom) ARGDW(pBBP->rclBand.right - pBBP->rclBand.left) ARGDW(pBBP->rclBand.bottom - pBBP->rclBand.top))); }
DBGP_IF(DBGP_AAHEADER, DBGP("Input: ptlBrushOrg=(%6ld, %6ld)" ARGDW(pBBP->ptlBrushOrg.x) ARGDW(pBBP->ptlBrushOrg.y)));
DBGP_IF(DBGP_AAHEADER, DBGP("Input: ptlSrcMask=(%6ld, %6ld)" ARGDW(pBBP->ptlSrcMask.x) ARGDW(pBBP->ptlSrcMask.y)));
if (BBP.rclDest.right < BBP.rclDest.left) {
XCHG(BBP.rclDest.left, BBP.rclDest.right, Tmp); AAHFlags |= AAHF_FLIP_X; }
if (BBP.rclDest.bottom < BBP.rclDest.top) {
XCHG(BBP.rclDest.top, BBP.rclDest.bottom, Tmp); AAHFlags |= AAHF_FLIP_Y; }
//
// The source RECT is always well ordered
//
if (BBP.rclSrc.right < BBP.rclSrc.left) {
XCHG(BBP.rclSrc.left, BBP.rclSrc.right, Tmp); AAHFlags ^= AAHF_FLIP_X; }
if (BBP.rclSrc.bottom < BBP.rclSrc.top) {
XCHG(BBP.rclSrc.top, BBP.rclSrc.bottom, Tmp); AAHFlags ^= AAHF_FLIP_Y; }
if ((BBP.rclSrc.left >= BBP.rclSrc.right) || (BBP.rclSrc.top >= BBP.rclSrc.bottom)) {
DBGP_IF(DBGP_AAHEADER, DBGP("EMPTY rclSrc: (%ld, %ld)-(%ld, %ld)=%ldx%ld" ARGDW(BBP.rclSrc.left) ARGDW(BBP.rclSrc.top) ARGDW(BBP.rclSrc.right) ARGDW(BBP.rclSrc.bottom) ARGDW(BBP.rclSrc.right - BBP.rclSrc.left) ARGDW(BBP.rclSrc.bottom - BBP.rclSrc.top)));
return(0); }
if ((BBP.rclDest.left >= BBP.rclDest.right) || (BBP.rclDest.top >= BBP.rclDest.bottom)) {
DBGP_IF(DBGP_AAHEADER, DBGP("EMPTY rclDest: (%ld, %ld)-(%ld, %ld)=%ldx%ld" ARGDW(BBP.rclDest.left) ARGDW(BBP.rclDest.top) ARGDW(BBP.rclDest.right) ARGDW(BBP.rclDest.bottom) ARGDW(BBP.rclDest.right - BBP.rclDest.left) ARGDW(BBP.rclDest.bottom - BBP.rclDest.top)));
return(0); }
//
// set the cxIn, cyIn for no sign
//
cxIn = BBP.rclSrc.right - BBP.rclSrc.left; cyIn = BBP.rclSrc.bottom - BBP.rclSrc.top; cxOut = BBP.rclDest.right - BBP.rclDest.left; cyOut = BBP.rclDest.bottom - BBP.rclDest.top; pAABBP->ptlBrushOrg = BBP.ptlBrushOrg;
if (((((cxOut * 1000) + 500) / cxIn) > 667) && ((((cyOut * 1000) + 500) / cyIn) > 667)) {
AAHFlags |= AAHF_DO_FIXUPDIB; }
if ((cxOut * cyOut) < (cxIn * cyIn)) {
AAHFlags |= AAHF_SHRINKING; AAHFlags |= AAHF_DO_DST_CLR_MAPPING;
} else {
AAHFlags |= AAHF_DO_SRC_CLR_MAPPING; }
if (cyIn == cyOut) {
pAABBP->AAMaskCYFunc = BltMask_CY; pAABBP->GetAVCYFunc = BltAV_CY; pAABBP->AABuildCYFunc = BuildBltAAInfo; pAABBP->CYFuncMode = AACYMODE_BLT;
} else if (cyIn < cyOut) {
pAABBP->AAMaskCYFunc = ExpandMask_CY; pAABBP->GetAVCYFunc = ExpandAV_CY; pAABBP->AABuildCYFunc = BuildExpandAAInfo;
if (cxOut > cxIn) {
if ((!(AAHFlags & AAHF_BBPF_AA_OFF)) && ((cyIn * FAST_MAX_CY) >= cyOut) && ((cxIn * FAST_MAX_CX) >= cxOut)) {
AAHFlags |= AAHF_FAST_EXP_AA; }
pAABBP->CYFuncMode = AACYMODE_EXPAND_EXPCX;
} else {
pAABBP->CYFuncMode = AACYMODE_EXPAND; }
#if DBG
if (!ExpExp) {
pAABBP->CYFuncMode = AACYMODE_EXPAND; }#endif
} else {
pAABBP->AAMaskCYFunc = ShrinkMask_CY; pAABBP->GetAVCYFunc = ShrinkAV_CY; pAABBP->AABuildCYFunc = BuildShrinkAAInfo; pAABBP->CYFuncMode = ((cxOut < cxIn) && (!GrayFunc)) ? AACYMODE_SHRINK_SRKCX : AACYMODE_SHRINK;#if DBG
if (!SrkSrk) {
pAABBP->CYFuncMode = AACYMODE_SHRINK; }#endif
}
DBGP_IF(DBGP_FUNC, DBGP("\n+++++ AACYFuncMode = %ld" ARGDW(pAABBP->CYFuncMode)));
if (cxIn == cxOut) {
pAABBP->CXFuncMode = AACXMODE_BLT; pAABBP->AAMaskCXFunc = BltMask_CX; pAABBP->GetAVCXFunc = BltAV_CX; pAABBP->AABuildCXFunc = BuildBltAAInfo; pAABBP->AACXFunc = (GrayFunc) ? (AACXFUNC)GrayCopyDIB_CX : (AACXFUNC)CopyDIB_CX;
DBGP_IF(DBGP_FUNC, DBGP("+++++ AACXFunc = CopyDIB_CX()"));
} else if (cxIn < cxOut) {
pAABBP->CXFuncMode = AACXMODE_EXPAND; pAABBP->AAMaskCXFunc = ExpandMask_CX; pAABBP->GetAVCXFunc = ExpandAV_CX; pAABBP->AABuildCXFunc = BuildExpandAAInfo; pAABBP->AACXFunc = (GrayFunc) ? (AACXFUNC)GrayExpandDIB_CX : (AACXFUNC)ExpandDIB_CX;
DBGP_IF(DBGP_FUNC, DBGP("+++++ AACXFunc = ExpandDIB_CX()"));
} else {
pAABBP->CXFuncMode = AACXMODE_SHRINK; pAABBP->AAMaskCXFunc = ShrinkMask_CX; AAHFlags |= AAHF_OR_AV; pAABBP->GetAVCXFunc = ShrinkAV_CX; pAABBP->AABuildCXFunc = BuildShrinkAAInfo; pAABBP->AACXFunc = (GrayFunc) ? (AACXFUNC)GrayShrinkDIB_CX : (AACXFUNC)ShrinkDIB_CX;
DBGP_IF(DBGP_FUNC, DBGP("+++++ AACXFunc = ShrinkDIB_CX()")); }
if (BBP.Flags & BBPF_TILE_SRC) {
pAABBP->CYFuncMode = AACYMODE_TILE; pAABBP->AAMaskCXFunc = BltMask_CX; pAABBP->AAMaskCYFunc = BltMask_CY; pAABBP->GetAVCXFunc = NULL; pAABBP->GetAVCYFunc = TileAV_CY; pAABBP->AABuildCYFunc = pAABBP->AABuildCXFunc = BuildTileAAInfo; pAABBP->AACXFunc = NULL;
DBGP_IF(DBGP_FUNC, DBGP("+++ TILE: TileBlt_CY(), AACXFunc = NULL")); }
pAABBP->AAHFlags = AAHFlags; pAABBP->rclSrc = BBP.rclSrc; pAABBP->ptlMask.x = BBP.ptlSrcMask.x - BBP.rclSrc.left; pAABBP->ptlMask.y = BBP.ptlSrcMask.y - BBP.rclSrc.top; rclSurf = BBP.rclDest;
if (BBP.Flags & BBPF_HAS_DEST_CLIPRECT) {
if (!IntersectRECTL(&rclSurf, &BBP.rclClip)) {
DBGP_IF(DBGP_AAHEADER, DBGP("rclClip=(%ld, %ld)-(%ld, %ld)=%ldx%ld < SURF=(%ld, %ld)-(%ld, %ld)=%ldx%ld" ARGDW(BBP.rclClip.left) ARGDW(BBP.rclClip.top) ARGDW(BBP.rclClip.right) ARGDW(BBP.rclClip.bottom) ARGDW(BBP.rclClip.right - BBP.rclClip.left) ARGDW(BBP.rclClip.bottom - BBP.rclClip.top) ARGDW(rclSurf.left) ARGDW(rclSurf.top) ARGDW(rclSurf.right) ARGDW(rclSurf.bottom) ARGDW(rclSurf.right - rclSurf.left) ARGDW(rclSurf.bottom - rclSurf.top)));
return(0); } }
if (BBP.Flags & BBPF_HAS_BANDRECT) {
ASSERT(BBP.rclBand.left >= 0); ASSERT(BBP.rclBand.top >= 0); ASSERT(BBP.rclBand.right > BBP.rclBand.left); ASSERT(BBP.rclBand.bottom > BBP.rclBand.top);
if (!IntersectRECTL(&rclSurf, &BBP.rclBand)) {
DBGP_IF(DBGP_AAHEADER, DBGP("rclBand=(%ld, %ld)-(%ld, %ld)=%ldx%ld < SURF=(%ld, %ld)-(%ld, %ld)=%ldx%ld" ARGDW(BBP.rclBand.left) ARGDW(BBP.rclBand.top) ARGDW(BBP.rclBand.right) ARGDW(BBP.rclBand.bottom) ARGDW(BBP.rclBand.right - BBP.rclBand.left) ARGDW(BBP.rclBand.bottom - BBP.rclBand.top) ARGDW(rclSurf.left) ARGDW(rclSurf.top) ARGDW(rclSurf.right) ARGDW(rclSurf.bottom) ARGDW(rclSurf.right - rclSurf.left) ARGDW(rclSurf.bottom - rclSurf.top)));
return(0); }
//
// 05-Aug-1998 Wed 19:38:56 updated -by- Daniel Chou (danielc)
// Fixed the banding problem when mirrored or upside down stretch
// The fixes is simple by offset all the Dest rects for the left/top
// of the band, and reset the cx/cy destination size according to the
// band size and then offset the brush origin according the the band's
// left/top position, after these all other codes should run the same
// excpet we do not need to check BBPF_HAS_BANDRECT at later time.
//
BBP.rclDest.left -= BBP.rclBand.left; BBP.rclDest.right -= BBP.rclBand.left; BBP.rclDest.top -= BBP.rclBand.top; BBP.rclDest.bottom -= BBP.rclBand.top; rclSurf.left -= BBP.rclBand.left; rclSurf.right -= BBP.rclBand.left; rclSurf.top -= BBP.rclBand.top; rclSurf.bottom -= BBP.rclBand.top; pAABBP->ptlBrushOrg.x -= BBP.rclBand.left; pAABBP->ptlBrushOrg.y -= BBP.rclBand.top;
ASSERT((BBP.rclBand.right - BBP.rclBand.left) <= pDstSI->Width); ASSERT((BBP.rclBand.bottom - BBP.rclBand.top) <= pDstSI->Height);
DBGP_IF(DBGP_AAHEADER, DBGP("BAND Output: Dest: %ld x %ld --> BAND: %ld x %ld" ARGDW(pDstSI->Width) ARGDW(pDstSI->Height) ARGDW(BBP.rclBand.right - BBP.rclBand.left) ARGDW(BBP.rclBand.bottom - BBP.rclBand.top))); }
rclPhyDst.left = rclPhyDst.top = 0; rclPhyDst.right = pDstSI->Width; rclPhyDst.bottom = pDstSI->Height;
if (!IntersectRECTL(&rclSurf, &rclPhyDst)) {
DBGP_IF(DBGP_AAHEADER, DBGP("PhyDest=(%ld, %ld)-(%ld, %ld)=%ldx%ld < SURF=(%ld, %ld)-(%ld, %ld)=%ldx%ld" ARGDW(rclPhyDst.left) ARGDW(rclPhyDst.top) ARGDW(rclPhyDst.right) ARGDW(rclPhyDst.bottom) ARGDW(rclPhyDst.right - rclPhyDst.left) ARGDW(rclPhyDst.bottom - rclPhyDst.top) ARGDW(rclSurf.left) ARGDW(rclSurf.top) ARGDW(rclSurf.right) ARGDW(rclSurf.bottom) ARGDW(rclSurf.right - rclSurf.left) ARGDW(rclSurf.bottom - rclSurf.top)));
return(0); }
//
// 10-Aug-1998 Mon 16:09:13 updated -by- Daniel Chou (danielc)
// flipping X computation: When we flip in X direction, we will first
// compute the destination original rectangle and final destination
// rectangle by compute its offset from right hand side so later at
// stretch computation is easier, after finished stretch computation
// (BuildExpand or BuildShrink) we will flip the rectangle back by
// substract it from ptlFlip.x.
//
if (AAHFlags & AAHF_FLIP_X) {
DBGP_IF(DBGP_AAHEADER, DBGP("*** FLIP X: rclDstOrg=(%6ld - %6ld)=%6ld" ARGDW(BBP.rclDest.left) ARGDW(BBP.rclDest.right) ARGDW(BBP.rclDest.right - BBP.rclDest.left)));
DBGP_IF(DBGP_AAHEADER, DBGP("*** FLIP X: rclDst=(%6ld - %6ld)=%6ld, ptlFlip.x=%ld" ARGDW(rclSurf.left) ARGDW(rclSurf.right) ARGDW(rclSurf.right - rclSurf.left) ARGDW(BBP.rclDest.right)));
Tmp = rclSurf.right - rclSurf.left; rclSurf.left = BBP.rclDest.right - rclSurf.right; rclSurf.right = rclSurf.left + Tmp; pAABBP->ptlFlip.x = BBP.rclDest.right; BBP.rclDest.right -= BBP.rclDest.left; BBP.rclDest.left = 0; }
//
// 10-Aug-1998 Mon 16:09:13 updated -by- Daniel Chou (danielc)
// flipping Y computation: When we flip in Y direction, we will first
// compute the destination original rectangle and final destination
// rectangle by compute its offset from bottom hand side so later at
// stretch computation is easier, after finished stretch computation
// (BuildExpand or BuildShrink) we will flip the rectangle back by
// substract it from ptlFlip.y.
//
if (AAHFlags & AAHF_FLIP_Y) {
DBGP_IF(DBGP_AAHEADER, DBGP("*** FLIP Y: rclDstOrg=(%6ld - %6ld)=%6ld" ARGDW(BBP.rclDest.top) ARGDW(BBP.rclDest.bottom) ARGDW(BBP.rclDest.bottom - BBP.rclDest.top)));
DBGP_IF(DBGP_AAHEADER, DBGP("*** FLIP Y: rclDst=(%6ld - %6ld)=%6ld, ptlFlip.y=%ld" ARGDW(rclSurf.top) ARGDW(rclSurf.bottom) ARGDW(rclSurf.bottom - rclSurf.top) ARGDW(BBP.rclDest.bottom)));
Tmp = rclSurf.bottom - rclSurf.top; rclSurf.top = BBP.rclDest.bottom - rclSurf.bottom; rclSurf.bottom = rclSurf.top + Tmp; pAABBP->ptlFlip.y = BBP.rclDest.bottom; BBP.rclDest.bottom -= BBP.rclDest.top; BBP.rclDest.top = 0; }
pAABBP->rclDstOrg = BBP.rclDest; pAABBP->rclDst = rclSurf;
DBGP_IF(DBGP_AAHEADER, DBGP("Output: rclSrc=(%6ld, %6ld)-(%6ld, %6ld)=%6ld x %6ld" ARGDW(pAABBP->rclSrc.left) ARGDW(pAABBP->rclSrc.top) ARGDW(pAABBP->rclSrc.right) ARGDW(pAABBP->rclSrc.bottom) ARGDW(pAABBP->rclSrc.right - pAABBP->rclSrc.left) ARGDW(pAABBP->rclSrc.bottom - pAABBP->rclSrc.top)));
DBGP_IF(DBGP_AAHEADER, DBGP("Output: rclDstOrg=(%6ld, %6ld)-(%6ld, %6ld)=%6ld x %6ld" ARGDW(pAABBP->rclDstOrg.left) ARGDW(pAABBP->rclDstOrg.top) ARGDW(pAABBP->rclDstOrg.right) ARGDW(pAABBP->rclDstOrg.bottom) ARGDW(pAABBP->rclDstOrg.right - pAABBP->rclDstOrg.left) ARGDW(pAABBP->rclDstOrg.bottom - pAABBP->rclDstOrg.top)));
DBGP_IF(DBGP_AAHEADER, DBGP("Output: rclDst=(%6ld, %6ld)-(%6ld, %6ld)=%6ld x %6ld" ARGDW(pAABBP->rclDst.left) ARGDW(pAABBP->rclDst.top) ARGDW(pAABBP->rclDst.right) ARGDW(pAABBP->rclDst.bottom) ARGDW(pAABBP->rclDst.right - pAABBP->rclDst.left) ARGDW(pAABBP->rclDst.bottom - pAABBP->rclDst.top)));
DBGP_IF(DBGP_AAHEADER, DBGP("Output: ptlBrushOrg=(%6ld, %6ld)" ARGDW(pAABBP->ptlBrushOrg.x) ARGDW(pAABBP->ptlBrushOrg.y)));
DBGP_IF(DBGP_AAHEADER, DBGP("Output: ptlSrcMask=(%6ld, %6ld)" ARGDW(pAABBP->ptlMask.x) ARGDW(pAABBP->ptlMask.y)));
return(1);
}
#if DBG
extern INT cCXMask;
LPSTRGetAACXFuncName( AACXFUNC AACXFunc )
/*++
Routine Description:
Arguments:
Return Value:
Author:
06-Jan-1999 Wed 19:11:27 created -by- Daniel Chou (danielc)
Revision History:
--*/
{ if (AACXFunc == (AACXFUNC)RepDIB_CX) {
return("RepDIB_CX");
} else if (AACXFunc == (AACXFUNC)SkipDIB_CX) {
return("SkipDIB_CX");
} else if (AACXFunc == (AACXFUNC)CopyDIB_CX) {
return("CopyDIB_CX");
} else if (AACXFunc == (AACXFUNC)ShrinkDIB_CX) {
return("ShrinkDIB_CX");
} else if (AACXFunc == (AACXFUNC)ExpandDIB_CX) {
return("ExpandDIB_CX");
} else if (AACXFunc == (AACXFUNC)GrayRepDIB_CX) {
return("GrayRepDIB_CX");
} else if (AACXFunc == (AACXFUNC)GraySkipDIB_CX) {
return("GraySkipDIB_CX");
} else if (AACXFunc == (AACXFUNC)GrayCopyDIB_CXGray) {
return("GrayCopyDIB_CXGray");
} else if (AACXFunc == (AACXFUNC)GrayCopyDIB_CX) {
return("GrayCopyDIB_CX");
} else if (AACXFunc == (AACXFUNC)GrayExpandDIB_CX) {
return("GrayExpandDIB_CX");
} else if (AACXFunc == (AACXFUNC)GrayShrinkDIB_CX) {
return("GrayShrinkDIB_CX");
} else {
DBGP("ERROR: Unknown AACXFUNC=%p, Function" ARGPTR(AACXFunc));
return("Unknown AACXFUNC"); }
}
LPSTRGetAACYFuncName( AACYFUNC AACYFunc )
/*++
Routine Description:
Arguments:
Return Value:
Author:
06-Jan-1999 Wed 19:11:27 created -by- Daniel Chou (danielc)
Revision History:
--*/
{
if (AACYFunc == (AACYFUNC)TileDIB_CY) {
return("TileDIB_CY");
} else if (AACYFunc == (AACYFUNC)RepDIB_CY) {
return("RepDIB_CY");
} else if (AACYFunc == (AACYFUNC)FastExpAA_CY) {
return("FastExpAA_CY");
} else if (AACYFunc == (AACYFUNC)SkipDIB_CY) {
return("SkipDIB_CY");
} else if (AACYFunc == (AACYFUNC)BltDIB_CY) {
return("BltDIB_CY");
} else if (AACYFunc == (AACYFUNC)ShrinkDIB_CY) {
return("ShrinkDIB_CY");
} else if (AACYFunc == (AACYFUNC)ShrinkDIB_CY_SrkCX) {
return("ShrinkDIB_CY_SrkCX");
} else if (AACYFunc == (AACYFUNC)ExpandDIB_CY_ExpCX) {
return("ExpandDIB_CY_ExpCX");
} else if (AACYFunc == (AACYFUNC)ExpandDIB_CY) {
return("ExpandDIB_CY");
} else if (AACYFunc == (AACYFUNC)GrayExpandDIB_CY_ExpCX) {
return("GrayExpandDIB_CY_ExpCX");
} else if (AACYFunc == (AACYFUNC)GrayExpandDIB_CY) {
return("GrayExpandDIB_CY");
} else if (AACYFunc == (AACYFUNC)GrayShrinkDIB_CY) {
return("GrayShrinkDIB_CY");
} else {
DBGP("ERROR: Unknown AACYFUNC=%p, Function" ARGPTR(AACYFunc));
return("Unknown AACYFUNC"); }}
#endif
LONGHTENTRYSetupAAHeader( PHALFTONERENDER pHR, PDEVICECOLORINFO pDCI, PAAHEADER pAAHdr, AACYFUNC *pAACYFunc )
/*++
Routine Description:
Arguments:
Return Value:
Author:
03-Apr-1998 Fri 04:27:16 created -by- Daniel Chou (danielc)
Revision History:
07-Aug-1998 Fri 19:28:21 updated -by- Daniel Chou (danielc) Fix the mirror/upside-down banding problems, the problem is when we flip the destination original size rectangle we did not flip the the real destination rectangle, this cause the the offset getting too big and its pointer offset passed the end of the bitmap, it also has problem that when we did not flip the destination rectangle the wrong portion of the bitmap will be in the band.
--*/
{ PAAINFO pAAInfoCX; PAAINFO pAAInfoCY; PBITBLTPARAMS pBBP; PHTSURFACEINFO pSrcSI; PHTSURFACEINFO pDstSI; PHTSURFACEINFO pMaskSI; LPBYTE pbExtra; AACYFUNC AACYFunc; AABBP aabbp; LONG iFree; LONG cbFreeBuf; LONG Top; LONG Bottom; LONG cyNext; LONG cxSize; LONG cbMaskSrc; LONG cIn; LONG cOut; LONG cbCYExtra; LONG cbCXExtra; LONG cbInBuf; LONG cbFUDI; LONG cbVGA256Xlate; LONG cbOutBuf; LONG cbAlphaBuf; LONG cbIdxBGR; LONG Result; LONG cbSrcPel; DWORD PrimAdjFlags; UINT DstSurfFmt;
pBBP = pHR->pBitbltParams; pSrcSI = pHR->pSrcSI; pDstSI = pHR->pDestSI; pMaskSI = pHR->pSrcMaskSI; DstSurfFmt = (UINT)pDstSI->SurfaceFormat; PrimAdjFlags = (DWORD)pHR->pDevClrAdj->PrimAdj.Flags; cbSrcPel = (pHR->pDevClrAdj->DMI.Flags & DMIF_GRAY) ? sizeof(BYTE) : sizeof(BGR8); DBGP_IF(DBGP_AAHEADER, DBGP("\nSrcSI=%ld x %ld [Format=%ld], DestSI=%ld x %ld [Format=%ld]" ARGDW(pSrcSI->Width) ARGDW(pSrcSI->Height) ARGDW(pSrcSI->SurfaceFormat) ARGDW(pDstSI->Width) ARGDW(pDstSI->Height) ARGDW(pDstSI->SurfaceFormat)));
aabbp.AAHFlags = (PrimAdjFlags & DCA_BBPF_AA_OFF) ? AAHF_BBPF_AA_OFF : 0;
if (pBBP->Flags & BBPF_TILE_SRC) {
aabbp.AAHFlags |= AAHF_TILE_SRC; aabbp.AAHFlags |= AAHF_BBPF_AA_OFF; }
if (ComputeAABBP(pBBP, pDstSI, &aabbp, cbSrcPel == sizeof(BYTE)) <= 0) {
return(0); }
cbCXExtra = sizeof(RGBLUTAA);
if (PrimAdjFlags & DCA_ALPHA_BLEND) {
aabbp.AAHFlags |= AAHF_ALPHA_BLEND;
if (PrimAdjFlags & DCA_CONST_ALPHA) {
aabbp.AAHFlags |= AAHF_CONST_ALPHA;
cbCXExtra += (AB_BGR_CA_SIZE + AB_CONST_SIZE);
} else {
if (PrimAdjFlags & DCA_AB_PREMUL_SRC) {
//
// Set the flag so we will try to compute premul's orginal src
//
pAAHdr->SrcSurfInfo.Flags |= AASIF_AB_PREMUL_SRC; }
if (PrimAdjFlags & DCA_AB_DEST) {
aabbp.AAHFlags |= AAHF_AB_DEST; }
cbCXExtra += AB_BGR_SIZE; } }
if (PrimAdjFlags & DCA_NO_MAPPING_TABLE) {
aabbp.AAHFlags &= ~AAHF_DO_CLR_MAPPING; }
ALIGN_MEM(cbCXExtra, cbCXExtra);
DBGP_IF(DBGP_LUT_MAP, DBGP("CXExtra=%ld (RGBLUTAA) + 0 (Mapping=%hs) = %ld" ARGDW(sizeof(RGBLUTAA)) ARGPTR((aabbp.AAHFlags & AAHF_DO_SRC_CLR_MAPPING) ? "SRC" : "DST") ARGDW(sizeof(RGBLUTAA) + 0)));
//
// This flag is set when reading the source bitmap, so the pointer will
// advanced to next scanline, this flag will not set for the destiantion
// since when blending and we reading from destination, we do not want to
// advanced the destination pointer because it will be done during output
//
ComputeInputColorInfo((LPBYTE)pSrcSI->pColorTriad->pColorTable, (UINT)pSrcSI->pColorTriad->BytesPerEntry, (UINT)pSrcSI->pColorTriad->PrimaryOrder, &(pHR->BFInfo), &(pAAHdr->SrcSurfInfo));
pAAHdr->SrcSurfInfo.Flags |= AASIF_INC_PB | ((cbSrcPel == sizeof(BYTE)) ? AASIF_GRAY : 0);
if (PrimAdjFlags & DCA_USE_ADDITIVE_PRIMS) {
aabbp.AAHFlags |= AAHF_ADDITIVE; }
if (pBBP->Flags & BBPF_TILE_SRC) {
aabbp.AAHFlags &= ~AAHF_DO_FIXUPDIB; }
if (aabbp.AAHFlags & AAHF_BBPF_AA_OFF) {
aabbp.AAHFlags &= ~AAHF_DO_FIXUPDIB; }
if (aabbp.AAHFlags & AAHF_DO_FIXUPDIB) {
CheckBMPNeedFixup(pDCI, pAAHdr, pSrcSI, &aabbp);
if (aabbp.AAHFlags & AAHF_SHRINKING) {
if (PrimAdjFlags & DCA_BBPF_AA_OFF) {
aabbp.AAHFlags |= AAHF_BBPF_AA_OFF;
} else {
aabbp.AAHFlags &= ~AAHF_BBPF_AA_OFF; } }
}
if (aabbp.AAHFlags & AAHF_BBPF_AA_OFF) {
aabbp.AAHFlags &= ~AAHF_FAST_EXP_AA; }
if (pMaskSI) {
aabbp.AAHFlags |= AAHF_HAS_MASK; }
if (!(pAAInfoCX = aabbp.AABuildCXFunc(pDCI, aabbp.AAHFlags, &aabbp.rclSrc.left, &aabbp.rclSrc.right, pSrcSI->Width, aabbp.rclDstOrg.left, aabbp.rclDstOrg.right, &aabbp.rclDst.left, &aabbp.rclDst.right, cbCXExtra))) {
//
// Remove cbCXExtra (use pDCI's rgbLUT and BGRMapTable, AlphaBlendBGR)
//
if (!(pAAInfoCX = aabbp.AABuildCXFunc(pDCI, aabbp.AAHFlags, &aabbp.rclSrc.left, &aabbp.rclSrc.right, pSrcSI->Width, aabbp.rclDstOrg.left, aabbp.rclDstOrg.right, &aabbp.rclDst.left, &aabbp.rclDst.right, cbCXExtra = 0))) {
return(HTERR_INSUFFICIENT_MEMORY); } }
// Bug 27036: reject empty rectangles
if(!pAAInfoCX->cIn || !pAAInfoCX->cOut) {#if 0
LONG crash = 1 ; // empty src or dest rectangle!
crash /= (pAAInfoCX->cIn * pAAInfoCX->cOut); // delete when debugging is complete.
if(crash) return 0 ;#endif
HTFreeMem(pAAInfoCX); return 0 ; }
if (cbCXExtra) {
pAAHdr->prgbLUT = (PRGBLUTAA)(pbExtra = pAAInfoCX->pbExtra); pbExtra += sizeof(RGBLUTAA);
ASSERT_MEM_ALIGN(pAAHdr->prgbLUT, sizeof(LONG));
if (aabbp.AAHFlags & AAHF_ALPHA_BLEND) {
pAAHdr->pAlphaBlendBGR = (LPBYTE)pbExtra;
if (PrimAdjFlags & DCA_CONST_ALPHA) {
pbExtra += (AB_BGR_CA_SIZE + AB_CONST_SIZE);
} else {
pbExtra += AB_BGR_SIZE; } }
} else {
ASSERT_MEM_ALIGN(&pDCI->rgbLUT, sizeof(LONG));
aabbp.AAHFlags |= AAHF_USE_DCI_DATA; pAAHdr->prgbLUT = &pDCI->rgbLUT;
if (aabbp.AAHFlags & AAHF_ALPHA_BLEND) {
pAAHdr->pAlphaBlendBGR = pDCI->pAlphaBlendBGR;
if (PrimAdjFlags & DCA_CONST_ALPHA) {
pAAHdr->pAlphaBlendBGR += AB_BGR_SIZE; } } }
pAAHdr->pIdxBGR = pAAHdr->prgbLUT->IdxBGR;
if (aabbp.AAHFlags & AAHF_FLIP_X) {
DBGP_IF(DBGP_AAHEADER, DBGP("X Dst=(%ld - %ld)=%ld change it to (%ld - %ld), ptlFlip.x=%ld" ARGDW(aabbp.rclDst.left) ARGDW(aabbp.rclDst.right) ARGDW(aabbp.rclDst.right - aabbp.rclDst.left) ARGDW(aabbp.ptlFlip.x - aabbp.rclDst.left - 1) ARGDW(aabbp.ptlFlip.x - aabbp.rclDst.right - 1) ARGDW(aabbp.ptlFlip.x)));
aabbp.rclDst.left = aabbp.ptlFlip.x - aabbp.rclDst.left - 1; aabbp.rclDst.right = aabbp.ptlFlip.x - aabbp.rclDst.right - 1; }
//
// cbCYExtra is for the input scan line, add one because we want to
// run it in DWORD mode so we need at least an extra byte at end
// of input buffer
//
cIn = pAAInfoCX->cIn; cOut = pAAInfoCX->cOut; cbInBuf = cIn + 6; // left extra=3, right extra=3
cbFreeBuf = cbCYExtra = 0; AACYFunc = NULL;
if (aabbp.AAHFlags & AAHF_BBPF_AA_OFF) {
switch (aabbp.CYFuncMode) {
case AACYMODE_SHRINK: case AACYMODE_SHRINK_SRKCX:
AACYFunc = (AACYFUNC)SkipDIB_CY; break;
case AACYMODE_EXPAND: case AACYMODE_EXPAND_EXPCX:
cbCYExtra = (aabbp.AAHFlags & AAHF_ALPHA_BLEND) ? (sizeof(BGR8) * (cOut + 6)) : 0;
AACYFunc = (AACYFUNC)RepDIB_CY; break;
case AACYMODE_BLT:
AACYFunc = (AACYFUNC)BltDIB_CY; break; }
if (AACYFunc) {
switch (aabbp.CXFuncMode) {
case AACXMODE_BLT:
aabbp.AACXFunc = (cbSrcPel == sizeof(BYTE)) ? (AACXFUNC)GrayCopyDIB_CXGray : (AACXFUNC)CopyDIB_CX; break;
case AACXMODE_SHRINK:
aabbp.AACXFunc = (cbSrcPel == sizeof(BYTE)) ? (AACXFUNC)GraySkipDIB_CX : (AACXFUNC)SkipDIB_CX; break;
case AACXMODE_EXPAND:
aabbp.AACXFunc = (cbSrcPel == sizeof(BYTE)) ? (AACXFUNC)GrayRepDIB_CX : (AACXFUNC)RepDIB_CX; break; }
aabbp.CYFuncMode = AACYMODE_NONE; } }
switch (aabbp.CYFuncMode) {
case AACYMODE_TILE:
AACYFunc = TileDIB_CY; cbCYExtra = (cbSrcPel == sizeof(BYTE)) ? (cIn * sizeof(WORD)) : 0; break;
case AACYMODE_BLT:
AACYFunc = (AACYFUNC)BltDIB_CY; break;
case AACYMODE_SHRINK:
//
// We need to make sure Off555Buf does not changed
//
cbFreeBuf = (sizeof(LONG) * 256 * 2);
if (cbSrcPel == sizeof(BYTE)) {
AACYFunc = GrayShrinkDIB_CY; cbCYExtra = (sizeof(LONG) * cOut * 3) + cbFreeBuf + ((cOut + 6) * cbSrcPel);
} else {
AACYFunc = ShrinkDIB_CY; cbCYExtra = (sizeof(RGBL) * cIn * 3) + cbFreeBuf + (cbInBuf * cbSrcPel); }
break;
case AACYMODE_SHRINK_SRKCX:
AACYFunc = ShrinkDIB_CY_SrkCX; cbFreeBuf = (sizeof(LONG) * 256 * 2); cbCYExtra = (sizeof(RGBL) * (pAAInfoCX->cAADone + 2) * 3) + cbFreeBuf; break;
case AACYMODE_EXPAND:
AACYFunc = (cbSrcPel == sizeof(BYTE)) ? GrayExpandDIB_CY : ExpandDIB_CY; cbCYExtra = (cbFreeBuf = (sizeof(LONG) * 256 * 4)) + ((cOut + 6) * cbSrcPel * 6); break;
case AACYMODE_EXPAND_EXPCX:
//
// This function use IputBufBeg to sharpening the input scanline so we
// need 4 extra BGR8 for running the expand pre-read
//
if (aabbp.AAHFlags & AAHF_FAST_EXP_AA) {
DBGP_IF(DBGP_AAHEADER, DBGP("Use FastExpAA_CY functions"));
cbCYExtra = (cbInBuf * 5 * cbSrcPel); AACYFunc = (AACYFUNC)FastExpAA_CY;
} else {
AACYFunc = (cbSrcPel == sizeof(BYTE)) ? GrayExpandDIB_CY_ExpCX : ExpandDIB_CY_ExpCX; cbCYExtra = (cbFreeBuf = (sizeof(LONG) * 256 * 4)) + (cbInBuf * cbSrcPel * 3) + ((cOut + 6) * cbSrcPel * 4); }
break; }
cbAlphaBuf = (aabbp.AAHFlags & AAHF_ALPHA_BLEND) ? cOut : 0; pAAHdr->DstSurfInfo.Flags = (cbSrcPel == sizeof(BYTE)) ? AASIF_GRAY : 0; pAAHdr->DstSurfInfo.cbCX = cbAlphaBuf * cbSrcPel;
ALIGN_MEM(cbAlphaBuf, (cbAlphaBuf + 2 + 6) * cbSrcPel); ALIGN_MEM(cbCYExtra, cbCYExtra);
//
// cbInBuf is for the input scan line of EXPAND/SHRINK mode, add one
// because we want to run it in DWORD mode so we need at least an extra
// byte at end of input buffer
//
// 26-Jun-1998 Fri 16:03:26 updated -by- Daniel Chou (danielc)
// The cbOutBuf is used only when we flipping in X direction, this is
// needed since the input/output buffer may collide into each other
//
ALIGN_MEM(cbInBuf, (cbInBuf + 2) * cbSrcPel); ALIGN_MEM(cbOutBuf, (cOut + (FAST_MAX_CX * 2)) * sizeof(BGRF));
cbMaskSrc = (aabbp.AAHFlags & AAHF_HAS_MASK) ? (ComputeBytesPerScanLine(BMF_1BPP, 4, cIn) + 4) : 0; cbMaskSrc = _ALIGN_MEM(cbMaskSrc);
if (cbInBuf < cbAlphaBuf) {
cbInBuf = cbAlphaBuf; }
if ((aabbp.AAHFlags & (AAHF_ALPHA_BLEND | AAHF_CONST_ALPHA)) == AAHF_ALPHA_BLEND) {
ALIGN_MEM(cbAlphaBuf, cOut);
} else {
cbAlphaBuf = 0; }
DBGP_IF(DBGP_FIXUPDIB, DBGP("** Allocate cIn=%ld, cOut=%ld, cbInBuf=%ld, cbOutBuf=%ld, cbMaskSrc=%ld" ARGDW(cIn) ARGDW(cOut) ARGDW(cbInBuf) ARGDW(cbOutBuf) ARGDW(cbMaskSrc)));
if ((DstSurfFmt == BMF_8BPP_VGA256) && (pHR->pXlate8BPP)) {
ALIGN_MEM(cbVGA256Xlate, SIZE_XLATE_666);
DBGP_IF((DBGP_AAHTPAT | DBGP_AAHT_MEM), DBGP("Allocate %ld bytes of Xlate8BPP" ARGDW(cbVGA256Xlate)));
} else {
cbVGA256Xlate = 0; }
if (aabbp.AAHFlags & AAHF_DO_FIXUPDIB) {
ALIGN_MEM(cbFUDI, (cIn + 4) * cbSrcPel);
} else {
cbFUDI = 0; }
if ((pAAHdr->SrcSurfInfo.Flags & AASIF_GRAY) && (pHR->BFInfo.Flags & BFIF_RGB_888) && (pAAHdr->SrcSurfInfo.AABFData.Flags & AABF_MASK_IS_ORDER) && (pHR->BFInfo.RGBOrder.Index != PRIMARY_ORDER_BGR)) {
//
// This is for mapping >= 16bpp source's IdxBGR to gray, when mapping
// to gray we will make IdxBGR to a correct source order so that it
// will optimized the source input function speed
//
// The AABF_MASK_IS_ORDER indicate the source is 8-bits each of Red,
// green and blue and it is only occuply lower 24-bits of either
// a 24-bits or a 32-bits data
//
ALIGN_MEM(cbIdxBGR, sizeof(LONG) * 256 * 3);
DBGP_IF(DBGP_AAHEADER, DBGP("Allocate gray non 24bits BGR [%ld] IDXBGR of %ld bytes" ARGDW(pHR->BFInfo.RGBOrder.Index) ARGDW(cbIdxBGR)));
} else {
cbIdxBGR = 0; }
if (pAAInfoCY = aabbp.AABuildCYFunc(pDCI, aabbp.AAHFlags, &aabbp.rclSrc.top, &aabbp.rclSrc.bottom, pSrcSI->Height, aabbp.rclDstOrg.top, aabbp.rclDstOrg.bottom, &aabbp.rclDst.top, &aabbp.rclDst.bottom, cbInBuf + cbOutBuf + cbMaskSrc + (cbFUDI * 6) + cbAlphaBuf + cbIdxBGR + cbVGA256Xlate + cbCYExtra)) { // Bug 27036: reject empty rectangles
if(!pAAInfoCY->cIn || !pAAInfoCY->cOut) {#if 0
LONG crash = 1 ; // empty src or dest rectangle!
crash /= (pAAInfoCY->cIn * pAAInfoCY->cOut); // delete when debugging is complete.
if(crash) return 0 ;#endif
HTFreeMem(pAAInfoCX); HTFreeMem(pAAInfoCY); return 0 ; }
pbExtra = pAAInfoCY->pbExtra + cbCYExtra; pAAHdr->Flags = aabbp.AAHFlags; pAAHdr->SrcSurfInfo.cbCX = cbSrcPel * cIn; pAAHdr->pInputBeg = (PBGR8)pbExtra; pbExtra += cbInBuf;
if (cbAlphaBuf) {
//
// 04-Aug-2000 Fri 10:31:45 updated -by- Daniel Chou (danielc)
// Since cbAlphaBuf is Memoey Aliged Adjusted, we want the
// pSrcAVEnd to be at exactly count of cOut not cbAlphaBuf
//
pAAHdr->pSrcAV = pAAHdr->pSrcAVBeg = (LPBYTE)pbExtra; pAAHdr->pSrcAVEnd = (LPBYTE)pbExtra + cOut; pbExtra += cbAlphaBuf; pAAHdr->SrcAVInc = sizeof(BYTE); }
if (cbFUDI) {
pAAHdr->pbFixupDIB = (LPBYTE)pbExtra; pAAHdr->FUDI.cbbgr = (DWORD)cbFUDI;
for (Top = 0; Top < 6; Top++) {
pAAHdr->FUDI.prgbD[Top] = (PBGR8)pbExtra; pbExtra += cbFUDI; } }
if (cbVGA256Xlate) {
pAAHdr->pXlate8BPP = pbExtra; pbExtra += cbVGA256Xlate; }
if (cbMaskSrc) {
pAAHdr->pMaskSrc = pbExtra; pbExtra += cbMaskSrc; }
if (cbIdxBGR) {
//
// The pIdxBGR is a local version that will be later re-arranged
// from pAAHdr->pIdxBGR to correct source byte order in
// SetGrayColorTable() function
//
pAAHdr->SrcSurfInfo.pIdxBGR = (PLONG)pbExtra; pbExtra += cbIdxBGR;
} else {
pAAHdr->SrcSurfInfo.pIdxBGR = pAAHdr->pIdxBGR; }
DBGP_IF(DBGP_AAHEADER, DBGP("cbInBuf=%ld, %p-%p" ARGDW(cbInBuf) ARGPTR(pAAHdr->pInputBeg) ARGPTR((LPBYTE)pAAHdr->pInputBeg + cbInBuf)));
//
// FAST_MAX_CX are added to both end of AABuf, this is needed when
// we want to process the output fast, since we may need to extended
// the computation to the neighbor pixels
//
pAAHdr->pOutputBeg = pAAHdr->pRealOutBeg = pAAHdr->pAABufBeg = (PBGRF)pbExtra + FAST_MAX_CX; pAAHdr->pRealOutEnd = pAAHdr->pOutputEnd = pAAHdr->pAABufBeg + cOut; pAAHdr->pAABufEnd = pAAHdr->pOutputEnd;
//
// Set the BGRF's Flags to 0xFF first, the 0xFF indicate that this
// pixel need to be output (masked).
//
FillMemory((LPBYTE)pAAHdr->pOutputBeg, (LPBYTE)pAAHdr->pOutputEnd - (LPBYTE)pAAHdr->pOutputBeg, PBGRF_MASK_FLAG);
//
// We mirror the image by composed the source the the AABuf in
// reverse way. Reading the source from left to right but when
// composed the source buffer (AABuf) we put it from right to left
//
if (aabbp.rclDst.left > aabbp.rclDst.right) {
XCHG(aabbp.rclDst.left, aabbp.rclDst.right, Result);
pAAHdr->pAABufBeg = pAAHdr->pOutputEnd - 1; pAAHdr->pAABufEnd = pAAHdr->pOutputBeg - 1; pAAHdr->AABufInc = -(LONG)sizeof(BGRF); pAAHdr->pSrcAVBeg = pAAHdr->pSrcAVEnd - 1; pAAHdr->pSrcAVEnd = pAAHdr->pSrcAV - 1; pAAHdr->SrcAVInc = -pAAHdr->SrcAVInc;
} else {
pAAHdr->AABufInc = (LONG)sizeof(BGRF); }
pAAHdr->ptlBrushOrg.x = aabbp.rclDst.left - aabbp.ptlBrushOrg.x;
DBGP_IF(DBGP_AAHEADER, DBGP("pInput=%p-%p (%ld), pAABuf=%p-%p (%ld), pOutput=%p-%p, DstLeft=%ld" ARGPTR(pAAHdr->pInputBeg) ARGPTR((LPBYTE)pAAHdr->pInputBeg + pAAHdr->SrcSurfInfo.cbCX) ARGL(pAAHdr->SrcSurfInfo.cbCX) ARGPTR(pAAHdr->pAABufBeg) ARGPTR(pAAHdr->pAABufEnd) ARGDW(pAAHdr->AABufInc) ARGPTR(pAAHdr->pOutputBeg) ARGPTR(pAAHdr->pOutputEnd) ARGDW(aabbp.rclDst.left)));
if (aabbp.AAHFlags & AAHF_FLIP_Y) {
DBGP_IF(DBGP_AAHEADER, DBGP("Y Dst=(%ld - %ld)=%ld change it to (%ld - %ld), ptlFlip.y=%ld" ARGDW(aabbp.rclDst.top) ARGDW(aabbp.rclDst.bottom) ARGDW(aabbp.rclDst.bottom - aabbp.rclDst.top) ARGDW(aabbp.ptlFlip.y - aabbp.rclDst.top - 1) ARGDW(aabbp.ptlFlip.y - aabbp.rclDst.bottom - 1) ARGDW(aabbp.ptlFlip.y)));
aabbp.rclDst.top = aabbp.ptlFlip.y - aabbp.rclDst.top - 1; aabbp.rclDst.bottom = aabbp.ptlFlip.y - aabbp.rclDst.bottom - 1; }
pAAHdr->ptlBrushOrg.y = aabbp.rclDst.top - aabbp.ptlBrushOrg.y;
DBGP_IF(DBGP_AAHEADER, DBGP("BrushOrg=(%ld, %ld) ---> (%ld, %ld)" ARGDW(pBBP->ptlBrushOrg.x) ARGDW(pBBP->ptlBrushOrg.y) ARGDW(pAAHdr->ptlBrushOrg.x) ARGDW(pAAHdr->ptlBrushOrg.y)));
pAAHdr->pAAInfoCX = pAAInfoCX; pAAHdr->pAAInfoCY = pAAInfoCY; pAAHdr->AACXFunc = aabbp.AACXFunc; pAAHdr->SrcSurfInfo.cx = pAAInfoCX->cIn; pAAHdr->SrcSurfInfo.cyOrg = pAAHdr->SrcSurfInfo.cy = pAAInfoCY->cIn;
if (aabbp.AAHFlags & AAHF_HAS_MASK) {
POINTL MaskEnd;
cyNext = cxSize = GET_PHTSI_CXSIZE(pMaskSI); aabbp.ptlMask.x += pAAInfoCX->Mask.iBeg; aabbp.ptlMask.y += pAAInfoCY->Mask.iBeg; MaskEnd.x = aabbp.ptlMask.x + pAAInfoCX->Mask.iSize; MaskEnd.y = aabbp.ptlMask.y + pAAInfoCY->Mask.iSize;
if ((aabbp.ptlMask.x < 0) || (aabbp.ptlMask.y < 0) || (MaskEnd.x > pMaskSI->Width) || (MaskEnd.y > pMaskSI->Height)) {
HTFreeMem(pAAInfoCX); HTFreeMem(pAAInfoCY);
return(HTERR_SRC_MASK_BITS_TOO_SMALL); }
pAAHdr->cyMaskNext = cyNext; pAAHdr->cyMaskIn = pAAInfoCY->Mask.iSize; iFree = ComputeByteOffset(BMF_1BPP, MaskEnd.x, &(pAAHdr->MaskBitOff)); cbFreeBuf = ComputeByteOffset(BMF_1BPP, aabbp.ptlMask.x, &(pAAHdr->MaskBitOff)); pAAHdr->cbMaskSrc = iFree - cbFreeBuf + 1; pAAHdr->pMaskIn = pMaskSI->pPlane + (aabbp.ptlMask.y * cxSize) + cbFreeBuf;
DBGP_IF(DBGP_MASK | DBGP_AAHEADER, DBGP("CX: iMaskBeg=%ld, iMaskSize=%ld, cMaskIn=%ld, cMaskOut=%ld" ARGDW(pAAInfoCX->Mask.iBeg) ARGDW(pAAInfoCX->Mask.iSize) ARGDW(pAAInfoCX->Mask.cIn) ARGDW(pAAInfoCX->Mask.cOut)));
DBGP_IF(DBGP_MASK | DBGP_AAHEADER, DBGP("CY: iMaskBeg=%ld, iMaskSize=%ld, cMaskIn=%ld, cMaskOut=%ld" ARGDW(pAAInfoCY->Mask.iBeg) ARGDW(pAAInfoCY->Mask.iSize) ARGDW(pAAInfoCY->Mask.cIn) ARGDW(pAAInfoCY->Mask.cOut)));
DBGP_IF(DBGP_MASK | DBGP_AAHEADER, DBGP("aabbp.ptlMask x=%ld - %ld, cb=%ld, MaskBitOff=%02lx [%ld]" ARGDW(aabbp.ptlMask.x) ARGDW(MaskEnd.x) ARGDW(pAAHdr->cbMaskSrc) ARGDW(pAAHdr->MaskBitOff) ARGDW(cbFreeBuf)));
//
// 0x01 in the source means use modified pixel, for the
// reason of or in the mask, we will use 0=0xFF, 1=0x00
// in the mask
//
if (pBBP->Flags & BBPF_INVERT_SRC_MASK) {
aabbp.AAHFlags |= AAHF_INVERT_MASK; }
pAAHdr->AAMaskCXFunc = aabbp.AAMaskCXFunc; pAAHdr->AAMaskCYFunc = aabbp.AAMaskCYFunc;
SETDBGVAR(cCXMask, 0);
DBGP_IF(DBGP_AAHEADER, DBGP("--- SrcMask=(%5ld, %5ld)->(%5ld, %5ld) ---" ARGDW(pBBP->ptlSrcMask.x) ARGDW(pBBP->ptlSrcMask.y) ARGDW(aabbp.ptlMask.x) ARGDW(aabbp.ptlMask.y))); }
cyNext = cxSize = GET_PHTSI_CXSIZE(pSrcSI);
pAAHdr->cyABNext = pAAHdr->SrcSurfInfo.cyNext = cyNext; pAAHdr->SrcSurfInfo.pbOrg = pAAHdr->SrcSurfInfo.pb = pSrcSI->pPlane + (aabbp.rclSrc.top * cxSize) + ComputeByteOffset((UINT)pSrcSI->SurfaceFormat, aabbp.rclSrc.left, &(pAAHdr->SrcSurfInfo.BitOffset));
pAAHdr->GetAVCXFunc = aabbp.GetAVCXFunc; pAAHdr->GetAVCYFunc = aabbp.GetAVCYFunc;
DBGP_IF(DBGP_AAHEADER, DBGP("** pIn: (%p-%p), Beg=(%4ld, %4ld)=%p [%5ld], XOff=%4ld:%ld, cbCYExtra=%p-%p (%ld)" ARGPTR(pSrcSI->pPlane) ARGPTR(pSrcSI->pPlane + (cxSize * pSrcSI->Height)) ARGDW(aabbp.rclSrc.left) ARGDW(aabbp.rclSrc.top) ARGPTR(pAAHdr->SrcSurfInfo.pb) ARGDW(pAAHdr->SrcSurfInfo.cyNext) ARGDW(ComputeByteOffset((UINT)pSrcSI->SurfaceFormat, aabbp.rclSrc.left, &(pAAHdr->SrcSurfInfo.BitOffset))) ARGDW(pAAHdr->SrcSurfInfo.BitOffset) ARGPTR(pAAInfoCY->pbExtra) ARGPTR(pAAInfoCY->pbExtra + cbCYExtra) ARGDW(cbCYExtra)));
//
// We do up-side-down output by writing to the destination scan lines
// in reverse order.
//
cxSize = GET_PHTSI_CXSIZE(pDstSI); cyNext = (aabbp.rclDst.top > aabbp.rclDst.bottom) ? -cxSize : cxSize; pAAHdr->DstSurfInfo.cyNext = cyNext; pAAHdr->DstSurfInfo.pbOrg = pAAHdr->DstSurfInfo.pb = pDstSI->pPlane + (aabbp.rclDst.top * cxSize) + ComputeByteOffset((UINT)DstSurfFmt, aabbp.rclDst.left, &(pAAHdr->DstSurfInfo.BitOffset)); pAAHdr->pOutLast = pAAHdr->DstSurfInfo.pb + (pAAHdr->DstSurfInfo.cyNext * pAAInfoCY->cOut); pAAHdr->DstSurfInfo.cx = cOut; pAAHdr->DstSurfInfo.cyOrg = pAAHdr->DstSurfInfo.cy = pAAInfoCY->cOut; pAAHdr->Flags = aabbp.AAHFlags; // Re-Save
*pAACYFunc = AACYFunc;
#if DBG
pAAHdr->pOutBeg = pDstSI->pPlane + (aabbp.rclDst.top * cxSize) + ComputeByteOffset((UINT)DstSurfFmt, aabbp.rclDst.left, (LPBYTE)&iFree); pAAHdr->pOutEnd = pDstSI->pPlane + (aabbp.rclDst.bottom * cxSize) + ComputeByteOffset((UINT)DstSurfFmt, aabbp.rclDst.right, (LPBYTE)&iFree);
if (pAAHdr->pOutBeg > pAAHdr->pOutEnd) {
pAAHdr->pOutEnd = pDstSI->pPlane + (aabbp.rclDst.top * cxSize) + ComputeByteOffset((UINT)DstSurfFmt, aabbp.rclDst.right, (LPBYTE)&iFree); pAAHdr->pOutBeg = pDstSI->pPlane + (aabbp.rclDst.bottom * cxSize) + ComputeByteOffset((UINT)DstSurfFmt, aabbp.rclDst.left, (LPBYTE)&iFree);
}#endif
//
// Check it out if we need to fixup the input source bitmap
//
if (aabbp.AAHFlags & AAHF_TILE_SRC) {
//
// Increase the source cy and adjust pIn now
//
pAAHdr->SrcSurfInfo.Flags |= AASIF_TILE_SRC;
DBGP_IF(DBGP_AAHEADER, DBGP("Advance pIn by iSrcBeg=%ld x %ld=%ld" ARGDW(pAAInfoCY->iSrcBeg) ARGDW(pAAHdr->SrcSurfInfo.cyNext) ARGDW(pAAInfoCY->iSrcBeg * pAAHdr->SrcSurfInfo.cyNext)));
pAAHdr->SrcSurfInfo.pb += pAAInfoCY->iSrcBeg * pAAHdr->SrcSurfInfo.cyNext; pAAHdr->SrcSurfInfo.cy -= pAAInfoCY->iSrcBeg; }
pAAHdr->pbgrfAB = (PBGRF)pAAHdr->SrcSurfInfo.pb; pAAHdr->cybgrfAB = pAAHdr->SrcSurfInfo.cy;
DBGP_IF(DBGP_AAHEADER, DBGP("** pOut: (%p-%p), Beg=(%4ld, %4ld)=%p-%p (%p-%p) [%5ld], XOff=%4ld:%ld, cbCYExtra=%5ld" ARGPTR(pDstSI->pPlane) ARGPTR(pDstSI->pPlane + (cxSize * pDstSI->Height)) ARGDW(aabbp.rclDst.left) ARGDW(aabbp.rclDst.top) ARGPTR(pAAHdr->DstSurfInfo.pb) ARGPTR(pAAHdr->pOutLast) ARGPTR(pAAHdr->pOutBeg) ARGPTR(pAAHdr->pOutEnd) ARGDW(pAAHdr->DstSurfInfo.cyNext) ARGDW(ComputeByteOffset((UINT)DstSurfFmt, aabbp.rclDst.left, &(pAAHdr->DstSurfInfo.BitOffset))) ARGDW(pAAHdr->DstSurfInfo.BitOffset) ARGDW(cbCYExtra)));
DBGP_IF(DBGP_AAHEADER, DBGP("--- BrushOrg=(%5ld, %5ld)->(%5ld, %5ld) ---" ARGDW(pBBP->ptlBrushOrg.x) ARGDW(pBBP->ptlBrushOrg.y) ARGDW(pAAHdr->ptlBrushOrg.x) ARGDW(pAAHdr->ptlBrushOrg.y)));
DBGP_IF(DBGP_AAHEADER, DBGP("pAAHdr=%p - %p, (%ld bytes)" ARGPTR(pAAHdr) ARGPTR((LPBYTE)pAAHdr + pAAHdr->cbAlloc) ARGDW(pAAHdr->cbAlloc)));
DBGP_IF(DBGP_AAHEADER, DBGP("pAAInfoCX=%p-%p (%ld), pAAInfoCY=%p-%p (%ld)" ARGPTR(pAAInfoCX) ARGPTR((LPBYTE)pAAInfoCX + pAAInfoCX->cbAlloc) ARGDW(pAAInfoCX->cbAlloc) ARGPTR(pAAInfoCY) ARGPTR((LPBYTE)pAAInfoCY + pAAInfoCY->cbAlloc) ARGDW(pAAInfoCY->cbAlloc)));
DBGP_IF(DBGP_FUNC, DBGP("AACYFunc=%hs, AACXFunc=%hs" ARGPTR(GetAACYFuncName(*pAACYFunc)) ARGPTR(GetAACXFuncName(aabbp.AACXFunc))));
return(1); }
HTFreeMem(pAAInfoCX);
return(HTERR_INSUFFICIENT_MEMORY);}
|
