/* $Xorg: miscanfill.h,v 1.3 2000/08/17 19:53:39 cpqbld Exp $ */ /* Copyright 1987, 1998 The Open Group All Rights Reserved. The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE OPEN GROUP BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. Except as contained in this notice, the name of The Open Group shall not be used in advertising or otherwise to promote the sale, use or other dealings in this Software without prior written authorization from The Open Group. */ #ifndef SCANFILLINCLUDED #define SCANFILLINCLUDED /* * scanfill.h * * Written by Brian Kelleher; Jan 1985 * * This file contains a few macros to help track * the edge of a filled object. The object is assumed * to be filled in scanline order, and thus the * algorithm used is an extension of Bresenham's line * drawing algorithm which assumes that y is always the * major axis. * Since these pieces of code are the same for any filled shape, * it is more convenient to gather the library in one * place, but since these pieces of code are also in * the inner loops of output primitives, procedure call * overhead is out of the question. * See the author for a derivation if needed. */ /* * In scan converting polygons, we want to choose those pixels * which are inside the polygon. Thus, we add .5 to the starting * x coordinate for both left and right edges. Now we choose the * first pixel which is inside the pgon for the left edge and the * first pixel which is outside the pgon for the right edge. * Draw the left pixel, but not the right. * * How to add .5 to the starting x coordinate: * If the edge is moving to the right, then subtract dy from the * error term from the general form of the algorithm. * If the edge is moving to the left, then add dy to the error term. * * The reason for the difference between edges moving to the left * and edges moving to the right is simple: If an edge is moving * to the right, then we want the algorithm to flip immediately. * If it is moving to the left, then we don't want it to flip until * we traverse an entire pixel. */ #define BRESINITPGON(dy, x1, x2, xStart, d, m, m1, incr1, incr2) { \ int dx; /* local storage */ \ \ /* \ * if the edge is horizontal, then it is ignored \ * and assumed not to be processed. Otherwise, do this stuff. \ */ \ if ((dy) != 0) { \ xStart = (x1); \ dx = (x2) - xStart; \ if (dx < 0) { \ m = dx / (dy); \ m1 = m - 1; \ incr1 = -2 * dx + 2 * (dy) * m1; \ incr2 = -2 * dx + 2 * (dy) * m; \ d = 2 * m * (dy) - 2 * dx - 2 * (dy); \ } else { \ m = dx / (dy); \ m1 = m + 1; \ incr1 = 2 * dx - 2 * (dy) * m1; \ incr2 = 2 * dx - 2 * (dy) * m; \ d = -2 * m * (dy) + 2 * dx; \ } \ } \ } #define BRESINCRPGON(d, minval, m, m1, incr1, incr2) { \ if (m1 > 0) { \ if (d > 0) { \ minval += m1; \ d += incr1; \ } \ else { \ minval += m; \ d += incr2; \ } \ } else {\ if (d >= 0) { \ minval += m1; \ d += incr1; \ } \ else { \ minval += m; \ d += incr2; \ } \ } \ } /* * This structure contains all of the information needed * to run the bresenham algorithm. * The variables may be hardcoded into the declarations * instead of using this structure to make use of * register declarations. */ typedef struct { int minor; /* minor axis */ int d; /* decision variable */ int m, m1; /* slope and slope+1 */ int incr1, incr2; /* error increments */ } BRESINFO; #define BRESINITPGONSTRUCT(dmaj, min1, min2, bres) \ BRESINITPGON(dmaj, min1, min2, bres.minor, bres.d, \ bres.m, bres.m1, bres.incr1, bres.incr2) #define BRESINCRPGONSTRUCT(bres) \ BRESINCRPGON(bres.d, bres.minor, bres.m, bres.m1, bres.incr1, bres.incr2) #endif