diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-07 09:06:44 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-07 09:06:44 +0000 |
commit | ed5640d8b587fbcfed7dd7967f3de04b37a76f26 (patch) | |
tree | 7a5f7c6c9d02226d7471cb3cc8fbbf631b415303 /sdext/source/pdfimport/xpdfwrapper/pnghelper.cxx | |
parent | Initial commit. (diff) | |
download | libreoffice-upstream/4%7.4.7.tar.xz libreoffice-upstream/4%7.4.7.zip |
Adding upstream version 4:7.4.7.upstream/4%7.4.7upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'sdext/source/pdfimport/xpdfwrapper/pnghelper.cxx')
-rw-r--r-- | sdext/source/pdfimport/xpdfwrapper/pnghelper.cxx | 404 |
1 files changed, 404 insertions, 0 deletions
diff --git a/sdext/source/pdfimport/xpdfwrapper/pnghelper.cxx b/sdext/source/pdfimport/xpdfwrapper/pnghelper.cxx new file mode 100644 index 000000000..798555fe2 --- /dev/null +++ b/sdext/source/pdfimport/xpdfwrapper/pnghelper.cxx @@ -0,0 +1,404 @@ +/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */ +/* + * This file is part of the LibreOffice project. + * + * This Source Code Form is subject to the terms of the Mozilla Public + * License, v. 2.0. If a copy of the MPL was not distributed with this + * file, You can obtain one at http://mozilla.org/MPL/2.0/. + * + * This file incorporates work covered by the following license notice: + * + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed + * with this work for additional information regarding copyright + * ownership. The ASF licenses this file to you under the Apache + * License, Version 2.0 (the "License"); you may not use this file + * except in compliance with the License. You may obtain a copy of + * the License at http://www.apache.org/licenses/LICENSE-2.0 . + */ + +#include "pnghelper.hxx" +#include <sal/macros.h> + +#include <zlib.h> + +using namespace pdfi; + +// checksum helpers, courtesy of libpng.org + +/* Table of CRCs of all 8-bit messages. */ +sal_uInt32 PngHelper::crc_table[256]; + +/* Flag: has the table been computed? Initially false. */ +bool PngHelper::bCRCTableInit = true; + +/* Make the table for a fast CRC. */ +void PngHelper::initCRCTable() +{ + for (sal_uInt32 n = 0; n < 256; n++) + { + sal_uInt32 c = n; + for (int k = 0; k < 8; k++) + { + if (c & 1) + c = 0xedb88320L ^ (c >> 1); + else + c = c >> 1; + } + crc_table[n] = c; + } + bCRCTableInit = false; +} + +/* Update a running CRC with the bytes buf[0..len-1]--the CRC + should be initialized to all 1's, and the transmitted value + is the 1's complement of the final running CRC (see the + crc() routine below)). */ + +void PngHelper::updateCRC( sal_uInt32& io_rCRC, const sal_uInt8* i_pBuf, size_t i_nLen ) +{ + if( bCRCTableInit ) + initCRCTable(); + + sal_uInt32 nCRC = io_rCRC; + for( size_t n = 0; n < i_nLen; n++ ) + nCRC = crc_table[(nCRC ^ i_pBuf[n]) & 0xff] ^ (nCRC >> 8); + io_rCRC = nCRC; +} + +sal_uInt32 PngHelper::getCRC( const sal_uInt8* i_pBuf, size_t i_nLen ) +{ + sal_uInt32 nCRC = 0xffffffff; + updateCRC( nCRC, i_pBuf, i_nLen ); + return nCRC ^ 0xffffffff; +} + +sal_uInt32 PngHelper::deflateBuffer( const Output_t* i_pBuf, size_t i_nLen, OutputBuffer& o_rOut ) +{ + size_t nOrigSize = o_rOut.size(); + + // prepare z stream + z_stream aStream; + aStream.zalloc = Z_NULL; + aStream.zfree = Z_NULL; + aStream.opaque = Z_NULL; + aStream.total_out = aStream.total_in = 0; + if (Z_OK != deflateInit(&aStream, Z_BEST_COMPRESSION)) + return 0; + aStream.avail_in = uInt(i_nLen); + aStream.next_in = i_pBuf; + + sal_uInt8 aOutBuf[ 32768 ]; + do + { + aStream.avail_out = sizeof( aOutBuf ); + aStream.next_out = aOutBuf; + + if( deflate( &aStream, Z_FINISH ) == Z_STREAM_ERROR ) + { + deflateEnd( &aStream ); + // scrap the data of this broken stream + o_rOut.resize( nOrigSize ); + return 0; + } + + // append compressed bytes + sal_uInt32 nCompressedBytes = sizeof( aOutBuf ) - aStream.avail_out; + if( nCompressedBytes ) + o_rOut.insert( o_rOut.end(), aOutBuf, aOutBuf+nCompressedBytes ); + + } while( aStream.avail_out == 0 ); + + // cleanup + deflateEnd( &aStream ); + + return sal_uInt32( o_rOut.size() - nOrigSize ); +} + +void PngHelper::appendFileHeader( OutputBuffer& o_rOutputBuf ) +{ + static const unsigned char aHeader[] = { 0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a }; + + o_rOutputBuf.insert( o_rOutputBuf.end(), aHeader, aHeader + SAL_N_ELEMENTS(aHeader) ); +} + +size_t PngHelper::startChunk( const char* pChunkName, OutputBuffer& o_rOutputBuf ) +{ + size_t nIndex = sal_uInt32( o_rOutputBuf.size() ); + o_rOutputBuf.insert( o_rOutputBuf.end(), 4, Output_t(0) ); + o_rOutputBuf.push_back( pChunkName[0] ); + o_rOutputBuf.push_back( pChunkName[1] ); + o_rOutputBuf.push_back( pChunkName[2] ); + o_rOutputBuf.push_back( pChunkName[3] ); + return nIndex; +} + +void PngHelper::set( sal_uInt32 i_nValue, OutputBuffer& o_rOutputBuf, size_t i_nIndex ) +{ + o_rOutputBuf[ i_nIndex ] = (i_nValue & 0xff000000) >> 24; + o_rOutputBuf[ i_nIndex+1 ] = (i_nValue & 0x00ff0000) >> 16; + o_rOutputBuf[ i_nIndex+2 ] = (i_nValue & 0x0000ff00) >> 8; + o_rOutputBuf[ i_nIndex+3 ] = (i_nValue & 0x000000ff); +} + +void PngHelper::endChunk( size_t nStart, OutputBuffer& o_rOutputBuf ) +{ + if( nStart+8 > o_rOutputBuf.size() ) + return; // something broken is going on + + // update chunk length + size_t nLen = o_rOutputBuf.size() - nStart; + sal_uInt32 nDataLen = sal_uInt32(nLen)-8; + set( nDataLen, o_rOutputBuf, nStart ); + + // append chunk crc + sal_uInt32 nChunkCRC = getCRC( reinterpret_cast<sal_uInt8*>(&o_rOutputBuf[nStart+4]), nLen-4 ); + append( nChunkCRC, o_rOutputBuf ); +} + +void PngHelper::appendIHDR( OutputBuffer& o_rOutputBuf, int width, int height, int depth, int colortype ) +{ + size_t nStart = startChunk( "IHDR", o_rOutputBuf ); + append( width, o_rOutputBuf ); + append( height, o_rOutputBuf ); + o_rOutputBuf.push_back( Output_t(depth) ); + o_rOutputBuf.push_back( Output_t(colortype) ); + o_rOutputBuf.push_back( 0 ); // compression method deflate + o_rOutputBuf.push_back( 0 ); // filtering method 0 (default) + o_rOutputBuf.push_back( 0 ); // no interlacing + endChunk( nStart, o_rOutputBuf ); +} + +void PngHelper::appendIEND( OutputBuffer& o_rOutputBuf ) +{ + size_t nStart = startChunk( "IEND", o_rOutputBuf ); + endChunk( nStart, o_rOutputBuf ); +} + +void PngHelper::createPng( OutputBuffer& o_rOutputBuf, + Stream* str, + int width, + int height, + GfxRGB const & zeroColor, + GfxRGB const & oneColor, + bool bIsMask + ) +{ + appendFileHeader( o_rOutputBuf ); + appendIHDR( o_rOutputBuf, width, height, 1, 3 ); + + // write palette + size_t nIdx = startChunk( "PLTE", o_rOutputBuf ); + // write colors 0 and 1 + o_rOutputBuf.push_back(colToByte(zeroColor.r)); + o_rOutputBuf.push_back(colToByte(zeroColor.g)); + o_rOutputBuf.push_back(colToByte(zeroColor.b)); + o_rOutputBuf.push_back(colToByte(oneColor.r)); + o_rOutputBuf.push_back(colToByte(oneColor.g)); + o_rOutputBuf.push_back(colToByte(oneColor.b)); + // end PLTE chunk + endChunk( nIdx, o_rOutputBuf ); + + if( bIsMask ) + { + // write tRNS chunk + nIdx = startChunk( "tRNS", o_rOutputBuf ); + o_rOutputBuf.push_back( 0xff ); + o_rOutputBuf.push_back( 0 ); + // end tRNS chunk + endChunk( nIdx, o_rOutputBuf ); + } + + // create scan line data buffer + OutputBuffer aScanlines; + int nLineSize = (width + 7)/8; + aScanlines.reserve( nLineSize * height + height ); + + str->reset(); + for( int y = 0; y < height; y++ ) + { + // determine filter type (none) for this scanline + aScanlines.push_back( 0 ); + for( int x = 0; x < nLineSize; x++ ) + aScanlines.push_back( str->getChar() ); + } + + // begin IDAT chunk for scanline data + nIdx = startChunk( "IDAT", o_rOutputBuf ); + // compress scanlines + deflateBuffer( aScanlines.data(), aScanlines.size(), o_rOutputBuf ); + // end IDAT chunk + endChunk( nIdx, o_rOutputBuf ); + + // output IEND + appendIEND( o_rOutputBuf ); +} + +void PngHelper::createPng( OutputBuffer& o_rOutputBuf, + Stream* str, + int width, int height, GfxImageColorMap* colorMap, + Stream* maskStr, + int maskWidth, int maskHeight, GfxImageColorMap* maskColorMap ) +{ + appendFileHeader( o_rOutputBuf ); + appendIHDR( o_rOutputBuf, width, height, 8, 6 ); // RGBA image + + // initialize stream + unsigned char *p, *pm; + GfxRGB rgb; + GfxGray alpha; + std::unique_ptr<ImageStream> imgStr( + new ImageStream(str, + width, + colorMap->getNumPixelComps(), + colorMap->getBits())); + imgStr->reset(); + + // create scan line data buffer + OutputBuffer aScanlines; + aScanlines.reserve( width*height*4 + height ); + + for( int y=0; y<height; ++y) + { + aScanlines.push_back( 0 ); + p = imgStr->getLine(); + for( int x=0; x<width; ++x) + { + colorMap->getRGB(p, &rgb); + aScanlines.push_back(colToByte(rgb.r)); + aScanlines.push_back(colToByte(rgb.g)); + aScanlines.push_back(colToByte(rgb.b)); + aScanlines.push_back( 0xff ); + + p +=colorMap->getNumPixelComps(); + } + } + + + // now fill in the mask data + + // CAUTION: originally this was done in one single loop + // it caused merry chaos; the reason is that maskStr and str are + // not independent streams, it happens that reading one advances + // the other, too. Hence the two passes are imperative ! + + // initialize mask stream + std::unique_ptr<ImageStream> imgStrMask( + new ImageStream(maskStr, + maskWidth, + maskColorMap->getNumPixelComps(), + maskColorMap->getBits())); + + imgStrMask->reset(); + for( int y = 0; y < maskHeight; ++y ) + { + pm = imgStrMask->getLine(); + for( int x = 0; x < maskWidth; ++x ) + { + maskColorMap->getGray(pm,&alpha); + pm += maskColorMap->getNumPixelComps(); + int nIndex = (y*height/maskHeight) * (width*4+1) + // mapped line + (x*width/maskWidth)*4 + 1 + 3 // mapped column + ; + aScanlines[ nIndex ] = colToByte(alpha); + } + } + + imgStr.reset(); + imgStrMask.reset(); + + // begind IDAT chunk for scanline data + size_t nIdx = startChunk( "IDAT", o_rOutputBuf ); + // compress scanlines + deflateBuffer( aScanlines.data(), aScanlines.size(), o_rOutputBuf ); + // end IDAT chunk + endChunk( nIdx, o_rOutputBuf ); + // output IEND + appendIEND( o_rOutputBuf ); +} + +// one bit mask; 0 bits opaque +void PngHelper::createPng( OutputBuffer& o_rOutputBuf, + Stream* str, + int width, int height, GfxImageColorMap* colorMap, + Stream* maskStr, + int maskWidth, int maskHeight, + bool maskInvert + ) +{ + appendFileHeader( o_rOutputBuf ); + appendIHDR( o_rOutputBuf, width, height, 8, 6 ); // RGBA image + + // initialize stream + unsigned char *p; + GfxRGB rgb; + std::unique_ptr<ImageStream> imgStr( + new ImageStream(str, + width, + colorMap->getNumPixelComps(), + colorMap->getBits())); + imgStr->reset(); + + // create scan line data buffer + OutputBuffer aScanlines; + aScanlines.reserve( width*height*4 + height ); + + for( int y=0; y<height; ++y) + { + aScanlines.push_back( 0 ); + p = imgStr->getLine(); + for( int x=0; x<width; ++x) + { + colorMap->getRGB(p, &rgb); + aScanlines.push_back(colToByte(rgb.r)); + aScanlines.push_back(colToByte(rgb.g)); + aScanlines.push_back(colToByte(rgb.b)); + aScanlines.push_back( 0xff ); + + p +=colorMap->getNumPixelComps(); + } + } + + + // now fill in the mask data + + // CAUTION: originally this was done in one single loop + // it caused merry chaos; the reason is that maskStr and str are + // not independent streams, it happens that reading one advances + // the other, too. Hence the two passes are imperative ! + + // initialize mask stream + std::unique_ptr<ImageStream> imgStrMask( + new ImageStream(maskStr, maskWidth, 1, 1)); + + imgStrMask->reset(); + for( int y = 0; y < maskHeight; ++y ) + { + for( int x = 0; x < maskWidth; ++x ) + { + unsigned char aPixel = 0; + imgStrMask->getPixel( &aPixel ); + int nIndex = (y*height/maskHeight) * (width*4+1) + // mapped line + (x*width/maskWidth)*4 + 1 + 3 // mapped column + ; + if( maskInvert ) + aScanlines[ nIndex ] = aPixel ? 0xff : 0x00; + else + aScanlines[ nIndex ] = aPixel ? 0x00 : 0xff; + } + } + + imgStr.reset(); + imgStrMask.reset(); + + // begind IDAT chunk for scanline data + size_t nIdx = startChunk( "IDAT", o_rOutputBuf ); + // compress scanlines + deflateBuffer( aScanlines.data(), aScanlines.size(), o_rOutputBuf ); + // end IDAT chunk + endChunk( nIdx, o_rOutputBuf ); + // output IEND + appendIEND( o_rOutputBuf ); +} + +/* vim:set shiftwidth=4 softtabstop=4 expandtab: */ |