From cca66b9ec4e494c1d919bff0f71a820d8afab1fa Mon Sep 17 00:00:00 2001 From: Daniel Baumann Date: Sun, 7 Apr 2024 20:24:48 +0200 Subject: Adding upstream version 1.2.2. Signed-off-by: Daniel Baumann --- src/livarot/BitLigne.cpp | 180 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 180 insertions(+) create mode 100644 src/livarot/BitLigne.cpp (limited to 'src/livarot/BitLigne.cpp') diff --git a/src/livarot/BitLigne.cpp b/src/livarot/BitLigne.cpp new file mode 100644 index 0000000..2e44359 --- /dev/null +++ b/src/livarot/BitLigne.cpp @@ -0,0 +1,180 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/** @file + * TODO: insert short description here + *//* + * Authors: + * see git history + * Fred + * + * Copyright (C) 2018 Authors + * Released under GNU GPL v2+, read the file 'COPYING' for more information. + */ + +#include "BitLigne.h" + +#include +#include +#include +#include +#include +#include +#include + +BitLigne::BitLigne(int ist,int ien,float iScale) +{ + scale=iScale; + invScale=1/iScale; + st=ist; + en=ien; + if ( en <= st ) en=st+1; + stBit=(int)floor(((float)st)*invScale); // round to pixel boundaries in the canvas + enBit=(int)ceil(((float)en)*invScale); + int nbBit=enBit-stBit; + if ( nbBit&31 ) { + nbInt=nbBit/32+1; + } else { + nbInt=nbBit/32; + } + nbInt+=1; + fullB=(uint32_t*)g_malloc(nbInt*sizeof(uint32_t)); + partB=(uint32_t*)g_malloc(nbInt*sizeof(uint32_t)); + + curMin=en; + curMax=st; +} +BitLigne::~BitLigne() +{ + g_free(fullB); + g_free(partB); +} + +void BitLigne::Reset() +{ + curMin=en; + curMax=st+1; + memset(fullB,0,nbInt*sizeof(uint32_t)); + memset(partB,0,nbInt*sizeof(uint32_t)); +} +int BitLigne::AddBord(float spos,float epos,bool full) +{ + if ( spos >= epos ) return 0; + + // separation of full and not entirely full bits is a bit useless + // the goal is to obtain a set of bits that are "on the edges" of the polygon, so that their coverage + // will be 1/2 on the average. in practice it's useless for anything but the even-odd fill rule + int ffBit,lfBit; // first and last bit of the portion of the line that is entirely covered + ffBit=(int)(ceil(invScale*spos)); + lfBit=(int)(floor(invScale*epos)); + int fpBit,lpBit; // first and last bit of the portion of the line that is not entirely but partially covered + fpBit=(int)(floor(invScale*spos)); + lpBit=(int)(ceil(invScale*epos)); + + // update curMin and curMax to reflect the start and end pixel that need to be updated on the canvas + if ( floor(spos) < curMin ) curMin=(int)floor(spos); + if ( ceil(epos) > curMax ) curMax=(int)ceil(epos); + + // clamp to the line + if ( ffBit < stBit ) ffBit=stBit; + if ( ffBit > enBit ) ffBit=enBit; + if ( lfBit < stBit ) lfBit=stBit; + if ( lfBit > enBit ) lfBit=enBit; + if ( fpBit < stBit ) fpBit=stBit; + if ( fpBit > enBit ) fpBit=enBit; + if ( lpBit < stBit ) lpBit=stBit; + if ( lpBit > enBit ) lpBit=enBit; + + // offset to get actual bit position in the array + ffBit-=stBit; + lfBit-=stBit; + fpBit-=stBit; + lpBit-=stBit; + + // get the end and start indices of the elements of fullB and partB that will receives coverage + int ffPos=ffBit>>5; + int lfPos=lfBit>>5; + int fpPos=fpBit>>5; + int lpPos=lpBit>>5; + // get bit numbers in the last and first changed elements of the fullB and partB arrays + int ffRem=ffBit&31; + int lfRem=lfBit&31; + int fpRem=fpBit&31; + int lpRem=lpBit&31; + // add the coverage + // note that the "full" bits are always a subset of the "not empty" bits, ie of the partial bits + // the function is a bit lame: since there is at most one bit that is partial but not full, or no full bit, + // it does 2 times the optimal amount of work when the coverage is full. but i'm too lazy to change that... + if ( fpPos == lpPos ) { // only one element of the arrays is modified + // compute the vector of changed bits in the element + uint32_t add=0xFFFFFFFF; + if ( lpRem < 32 ) {add>>=32-lpRem;add<<=32-lpRem; } + if ( lpRem <= 0 ) add=0; + if ( fpRem > 0) {add<<=fpRem;add>>=fpRem;} + // and put it in the line + fullB[fpPos]&=~(add); // partial is exclusive from full, so partial bits are removed from fullB + partB[fpPos]|=add; // and added to partB + if ( full ) { // if the coverage is full, add the vector of full bits + if ( ffBit <= lfBit ) { + add=0xFFFFFFFF; + if ( lfRem < 32 ) {add>>=32-lfRem;add<<=32-lfRem;} + if ( lfRem <= 0 ) add=0; + if ( ffRem > 0 ) {add<<=ffRem;add>>=ffRem;} + fullB[ffPos]|=add; + partB[ffPos]&=~(add); + } + } + } else { + // first and last elements are differents, so add what appropriate to each + uint32_t add=0xFFFFFFFF; + if ( fpRem > 0 ) {add<<=fpRem;add>>=fpRem;} + fullB[fpPos]&=~(add); + partB[fpPos]|=add; + + add=0xFFFFFFFF; + if ( lpRem < 32 ) {add>>=32-lpRem;add<<=32-lpRem;} + if ( lpRem <= 0 ) add=0; + fullB[lpPos]&=~(add); + partB[lpPos]|=add; + + // and fill what's in between with partial bits + if ( lpPos > fpPos+1 ) memset(fullB+(fpPos+1),0x00,(lpPos-fpPos-1)*sizeof(uint32_t)); + if ( lpPos > fpPos+1 ) memset(partB+(fpPos+1),0xFF,(lpPos-fpPos-1)*sizeof(uint32_t)); + + if ( full ) { // is the coverage is full, do your magic + if ( ffBit <= lfBit ) { + if ( ffPos == lfPos ) { + add=0xFFFFFFFF; + if ( lfRem < 32 ) {add>>=32-lfRem;add<<=32-lfRem;} + if ( lfRem <= 0 ) add=0; + if ( ffRem > 0 ) {add<<=ffRem;add>>=ffRem;} + fullB[ffPos]|=add; + partB[ffPos]&=~(add); + } else { + add=0xFFFFFFFF; + if ( ffRem > 0 ) {add<<=ffRem;add>>=ffRem;} + fullB[ffPos]|=add; + partB[ffPos]&=~add; + + add=0xFFFFFFFF; + if ( lfRem < 32 ) {add>>=32-lfRem;add<<=32-lfRem;} + if ( lfRem <= 0 ) add=0; + fullB[lfPos]|=add; + partB[lfPos]&=~add; + + if ( lfPos > ffPos+1 ) memset(fullB+(ffPos+1),0xFF,(lfPos-ffPos-1)*sizeof(uint32_t)); + if ( lfPos > ffPos+1 ) memset(partB+(ffPos+1),0x00,(lfPos-ffPos-1)*sizeof(uint32_t)); + } + } + } + } + return 0; +} + + +void BitLigne::Affiche() +{ + for (int i=0;i