| 2 |
- |
1 |
/* Copyright (C) 1998, 1999 Aladdin Enterprises. All rights reserved.
|
|
|
2 |
|
|
|
3 |
This software is provided AS-IS with no warranty, either express or
|
|
|
4 |
implied.
|
|
|
5 |
|
|
|
6 |
This software is distributed under license and may not be copied,
|
|
|
7 |
modified or distributed except as expressly authorized under the terms
|
|
|
8 |
of the license contained in the file LICENSE in this distribution.
|
|
|
9 |
|
|
|
10 |
For more information about licensing, please refer to
|
|
|
11 |
http://www.ghostscript.com/licensing/. For information on
|
|
|
12 |
commercial licensing, go to http://www.artifex.com/licensing/ or
|
|
|
13 |
contact Artifex Software, Inc., 101 Lucas Valley Road #110,
|
|
|
14 |
San Rafael, CA 94903, U.S.A., +1(415)492-9861.
|
|
|
15 |
*/
|
|
|
16 |
|
|
|
17 |
/* $Id: gxshade6.c,v 1.100 2005/05/25 15:57:58 igor Exp $ */
|
|
|
18 |
/* Rendering for Coons and tensor patch shadings */
|
|
|
19 |
/*
|
|
|
20 |
A contiguous non-overlapping decomposition
|
|
|
21 |
of a tensor shading into linear color triangles.
|
|
|
22 |
*/
|
|
|
23 |
#include "memory_.h"
|
|
|
24 |
#include "gx.h"
|
|
|
25 |
#include "gserrors.h"
|
|
|
26 |
#include "gsmatrix.h" /* for gscoord.h */
|
|
|
27 |
#include "gscoord.h"
|
|
|
28 |
#include "gxcspace.h"
|
|
|
29 |
#include "gxdcolor.h"
|
|
|
30 |
#include "gxistate.h"
|
|
|
31 |
#include "gxshade.h"
|
|
|
32 |
#include "gxshade4.h"
|
|
|
33 |
#include "gxdevcli.h"
|
|
|
34 |
#include "gxarith.h"
|
|
|
35 |
#include "gzpath.h"
|
|
|
36 |
#include "stdint_.h"
|
|
|
37 |
#include "math_.h"
|
|
|
38 |
#include "vdtrace.h"
|
|
|
39 |
#include <assert.h>
|
|
|
40 |
|
|
|
41 |
#define VD_TRACE_TENSOR_PATCH 1
|
|
|
42 |
|
|
|
43 |
#if NOFILL_TEST
|
|
|
44 |
static bool dbg_nofill = false;
|
|
|
45 |
#endif
|
|
|
46 |
#if SKIP_TEST
|
|
|
47 |
static int dbg_patch_cnt = 0;
|
|
|
48 |
static int dbg_quad_cnt = 0;
|
|
|
49 |
static int dbg_triangle_cnt = 0;
|
|
|
50 |
static int dbg_wedge_triangle_cnt = 0;
|
|
|
51 |
#endif
|
|
|
52 |
|
|
|
53 |
static int min_linear_grades = 255; /* The minimal number of device color grades,
|
|
|
54 |
required to apply linear color device functions. */
|
|
|
55 |
|
|
|
56 |
/* ================ Utilities ================ */
|
|
|
57 |
|
|
|
58 |
/* Get colors for patch vertices. */
|
|
|
59 |
private int
|
|
|
60 |
shade_next_colors(shade_coord_stream_t * cs, patch_curve_t * curves,
|
|
|
61 |
int num_vertices)
|
|
|
62 |
{
|
|
|
63 |
int i, code = 0;
|
|
|
64 |
|
|
|
65 |
for (i = 0; i < num_vertices && code >= 0; ++i) {
|
|
|
66 |
curves[i].vertex.cc[1] = 0; /* safety. (patch_fill may assume 2 arguments) */
|
|
|
67 |
code = shade_next_color(cs, curves[i].vertex.cc);
|
|
|
68 |
}
|
|
|
69 |
return code;
|
|
|
70 |
}
|
|
|
71 |
|
|
|
72 |
/* Get a Bezier or tensor patch element. */
|
|
|
73 |
private int
|
|
|
74 |
shade_next_curve(shade_coord_stream_t * cs, patch_curve_t * curve)
|
|
|
75 |
{
|
|
|
76 |
int code = shade_next_coords(cs, &curve->vertex.p, 1);
|
|
|
77 |
|
|
|
78 |
if (code >= 0)
|
|
|
79 |
code = shade_next_coords(cs, curve->control,
|
|
|
80 |
countof(curve->control));
|
|
|
81 |
return code;
|
|
|
82 |
}
|
|
|
83 |
|
|
|
84 |
/*
|
|
|
85 |
* Parse the next patch out of the input stream. Return 1 if done,
|
|
|
86 |
* 0 if patch, <0 on error.
|
|
|
87 |
*/
|
|
|
88 |
private int
|
|
|
89 |
shade_next_patch(shade_coord_stream_t * cs, int BitsPerFlag,
|
|
|
90 |
patch_curve_t curve[4], gs_fixed_point interior[4] /* 0 for Coons patch */ )
|
|
|
91 |
{
|
|
|
92 |
int flag = shade_next_flag(cs, BitsPerFlag);
|
|
|
93 |
int num_colors, code;
|
|
|
94 |
|
|
|
95 |
if (flag < 0) {
|
|
|
96 |
if (!cs->is_eod(cs))
|
|
|
97 |
return_error(gs_error_rangecheck);
|
|
|
98 |
return 1; /* no more data */
|
|
|
99 |
}
|
|
|
100 |
switch (flag & 3) {
|
|
|
101 |
default:
|
|
|
102 |
return_error(gs_error_rangecheck); /* not possible */
|
|
|
103 |
case 0:
|
|
|
104 |
if ((code = shade_next_curve(cs, &curve[0])) < 0 ||
|
|
|
105 |
(code = shade_next_coords(cs, &curve[1].vertex.p, 1)) < 0
|
|
|
106 |
)
|
|
|
107 |
return code;
|
|
|
108 |
num_colors = 4;
|
|
|
109 |
goto vx;
|
|
|
110 |
case 1:
|
|
|
111 |
curve[0] = curve[1], curve[1].vertex = curve[2].vertex;
|
|
|
112 |
goto v3;
|
|
|
113 |
case 2:
|
|
|
114 |
curve[0] = curve[2], curve[1].vertex = curve[3].vertex;
|
|
|
115 |
goto v3;
|
|
|
116 |
case 3:
|
|
|
117 |
curve[1].vertex = curve[0].vertex, curve[0] = curve[3];
|
|
|
118 |
v3: num_colors = 2;
|
|
|
119 |
vx: if ((code = shade_next_coords(cs, curve[1].control, 2)) < 0 ||
|
|
|
120 |
(code = shade_next_curve(cs, &curve[2])) < 0 ||
|
|
|
121 |
(code = shade_next_curve(cs, &curve[3])) < 0 ||
|
|
|
122 |
(interior != 0 &&
|
|
|
123 |
(code = shade_next_coords(cs, interior, 4)) < 0) ||
|
|
|
124 |
(code = shade_next_colors(cs, &curve[4 - num_colors],
|
|
|
125 |
num_colors)) < 0
|
|
|
126 |
)
|
|
|
127 |
return code;
|
|
|
128 |
}
|
|
|
129 |
return 0;
|
|
|
130 |
}
|
|
|
131 |
|
|
|
132 |
int
|
|
|
133 |
init_patch_fill_state(patch_fill_state_t *pfs)
|
|
|
134 |
{
|
|
|
135 |
/* Warning : pfs->Function must be set in advance. */
|
|
|
136 |
const gs_color_space *pcs = pfs->direct_space;
|
|
|
137 |
gs_client_color fcc0, fcc1;
|
|
|
138 |
int i, code;
|
|
|
139 |
|
|
|
140 |
for (i = 0; i < pfs->num_components; i++) {
|
|
|
141 |
fcc0.paint.values[i] = -1000000;
|
|
|
142 |
fcc1.paint.values[i] = 1000000;
|
|
|
143 |
}
|
|
|
144 |
pcs->type->restrict_color(&fcc0, pcs);
|
|
|
145 |
pcs->type->restrict_color(&fcc1, pcs);
|
|
|
146 |
for (i = 0; i < pfs->num_components; i++)
|
|
|
147 |
pfs->color_domain.paint.values[i] = max(fcc1.paint.values[i] - fcc0.paint.values[i], 1);
|
|
|
148 |
pfs->vectorization = false; /* A stub for a while. Will use with pclwrite. */
|
|
|
149 |
pfs->maybe_self_intersecting = true;
|
|
|
150 |
pfs->monotonic_color = (pfs->Function == NULL);
|
|
|
151 |
pfs->linear_color = false;
|
|
|
152 |
pfs->inside = false;
|
|
|
153 |
pfs->n_color_args = 1;
|
|
|
154 |
pfs->fixed_flat = float2fixed(pfs->pis->flatness);
|
|
|
155 |
pfs->smoothness = pfs->pis->smoothness;
|
|
|
156 |
# if LAZY_WEDGES
|
|
|
157 |
code = wedge_vertex_list_elem_buffer_alloc(pfs);
|
|
|
158 |
if (code < 0)
|
|
|
159 |
return code;
|
|
|
160 |
# endif
|
|
|
161 |
pfs->max_small_coord = 1 << ((sizeof(int64_t) * 8 - 1/*sign*/) / 3);
|
|
|
162 |
return 0;
|
|
|
163 |
}
|
|
|
164 |
|
|
|
165 |
void
|
|
|
166 |
term_patch_fill_state(patch_fill_state_t *pfs)
|
|
|
167 |
{
|
|
|
168 |
# if LAZY_WEDGES
|
|
|
169 |
wedge_vertex_list_elem_buffer_free(pfs);
|
|
|
170 |
# endif
|
|
|
171 |
}
|
|
|
172 |
|
|
|
173 |
/* Resolve a patch color using the Function if necessary. */
|
|
|
174 |
inline private void
|
|
|
175 |
patch_resolve_color_inline(patch_color_t * ppcr, const patch_fill_state_t *pfs)
|
|
|
176 |
{
|
|
|
177 |
if (pfs->Function) {
|
|
|
178 |
const gs_color_space *pcs = pfs->direct_space;
|
|
|
179 |
|
|
|
180 |
gs_function_evaluate(pfs->Function, ppcr->t, ppcr->cc.paint.values);
|
|
|
181 |
pcs->type->restrict_color(&ppcr->cc, pcs);
|
|
|
182 |
}
|
|
|
183 |
}
|
|
|
184 |
|
|
|
185 |
void
|
|
|
186 |
patch_resolve_color(patch_color_t * ppcr, const patch_fill_state_t *pfs)
|
|
|
187 |
{
|
|
|
188 |
patch_resolve_color_inline(ppcr, pfs);
|
|
|
189 |
}
|
|
|
190 |
|
|
|
191 |
|
|
|
192 |
/*
|
|
|
193 |
* Calculate the interpolated color at a given point.
|
|
|
194 |
* Note that we must do this twice for bilinear interpolation.
|
|
|
195 |
* We use the name ppcr rather than ppc because an Apple compiler defines
|
|
|
196 |
* ppc when compiling on PowerPC systems (!).
|
|
|
197 |
*/
|
|
|
198 |
private void
|
|
|
199 |
patch_interpolate_color(patch_color_t * ppcr, const patch_color_t * ppc0,
|
|
|
200 |
const patch_color_t * ppc1, const patch_fill_state_t *pfs, floatp t)
|
|
|
201 |
{
|
|
|
202 |
/* The old code gives -IND on Intel. */
|
|
|
203 |
if (pfs->Function) {
|
|
|
204 |
ppcr->t[0] = ppc0->t[0] * (1 - t) + t * ppc1->t[0];
|
|
|
205 |
ppcr->t[1] = ppc0->t[1] * (1 - t) + t * ppc1->t[1];
|
|
|
206 |
patch_resolve_color_inline(ppcr, pfs);
|
|
|
207 |
} else {
|
|
|
208 |
int ci;
|
|
|
209 |
|
|
|
210 |
for (ci = pfs->num_components - 1; ci >= 0; --ci)
|
|
|
211 |
ppcr->cc.paint.values[ci] =
|
|
|
212 |
ppc0->cc.paint.values[ci] * (1 - t) + t * ppc1->cc.paint.values[ci];
|
|
|
213 |
}
|
|
|
214 |
}
|
|
|
215 |
|
|
|
216 |
/* ================ Specific shadings ================ */
|
|
|
217 |
|
|
|
218 |
/*
|
|
|
219 |
* The curves are stored in a clockwise or counter-clockwise order that maps
|
|
|
220 |
* to the patch definition in a non-intuitive way. The documentation
|
|
|
221 |
* (pp. 277-281 of the PostScript Language Reference Manual, Third Edition)
|
|
|
222 |
* says that the curves are in the order D1, C2, D2, C1.
|
|
|
223 |
*/
|
|
|
224 |
/* The starting points of the curves: */
|
|
|
225 |
#define D1START 0
|
|
|
226 |
#define C2START 1
|
|
|
227 |
#define D2START 3
|
|
|
228 |
#define C1START 0
|
|
|
229 |
/* The control points of the curves (x means reversed order): */
|
|
|
230 |
#define D1CTRL 0
|
|
|
231 |
#define C2CTRL 1
|
|
|
232 |
#define D2XCTRL 2
|
|
|
233 |
#define C1XCTRL 3
|
|
|
234 |
/* The end points of the curves: */
|
|
|
235 |
#define D1END 1
|
|
|
236 |
#define C2END 2
|
|
|
237 |
#define D2END 2
|
|
|
238 |
#define C1END 3
|
|
|
239 |
|
|
|
240 |
/* ---------------- Common code ---------------- */
|
|
|
241 |
|
|
|
242 |
/* Evaluate a curve at a given point. */
|
|
|
243 |
private void
|
|
|
244 |
curve_eval(gs_fixed_point * pt, const gs_fixed_point * p0,
|
|
|
245 |
const gs_fixed_point * p1, const gs_fixed_point * p2,
|
|
|
246 |
const gs_fixed_point * p3, floatp t)
|
|
|
247 |
{
|
|
|
248 |
fixed a, b, c, d;
|
|
|
249 |
fixed t01, t12;
|
|
|
250 |
|
|
|
251 |
d = p0->x;
|
|
|
252 |
curve_points_to_coefficients(d, p1->x, p2->x, p3->x,
|
|
|
253 |
a, b, c, t01, t12);
|
|
|
254 |
pt->x = (fixed) (((a * t + b) * t + c) * t + d);
|
|
|
255 |
d = p0->y;
|
|
|
256 |
curve_points_to_coefficients(d, p1->y, p2->y, p3->y,
|
|
|
257 |
a, b, c, t01, t12);
|
|
|
258 |
pt->y = (fixed) (((a * t + b) * t + c) * t + d);
|
|
|
259 |
if_debug3('2', "[2]t=%g => (%g,%g)\n", t, fixed2float(pt->x),
|
|
|
260 |
fixed2float(pt->y));
|
|
|
261 |
}
|
|
|
262 |
|
|
|
263 |
/* ---------------- Coons patch shading ---------------- */
|
|
|
264 |
|
|
|
265 |
/* Calculate the device-space coordinate corresponding to (u,v). */
|
|
|
266 |
private void
|
|
|
267 |
Cp_transform(gs_fixed_point * pt, const patch_curve_t curve[4],
|
|
|
268 |
const gs_fixed_point ignore_interior[4], floatp u, floatp v)
|
|
|
269 |
{
|
|
|
270 |
double co_u = 1.0 - u, co_v = 1.0 - v;
|
|
|
271 |
gs_fixed_point c1u, d1v, c2u, d2v;
|
|
|
272 |
|
|
|
273 |
curve_eval(&c1u, &curve[C1START].vertex.p,
|
|
|
274 |
&curve[C1XCTRL].control[1], &curve[C1XCTRL].control[0],
|
|
|
275 |
&curve[C1END].vertex.p, u);
|
|
|
276 |
curve_eval(&d1v, &curve[D1START].vertex.p,
|
|
|
277 |
&curve[D1CTRL].control[0], &curve[D1CTRL].control[1],
|
|
|
278 |
&curve[D1END].vertex.p, v);
|
|
|
279 |
curve_eval(&c2u, &curve[C2START].vertex.p,
|
|
|
280 |
&curve[C2CTRL].control[0], &curve[C2CTRL].control[1],
|
|
|
281 |
&curve[C2END].vertex.p, u);
|
|
|
282 |
curve_eval(&d2v, &curve[D2START].vertex.p,
|
|
|
283 |
&curve[D2XCTRL].control[1], &curve[D2XCTRL].control[0],
|
|
|
284 |
&curve[D2END].vertex.p, v);
|
|
|
285 |
#define COMPUTE_COORD(xy)\
|
|
|
286 |
pt->xy = (fixed)\
|
|
|
287 |
((co_v * c1u.xy + v * c2u.xy) + (co_u * d1v.xy + u * d2v.xy) -\
|
|
|
288 |
(co_v * (co_u * curve[C1START].vertex.p.xy +\
|
|
|
289 |
u * curve[C1END].vertex.p.xy) +\
|
|
|
290 |
v * (co_u * curve[C2START].vertex.p.xy +\
|
|
|
291 |
u * curve[C2END].vertex.p.xy)))
|
|
|
292 |
COMPUTE_COORD(x);
|
|
|
293 |
COMPUTE_COORD(y);
|
|
|
294 |
#undef COMPUTE_COORD
|
|
|
295 |
if_debug4('2', "[2](u=%g,v=%g) => (%g,%g)\n",
|
|
|
296 |
u, v, fixed2float(pt->x), fixed2float(pt->y));
|
|
|
297 |
}
|
|
|
298 |
|
|
|
299 |
int
|
|
|
300 |
gs_shading_Cp_fill_rectangle(const gs_shading_t * psh0, const gs_rect * rect,
|
|
|
301 |
const gs_fixed_rect * rect_clip,
|
|
|
302 |
gx_device * dev, gs_imager_state * pis)
|
|
|
303 |
{
|
|
|
304 |
const gs_shading_Cp_t * const psh = (const gs_shading_Cp_t *)psh0;
|
|
|
305 |
patch_fill_state_t state;
|
|
|
306 |
shade_coord_stream_t cs;
|
|
|
307 |
patch_curve_t curve[4];
|
|
|
308 |
int code;
|
|
|
309 |
|
|
|
310 |
code = mesh_init_fill_state((mesh_fill_state_t *) &state,
|
|
|
311 |
(const gs_shading_mesh_t *)psh0, rect_clip, dev, pis);
|
|
|
312 |
if (code < 0)
|
|
|
313 |
return code;
|
|
|
314 |
state.Function = psh->params.Function;
|
|
|
315 |
code = init_patch_fill_state(&state);
|
|
|
316 |
if(code < 0)
|
|
|
317 |
return code;
|
|
|
318 |
if (VD_TRACE_TENSOR_PATCH && vd_allowed('s')) {
|
|
|
319 |
vd_get_dc('s');
|
|
|
320 |
vd_set_shift(0, 0);
|
|
|
321 |
vd_set_scale(0.01);
|
|
|
322 |
vd_set_origin(0, 0);
|
|
|
323 |
/* vd_erase(RGB(192, 192, 192)); */
|
|
|
324 |
}
|
|
|
325 |
|
|
|
326 |
curve[0].straight = curve[1].straight = curve[2].straight = curve[3].straight = false;
|
|
|
327 |
shade_next_init(&cs, (const gs_shading_mesh_params_t *)&psh->params, pis);
|
|
|
328 |
while ((code = shade_next_patch(&cs, psh->params.BitsPerFlag,
|
|
|
329 |
curve, NULL)) == 0 &&
|
|
|
330 |
(code = patch_fill(&state, curve, NULL, Cp_transform)) >= 0
|
|
|
331 |
) {
|
|
|
332 |
DO_NOTHING;
|
|
|
333 |
}
|
|
|
334 |
if (VD_TRACE_TENSOR_PATCH && vd_allowed('s'))
|
|
|
335 |
vd_release_dc;
|
|
|
336 |
term_patch_fill_state(&state);
|
|
|
337 |
return min(code, 0);
|
|
|
338 |
}
|
|
|
339 |
|
|
|
340 |
/* ---------------- Tensor product patch shading ---------------- */
|
|
|
341 |
|
|
|
342 |
/* Calculate the device-space coordinate corresponding to (u,v). */
|
|
|
343 |
private void
|
|
|
344 |
Tpp_transform(gs_fixed_point * pt, const patch_curve_t curve[4],
|
|
|
345 |
const gs_fixed_point interior[4], floatp u, floatp v)
|
|
|
346 |
{
|
|
|
347 |
double Bu[4], Bv[4];
|
|
|
348 |
gs_fixed_point pts[4][4];
|
|
|
349 |
int i, j;
|
|
|
350 |
double x = 0, y = 0;
|
|
|
351 |
|
|
|
352 |
/* Compute the Bernstein polynomials of u and v. */
|
|
|
353 |
{
|
|
|
354 |
double u2 = u * u, co_u = 1.0 - u, co_u2 = co_u * co_u;
|
|
|
355 |
double v2 = v * v, co_v = 1.0 - v, co_v2 = co_v * co_v;
|
|
|
356 |
|
|
|
357 |
Bu[0] = co_u * co_u2, Bu[1] = 3 * u * co_u2,
|
|
|
358 |
Bu[2] = 3 * u2 * co_u, Bu[3] = u * u2;
|
|
|
359 |
Bv[0] = co_v * co_v2, Bv[1] = 3 * v * co_v2,
|
|
|
360 |
Bv[2] = 3 * v2 * co_v, Bv[3] = v * v2;
|
|
|
361 |
}
|
|
|
362 |
|
|
|
363 |
/* Arrange the points into an indexable order. */
|
|
|
364 |
pts[0][0] = curve[0].vertex.p;
|
|
|
365 |
pts[0][1] = curve[0].control[0];
|
|
|
366 |
pts[0][2] = curve[0].control[1];
|
|
|
367 |
pts[0][3] = curve[1].vertex.p;
|
|
|
368 |
pts[1][3] = curve[1].control[0];
|
|
|
369 |
pts[2][3] = curve[1].control[1];
|
|
|
370 |
pts[3][3] = curve[2].vertex.p;
|
|
|
371 |
pts[3][2] = curve[2].control[0];
|
|
|
372 |
pts[3][1] = curve[2].control[1];
|
|
|
373 |
pts[3][0] = curve[3].vertex.p;
|
|
|
374 |
pts[2][0] = curve[3].control[0];
|
|
|
375 |
pts[1][0] = curve[3].control[1];
|
|
|
376 |
pts[1][1] = interior[0];
|
|
|
377 |
pts[2][1] = interior[1];
|
|
|
378 |
pts[2][2] = interior[2];
|
|
|
379 |
pts[1][2] = interior[3];
|
|
|
380 |
|
|
|
381 |
/* Now compute the actual point. */
|
|
|
382 |
for (i = 0; i < 4; ++i)
|
|
|
383 |
for (j = 0; j < 4; ++j) {
|
|
|
384 |
double coeff = Bu[i] * Bv[j];
|
|
|
385 |
|
|
|
386 |
x += pts[i][j].x * coeff, y += pts[i][j].y * coeff;
|
|
|
387 |
}
|
|
|
388 |
pt->x = (fixed)x, pt->y = (fixed)y;
|
|
|
389 |
}
|
|
|
390 |
|
|
|
391 |
int
|
|
|
392 |
gs_shading_Tpp_fill_rectangle(const gs_shading_t * psh0, const gs_rect * rect,
|
|
|
393 |
const gs_fixed_rect * rect_clip,
|
|
|
394 |
gx_device * dev, gs_imager_state * pis)
|
|
|
395 |
{
|
|
|
396 |
const gs_shading_Tpp_t * const psh = (const gs_shading_Tpp_t *)psh0;
|
|
|
397 |
patch_fill_state_t state;
|
|
|
398 |
shade_coord_stream_t cs;
|
|
|
399 |
patch_curve_t curve[4];
|
|
|
400 |
gs_fixed_point interior[4];
|
|
|
401 |
int code;
|
|
|
402 |
|
|
|
403 |
code = mesh_init_fill_state((mesh_fill_state_t *) & state,
|
|
|
404 |
(const gs_shading_mesh_t *)psh0, rect_clip, dev, pis);
|
|
|
405 |
if (code < 0)
|
|
|
406 |
return code;
|
|
|
407 |
state.Function = psh->params.Function;
|
|
|
408 |
code = init_patch_fill_state(&state);
|
|
|
409 |
if(code < 0)
|
|
|
410 |
return code;
|
|
|
411 |
if (VD_TRACE_TENSOR_PATCH && vd_allowed('s')) {
|
|
|
412 |
vd_get_dc('s');
|
|
|
413 |
vd_set_shift(0, 0);
|
|
|
414 |
vd_set_scale(0.01);
|
|
|
415 |
vd_set_origin(0, 0);
|
|
|
416 |
/* vd_erase(RGB(192, 192, 192)); */
|
|
|
417 |
}
|
|
|
418 |
curve[0].straight = curve[1].straight = curve[2].straight = curve[3].straight = false;
|
|
|
419 |
shade_next_init(&cs, (const gs_shading_mesh_params_t *)&psh->params, pis);
|
|
|
420 |
while ((code = shade_next_patch(&cs, psh->params.BitsPerFlag,
|
|
|
421 |
curve, interior)) == 0) {
|
|
|
422 |
/*
|
|
|
423 |
* The order of points appears to be consistent with that for Coons
|
|
|
424 |
* patches, which is different from that documented in Red Book 3.
|
|
|
425 |
*/
|
|
|
426 |
gs_fixed_point swapped_interior[4];
|
|
|
427 |
|
|
|
428 |
swapped_interior[0] = interior[0];
|
|
|
429 |
swapped_interior[1] = interior[3];
|
|
|
430 |
swapped_interior[2] = interior[2];
|
|
|
431 |
swapped_interior[3] = interior[1];
|
|
|
432 |
code = patch_fill(&state, curve, swapped_interior, Tpp_transform);
|
|
|
433 |
if (code < 0)
|
|
|
434 |
break;
|
|
|
435 |
}
|
|
|
436 |
term_patch_fill_state(&state);
|
|
|
437 |
if (VD_TRACE_TENSOR_PATCH && vd_allowed('s'))
|
|
|
438 |
vd_release_dc;
|
|
|
439 |
return min(code, 0);
|
|
|
440 |
}
|
|
|
441 |
|
|
|
442 |
/*
|
|
|
443 |
This algorithm performs a decomposition of the shading area
|
|
|
444 |
into a set of constant color trapezoids, some of which
|
|
|
445 |
may use the transpozed coordinate system.
|
|
|
446 |
|
|
|
447 |
The target device assumes semi-open intrvals by X to be painted
|
|
|
448 |
(See PLRM3, 7.5. Scan conversion details), i.e.
|
|
|
449 |
it doesn't paint pixels which falls exactly to the right side.
|
|
|
450 |
Note that with raster devices the algorithm doesn't paint pixels,
|
|
|
451 |
whigh are partially covered by the shading area,
|
|
|
452 |
but which's centers are outside the area.
|
|
|
453 |
|
|
|
454 |
Pixels inside a monotonic part of the shading area are painted
|
|
|
455 |
at once, but some exceptions may happen :
|
|
|
456 |
|
|
|
457 |
- While flattening boundaries of a subpatch,
|
|
|
458 |
to keep the plane coverage contiguity we insert wedges
|
|
|
459 |
between neighbor subpatches, which use a different
|
|
|
460 |
flattening factor. With non-monotonic curves
|
|
|
461 |
those wedges may overlap or be self-overlapping, and a pixel
|
|
|
462 |
is painted so many times as many wedges cover it. Fortunately
|
|
|
463 |
the area of most wedges is zero or extremily small.
|
|
|
464 |
|
|
|
465 |
- Since quazi-horizontal wedges may have a non-constant color,
|
|
|
466 |
they can't decompose into constant color trapezoids with
|
|
|
467 |
keeping the coverage contiguity. To represent them we
|
|
|
468 |
apply the XY plane transposition. But with the transposition
|
|
|
469 |
a semiopen interval can met a non-transposed one,
|
|
|
470 |
so that some lines are not covered. Therefore we emulate
|
|
|
471 |
closed intervals with expanding the transposed trapesoids in
|
|
|
472 |
fixed_epsilon, and pixels at that boundary may be painted twice.
|
|
|
473 |
|
|
|
474 |
- A boundary of a monotonic area can't compute in XY
|
|
|
475 |
preciselly due to high order polynomial equations.
|
|
|
476 |
Therefore the subdivision near the monotonity boundary
|
|
|
477 |
may paint some pixels twice within same monotonic part.
|
|
|
478 |
|
|
|
479 |
Non-monotonic areas slow down due to a tinny subdivision required.
|
|
|
480 |
|
|
|
481 |
The target device may be either raster or vector.
|
|
|
482 |
Vector devices should preciselly pass trapezoids to the output.
|
|
|
483 |
Note that ends of sides of a trapesoid are not necessary
|
|
|
484 |
the trapezoid's vertices. Converting this thing into
|
|
|
485 |
an exact quadrangle may cause an arithmetic error,
|
|
|
486 |
and the rounding must be done so that the coverage
|
|
|
487 |
contiguity is not lost.
|
|
|
488 |
|
|
|
489 |
When a device passes a trapezoid to it's output,
|
|
|
490 |
a regular rounding would keep the coverage contiguity,
|
|
|
491 |
except for the transposed trapesoids.
|
|
|
492 |
If a transposed trapezoid is being transposed back,
|
|
|
493 |
it doesn't become a canonic trapezoid, and a further
|
|
|
494 |
decomposition is neccessary. But rounding errors here
|
|
|
495 |
would break the coverage contiguity at boundaries
|
|
|
496 |
of the tansposed part of the area.
|
|
|
497 |
|
|
|
498 |
Devices, which have no transposed trapezoids and represent
|
|
|
499 |
trapezoids only with 8 coordinates of vertices of the quadrangle
|
|
|
500 |
(pclwrite is an example) may apply the backward transposition,
|
|
|
501 |
and a clipping instead the further decomposition.
|
|
|
502 |
Note that many clip regions may appear for all wedges.
|
|
|
503 |
Note that in some cases the adjustment of the right side to be
|
|
|
504 |
withdrown before the backward transposition.
|
|
|
505 |
*/
|
|
|
506 |
/* We believe that a multiplication of 32-bit integers with a
|
|
|
507 |
64-bit result is performed by modern platforms performs
|
|
|
508 |
in hardware level. Therefore we widely use it here,
|
|
|
509 |
but we minimize the usage of a multiplication of longer integers.
|
|
|
510 |
|
|
|
511 |
Unfortunately we do need a multiplication of long integers
|
|
|
512 |
in intersection_of_small_bars, because solving the linear system
|
|
|
513 |
requires tripple multiples of 'fixed'. Therefore we retain
|
|
|
514 |
of it's usage in the algorithm of the main branch.
|
|
|
515 |
Configuration macro QUADRANGLES prevents it.
|
|
|
516 |
*/
|
|
|
517 |
|
|
|
518 |
typedef struct {
|
|
|
519 |
gs_fixed_point pole[4][4]; /* [v][u] */
|
|
|
520 |
patch_color_t c[2][2]; /* [v][u] */
|
|
|
521 |
} tensor_patch;
|
|
|
522 |
|
|
|
523 |
typedef struct {
|
|
|
524 |
const shading_vertex_t *p[2][2]; /* [v][u] */
|
|
|
525 |
wedge_vertex_list_t *l0001, *l0111, *l1110, *l1000;
|
|
|
526 |
} quadrangle_patch;
|
|
|
527 |
|
|
|
528 |
typedef enum {
|
|
|
529 |
interpatch_padding = 1, /* A Padding between patches for poorly designed documents. */
|
|
|
530 |
inpatch_wedge = 2 /* Wedges while a patch decomposition. */
|
|
|
531 |
} wedge_type_t;
|
|
|
532 |
|
|
|
533 |
int
|
|
|
534 |
wedge_vertex_list_elem_buffer_alloc(patch_fill_state_t *pfs)
|
|
|
535 |
{
|
|
|
536 |
const int max_level = LAZY_WEDGES_MAX_LEVEL;
|
|
|
537 |
gs_memory_t *memory = pfs->pis->memory;
|
|
|
538 |
|
|
|
539 |
pfs->wedge_vertex_list_elem_count_max = max_level * (1 << max_level);
|
|
|
540 |
pfs->wedge_vertex_list_elem_buffer = (wedge_vertex_list_elem_t *)gs_alloc_bytes(memory,
|
|
|
541 |
sizeof(wedge_vertex_list_elem_t) * pfs->wedge_vertex_list_elem_count_max,
|
|
|
542 |
"alloc_wedge_vertex_list_elem_buffer");
|
|
|
543 |
if (pfs->wedge_vertex_list_elem_buffer == NULL)
|
|
|
544 |
return_error(gs_error_VMerror);
|
|
|
545 |
pfs->free_wedge_vertex = NULL;
|
|
|
546 |
pfs->wedge_vertex_list_elem_count = 0;
|
|
|
547 |
return 0;
|
|
|
548 |
}
|
|
|
549 |
|
|
|
550 |
void
|
|
|
551 |
wedge_vertex_list_elem_buffer_free(patch_fill_state_t *pfs)
|
|
|
552 |
{
|
|
|
553 |
gs_memory_t *memory = pfs->pis->memory;
|
|
|
554 |
|
|
|
555 |
gs_free_object(memory, pfs->wedge_vertex_list_elem_buffer,
|
|
|
556 |
"wedge_vertex_list_elem_buffer_free");
|
|
|
557 |
pfs->wedge_vertex_list_elem_buffer = NULL;
|
|
|
558 |
pfs->free_wedge_vertex = NULL;
|
|
|
559 |
}
|
|
|
560 |
|
|
|
561 |
private inline wedge_vertex_list_elem_t *
|
|
|
562 |
wedge_vertex_list_elem_reserve(patch_fill_state_t *pfs)
|
|
|
563 |
{
|
|
|
564 |
wedge_vertex_list_elem_t *e = pfs->free_wedge_vertex;
|
|
|
565 |
|
|
|
566 |
if (e != NULL) {
|
|
|
567 |
pfs->free_wedge_vertex = e->next;
|
|
|
568 |
return e;
|
|
|
569 |
}
|
|
|
570 |
if (pfs->wedge_vertex_list_elem_count < pfs->wedge_vertex_list_elem_count_max)
|
|
|
571 |
return pfs->wedge_vertex_list_elem_buffer + pfs->wedge_vertex_list_elem_count++;
|
|
|
572 |
return NULL;
|
|
|
573 |
}
|
|
|
574 |
|
|
|
575 |
private inline void
|
|
|
576 |
wedge_vertex_list_elem_release(patch_fill_state_t *pfs, wedge_vertex_list_elem_t *e)
|
|
|
577 |
{
|
|
|
578 |
e->next = pfs->free_wedge_vertex;
|
|
|
579 |
pfs->free_wedge_vertex = e;
|
|
|
580 |
}
|
|
|
581 |
|
|
|
582 |
private inline void
|
|
|
583 |
release_triangle_wedge_vertex_list_elem(patch_fill_state_t *pfs,
|
|
|
584 |
wedge_vertex_list_elem_t *beg, wedge_vertex_list_elem_t *end)
|
|
|
585 |
{
|
|
|
586 |
wedge_vertex_list_elem_t *e = beg->next;
|
|
|
587 |
|
|
|
588 |
assert(beg->next->next == end);
|
|
|
589 |
beg->next = end;
|
|
|
590 |
end->prev = beg;
|
|
|
591 |
wedge_vertex_list_elem_release(pfs, e);
|
|
|
592 |
}
|
|
|
593 |
|
|
|
594 |
private inline void
|
|
|
595 |
release_wedge_vertex_list_interval(patch_fill_state_t *pfs,
|
|
|
596 |
wedge_vertex_list_elem_t *beg, wedge_vertex_list_elem_t *end)
|
|
|
597 |
{
|
|
|
598 |
wedge_vertex_list_elem_t *e = beg->next, *ee;
|
|
|
599 |
|
|
|
600 |
beg->next = end;
|
|
|
601 |
end->prev = beg;
|
|
|
602 |
for (; e != end; e = ee) {
|
|
|
603 |
ee = e->next;
|
|
|
604 |
wedge_vertex_list_elem_release(pfs, e);
|
|
|
605 |
}
|
|
|
606 |
}
|
|
|
607 |
|
|
|
608 |
private inline void
|
|
|
609 |
release_wedge_vertex_list(patch_fill_state_t *pfs, wedge_vertex_list_t *ll, int n)
|
|
|
610 |
{
|
|
|
611 |
int i;
|
|
|
612 |
|
|
|
613 |
for (i = 0; i < n; i++) {
|
|
|
614 |
wedge_vertex_list_t *l = ll + i;
|
|
|
615 |
|
|
|
616 |
if (l->beg != NULL) {
|
|
|
617 |
assert(l->end != NULL);
|
|
|
618 |
release_wedge_vertex_list_interval(pfs, l->beg, l->end);
|
|
|
619 |
wedge_vertex_list_elem_release(pfs, l->beg);
|
|
|
620 |
wedge_vertex_list_elem_release(pfs, l->end);
|
|
|
621 |
l->beg = l->end = NULL;
|
|
|
622 |
} else
|
|
|
623 |
assert(l->end == NULL);
|
|
|
624 |
}
|
|
|
625 |
}
|
|
|
626 |
|
|
|
627 |
private inline wedge_vertex_list_elem_t *
|
|
|
628 |
wedge_vertex_list_find(wedge_vertex_list_elem_t *beg, const wedge_vertex_list_elem_t *end,
|
|
|
629 |
int level)
|
|
|
630 |
{
|
|
|
631 |
wedge_vertex_list_elem_t *e = beg;
|
|
|
632 |
|
|
|
633 |
assert(beg != NULL && end != NULL);
|
|
|
634 |
for (; e != end; e = e->next)
|
|
|
635 |
if (e->level == level)
|
|
|
636 |
return e;
|
|
|
637 |
return NULL;
|
|
|
638 |
}
|
|
|
639 |
|
|
|
640 |
private inline void
|
|
|
641 |
init_wedge_vertex_list(wedge_vertex_list_t *l, int n)
|
|
|
642 |
{
|
|
|
643 |
memset(l, 0, sizeof(*l) * n);
|
|
|
644 |
}
|
|
|
645 |
|
|
|
646 |
private void
|
|
|
647 |
draw_patch(const tensor_patch *p, bool interior, ulong rgbcolor)
|
|
|
648 |
{
|
|
|
649 |
#ifdef DEBUG
|
|
|
650 |
#if 0 /* Disabled for a better view with a specific purpose.
|
|
|
651 |
Feel free to enable fo needed. */
|
|
|
652 |
int i, step = (interior ? 1 : 3);
|
|
|
653 |
|
|
|
654 |
for (i = 0; i < 4; i += step) {
|
|
|
655 |
vd_curve(p->pole[i][0].x, p->pole[i][0].y,
|
|
|
656 |
p->pole[i][1].x, p->pole[i][1].y,
|
|
|
657 |
p->pole[i][2].x, p->pole[i][2].y,
|
|
|
658 |
p->pole[i][3].x, p->pole[i][3].y,
|
|
|
659 |
0, rgbcolor);
|
|
|
660 |
vd_curve(p->pole[0][i].x, p->pole[0][i].y,
|
|
|
661 |
p->pole[1][i].x, p->pole[1][i].y,
|
|
|
662 |
p->pole[2][i].x, p->pole[2][i].y,
|
|
|
663 |
p->pole[3][i].x, p->pole[3][i].y,
|
|
|
664 |
0, rgbcolor);
|
|
|
665 |
}
|
|
|
666 |
#endif
|
|
|
667 |
#endif
|
|
|
668 |
}
|
|
|
669 |
|
|
|
670 |
private inline void
|
|
|
671 |
draw_triangle(const gs_fixed_point *p0, const gs_fixed_point *p1,
|
|
|
672 |
const gs_fixed_point *p2, ulong rgbcolor)
|
|
|
673 |
{
|
|
|
674 |
#ifdef DEBUG
|
|
|
675 |
if (!vd_enabled)
|
|
|
676 |
return;
|
|
|
677 |
vd_quad(p0->x, p0->y, p0->x, p0->y, p1->x, p1->y, p2->x, p2->y, 0, rgbcolor);
|
|
|
678 |
#endif
|
|
|
679 |
}
|
|
|
680 |
|
|
|
681 |
private inline void
|
|
|
682 |
draw_quadrangle(const quadrangle_patch *p, ulong rgbcolor)
|
|
|
683 |
{
|
|
|
684 |
#ifdef DEBUG
|
|
|
685 |
vd_quad(p->p[0][0]->p.x, p->p[0][0]->p.y,
|
|
|
686 |
p->p[0][1]->p.x, p->p[0][1]->p.y,
|
|
|
687 |
p->p[1][1]->p.x, p->p[1][1]->p.y,
|
|
|
688 |
p->p[1][0]->p.x, p->p[1][0]->p.y,
|
|
|
689 |
0, rgbcolor);
|
|
|
690 |
#endif
|
|
|
691 |
}
|
|
|
692 |
|
|
|
693 |
private inline int
|
|
|
694 |
curve_samples(patch_fill_state_t *pfs,
|
|
|
695 |
const gs_fixed_point *pole, int pole_step, fixed fixed_flat)
|
|
|
696 |
{
|
|
|
697 |
curve_segment s;
|
|
|
698 |
int k;
|
|
|
699 |
|
|
|
700 |
s.p1.x = pole[pole_step].x;
|
|
|
701 |
s.p1.y = pole[pole_step].y;
|
|
|
702 |
s.p2.x = pole[pole_step * 2].x;
|
|
|
703 |
s.p2.y = pole[pole_step * 2].y;
|
|
|
704 |
s.pt.x = pole[pole_step * 3].x;
|
|
|
705 |
s.pt.y = pole[pole_step * 3].y;
|
|
|
706 |
k = gx_curve_log2_samples(pole[0].x, pole[0].y, &s, fixed_flat);
|
|
|
707 |
{
|
|
|
708 |
# if LAZY_WEDGES || QUADRANGLES
|
|
|
709 |
int k1;
|
|
|
710 |
fixed L = any_abs(pole[1].x - pole[0].x) + any_abs(pole[1].y - pole[0].y) +
|
|
|
711 |
any_abs(pole[2].x - pole[1].x) + any_abs(pole[2].y - pole[1].y) +
|
|
|
712 |
any_abs(pole[3].x - pole[2].x) + any_abs(pole[3].y - pole[2].y);
|
|
|
713 |
# endif
|
|
|
714 |
|
|
|
715 |
# if LAZY_WEDGES
|
|
|
716 |
/* Restrict lengths for a reasonable memory consumption : */
|
|
|
717 |
k1 = ilog2(L / fixed_1 / (1 << (LAZY_WEDGES_MAX_LEVEL - 1)));
|
|
|
718 |
k = max(k, k1);
|
|
|
719 |
# endif
|
|
|
720 |
# if QUADRANGLES
|
|
|
721 |
/* Restrict lengths for intersection_of_small_bars : */
|
|
|
722 |
k = max(k, ilog2(L) - ilog2(pfs->max_small_coord));
|
|
|
723 |
# endif
|
|
|
724 |
}
|
|
|
725 |
return 1 << k;
|
|
|
726 |
}
|
|
|
727 |
|
|
|
728 |
private bool
|
|
|
729 |
intersection_of_small_bars(const gs_fixed_point q[4], int i0, int i1, int i2, int i3, fixed *ry, fixed *ey)
|
|
|
730 |
{
|
|
|
731 |
/* This function is only used with QUADRANGLES. */
|
|
|
732 |
fixed dx1 = q[i1].x - q[i0].x, dy1 = q[i1].y - q[i0].y;
|
|
|
733 |
fixed dx2 = q[i2].x - q[i0].x, dy2 = q[i2].y - q[i0].y;
|
|
|
734 |
fixed dx3 = q[i3].x - q[i0].x, dy3 = q[i3].y - q[i0].y;
|
|
|
735 |
int64_t vp2a, vp2b, vp3a, vp3b;
|
|
|
736 |
int s2, s3;
|
|
|
737 |
|
|
|
738 |
if (dx1 == 0 && dy1 == 0)
|
|
|
739 |
return false; /* Zero length bars are out of interest. */
|
|
|
740 |
if (dx2 == 0 && dy2 == 0)
|
|
|
741 |
return false; /* Contacting ends are out of interest. */
|
|
|
742 |
if (dx3 == 0 && dy3 == 0)
|
|
|
743 |
return false; /* Contacting ends are out of interest. */
|
|
|
744 |
if (dx2 == dx1 && dy2 == dy1)
|
|
|
745 |
return false; /* Contacting ends are out of interest. */
|
|
|
746 |
if (dx3 == dx1 && dy3 == dy1)
|
|
|
747 |
return false; /* Contacting ends are out of interest. */
|
|
|
748 |
if (dx2 == dx3 && dy2 == dy3)
|
|
|
749 |
return false; /* Zero length bars are out of interest. */
|
|
|
750 |
vp2a = (int64_t)dx1 * dy2;
|
|
|
751 |
vp2b = (int64_t)dy1 * dx2;
|
|
|
752 |
/* vp2 = vp2a - vp2b; It can overflow int64_t, but we only need the sign. */
|
|
|
753 |
if (vp2a > vp2b)
|
|
|
754 |
s2 = 1;
|
|
|
755 |
else if (vp2a < vp2b)
|
|
|
756 |
s2 = -1;
|
|
|
757 |
else
|
|
|
758 |
s2 = 0;
|
|
|
759 |
vp3a = (int64_t)dx1 * dy3;
|
|
|
760 |
vp3b = (int64_t)dy1 * dx3;
|
|
|
761 |
/* vp3 = vp3a - vp3b; It can overflow int64_t, but we only need the sign. */
|
|
|
762 |
if (vp3a > vp3b)
|
|
|
763 |
s3 = 1;
|
|
|
764 |
else if (vp3a < vp3b)
|
|
|
765 |
s3 = -1;
|
|
|
766 |
else
|
|
|
767 |
s3 = 0;
|
|
|
768 |
if (s2 == 0) {
|
|
|
769 |
if (s3 == 0)
|
|
|
770 |
return false; /* Collinear bars - out of interest. */
|
|
|
771 |
if (0 <= dx2 && dx2 <= dx1 && 0 <= dy2 && dy2 <= dy1) {
|
|
|
772 |
/* The start of the bar 2 is in the bar 1. */
|
|
|
773 |
*ry = q[i2].y;
|
|
|
774 |
*ey = 0;
|
|
|
775 |
return true;
|
|
|
776 |
}
|
|
|
777 |
} else if (s3 == 0) {
|
|
|
778 |
if (0 <= dx3 && dx3 <= dx1 && 0 <= dy3 && dy3 <= dy1) {
|
|
|
779 |
/* The end of the bar 2 is in the bar 1. */
|
|
|
780 |
*ry = q[i3].y;
|
|
|
781 |
*ey = 0;
|
|
|
782 |
return true;
|
|
|
783 |
}
|
|
|
784 |
} else if (s2 * s3 < 0) {
|
|
|
785 |
/* The intersection definitely exists, so the determinant isn't zero. */
|
|
|
786 |
fixed d23x = dx3 - dx2, d23y = dy3 - dy2;
|
|
|
787 |
int64_t det = (int64_t)dx1 * d23y - (int64_t)dy1 * d23x;
|
|
|
788 |
int64_t mul = (int64_t)dx2 * d23y - (int64_t)dy2 * d23x;
|
|
|
789 |
# define USE_DOUBLE 0
|
|
|
790 |
# define USE_INT64_T (1 || !USE_DOUBLE)
|
|
|
791 |
# if USE_DOUBLE
|
|
|
792 |
{
|
|
|
793 |
/* Assuming big bars. Not a good thing due to 'double'. */
|
|
|
794 |
/* The determinant can't compute in double due to
|
|
|
795 |
possible loss of all significant bits when subtracting the
|
|
|
796 |
trucnated prodicts. But after we subtract in int64_t,
|
|
|
797 |
it converts to 'double' with a reasonable truncation. */
|
|
|
798 |
double dy = dy1 * (double)mul / (double)det;
|
|
|
799 |
fixed iy;
|
|
|
800 |
|
|
|
801 |
if (dy1 > 0 && dy >= dy1)
|
|
|
802 |
return false; /* Outside the bar 1. */
|
|
|
803 |
if (dy1 < 0 && dy <= dy1)
|
|
|
804 |
return false; /* Outside the bar 1. */
|
|
|
805 |
if (dy2 < dy3) {
|
|
|
806 |
if (dy <= dy2 || dy >= dy3)
|
|
|
807 |
return false; /* Outside the bar 2. */
|
|
|
808 |
} else {
|
|
|
809 |
if (dy >= dy2 || dy <= dy3)
|
|
|
810 |
return false; /* Outside the bar 2. */
|
|
|
811 |
}
|
|
|
812 |
iy = (int)floor(dy);
|
|
|
813 |
*ry = q[i0].y + iy;
|
|
|
814 |
*ey = (dy > iy ? 1 : 0);
|
|
|
815 |
}
|
|
|
816 |
# endif
|
|
|
817 |
# if USE_INT64_T
|
|
|
818 |
{
|
|
|
819 |
/* Assuming small bars : cubes of coordinates must fit into int64_t.
|
|
|
820 |
curve_samples must provide that. */
|
|
|
821 |
int64_t num = dy1 * mul, iiy;
|
|
|
822 |
fixed iy;
|
|
|
823 |
fixed pry, pey;
|
|
|
824 |
|
|
|
825 |
{ /* Likely when called form wedge_trap_decompose or constant_color_quadrangle,
|
|
|
826 |
we always have det > 0 && num >= 0, but we check here for a safety reason. */
|
|
|
827 |
if (det < 0)
|
|
|
828 |
{num = -num; det = -det;}
|
|
|
829 |
if(num >= 0)
|
|
|
830 |
iiy = num / det;
|
|
|
831 |
else
|
|
|
832 |
iiy = (num - det + 1) / det;
|
|
|
833 |
iy = (fixed)iiy;
|
|
|
834 |
if (iy != iiy) {
|
|
|
835 |
/* If it is inside the bars, it must fit into fixed. */
|
|
|
836 |
return false;
|
|
|
837 |
}
|
|
|
838 |
}
|
|
|
839 |
if (dy1 > 0 && iy >= dy1)
|
|
|
840 |
return false; /* Outside the bar 1. */
|
|
|
841 |
if (dy1 < 0 && iy <= dy1)
|
|
|
842 |
return false; /* Outside the bar 1. */
|
|
|
843 |
if (dy2 < dy3) {
|
|
|
844 |
if (iy <= dy2 || iy >= dy3)
|
|
|
845 |
return false; /* Outside the bar 2. */
|
|
|
846 |
} else {
|
|
|
847 |
if (iy >= dy2 || iy <= dy3)
|
|
|
848 |
return false; /* Outside the bar 2. */
|
|
|
849 |
}
|
|
|
850 |
pry = q[i0].y + (fixed)iy;
|
|
|
851 |
pey = (iy * det < num ? 1 : 0);
|
|
|
852 |
# if USE_DOUBLE && USE_INT64_T
|
|
|
853 |
assert(*ry == pry);
|
|
|
854 |
assert(*ey == pey);
|
|
|
855 |
# endif
|
|
|
856 |
*ry = pry;
|
|
|
857 |
*ey = pey;
|
|
|
858 |
}
|
|
|
859 |
# endif
|
|
|
860 |
return true;
|
|
|
861 |
}
|
|
|
862 |
return false;
|
|
|
863 |
}
|
|
|
864 |
|
|
|
865 |
private inline void
|
|
|
866 |
adjust_swapped_boundary(fixed *b, bool swap_axes)
|
|
|
867 |
{
|
|
|
868 |
if (swap_axes) {
|
|
|
869 |
/* Sinse the rasterizer algorithm assumes semi-open interval
|
|
|
870 |
when computing pixel coverage, we should expand
|
|
|
871 |
the right side of the area. Otherwise a dropout can happen :
|
|
|
872 |
if the left neighbour is painted with !swap_axes,
|
|
|
873 |
the left side of this area appears to be the left side
|
|
|
874 |
of the neighbour area, and the side is not included
|
|
|
875 |
into both areas.
|
|
|
876 |
*/
|
|
|
877 |
*b += fixed_epsilon;
|
|
|
878 |
}
|
|
|
879 |
}
|
|
|
880 |
|
|
|
881 |
private inline void
|
|
|
882 |
make_trapezoid(const gs_fixed_point q[4],
|
|
|
883 |
int vi0, int vi1, int vi2, int vi3, fixed ybot, fixed ytop,
|
|
|
884 |
bool swap_axes, bool orient, gs_fixed_edge *le, gs_fixed_edge *re)
|
|
|
885 |
{
|
|
|
886 |
if (!orient) {
|
|
|
887 |
le->start = q[vi0];
|
|
|
888 |
le->end = q[vi1];
|
|
|
889 |
re->start = q[vi2];
|
|
|
890 |
re->end = q[vi3];
|
|
|
891 |
} else {
|
|
|
892 |
le->start = q[vi2];
|
|
|
893 |
le->end = q[vi3];
|
|
|
894 |
re->start = q[vi0];
|
|
|
895 |
re->end = q[vi1];
|
|
|
896 |
}
|
|
|
897 |
adjust_swapped_boundary(&re->start.x, swap_axes);
|
|
|
898 |
adjust_swapped_boundary(&re->end.x, swap_axes);
|
|
|
899 |
}
|
|
|
900 |
|
|
|
901 |
private inline int
|
|
|
902 |
gx_shade_trapezoid(patch_fill_state_t *pfs, const gs_fixed_point q[4],
|
|
|
903 |
int vi0, int vi1, int vi2, int vi3, fixed ybot0, fixed ytop0,
|
|
|
904 |
bool swap_axes, const gx_device_color *pdevc, bool orient)
|
|
|
905 |
{
|
|
|
906 |
gs_fixed_edge le, re;
|
|
|
907 |
int code;
|
|
|
908 |
fixed ybot = max(ybot0, swap_axes ? pfs->rect.p.x : pfs->rect.p.y);
|
|
|
909 |
fixed ytop = min(ytop0, swap_axes ? pfs->rect.q.x : pfs->rect.q.y);
|
|
|
910 |
vd_save;
|
|
|
911 |
|
|
|
912 |
if (ybot > ytop)
|
|
|
913 |
return 0;
|
|
|
914 |
# if NOFILL_TEST
|
|
|
915 |
if (dbg_nofill)
|
|
|
916 |
return 0;
|
|
|
917 |
# endif
|
|
|
918 |
make_trapezoid(q, vi0, vi1, vi2, vi3, ybot, ytop, swap_axes, orient, &le, &re);
|
|
|
919 |
if (!VD_TRACE_DOWN)
|
|
|
920 |
vd_disable;
|
|
|
921 |
code = dev_proc(pfs->dev, fill_trapezoid)(pfs->dev,
|
|
|
922 |
&le, &re, ybot, ytop, swap_axes, pdevc, pfs->pis->log_op);
|
|
|
923 |
vd_restore;
|
|
|
924 |
return code;
|
|
|
925 |
}
|
|
|
926 |
|
|
|
927 |
private int
|
|
|
928 |
patch_color_to_device_color(const patch_fill_state_t *pfs, const patch_color_t *c, gx_device_color *pdevc)
|
|
|
929 |
{
|
|
|
930 |
/* A code fragment copied from mesh_fill_triangle. */
|
|
|
931 |
gs_client_color fcc;
|
|
|
932 |
const gs_color_space *pcs = pfs->direct_space;
|
|
|
933 |
|
|
|
934 |
memcpy(fcc.paint.values, c->cc.paint.values,
|
|
|
935 |
sizeof(fcc.paint.values[0]) * pfs->num_components);
|
|
|
936 |
return pcs->type->remap_color(&fcc, pcs, pdevc, pfs->pis,
|
|
|
937 |
pfs->dev, gs_color_select_texture);
|
|
|
938 |
}
|
|
|
939 |
|
|
|
940 |
private inline double
|
|
|
941 |
color_span(const patch_fill_state_t *pfs, const patch_color_t *c0, const patch_color_t *c1)
|
|
|
942 |
{
|
|
|
943 |
int n = pfs->num_components, i;
|
|
|
944 |
double m;
|
|
|
945 |
|
|
|
946 |
/* Dont want to copy colors, which are big things. */
|
|
|
947 |
m = any_abs(c1->cc.paint.values[0] - c0->cc.paint.values[0]) / pfs->color_domain.paint.values[0];
|
|
|
948 |
for (i = 1; i < n; i++)
|
|
|
949 |
m = max(m, any_abs(c1->cc.paint.values[i] - c0->cc.paint.values[i]) / pfs->color_domain.paint.values[i]);
|
|
|
950 |
return m;
|
|
|
951 |
}
|
|
|
952 |
|
|
|
953 |
private inline void
|
|
|
954 |
color_diff(const patch_fill_state_t *pfs, const patch_color_t *c0, const patch_color_t *c1, patch_color_t *d)
|
|
|
955 |
{
|
|
|
956 |
int n = pfs->num_components, i;
|
|
|
957 |
|
|
|
958 |
for (i = 0; i < n; i++)
|
|
|
959 |
d->cc.paint.values[i] = c1->cc.paint.values[i] - c0->cc.paint.values[i];
|
|
|
960 |
}
|
|
|
961 |
|
|
|
962 |
private inline double
|
|
|
963 |
color_norm(const patch_fill_state_t *pfs, const patch_color_t *c)
|
|
|
964 |
{
|
|
|
965 |
int n = pfs->num_components, i;
|
|
|
966 |
double m;
|
|
|
967 |
|
|
|
968 |
m = any_abs(c->cc.paint.values[0]) / pfs->color_domain.paint.values[0];
|
|
|
969 |
for (i = 1; i < n; i++)
|
|
|
970 |
m = max(m, any_abs(c->cc.paint.values[i]) / pfs->color_domain.paint.values[i]);
|
|
|
971 |
return m;
|
|
|
972 |
}
|
|
|
973 |
|
|
|
974 |
private inline int
|
|
|
975 |
isnt_color_monotonic(const patch_fill_state_t *pfs, const patch_color_t *c0, const patch_color_t *c1)
|
|
|
976 |
{ /* checks whether the color is monotonic in the n-dimensional interval,
|
|
|
977 |
where n is the number of parameters in c0->t, c1->t.
|
|
|
978 |
returns : 0 = monotonic,
|
|
|
979 |
bit 0 = not or don't know by t0,
|
|
|
980 |
bit 1 = not or don't know by t1,
|
|
|
981 |
<0 = error. */
|
|
|
982 |
/* When pfs->Function is not set, the color is monotonic.
|
|
|
983 |
In this case do not call this function because
|
|
|
984 |
it doesn't check whether pfs->Function is set.
|
|
|
985 |
Actually pfs->monotonic_color prevents that.
|
|
|
986 |
*/
|
|
|
987 |
uint mask;
|
|
|
988 |
int code = gs_function_is_monotonic(pfs->Function, c0->t, c1->t, &mask);
|
|
|
989 |
|
|
|
990 |
if (code >= 0)
|
|
|
991 |
return mask;
|
|
|
992 |
return code;
|
|
|
993 |
}
|
|
|
994 |
|
|
|
995 |
private inline bool
|
|
|
996 |
covers_pixel_centers(fixed ybot, fixed ytop)
|
|
|
997 |
{
|
|
|
998 |
return fixed_pixround(ybot) < fixed_pixround(ytop);
|
|
|
999 |
}
|
|
|
1000 |
|
|
|
1001 |
private inline int
|
|
|
1002 |
constant_color_trapezoid(patch_fill_state_t *pfs, gs_fixed_edge *le, gs_fixed_edge *re,
|
|
|
1003 |
fixed ybot, fixed ytop, bool swap_axes, const patch_color_t *c)
|
|
|
1004 |
{
|
|
|
1005 |
patch_color_t c1 = *c;
|
|
|
1006 |
gx_device_color dc;
|
|
|
1007 |
int code;
|
|
|
1008 |
vd_save;
|
|
|
1009 |
|
|
|
1010 |
# if NOFILL_TEST
|
|
|
1011 |
/* if (dbg_nofill)
|
|
|
1012 |
return 0; */
|
|
|
1013 |
# endif
|
|
|
1014 |
code = patch_color_to_device_color(pfs, &c1, &dc);
|
|
|
1015 |
if (code < 0)
|
|
|
1016 |
return code;
|
|
|
1017 |
if (!VD_TRACE_DOWN)
|
|
|
1018 |
vd_disable;
|
|
|
1019 |
code = dev_proc(pfs->dev, fill_trapezoid)(pfs->dev,
|
|
|
1020 |
le, re, ybot, ytop, swap_axes, &dc, pfs->pis->log_op);
|
|
|
1021 |
vd_restore;
|
|
|
1022 |
return code;
|
|
|
1023 |
}
|
|
|
1024 |
|
|
|
1025 |
private inline void
|
|
|
1026 |
dc2fc(const patch_fill_state_t *pfs, gx_color_index c,
|
|
|
1027 |
frac31 fc[GX_DEVICE_COLOR_MAX_COMPONENTS])
|
|
|
1028 |
{
|
|
|
1029 |
int j;
|
|
|
1030 |
const gx_device_color_info *cinfo = &pfs->dev->color_info;
|
|
|
1031 |
|
|
|
1032 |
for (j = 0; j < cinfo->num_components; j++) {
|
|
|
1033 |
int shift = cinfo->comp_shift[j];
|
|
|
1034 |
int bits = cinfo->comp_bits[j];
|
|
|
1035 |
|
|
|
1036 |
fc[j] = ((c >> shift) & ((1 << bits) - 1)) << (sizeof(frac31) * 8 - 1 - bits);
|
|
|
1037 |
}
|
|
|
1038 |
}
|
|
|
1039 |
|
|
|
1040 |
private inline float
|
|
|
1041 |
function_linearity(const patch_fill_state_t *pfs, const patch_color_t *c0, const patch_color_t *c1)
|
|
|
1042 |
{
|
|
|
1043 |
float smoothness = max(pfs->smoothness, 1.0 / min_linear_grades), s = 0;
|
|
|
1044 |
/* Restrict the smoothness with 1/min_linear_grades, because cs_is_linear
|
|
|
1045 |
can't provide a better precision due to the color
|
|
|
1046 |
representation with integers.
|
|
|
1047 |
*/
|
|
|
1048 |
|
|
|
1049 |
if (pfs->Function != NULL) {
|
|
|
1050 |
patch_color_t c;
|
|
|
1051 |
const float q[2] = {(float)0.3, (float)0.7};
|
|
|
1052 |
int i, j;
|
|
|
1053 |
|
|
|
1054 |
for (j = 0; j < count_of(q); j++) {
|
|
|
1055 |
c.t[0] = c0->t[0] * (1 - q[j]) + c1->t[0] * q[j];
|
|
|
1056 |
c.t[1] = c0->t[1] * (1 - q[j]) + c1->t[1] * q[j];
|
|
|
1057 |
patch_resolve_color_inline(&c, pfs);
|
|
|
1058 |
for (i = 0; i < pfs->num_components; i++) {
|
|
|
1059 |
float v = c0->cc.paint.values[i] * (1 - q[j]) + c1->cc.paint.values[i] * q[j];
|
|
|
1060 |
float d = v - c.cc.paint.values[i];
|
|
|
1061 |
float s1 = any_abs(d) / pfs->color_domain.paint.values[i];
|
|
|
1062 |
|
|
|
1063 |
if (s1 > smoothness)
|
|
|
1064 |
return s1;
|
|
|
1065 |
if (s < s1)
|
|
|
1066 |
s = s1;
|
|
|
1067 |
}
|
|
|
1068 |
}
|
|
|
1069 |
}
|
|
|
1070 |
return s;
|
|
|
1071 |
}
|
|
|
1072 |
|
|
|
1073 |
private inline int
|
|
|
1074 |
is_color_linear(const patch_fill_state_t *pfs, const patch_color_t *c0, const patch_color_t *c1)
|
|
|
1075 |
{ /* returns : 1 = linear, 0 = unlinear, <0 = error. */
|
|
|
1076 |
if (pfs->unlinear)
|
|
|
1077 |
return 1; /* Disable this check. */
|
|
|
1078 |
else {
|
|
|
1079 |
gs_direct_color_space *cs =
|
|
|
1080 |
(gs_direct_color_space *)pfs->direct_space; /* break 'const'. */
|
|
|
1081 |
int code;
|
|
|
1082 |
float smoothness = max(pfs->smoothness, 1.0 / min_linear_grades);
|
|
|
1083 |
/* Restrict the smoothness with 1/min_linear_grades, because cs_is_linear
|
|
|
1084 |
can't provide a better precision due to the color
|
|
|
1085 |
representation with integers.
|
|
|
1086 |
*/
|
|
|
1087 |
float s = function_linearity(pfs, c0, c1);
|
|
|
1088 |
|
|
|
1089 |
if (s > smoothness)
|
|
|
1090 |
return 0;
|
|
|
1091 |
code = cs_is_linear(cs, pfs->pis, pfs->dev,
|
|
|
1092 |
&c0->cc, &c1->cc, NULL, NULL, smoothness - s);
|
|
|
1093 |
if (code <= 0)
|
|
|
1094 |
return code;
|
|
|
1095 |
return 1;
|
|
|
1096 |
}
|
|
|
1097 |
}
|
|
|
1098 |
|
|
|
1099 |
private int
|
|
|
1100 |
decompose_linear_color(patch_fill_state_t *pfs, gs_fixed_edge *le, gs_fixed_edge *re,
|
|
|
1101 |
fixed ybot, fixed ytop, bool swap_axes, const patch_color_t *c0,
|
|
|
1102 |
const patch_color_t *c1, int level)
|
|
|
1103 |
{
|
|
|
1104 |
/* Assuming a very narrow trapezoid - ignore the transversal color variation. */
|
|
|
1105 |
/* Assuming the XY span is restricted with curve_samples.
|
|
|
1106 |
It is important for intersection_of_small_bars to compute faster. */
|
|
|
1107 |
int code;
|
|
|
1108 |
patch_color_t c;
|
|
|
1109 |
|
|
|
1110 |
if (level > 100)
|
|
|
1111 |
return_error(gs_error_unregistered); /* Must not happen. */
|
|
|
1112 |
/* Use the recursive decomposition due to isnt_color_monotonic
|
|
|
1113 |
based on fn_is_monotonic_proc_t is_monotonic,
|
|
|
1114 |
which applies to intervals. */
|
|
|
1115 |
patch_interpolate_color(&c, c0, c1, pfs, 0.5);
|
|
|
1116 |
if (ytop - ybot < fixed_1 / 2) /* Prevent an infinite color decomposition. */
|
|
|
1117 |
return constant_color_trapezoid(pfs, le, re, ybot, ytop, swap_axes, &c);
|
|
|
1118 |
else {
|
|
|
1119 |
bool monotonic_color_save = pfs->monotonic_color;
|
|
|
1120 |
bool linear_color_save = pfs->linear_color;
|
|
|
1121 |
|
|
|
1122 |
if (!pfs->monotonic_color) {
|
|
|
1123 |
code = isnt_color_monotonic(pfs, c0, c1);
|
|
|
1124 |
if (code < 0)
|
|
|
1125 |
return code;
|
|
|
1126 |
if (!code)
|
|
|
1127 |
pfs->monotonic_color = true;
|
|
|
1128 |
}
|
|
|
1129 |
if (pfs->monotonic_color && !pfs->linear_color) {
|
|
|
1130 |
code = is_color_linear(pfs, c0, c1);
|
|
|
1131 |
if (code < 0)
|
|
|
1132 |
return code;
|
|
|
1133 |
if (code > 0)
|
|
|
1134 |
pfs->linear_color = true;
|
|
|
1135 |
}
|
|
|
1136 |
if (!pfs->unlinear && pfs->linear_color) {
|
|
|
1137 |
gx_device *pdev = pfs->dev;
|
|
|
1138 |
frac31 fc[2][GX_DEVICE_COLOR_MAX_COMPONENTS];
|
|
|
1139 |
gs_fill_attributes fa;
|
|
|
1140 |
gx_device_color dc[2];
|
|
|
1141 |
gs_fixed_rect clip;
|
|
|
1142 |
int code;
|
|
|
1143 |
|
|
|
1144 |
clip = pfs->rect;
|
|
|
1145 |
if (swap_axes) {
|
|
|
1146 |
fixed v;
|
|
|
1147 |
|
|
|
1148 |
v = clip.p.x; clip.p.x = clip.p.y; clip.p.y = v;
|
|
|
1149 |
v = clip.q.x; clip.q.x = clip.q.y; clip.q.y = v;
|
|
|
1150 |
/* Don't need adjust_swapped_boundary here. */
|
|
|
1151 |
}
|
|
|
1152 |
clip.p.y = max(clip.p.y, ybot);
|
|
|
1153 |
clip.q.y = min(clip.q.y, ytop);
|
|
|
1154 |
fa.clip = &clip;
|
|
|
1155 |
fa.ht = NULL;
|
|
|
1156 |
fa.swap_axes = swap_axes;
|
|
|
1157 |
fa.lop = 0;
|
|
|
1158 |
fa.ystart = ybot;
|
|
|
1159 |
fa.yend = ytop;
|
|
|
1160 |
code = patch_color_to_device_color(pfs, c0, &dc[0]);
|
|
|
1161 |
if (code < 0)
|
|
|
1162 |
return code;
|
|
|
1163 |
if (dc[0].type == &gx_dc_type_data_pure) {
|
|
|
1164 |
dc2fc(pfs, dc[0].colors.pure, fc[0]);
|
|
|
1165 |
code = patch_color_to_device_color(pfs, c1, &dc[1]);
|
|
|
1166 |
if (code < 0)
|
|
|
1167 |
return code;
|
|
|
1168 |
dc2fc(pfs, dc[1].colors.pure, fc[1]);
|
|
|
1169 |
code = dev_proc(pdev, fill_linear_color_trapezoid)(pdev, &fa,
|
|
|
1170 |
&le->start, &le->end, &re->start, &re->end,
|
|
|
1171 |
fc[0], fc[1], NULL, NULL);
|
|
|
1172 |
if (code == 1) {
|
|
|
1173 |
pfs->monotonic_color = monotonic_color_save;
|
|
|
1174 |
pfs->linear_color = linear_color_save;
|
|
|
1175 |
return 0; /* The area is filled. */
|
|
|
1176 |
}
|
|
|
1177 |
if (code < 0)
|
|
|
1178 |
return code;
|
|
|
1179 |
else /* code == 0, the device requested to decompose the area. */
|
|
|
1180 |
return_error(gs_error_unregistered); /* Must not happen. */
|
|
|
1181 |
}
|
|
|
1182 |
}
|
|
|
1183 |
if (!pfs->unlinear || !pfs->linear_color ||
|
|
|
1184 |
color_span(pfs, c0, c1) > pfs->smoothness) {
|
|
|
1185 |
fixed y = (ybot + ytop) / 2;
|
|
|
1186 |
|
|
|
1187 |
code = decompose_linear_color(pfs, le, re, ybot, y, swap_axes, c0, &c, level + 1);
|
|
|
1188 |
if (code >= 0)
|
|
|
1189 |
code = decompose_linear_color(pfs, le, re, y, ytop, swap_axes, &c, c1, level + 1);
|
|
|
1190 |
} else
|
|
|
1191 |
code = constant_color_trapezoid(pfs, le, re, ybot, ytop, swap_axes, &c);
|
|
|
1192 |
pfs->monotonic_color = monotonic_color_save;
|
|
|
1193 |
pfs->linear_color = linear_color_save;
|
|
|
1194 |
return code;
|
|
|
1195 |
}
|
|
|
1196 |
}
|
|
|
1197 |
|
|
|
1198 |
private inline int
|
|
|
1199 |
linear_color_trapezoid(patch_fill_state_t *pfs, gs_fixed_point q[4], int i0, int i1, int i2, int i3,
|
|
|
1200 |
fixed ybot, fixed ytop, bool swap_axes, const patch_color_t *c0, const patch_color_t *c1,
|
|
|
1201 |
bool orient)
|
|
|
1202 |
{
|
|
|
1203 |
/* Assuming a very narrow trapezoid - ignore the transversal color change. */
|
|
|
1204 |
gs_fixed_edge le, re;
|
|
|
1205 |
|
|
|
1206 |
make_trapezoid(q, i0, i1, i2, i3, ybot, ytop, swap_axes, orient, &le, &re);
|
|
|
1207 |
return decompose_linear_color(pfs, &le, &re, ybot, ytop, swap_axes, c0, c1, 0);
|
|
|
1208 |
}
|
|
|
1209 |
|
|
|
1210 |
private int
|
|
|
1211 |
wedge_trap_decompose(patch_fill_state_t *pfs, gs_fixed_point q[4],
|
|
|
1212 |
fixed ybot, fixed ytop, const patch_color_t *c0, const patch_color_t *c1,
|
|
|
1213 |
bool swap_axes, bool self_intersecting)
|
|
|
1214 |
{
|
|
|
1215 |
/* Assuming a very narrow trapezoid - ignore the transversal color change. */
|
|
|
1216 |
fixed dx1, dy1, dx2, dy2;
|
|
|
1217 |
bool orient;
|
|
|
1218 |
|
|
|
1219 |
if (!pfs->vectorization && !covers_pixel_centers(ybot, ytop))
|
|
|
1220 |
return 0;
|
|
|
1221 |
if (ybot == ytop)
|
|
|
1222 |
return 0;
|
|
|
1223 |
dx1 = q[1].x - q[0].x, dy1 = q[1].y - q[0].y;
|
|
|
1224 |
dx2 = q[2].x - q[0].x, dy2 = q[2].y - q[0].y;
|
|
|
1225 |
#if 1
|
|
|
1226 |
if (!swap_axes)
|
|
|
1227 |
vd_quad(q[0].x, q[0].y, q[1].x, q[1].y, q[3].x, q[3].y, q[2].x, q[2].y, 0, RGB(255, 0, 0));
|
|
|
1228 |
else
|
|
|
1229 |
vd_quad(q[0].y, q[0].x, q[1].y, q[1].x, q[3].y, q[3].x, q[2].y, q[2].x, 0, RGB(255, 0, 0));
|
|
|
1230 |
#endif
|
|
|
1231 |
if ((int64_t)dx1 * dy2 != (int64_t)dy1 * dx2) {
|
|
|
1232 |
orient = ((int64_t)dx1 * dy2 > (int64_t)dy1 * dx2);
|
|
|
1233 |
return linear_color_trapezoid(pfs, q, 0, 1, 2, 3, ybot, ytop, swap_axes, c0, c1, orient);
|
|
|
1234 |
} else {
|
|
|
1235 |
fixed dx3 = q[3].x - q[0].x, dy3 = q[3].y - q[0].y;
|
|
|
1236 |
|
|
|
1237 |
orient = ((int64_t)dx1 * dy3 > (int64_t)dy1 * dx3);
|
|
|
1238 |
return linear_color_trapezoid(pfs, q, 0, 1, 2, 3, ybot, ytop, swap_axes, c0, c1, orient);
|
|
|
1239 |
}
|
|
|
1240 |
}
|
|
|
1241 |
|
|
|
1242 |
private inline int
|
|
|
1243 |
fill_wedge_trap(patch_fill_state_t *pfs, const gs_fixed_point *p0, const gs_fixed_point *p1,
|
|
|
1244 |
const gs_fixed_point *q0, const gs_fixed_point *q1, const patch_color_t *c0, const patch_color_t *c1,
|
|
|
1245 |
bool swap_axes, bool self_intersecting)
|
|
|
1246 |
{
|
|
|
1247 |
/* We assume that the width of the wedge is close to zero,
|
|
|
1248 |
so we can ignore the slope when computing transversal distances. */
|
|
|
1249 |
gs_fixed_point p[4];
|
|
|
1250 |
const patch_color_t *cc0, *cc1;
|
|
|
1251 |
|
|
|
1252 |
if (p0->y < p1->y) {
|
|
|
1253 |
p[2] = *p0;
|
|
|
1254 |
p[3] = *p1;
|
|
|
1255 |
cc0 = c0;
|
|
|
1256 |
cc1 = c1;
|
|
|
1257 |
} else {
|
|
|
1258 |
p[2] = *p1;
|
|
|
1259 |
p[3] = *p0;
|
|
|
1260 |
cc0 = c1;
|
|
|
1261 |
cc1 = c0;
|
|
|
1262 |
}
|
|
|
1263 |
p[0] = *q0;
|
|
|
1264 |
p[1] = *q1;
|
|
|
1265 |
return wedge_trap_decompose(pfs, p, p[2].y, p[3].y, cc0, cc1, swap_axes, self_intersecting);
|
|
|
1266 |
}
|
|
|
1267 |
|
|
|
1268 |
private void
|
|
|
1269 |
split_curve_s(const gs_fixed_point *pole, gs_fixed_point *q0, gs_fixed_point *q1, int pole_step)
|
|
|
1270 |
{
|
|
|
1271 |
/* This copies a code fragment from split_curve_midpoint,
|
|
|
1272 |
substituting another data type.
|
|
|
1273 |
*/
|
|
|
1274 |
/*
|
|
|
1275 |
* We have to define midpoint carefully to avoid overflow.
|
|
|
1276 |
* (If it overflows, something really pathological is going
|
|
|
1277 |
* on, but we could get infinite recursion that way....)
|
|
|
1278 |
*/
|
|
|
1279 |
#define midpoint(a,b)\
|
|
|
1280 |
(arith_rshift_1(a) + arith_rshift_1(b) + (((a) | (b)) & 1))
|
|
|
1281 |
fixed x12 = midpoint(pole[1 * pole_step].x, pole[2 * pole_step].x);
|
|
|
1282 |
fixed y12 = midpoint(pole[1 * pole_step].y, pole[2 * pole_step].y);
|
|
|
1283 |
|
|
|
1284 |
/* q[0] and q[1] must not be the same as pole. */
|
|
|
1285 |
q0[1 * pole_step].x = midpoint(pole[0 * pole_step].x, pole[1 * pole_step].x);
|
|
|
1286 |
q0[1 * pole_step].y = midpoint(pole[0 * pole_step].y, pole[1 * pole_step].y);
|
|
|
1287 |
q1[2 * pole_step].x = midpoint(pole[2 * pole_step].x, pole[3 * pole_step].x);
|
|
|
1288 |
q1[2 * pole_step].y = midpoint(pole[2 * pole_step].y, pole[3 * pole_step].y);
|
|
|
1289 |
q0[2 * pole_step].x = midpoint(q0[1 * pole_step].x, x12);
|
|
|
1290 |
q0[2 * pole_step].y = midpoint(q0[1 * pole_step].y, y12);
|
|
|
1291 |
q1[1 * pole_step].x = midpoint(x12, q1[2 * pole_step].x);
|
|
|
1292 |
q1[1 * pole_step].y = midpoint(y12, q1[2 * pole_step].y);
|
|
|
1293 |
q0[0 * pole_step].x = pole[0 * pole_step].x;
|
|
|
1294 |
q0[0 * pole_step].y = pole[0 * pole_step].y;
|
|
|
1295 |
q0[3 * pole_step].x = q1[0 * pole_step].x = midpoint(q0[2 * pole_step].x, q1[1 * pole_step].x);
|
|
|
1296 |
q0[3 * pole_step].y = q1[0 * pole_step].y = midpoint(q0[2 * pole_step].y, q1[1 * pole_step].y);
|
|
|
1297 |
q1[3 * pole_step].x = pole[3 * pole_step].x;
|
|
|
1298 |
q1[3 * pole_step].y = pole[3 * pole_step].y;
|
|
|
1299 |
#undef midpoint
|
|
|
1300 |
}
|
|
|
1301 |
|
|
|
1302 |
private void
|
|
|
1303 |
split_curve(const gs_fixed_point pole[4], gs_fixed_point q0[4], gs_fixed_point q1[4])
|
|
|
1304 |
{
|
|
|
1305 |
split_curve_s(pole, q0, q1, 1);
|
|
|
1306 |
}
|
|
|
1307 |
|
|
|
1308 |
|
|
|
1309 |
private void
|
|
|
1310 |
generate_inner_vertices(gs_fixed_point *p, const gs_fixed_point pole[4], int k)
|
|
|
1311 |
{
|
|
|
1312 |
/* Recure to get exactly same points as when devided a patch. */
|
|
|
1313 |
/* An iteration can't give them preciselly. */
|
|
|
1314 |
if (k > 1) {
|
|
|
1315 |
gs_fixed_point q[2][4];
|
|
|
1316 |
|
|
|
1317 |
split_curve(pole, q[0], q[1]);
|
|
|
1318 |
p[k / 2] = q[0][3];
|
|
|
1319 |
generate_inner_vertices(p, q[0], k / 2);
|
|
|
1320 |
generate_inner_vertices(p + k / 2, q[1], k / 2);
|
|
|
1321 |
}
|
|
|
1322 |
}
|
|
|
1323 |
|
|
|
1324 |
private inline void
|
|
|
1325 |
do_swap_axes(gs_fixed_point *p, int k)
|
|
|
1326 |
{
|
|
|
1327 |
int i;
|
|
|
1328 |
|
|
|
1329 |
for (i = 0; i < k; i++) {
|
|
|
1330 |
p[i].x ^= p[i].y; p[i].y ^= p[i].x; p[i].x ^= p[i].y;
|
|
|
1331 |
}
|
|
|
1332 |
}
|
|
|
1333 |
|
|
|
1334 |
private inline void
|
|
|
1335 |
y_extreme_vertice(gs_fixed_point *q, const gs_fixed_point *p, int k, int minmax)
|
|
|
1336 |
{
|
|
|
1337 |
int i;
|
|
|
1338 |
gs_fixed_point r = *p;
|
|
|
1339 |
|
|
|
1340 |
for (i = 1; i < k; i++)
|
|
|
1341 |
if ((p[i].y - r.y) * minmax > 0)
|
|
|
1342 |
r = p[i];
|
|
|
1343 |
*q = r;
|
|
|
1344 |
}
|
|
|
1345 |
|
|
|
1346 |
private inline fixed
|
|
|
1347 |
span_x(const gs_fixed_point *p, int k)
|
|
|
1348 |
{
|
|
|
1349 |
int i;
|
|
|
1350 |
fixed xmin = p[0].x, xmax = p[0].x;
|
|
|
1351 |
|
|
|
1352 |
for (i = 1; i < k; i++) {
|
|
|
1353 |
xmin = min(xmin, p[i].x);
|
|
|
1354 |
xmax = max(xmax, p[i].x);
|
|
|
1355 |
}
|
|
|
1356 |
return xmax - xmin;
|
|
|
1357 |
}
|
|
|
1358 |
|
|
|
1359 |
private inline fixed
|
|
|
1360 |
span_y(const gs_fixed_point *p, int k)
|
|
|
1361 |
{
|
|
|
1362 |
int i;
|
|
|
1363 |
fixed ymin = p[0].y, ymax = p[0].y;
|
|
|
1364 |
|
|
|
1365 |
for (i = 1; i < k; i++) {
|
|
|
1366 |
ymin = min(ymin, p[i].y);
|
|
|
1367 |
ymax = max(ymax, p[i].y);
|
|
|
1368 |
}
|
|
|
1369 |
return ymax - ymin;
|
|
|
1370 |
}
|
|
|
1371 |
|
|
|
1372 |
private inline void
|
|
|
1373 |
draw_wedge(const gs_fixed_point *p, int n)
|
|
|
1374 |
{
|
|
|
1375 |
#ifdef DEBUG
|
|
|
1376 |
int i;
|
|
|
1377 |
|
|
|
1378 |
if (!vd_enabled)
|
|
|
1379 |
return;
|
|
|
1380 |
vd_setlinewidth(4);
|
|
|
1381 |
vd_setcolor(RGB(255, 0, 0));
|
|
|
1382 |
vd_beg_path;
|
|
|
1383 |
vd_moveto(p[0].x, p[0].y);
|
|
|
1384 |
for (i = 1; i < n; i++)
|
|
|
1385 |
vd_lineto(p[i].x, p[i].y);
|
|
|
1386 |
vd_closepath;
|
|
|
1387 |
vd_end_path;
|
|
|
1388 |
vd_fill;
|
|
|
1389 |
/*vd_stroke;*/
|
|
|
1390 |
#endif
|
|
|
1391 |
}
|
|
|
1392 |
|
|
|
1393 |
private inline fixed
|
|
|
1394 |
manhattan_dist(const gs_fixed_point *p0, const gs_fixed_point *p1)
|
|
|
1395 |
{
|
|
|
1396 |
fixed dx = any_abs(p1->x - p0->x), dy = any_abs(p1->y - p0->y);
|
|
|
1397 |
|
|
|
1398 |
return max(dx, dy);
|
|
|
1399 |
}
|
|
|
1400 |
|
|
|
1401 |
private inline void
|
|
|
1402 |
create_wedge_vertex_list(patch_fill_state_t *pfs, wedge_vertex_list_t *l,
|
|
|
1403 |
const gs_fixed_point *p0, const gs_fixed_point *p1)
|
|
|
1404 |
{
|
|
|
1405 |
assert(l->end == NULL);
|
|
|
1406 |
l->beg = wedge_vertex_list_elem_reserve(pfs);
|
|
|
1407 |
l->end = wedge_vertex_list_elem_reserve(pfs);
|
|
|
1408 |
assert(l->beg != NULL);
|
|
|
1409 |
assert(l->end != NULL);
|
|
|
1410 |
l->beg->prev = l->end->next = NULL;
|
|
|
1411 |
l->beg->next = l->end;
|
|
|
1412 |
l->end->prev = l->beg;
|
|
|
1413 |
l->beg->p = *p0;
|
|
|
1414 |
l->end->p = *p1;
|
|
|
1415 |
l->beg->level = l->end->level = 0;
|
|
|
1416 |
}
|
|
|
1417 |
|
|
|
1418 |
private inline wedge_vertex_list_elem_t *
|
|
|
1419 |
insert_wedge_vertex_list_elem(patch_fill_state_t *pfs, wedge_vertex_list_t *l, const gs_fixed_point *p)
|
|
|
1420 |
{
|
|
|
1421 |
wedge_vertex_list_elem_t *e = wedge_vertex_list_elem_reserve(pfs);
|
|
|
1422 |
|
|
|
1423 |
/* We have got enough free elements due to the preliminary decomposition
|
|
|
1424 |
of curves to LAZY_WEDGES_MAX_LEVEL, see curve_samples. */
|
|
|
1425 |
assert(e != NULL);
|
|
|
1426 |
assert(l->beg->next == l->end);
|
|
|
1427 |
assert(l->end->prev == l->beg);
|
|
|
1428 |
e->next = l->end;
|
|
|
1429 |
e->prev = l->beg;
|
|
|
1430 |
e->p = *p;
|
|
|
1431 |
e->level = max(l->beg->level, l->end->level) + 1;
|
|
|
1432 |
e->divide_count = 0;
|
|
|
1433 |
l->beg->next = l->end->prev = e;
|
|
|
1434 |
{ int sx = l->beg->p.x < l->end->p.x ? 1 : -1;
|
|
|
1435 |
int sy = l->beg->p.y < l->end->p.y ? 1 : -1;
|
|
|
1436 |
|
|
|
1437 |
assert((p->x - l->beg->p.x) * sx >= 0);
|
|
|
1438 |
assert((p->y - l->beg->p.y) * sy >= 0);
|
|
|
1439 |
assert((l->end->p.x - p->x) * sx >= 0);
|
|
|
1440 |
assert((l->end->p.y - p->y) * sy >= 0);
|
|
|
1441 |
}
|
|
|
1442 |
return e;
|
|
|
1443 |
}
|
|
|
1444 |
|
|
|
1445 |
private inline wedge_vertex_list_elem_t *
|
|
|
1446 |
open_wedge_median(patch_fill_state_t *pfs, wedge_vertex_list_t *l,
|
|
|
1447 |
const gs_fixed_point *p0, const gs_fixed_point *p1, const gs_fixed_point *pm)
|
|
|
1448 |
{
|
|
|
1449 |
wedge_vertex_list_elem_t *e;
|
|
|
1450 |
|
|
|
1451 |
if (!l->last_side) {
|
|
|
1452 |
if (l->beg == NULL)
|
|
|
1453 |
create_wedge_vertex_list(pfs, l, p0, p1);
|
|
|
1454 |
assert(l->beg->p.x == p0->x);
|
|
|
1455 |
assert(l->beg->p.y == p0->y);
|
|
|
1456 |
assert(l->end->p.x == p1->x);
|
|
|
1457 |
assert(l->end->p.y == p1->y);
|
|
|
1458 |
e = insert_wedge_vertex_list_elem(pfs, l, pm);
|
|
|
1459 |
e->divide_count++;
|
|
|
1460 |
return e;
|
|
|
1461 |
} else {
|
|
|
1462 |
if (l->beg == NULL) {
|
|
|
1463 |
create_wedge_vertex_list(pfs, l, p1, p0);
|
|
|
1464 |
e = insert_wedge_vertex_list_elem(pfs, l, pm);
|
|
|
1465 |
e->divide_count++;
|
|
|
1466 |
return e;
|
|
|
1467 |
}
|
|
|
1468 |
assert(l->beg->p.x == p1->x);
|
|
|
1469 |
assert(l->beg->p.y == p1->y);
|
|
|
1470 |
assert(l->end->p.x == p0->x);
|
|
|
1471 |
assert(l->end->p.y == p0->y);
|
|
|
1472 |
if (l->beg->next == l->end) {
|
|
|
1473 |
e = insert_wedge_vertex_list_elem(pfs, l, pm);
|
|
|
1474 |
e->divide_count++;
|
|
|
1475 |
return e;
|
|
|
1476 |
} else {
|
|
|
1477 |
e = wedge_vertex_list_find(l->beg, l->end,
|
|
|
1478 |
max(l->beg->level, l->end->level) + 1);
|
|
|
1479 |
assert(e != NULL);
|
|
|
1480 |
assert(e->p.x == pm->x && e->p.y == pm->y);
|
|
|
1481 |
e->divide_count++;
|
|
|
1482 |
return e;
|
|
|
1483 |
}
|
|
|
1484 |
}
|
|
|
1485 |
}
|
|
|
1486 |
|
|
|
1487 |
private inline void
|
|
|
1488 |
make_wedge_median(patch_fill_state_t *pfs, wedge_vertex_list_t *l,
|
|
|
1489 |
wedge_vertex_list_t *l0, bool forth,
|
|
|
1490 |
const gs_fixed_point *p0, const gs_fixed_point *p1, const gs_fixed_point *pm)
|
|
|
1491 |
{
|
|
|
1492 |
l->last_side = l0->last_side;
|
|
|
1493 |
if (!l->last_side ^ !forth) {
|
|
|
1494 |
l->end = open_wedge_median(pfs, l0, p0, p1, pm);
|
|
|
1495 |
l->beg = l0->beg;
|
|
|
1496 |
} else {
|
|
|
1497 |
l->beg = open_wedge_median(pfs, l0, p0, p1, pm);
|
|
|
1498 |
l->end = l0->end;
|
|
|
1499 |
}
|
|
|
1500 |
}
|
|
|
1501 |
|
|
|
1502 |
private int fill_wedge_from_list(patch_fill_state_t *pfs, const wedge_vertex_list_t *l,
|
|
|
1503 |
const patch_color_t *c0, const patch_color_t *c1);
|
|
|
1504 |
|
|
|
1505 |
private inline int
|
|
|
1506 |
close_wedge_median(patch_fill_state_t *pfs, wedge_vertex_list_t *l,
|
|
|
1507 |
const patch_color_t *c0, const patch_color_t *c1)
|
|
|
1508 |
{
|
|
|
1509 |
int code;
|
|
|
1510 |
|
|
|
1511 |
if (!l->last_side)
|
|
|
1512 |
return 0;
|
|
|
1513 |
code = fill_wedge_from_list(pfs, l, c1, c0);
|
|
|
1514 |
if (code < 0)
|
|
|
1515 |
return code;
|
|
|
1516 |
release_wedge_vertex_list_interval(pfs, l->beg, l->end);
|
|
|
1517 |
return 0;
|
|
|
1518 |
}
|
|
|
1519 |
|
|
|
1520 |
private inline void
|
|
|
1521 |
move_wedge(wedge_vertex_list_t *l, const wedge_vertex_list_t *l0, bool forth)
|
|
|
1522 |
{
|
|
|
1523 |
if (!l->last_side ^ !forth) {
|
|
|
1524 |
l->beg = l->end;
|
|
|
1525 |
l->end = l0->end;
|
|
|
1526 |
} else {
|
|
|
1527 |
l->end = l->beg;
|
|
|
1528 |
l->beg = l0->beg;
|
|
|
1529 |
}
|
|
|
1530 |
}
|
|
|
1531 |
|
|
|
1532 |
private inline int
|
|
|
1533 |
fill_triangle_wedge_aux(patch_fill_state_t *pfs,
|
|
|
1534 |
const shading_vertex_t *q0, const shading_vertex_t *q1, const shading_vertex_t *q2)
|
|
|
1535 |
{ int code;
|
|
|
1536 |
const gs_fixed_point *p0, *p1, *p2;
|
|
|
1537 |
gs_fixed_point qq0, qq1, qq2;
|
|
|
1538 |
fixed dx = any_abs(q0->p.x - q1->p.x), dy = any_abs(q0->p.y - q1->p.y);
|
|
|
1539 |
bool swap_axes;
|
|
|
1540 |
|
|
|
1541 |
# if SKIP_TEST
|
|
|
1542 |
dbg_wedge_triangle_cnt++;
|
|
|
1543 |
# endif
|
|
|
1544 |
if (dx > dy) {
|
|
|
1545 |
swap_axes = true;
|
|
|
1546 |
qq0.x = q0->p.y;
|
|
|
1547 |
qq0.y = q0->p.x;
|
|
|
1548 |
qq1.x = q1->p.y;
|
|
|
1549 |
qq1.y = q1->p.x;
|
|
|
1550 |
qq2.x = q2->p.y;
|
|
|
1551 |
qq2.y = q2->p.x;
|
|
|
1552 |
p0 = &qq0;
|
|
|
1553 |
p1 = &qq1;
|
|
|
1554 |
p2 = &qq2;
|
|
|
1555 |
} else {
|
|
|
1556 |
swap_axes = false;
|
|
|
1557 |
p0 = &q0->p;
|
|
|
1558 |
p1 = &q1->p;
|
|
|
1559 |
p2 = &q2->p;
|
|
|
1560 |
}
|
|
|
1561 |
/* We decompose the thin triangle into 2 thin trapezoids.
|
|
|
1562 |
An optimization with decomposing into 2 triangles
|
|
|
1563 |
appears low useful, because the self_intersecting argument
|
|
|
1564 |
with inline expansion does that job perfectly. */
|
|
|
1565 |
if (p0->y < p1->y) {
|
|
|
1566 |
code = fill_wedge_trap(pfs, p0, p2, p0, p1, &q0->c, &q2->c, swap_axes, false);
|
|
|
1567 |
if (code < 0)
|
|
|
1568 |
return code;
|
|
|
1569 |
return fill_wedge_trap(pfs, p2, p1, p0, p1, &q2->c, &q1->c, swap_axes, false);
|
|
|
1570 |
} else {
|
|
|
1571 |
code = fill_wedge_trap(pfs, p0, p2, p1, p0, &q0->c, &q2->c, swap_axes, false);
|
|
|
1572 |
if (code < 0)
|
|
|
1573 |
return code;
|
|
|
1574 |
return fill_wedge_trap(pfs, p2, p1, p1, p0, &q2->c, &q1->c, swap_axes, false);
|
|
|
1575 |
}
|
|
|
1576 |
}
|
|
|
1577 |
|
|
|
1578 |
private inline int
|
|
|
1579 |
try_device_linear_color(patch_fill_state_t *pfs, bool wedge,
|
|
|
1580 |
const shading_vertex_t *p0, const shading_vertex_t *p1,
|
|
|
1581 |
const shading_vertex_t *p2)
|
|
|
1582 |
{
|
|
|
1583 |
/* Returns :
|
|
|
1584 |
<0 - error;
|
|
|
1585 |
|
|
|
1586 |
1 - decompose to linear color areas;
|
|
|
1587 |
2 - decompose to constant color areas;
|
|
|
1588 |
*/
|
|
|
1589 |
int code;
|
|
|
1590 |
|
|
|
1591 |
if (pfs->unlinear)
|
|
|
1592 |
return 2;
|
|
|
1593 |
if (!wedge) {
|
|
|
1594 |
gs_direct_color_space *cs =
|
|
|
1595 |
(gs_direct_color_space *)pfs->direct_space; /* break 'const'. */
|
|
|
1596 |
float smoothness = max(pfs->smoothness, 1.0 / min_linear_grades);
|
|
|
1597 |
/* Restrict the smoothness with 1/min_linear_grades, because cs_is_linear
|
|
|
1598 |
can't provide a better precision due to the color
|
|
|
1599 |
representation with integers.
|
|
|
1600 |
*/
|
|
|
1601 |
float s0, s1, s2, s01, s012;
|
|
|
1602 |
|
|
|
1603 |
s0 = function_linearity(pfs, &p0->c, &p1->c);
|
|
|
1604 |
if (s0 > smoothness)
|
|
|
1605 |
return 1;
|
|
|
1606 |
s1 = function_linearity(pfs, &p1->c, &p2->c);
|
|
|
1607 |
if (s1 > smoothness)
|
|
|
1608 |
return 1;
|
|
|
1609 |
s2 = function_linearity(pfs, &p2->c, &p0->c);
|
|
|
1610 |
if (s2 > smoothness)
|
|
|
1611 |
return 1;
|
|
|
1612 |
/* fixme: check an inner color ? */
|
|
|
1613 |
s01 = max(s0, s1);
|
|
|
1614 |
s012 = max(s01, s2);
|
|
|
1615 |
code = cs_is_linear(cs, pfs->pis, pfs->dev,
|
|
|
1616 |
&p0->c.cc, &p1->c.cc, &p2->c.cc, NULL, smoothness - s012);
|
|
|
1617 |
if (code < 0)
|
|
|
1618 |
return code;
|
|
|
1619 |
if (code == 0)
|
|
|
1620 |
return 1;
|
|
|
1621 |
}
|
|
|
1622 |
{ gx_device *pdev = pfs->dev;
|
|
|
1623 |
frac31 fc[3][GX_DEVICE_COLOR_MAX_COMPONENTS];
|
|
|
1624 |
gs_fill_attributes fa;
|
|
|
1625 |
gx_device_color dc[3];
|
|
|
1626 |
vd_save;
|
|
|
1627 |
|
|
|
1628 |
fa.clip = &pfs->rect;
|
|
|
1629 |
fa.ht = NULL;
|
|
|
1630 |
fa.swap_axes = false;
|
|
|
1631 |
fa.lop = 0;
|
|
|
1632 |
code = patch_color_to_device_color(pfs, &p0->c, &dc[0]);
|
|
|
1633 |
if (code < 0)
|
|
|
1634 |
return code;
|
|
|
1635 |
if (dc[0].type != &gx_dc_type_data_pure)
|
|
|
1636 |
return 2;
|
|
|
1637 |
dc2fc(pfs, dc[0].colors.pure, fc[0]);
|
|
|
1638 |
if (!wedge) {
|
|
|
1639 |
code = patch_color_to_device_color(pfs, &p1->c, &dc[1]);
|
|
|
1640 |
if (code < 0)
|
|
|
1641 |
return code;
|
|
|
1642 |
dc2fc(pfs, dc[1].colors.pure, fc[1]);
|
|
|
1643 |
}
|
|
|
1644 |
code = patch_color_to_device_color(pfs, &p2->c, &dc[2]);
|
|
|
1645 |
if (code < 0)
|
|
|
1646 |
return code;
|
|
|
1647 |
dc2fc(pfs, dc[2].colors.pure, fc[2]);
|
|
|
1648 |
draw_triangle(&p0->p, &p1->p, &p2->p, RGB(255, 0, 0));
|
|
|
1649 |
if (!VD_TRACE_DOWN)
|
|
|
1650 |
vd_disable;
|
|
|
1651 |
code = dev_proc(pdev, fill_linear_color_triangle)(pdev, &fa,
|
|
|
1652 |
&p0->p, &p1->p, &p2->p,
|
|
|
1653 |
fc[0], (wedge ? NULL : fc[1]), fc[2]);
|
|
|
1654 |
vd_restore;
|
|
|
1655 |
if (code == 1)
|
|
|
1656 |
return 0; /* The area is filled. */
|
|
|
1657 |
if (code < 0)
|
|
|
1658 |
return code;
|
|
|
1659 |
else /* code == 0, the device requested to decompose the area. */
|
|
|
1660 |
return 1;
|
|
|
1661 |
}
|
|
|
1662 |
}
|
|
|
1663 |
|
|
|
1664 |
private inline int
|
|
|
1665 |
fill_triangle_wedge(patch_fill_state_t *pfs,
|
|
|
1666 |
const shading_vertex_t *q0, const shading_vertex_t *q1, const shading_vertex_t *q2)
|
|
|
1667 |
{
|
|
|
1668 |
if ((int64_t)(q1->p.x - q0->p.x) * (q2->p.y - q0->p.y) ==
|
|
|
1669 |
(int64_t)(q1->p.y - q0->p.y) * (q2->p.x - q0->p.x))
|
|
|
1670 |
return 0; /* Zero area. */
|
|
|
1671 |
draw_triangle(&q0->p, &q1->p, &q2->p, RGB(255, 255, 0));
|
|
|
1672 |
/*
|
|
|
1673 |
Can't apply try_device_linear_color here
|
|
|
1674 |
because didn't check is_color_linear.
|
|
|
1675 |
Maybe need a decomposition.
|
|
|
1676 |
Do same as for 'unlinear', and branch later.
|
|
|
1677 |
*/
|
|
|
1678 |
return fill_triangle_wedge_aux(pfs, q0, q1, q2);
|
|
|
1679 |
}
|
|
|
1680 |
|
|
|
1681 |
private inline int
|
|
|
1682 |
fill_triangle_wedge_from_list(patch_fill_state_t *pfs,
|
|
|
1683 |
const wedge_vertex_list_elem_t *beg, const wedge_vertex_list_elem_t *end,
|
|
|
1684 |
const wedge_vertex_list_elem_t *mid,
|
|
|
1685 |
const patch_color_t *c0, const patch_color_t *c1)
|
|
|
1686 |
{
|
|
|
1687 |
shading_vertex_t p[3];
|
|
|
1688 |
|
|
|
1689 |
p[0].p = beg->p;
|
|
|
1690 |
p[0].c = *c0; /* fixme : unhappy copying colors. */
|
|
|
1691 |
p[1].p = end->p;
|
|
|
1692 |
p[1].c = *c1;
|
|
|
1693 |
p[2].p = mid->p;
|
|
|
1694 |
patch_interpolate_color(&p[2].c, c0, c1, pfs, 0.5);
|
|
|
1695 |
return fill_triangle_wedge(pfs, &p[0], &p[1], &p[2]);
|
|
|
1696 |
}
|
|
|
1697 |
|
|
|
1698 |
private int
|
|
|
1699 |
fill_wedge_from_list_rec(patch_fill_state_t *pfs,
|
|
|
1700 |
wedge_vertex_list_elem_t *beg, const wedge_vertex_list_elem_t *end,
|
|
|
1701 |
int level, const patch_color_t *c0, const patch_color_t *c1)
|
|
|
1702 |
{
|
|
|
1703 |
if (beg->next == end)
|
|
|
1704 |
return 0;
|
|
|
1705 |
else if (beg->next->next == end) {
|
|
|
1706 |
assert(beg->next->divide_count == 1 || beg->next->divide_count == 2);
|
|
|
1707 |
if (beg->next->divide_count != 1)
|
|
|
1708 |
return 0;
|
|
|
1709 |
return fill_triangle_wedge_from_list(pfs, beg, end, beg->next, c0, c1);
|
|
|
1710 |
} else {
|
|
|
1711 |
gs_fixed_point p;
|
|
|
1712 |
wedge_vertex_list_elem_t *e;
|
|
|
1713 |
patch_color_t c;
|
|
|
1714 |
int code;
|
|
|
1715 |
|
|
|
1716 |
p.x = (beg->p.x + end->p.x) / 2;
|
|
|
1717 |
p.y = (beg->p.y + end->p.y) / 2;
|
|
|
1718 |
e = wedge_vertex_list_find(beg, end, level + 1);
|
|
|
1719 |
assert(e != NULL);
|
|
|
1720 |
assert(e->p.x == p.x && e->p.y == p.y);
|
|
|
1721 |
patch_interpolate_color(&c, c0, c1, pfs, 0.5);
|
|
|
1722 |
code = fill_wedge_from_list_rec(pfs, beg, e, level + 1, c0, &c);
|
|
|
1723 |
if (code < 0)
|
|
|
1724 |
return code;
|
|
|
1725 |
code = fill_wedge_from_list_rec(pfs, e, end, level + 1, &c, c1);
|
|
|
1726 |
if (code < 0)
|
|
|
1727 |
return code;
|
|
|
1728 |
assert(e->divide_count == 1 || e->divide_count == 2);
|
|
|
1729 |
if (e->divide_count != 1)
|
|
|
1730 |
return 0;
|
|
|
1731 |
return fill_triangle_wedge_from_list(pfs, beg, end, e, c0, c1);
|
|
|
1732 |
}
|
|
|
1733 |
}
|
|
|
1734 |
|
|
|
1735 |
private int
|
|
|
1736 |
fill_wedge_from_list(patch_fill_state_t *pfs, const wedge_vertex_list_t *l,
|
|
|
1737 |
const patch_color_t *c0, const patch_color_t *c1)
|
|
|
1738 |
{
|
|
|
1739 |
return fill_wedge_from_list_rec(pfs, l->beg, l->end,
|
|
|
1740 |
max(l->beg->level, l->end->level), c0, c1);
|
|
|
1741 |
}
|
|
|
1742 |
|
|
|
1743 |
private inline int
|
|
|
1744 |
terminate_wedge_vertex_list(patch_fill_state_t *pfs, wedge_vertex_list_t *l,
|
|
|
1745 |
const patch_color_t *c0, const patch_color_t *c1)
|
|
|
1746 |
{
|
|
|
1747 |
if (l->beg != NULL) {
|
|
|
1748 |
int code = fill_wedge_from_list(pfs, l, c0, c1);
|
|
|
1749 |
|
|
|
1750 |
if (code < 0)
|
|
|
1751 |
return code;
|
|
|
1752 |
release_wedge_vertex_list(pfs, l, 1);
|
|
|
1753 |
}
|
|
|
1754 |
return 0;
|
|
|
1755 |
}
|
|
|
1756 |
|
|
|
1757 |
private int
|
|
|
1758 |
wedge_by_triangles(patch_fill_state_t *pfs, int ka,
|
|
|
1759 |
const gs_fixed_point pole[4], const patch_color_t *c0, const patch_color_t *c1)
|
|
|
1760 |
{ /* Assuming ka >= 2, see fill_wedges. */
|
|
|
1761 |
gs_fixed_point q[2][4];
|
|
|
1762 |
shading_vertex_t p[3];
|
|
|
1763 |
int code;
|
|
|
1764 |
|
|
|
1765 |
split_curve(pole, q[0], q[1]);
|
|
|
1766 |
p[0].p = pole[0];
|
|
|
1767 |
p[0].c = *c0; /* fixme : unhappy copying colors. */
|
|
|
1768 |
p[1].p = pole[3];
|
|
|
1769 |
p[1].c = *c1;
|
|
|
1770 |
p[2].p = q[0][3];
|
|
|
1771 |
patch_interpolate_color(&p[2].c, c0, c1, pfs, 0.5);
|
|
|
1772 |
code = fill_triangle_wedge(pfs, &p[0], &p[1], &p[2]);
|
|
|
1773 |
if (code < 0)
|
|
|
1774 |
return code;
|
|
|
1775 |
if (ka == 2)
|
|
|
1776 |
return 0;
|
|
|
1777 |
code = wedge_by_triangles(pfs, ka / 2, q[0], c0, &p[2].c);
|
|
|
1778 |
if (code < 0)
|
|
|
1779 |
return code;
|
|
|
1780 |
return wedge_by_triangles(pfs, ka / 2, q[1], &p[2].c, c1);
|
|
|
1781 |
}
|
|
|
1782 |
|
|
|
1783 |
private inline bool
|
|
|
1784 |
is_linear_color_applicable(const patch_fill_state_t *pfs)
|
|
|
1785 |
{
|
|
|
1786 |
if (!USE_LINEAR_COLOR_PROCS)
|
|
|
1787 |
return false;
|
|
|
1788 |
if (pfs->dev->color_info.separable_and_linear != GX_CINFO_SEP_LIN)
|
|
|
1789 |
return false;
|
|
|
1790 |
if (gx_get_cmap_procs(pfs->pis, pfs->dev)->is_halftoned(pfs->pis, pfs->dev))
|
|
|
1791 |
return false;
|
|
|
1792 |
return true;
|
|
|
1793 |
}
|
|
|
1794 |
|
|
|
1795 |
int
|
|
|
1796 |
mesh_padding(patch_fill_state_t *pfs, const gs_fixed_point *p0, const gs_fixed_point *p1,
|
|
|
1797 |
const patch_color_t *c0, const patch_color_t *c1)
|
|
|
1798 |
{
|
|
|
1799 |
gs_fixed_point q0, q1;
|
|
|
1800 |
const patch_color_t *cc0, *cc1;
|
|
|
1801 |
fixed dx = p1->x - p0->x;
|
|
|
1802 |
fixed dy = p1->y - p0->y;
|
|
|
1803 |
bool swap_axes = (any_abs(dx) > any_abs(dy));
|
|
|
1804 |
gs_fixed_edge le, re;
|
|
|
1805 |
const fixed adjust = INTERPATCH_PADDING;
|
|
|
1806 |
|
|
|
1807 |
pfs->unlinear = !is_linear_color_applicable(pfs);
|
|
|
1808 |
if (swap_axes) {
|
|
|
1809 |
if (p0->x < p1->x) {
|
|
|
1810 |
q0.x = p0->y;
|
|
|
1811 |
q0.y = p0->x;
|
|
|
1812 |
q1.x = p1->y;
|
|
|
1813 |
q1.y = p1->x;
|
|
|
1814 |
cc0 = c0;
|
|
|
1815 |
cc1 = c1;
|
|
|
1816 |
} else {
|
|
|
1817 |
q0.x = p1->y;
|
|
|
1818 |
q0.y = p1->x;
|
|
|
1819 |
q1.x = p0->y;
|
|
|
1820 |
q1.y = p0->x;
|
|
|
1821 |
cc0 = c1;
|
|
|
1822 |
cc1 = c0;
|
|
|
1823 |
}
|
|
|
1824 |
} else if (p0->y < p1->y) {
|
|
|
1825 |
q0 = *p0;
|
|
|
1826 |
q1 = *p1;
|
|
|
1827 |
cc0 = c0;
|
|
|
1828 |
cc1 = c1;
|
|
|
1829 |
} else {
|
|
|
1830 |
q0 = *p1;
|
|
|
1831 |
q1 = *p0;
|
|
|
1832 |
cc0 = c1;
|
|
|
1833 |
cc1 = c0;
|
|
|
1834 |
}
|
|
|
1835 |
le.start.x = q0.x - adjust;
|
|
|
1836 |
re.start.x = q0.x + adjust;
|
|
|
1837 |
le.start.y = re.start.y = q0.y - adjust;
|
|
|
1838 |
le.end.x = q1.x - adjust;
|
|
|
1839 |
re.end.x = q1.x + adjust;
|
|
|
1840 |
le.end.y = re.end.y = q1.y + adjust;
|
|
|
1841 |
adjust_swapped_boundary(&re.start.x, swap_axes);
|
|
|
1842 |
adjust_swapped_boundary(&re.end.x, swap_axes);
|
|
|
1843 |
return decompose_linear_color(pfs, &le, &re, le.start.y, le.end.y, swap_axes, cc0, cc1, 0);
|
|
|
1844 |
/* fixme : for a better performance and quality, we would like to
|
|
|
1845 |
consider the bar as an oriented one and to know at what side of it the spot resides.
|
|
|
1846 |
If we know that, we could expand only to outside the spot.
|
|
|
1847 |
Note that if the boundary has a self-intersection,
|
|
|
1848 |
we still need to expand to both directions.
|
|
|
1849 |
*/
|
|
|
1850 |
}
|
|
|
1851 |
|
|
|
1852 |
private int
|
|
|
1853 |
fill_wedges_aux(patch_fill_state_t *pfs, int k, int ka,
|
|
|
1854 |
const gs_fixed_point pole[4], const patch_color_t *c0, const patch_color_t *c1,
|
|
|
1855 |
int wedge_type)
|
|
|
1856 |
{
|
|
|
1857 |
int code;
|
|
|
1858 |
|
|
|
1859 |
if (k > 1) {
|
|
|
1860 |
gs_fixed_point q[2][4];
|
|
|
1861 |
patch_color_t c;
|
|
|
1862 |
|
|
|
1863 |
patch_interpolate_color(&c, c0, c1, pfs, 0.5);
|
|
|
1864 |
split_curve(pole, q[0], q[1]);
|
|
|
1865 |
code = fill_wedges_aux(pfs, k / 2, ka, q[0], c0, &c, wedge_type);
|
|
|
1866 |
if (code < 0)
|
|
|
1867 |
return code;
|
|
|
1868 |
return fill_wedges_aux(pfs, k / 2, ka, q[1], &c, c1, wedge_type);
|
|
|
1869 |
} else {
|
|
|
1870 |
if (INTERPATCH_PADDING && (wedge_type & interpatch_padding)) {
|
|
|
1871 |
vd_bar(pole[0].x, pole[0].y, pole[3].x, pole[3].y, 0, RGB(255, 0, 0));
|
|
|
1872 |
code = mesh_padding(pfs, &pole[0], &pole[3], c0, c1);
|
|
|
1873 |
if (code < 0)
|
|
|
1874 |
return code;
|
|
|
1875 |
}
|
|
|
1876 |
if (ka >= 2 && (wedge_type & inpatch_wedge))
|
|
|
1877 |
return wedge_by_triangles(pfs, ka, pole, c0, c1);
|
|
|
1878 |
return 0;
|
|
|
1879 |
}
|
|
|
1880 |
}
|
|
|
1881 |
|
|
|
1882 |
private int
|
|
|
1883 |
fill_wedges(patch_fill_state_t *pfs, int k0, int k1,
|
|
|
1884 |
const gs_fixed_point *pole, int pole_step,
|
|
|
1885 |
const patch_color_t *c0, const patch_color_t *c1, int wedge_type)
|
|
|
1886 |
{
|
|
|
1887 |
/* Generate wedges between 2 variants of a curve flattening. */
|
|
|
1888 |
/* k0, k1 is a power of 2. */
|
|
|
1889 |
gs_fixed_point p[4];
|
|
|
1890 |
|
|
|
1891 |
if (!(wedge_type & interpatch_padding) && k0 == k1)
|
|
|
1892 |
return 0; /* Wedges are zero area. */
|
|
|
1893 |
if (k0 > k1) {
|
|
|
1894 |
k0 ^= k1; k1 ^= k0; k0 ^= k1;
|
|
|
1895 |
}
|
|
|
1896 |
p[0] = pole[0];
|
|
|
1897 |
p[1] = pole[pole_step];
|
|
|
1898 |
p[2] = pole[pole_step * 2];
|
|
|
1899 |
p[3] = pole[pole_step * 3];
|
|
|
1900 |
return fill_wedges_aux(pfs, k0, k1 / k0, p, c0, c1, wedge_type);
|
|
|
1901 |
}
|
|
|
1902 |
|
|
|
1903 |
private inline void
|
|
|
1904 |
make_vertices(gs_fixed_point q[4], const quadrangle_patch *p)
|
|
|
1905 |
{
|
|
|
1906 |
q[0] = p->p[0][0]->p;
|
|
|
1907 |
q[1] = p->p[0][1]->p;
|
|
|
1908 |
q[2] = p->p[1][1]->p;
|
|
|
1909 |
q[3] = p->p[1][0]->p;
|
|
|
1910 |
}
|
|
|
1911 |
|
|
|
1912 |
private inline void
|
|
|
1913 |
wrap_vertices_by_y(gs_fixed_point q[4], const gs_fixed_point s[4])
|
|
|
1914 |
{
|
|
|
1915 |
fixed y = s[0].y;
|
|
|
1916 |
int i = 0;
|
|
|
1917 |
|
|
|
1918 |
if (y > s[1].y)
|
|
|
1919 |
i = 1, y = s[1].y;
|
|
|
1920 |
if (y > s[2].y)
|
|
|
1921 |
i = 2, y = s[2].y;
|
|
|
1922 |
if (y > s[3].y)
|
|
|
1923 |
i = 3, y = s[3].y;
|
|
|
1924 |
q[0] = s[(i + 0) % 4];
|
|
|
1925 |
q[1] = s[(i + 1) % 4];
|
|
|
1926 |
q[2] = s[(i + 2) % 4];
|
|
|
1927 |
q[3] = s[(i + 3) % 4];
|
|
|
1928 |
}
|
|
|
1929 |
|
|
|
1930 |
private int
|
|
|
1931 |
ordered_triangle(patch_fill_state_t *pfs, gs_fixed_edge *le, gs_fixed_edge *re, patch_color_t *c)
|
|
|
1932 |
{
|
|
|
1933 |
gs_fixed_edge ue;
|
|
|
1934 |
int code;
|
|
|
1935 |
gx_device_color dc;
|
|
|
1936 |
vd_save;
|
|
|
1937 |
|
|
|
1938 |
# if NOFILL_TEST
|
|
|
1939 |
if (dbg_nofill)
|
|
|
1940 |
return 0;
|
|
|
1941 |
# endif
|
|
|
1942 |
if (!VD_TRACE_DOWN)
|
|
|
1943 |
vd_disable;
|
|
|
1944 |
code = patch_color_to_device_color(pfs, c, &dc);
|
|
|
1945 |
if (code < 0)
|
|
|
1946 |
return code;
|
|
|
1947 |
if (le->end.y < re->end.y) {
|
|
|
1948 |
code = dev_proc(pfs->dev, fill_trapezoid)(pfs->dev,
|
|
|
1949 |
le, re, le->start.y, le->end.y, false, &dc, pfs->pis->log_op);
|
|
|
1950 |
if (code >= 0) {
|
|
|
1951 |
ue.start = le->end;
|
|
|
1952 |
ue.end = re->end;
|
|
|
1953 |
code = dev_proc(pfs->dev, fill_trapezoid)(pfs->dev,
|
|
|
1954 |
&ue, re, le->end.y, re->end.y, false, &dc, pfs->pis->log_op);
|
|
|
1955 |
}
|
|
|
1956 |
} else if (le->end.y > re->end.y) {
|
|
|
1957 |
code = dev_proc(pfs->dev, fill_trapezoid)(pfs->dev,
|
|
|
1958 |
le, re, le->start.y, re->end.y, false, &dc, pfs->pis->log_op);
|
|
|
1959 |
if (code >= 0) {
|
|
|
1960 |
ue.start = re->end;
|
|
|
1961 |
ue.end = le->end;
|
|
|
1962 |
code = dev_proc(pfs->dev, fill_trapezoid)(pfs->dev,
|
|
|
1963 |
le, &ue, re->end.y, le->end.y, false, &dc, pfs->pis->log_op);
|
|
|
1964 |
}
|
|
|
1965 |
} else
|
|
|
1966 |
code = dev_proc(pfs->dev, fill_trapezoid)(pfs->dev,
|
|
|
1967 |
le, re, le->start.y, le->end.y, false, &dc, pfs->pis->log_op);
|
|
|
1968 |
vd_restore;
|
|
|
1969 |
return code;
|
|
|
1970 |
}
|
|
|
1971 |
|
|
|
1972 |
private int
|
|
|
1973 |
constant_color_triangle(patch_fill_state_t *pfs,
|
|
|
1974 |
const shading_vertex_t *p0, const shading_vertex_t *p1, const shading_vertex_t *p2)
|
|
|
1975 |
{
|
|
|
1976 |
patch_color_t c, cc;
|
|
|
1977 |
gs_fixed_edge le, re;
|
|
|
1978 |
fixed dx0, dy0, dx1, dy1;
|
|
|
1979 |
const shading_vertex_t *pp;
|
|
|
1980 |
int i;
|
|
|
1981 |
|
|
|
1982 |
draw_triangle(&p0->p, &p1->p, &p2->p, RGB(255, 0, 0));
|
|
|
1983 |
patch_interpolate_color(&c, &p0->c, &p1->c, pfs, 0.5);
|
|
|
1984 |
patch_interpolate_color(&cc, &p2->c, &c, pfs, 0.5);
|
|
|
1985 |
for (i = 0; i < 3; i++) {
|
|
|
1986 |
/* fixme : does optimizer compiler expand this cycle ? */
|
|
|
1987 |
if (p0->p.y <= p1->p.y && p0->p.y <= p2->p.y) {
|
|
|
1988 |
le.start = re.start = p0->p;
|
|
|
1989 |
le.end = p1->p;
|
|
|
1990 |
re.end = p2->p;
|
|
|
1991 |
|
|
|
1992 |
dx0 = le.end.x - le.start.x;
|
|
|
1993 |
dy0 = le.end.y - le.start.y;
|
|
|
1994 |
dx1 = re.end.x - re.start.x;
|
|
|
1995 |
dy1 = re.end.y - re.start.y;
|
|
|
1996 |
if ((int64_t)dx0 * dy1 < (int64_t)dy0 * dx1)
|
|
|
1997 |
return ordered_triangle(pfs, &le, &re, &c);
|
|
|
1998 |
else
|
|
|
1999 |
return ordered_triangle(pfs, &re, &le, &c);
|
|
|
2000 |
}
|
|
|
2001 |
pp = p0; p0 = p1; p1 = p2; p2 = pp;
|
|
|
2002 |
}
|
|
|
2003 |
return 0;
|
|
|
2004 |
}
|
|
|
2005 |
|
|
|
2006 |
|
|
|
2007 |
private int
|
|
|
2008 |
constant_color_quadrangle(patch_fill_state_t *pfs, const quadrangle_patch *p, bool self_intersecting)
|
|
|
2009 |
{
|
|
|
2010 |
/* Assuming the XY span is restricted with curve_samples.
|
|
|
2011 |
It is important for intersection_of_small_bars to compute faster. */
|
|
|
2012 |
gs_fixed_point q[4];
|
|
|
2013 |
fixed ry, ey;
|
|
|
2014 |
int code;
|
|
|
2015 |
bool swap_axes = false;
|
|
|
2016 |
gx_device_color dc;
|
|
|
2017 |
patch_color_t c1, c2, c;
|
|
|
2018 |
bool orient;
|
|
|
2019 |
|
|
|
2020 |
draw_quadrangle(p, RGB(0, 255, 0));
|
|
|
2021 |
patch_interpolate_color(&c1, &p->p[0][0]->c, &p->p[0][1]->c, pfs, 0.5);
|
|
|
2022 |
patch_interpolate_color(&c2, &p->p[1][0]->c, &p->p[1][1]->c, pfs, 0.5);
|
|
|
2023 |
patch_interpolate_color(&c, &c1, &c2, pfs, 0.5);
|
|
|
2024 |
code = patch_color_to_device_color(pfs, &c, &dc);
|
|
|
2025 |
if (code < 0)
|
|
|
2026 |
return code;
|
|
|
2027 |
{ gs_fixed_point qq[4];
|
|
|
2028 |
|
|
|
2029 |
make_vertices(qq, p);
|
|
|
2030 |
# if 0 /* Swapping axes may improve the precision,
|
|
|
2031 |
but slows down due to the area expantion needed
|
|
|
2032 |
in gx_shade_trapezoid. */
|
|
|
2033 |
dx = span_x(qq, 4);
|
|
|
2034 |
dy = span_y(qq, 4);
|
|
|
2035 |
if (dy < dx) {
|
|
|
2036 |
do_swap_axes(qq, 4);
|
|
|
2037 |
swap_axes = true;
|
|
|
2038 |
}
|
|
|
2039 |
# endif
|
|
|
2040 |
wrap_vertices_by_y(q, qq);
|
|
|
2041 |
}
|
|
|
2042 |
{ fixed dx1 = q[1].x - q[0].x, dy1 = q[1].y - q[0].y;
|
|
|
2043 |
fixed dx3 = q[3].x - q[0].x, dy3 = q[3].y - q[0].y;
|
|
|
2044 |
int64_t g13 = (int64_t)dx1 * dy3, h13 = (int64_t)dy1 * dx3;
|
|
|
2045 |
|
|
|
2046 |
if (g13 == h13) {
|
|
|
2047 |
fixed dx2 = q[2].x - q[0].x, dy2 = q[2].y - q[0].y;
|
|
|
2048 |
int64_t g23 = (int64_t)dx2 * dy3, h23 = (int64_t)dy2 * dx3;
|
|
|
2049 |
|
|
|
2050 |
if (dx1 == 0 && dy1 == 0 && g23 == h23)
|
|
|
2051 |
return 0;
|
|
|
2052 |
if (g23 != h23) {
|
|
|
2053 |
orient = (g23 > h23);
|
|
|
2054 |
if (q[2].y <= q[3].y) {
|
|
|
2055 |
if ((code = gx_shade_trapezoid(pfs, q, 1, 2, 0, 3, q[1].y, q[2].y, swap_axes, &dc, orient)) < 0)
|
|
|
2056 |
return code;
|
|
|
2057 |
return gx_shade_trapezoid(pfs, q, 2, 3, 0, 3, q[2].y, q[3].y, swap_axes, &dc, orient);
|
|
|
2058 |
} else {
|
|
|
2059 |
if ((code = gx_shade_trapezoid(pfs, q, 1, 2, 0, 3, q[1].y, q[3].y, swap_axes, &dc, orient)) < 0)
|
|
|
2060 |
return code;
|
|
|
2061 |
return gx_shade_trapezoid(pfs, q, 1, 2, 3, 2, q[3].y, q[2].y, swap_axes, &dc, orient);
|
|
|
2062 |
}
|
|
|
2063 |
} else {
|
|
|
2064 |
int64_t g12 = (int64_t)dx1 * dy2, h12 = (int64_t)dy1 * dx2;
|
|
|
2065 |
|
|
|
2066 |
if (dx3 == 0 && dy3 == 0 && g12 == h12)
|
|
|
2067 |
return 0;
|
|
|
2068 |
orient = (g12 > h12);
|
|
|
2069 |
if (q[1].y <= q[2].y) {
|
|
|
2070 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 3, 2, q[0].y, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2071 |
return code;
|
|
|
2072 |
return gx_shade_trapezoid(pfs, q, 1, 2, 3, 2, q[1].y, q[2].y, swap_axes, &dc, orient);
|
|
|
2073 |
} else {
|
|
|
2074 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 3, 2, q[0].y, q[2].y, swap_axes, &dc, orient)) < 0)
|
|
|
2075 |
return code;
|
|
|
2076 |
return gx_shade_trapezoid(pfs, q, 0, 1, 2, 1, q[2].y, q[1].y, swap_axes, &dc, orient);
|
|
|
2077 |
}
|
|
|
2078 |
}
|
|
|
2079 |
}
|
|
|
2080 |
orient = ((int64_t)dx1 * dy3 > (int64_t)dy1 * dx3);
|
|
|
2081 |
}
|
|
|
2082 |
if (q[1].y <= q[2].y && q[2].y <= q[3].y) {
|
|
|
2083 |
if (self_intersecting && intersection_of_small_bars(q, 0, 3, 1, 2, &ry, &ey)) {
|
|
|
2084 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2085 |
return code;
|
|
|
2086 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 3, 1, 2, q[1].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2087 |
return code;
|
|
|
2088 |
if ((code = gx_shade_trapezoid(pfs, q, 1, 2, 0, 3, ry, q[2].y, swap_axes, &dc, orient)) < 0)
|
|
|
2089 |
return code;
|
|
|
2090 |
return gx_shade_trapezoid(pfs, q, 0, 3, 2, 3, q[2].y, q[3].y, swap_axes, &dc, orient);
|
|
|
2091 |
} else {
|
|
|
2092 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2093 |
return code;
|
|
|
2094 |
if ((code = gx_shade_trapezoid(pfs, q, 1, 2, 0, 3, q[1].y, q[2].y, swap_axes, &dc, orient)) < 0)
|
|
|
2095 |
return code;
|
|
|
2096 |
return gx_shade_trapezoid(pfs, q, 2, 3, 0, 3, q[2].y, q[3].y, swap_axes, &dc, orient);
|
|
|
2097 |
}
|
|
|
2098 |
} else if (q[1].y <= q[3].y && q[3].y <= q[2].y) {
|
|
|
2099 |
if (self_intersecting && intersection_of_small_bars(q, 0, 3, 1, 2, &ry, &ey)) {
|
|
|
2100 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2101 |
return code;
|
|
|
2102 |
if ((code = gx_shade_trapezoid(pfs, q, 1, 2, 0, 3, q[1].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2103 |
return code;
|
|
|
2104 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 3, 1, 2, ry, q[3].y, swap_axes, &dc, orient)) < 0)
|
|
|
2105 |
return code;
|
|
|
2106 |
return gx_shade_trapezoid(pfs, q, 3, 2, 1, 2, q[3].y, q[2].y, swap_axes, &dc, orient);
|
|
|
2107 |
} else {
|
|
|
2108 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2109 |
return code;
|
|
|
2110 |
if ((code = gx_shade_trapezoid(pfs, q, 1, 2, 0, 3, q[1].y, q[3].y, swap_axes, &dc, orient)) < 0)
|
|
|
2111 |
return code;
|
|
|
2112 |
return gx_shade_trapezoid(pfs, q, 1, 2, 3, 2, q[3].y, q[2].y, swap_axes, &dc, orient);
|
|
|
2113 |
}
|
|
|
2114 |
} else if (q[2].y <= q[1].y && q[1].y <= q[3].y) {
|
|
|
2115 |
if (self_intersecting && intersection_of_small_bars(q, 0, 1, 2, 3, &ry, &ey)) {
|
|
|
2116 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2117 |
return code;
|
|
|
2118 |
if ((code = gx_shade_trapezoid(pfs, q, 2, 1, 2, 3, q[2].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2119 |
return code;
|
|
|
2120 |
if ((code = gx_shade_trapezoid(pfs, q, 2, 1, 0, 1, ry, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2121 |
return code;
|
|
|
2122 |
return gx_shade_trapezoid(pfs, q, 2, 3, 0, 3, ry, q[3].y, swap_axes, &dc, orient);
|
|
|
2123 |
} else if (self_intersecting && intersection_of_small_bars(q, 0, 3, 1, 2, &ry, &ey)) {
|
|
|
2124 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2125 |
return code;
|
|
|
2126 |
if ((code = gx_shade_trapezoid(pfs, q, 2, 1, 2, 3, q[2].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2127 |
return code;
|
|
|
2128 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 2, 1, ry, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2129 |
return code;
|
|
|
2130 |
return gx_shade_trapezoid(pfs, q, 0, 3, 2, 3, ry, q[3].y, swap_axes, &dc, orient);
|
|
|
2131 |
} else {
|
|
|
2132 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2133 |
return code;
|
|
|
2134 |
if ((code = gx_shade_trapezoid(pfs, q, 2, 3, 2, 1, q[2].y, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2135 |
return code;
|
|
|
2136 |
return gx_shade_trapezoid(pfs, q, 2, 3, 0, 3, q[1].y, q[3].y, swap_axes, &dc, orient);
|
|
|
2137 |
}
|
|
|
2138 |
} else if (q[2].y <= q[3].y && q[3].y <= q[1].y) {
|
|
|
2139 |
if (self_intersecting && intersection_of_small_bars(q, 0, 1, 2, 3, &ry, &ey)) {
|
|
|
2140 |
/* Same code as someone above. */
|
|
|
2141 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2142 |
return code;
|
|
|
2143 |
if ((code = gx_shade_trapezoid(pfs, q, 2, 1, 2, 3, q[2].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2144 |
return code;
|
|
|
2145 |
if ((code = gx_shade_trapezoid(pfs, q, 2, 1, 0, 1, ry, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2146 |
return code;
|
|
|
2147 |
return gx_shade_trapezoid(pfs, q, 2, 3, 0, 3, ry, q[3].y, swap_axes, &dc, orient);
|
|
|
2148 |
} else if (self_intersecting && intersection_of_small_bars(q, 0, 3, 2, 1, &ry, &ey)) {
|
|
|
2149 |
/* Same code as someone above. */
|
|
|
2150 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2151 |
return code;
|
|
|
2152 |
if ((code = gx_shade_trapezoid(pfs, q, 2, 1, 2, 3, q[2].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2153 |
return code;
|
|
|
2154 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 2, 1, ry, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2155 |
return code;
|
|
|
2156 |
return gx_shade_trapezoid(pfs, q, 0, 3, 2, 3, ry, q[3].y, swap_axes, &dc, orient);
|
|
|
2157 |
} else {
|
|
|
2158 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[2].y, swap_axes, &dc, orient)) < 0)
|
|
|
2159 |
return code;
|
|
|
2160 |
if ((code = gx_shade_trapezoid(pfs, q, 2, 3, 0, 3, q[2].y, q[3].y, swap_axes, &dc, orient)) < 0)
|
|
|
2161 |
return code;
|
|
|
2162 |
return gx_shade_trapezoid(pfs, q, 0, 1, 2, 1, q[2].y, q[1].y, swap_axes, &dc, orient);
|
|
|
2163 |
}
|
|
|
2164 |
} else if (q[3].y <= q[1].y && q[1].y <= q[2].y) {
|
|
|
2165 |
if (self_intersecting && intersection_of_small_bars(q, 0, 1, 3, 2, &ry, &ey)) {
|
|
|
2166 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[3].y, swap_axes, &dc, orient)) < 0)
|
|
|
2167 |
return code;
|
|
|
2168 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 3, 2, q[3].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2169 |
return code;
|
|
|
2170 |
if ((code = gx_shade_trapezoid(pfs, q, 3, 2, 0, 1, ry, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2171 |
return code;
|
|
|
2172 |
return gx_shade_trapezoid(pfs, q, 3, 2, 1, 2, q[1].y, q[2].y, swap_axes, &dc, orient);
|
|
|
2173 |
} else {
|
|
|
2174 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[3].y, swap_axes, &dc, orient)) < 0)
|
|
|
2175 |
return code;
|
|
|
2176 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 3, 2, q[3].y, q[1].y, swap_axes, &dc, orient)) < 0)
|
|
|
2177 |
return code;
|
|
|
2178 |
return gx_shade_trapezoid(pfs, q, 1, 2, 3, 2, q[1].y, q[2].y, swap_axes, &dc, orient);
|
|
|
2179 |
}
|
|
|
2180 |
} else if (q[3].y <= q[2].y && q[2].y <= q[1].y) {
|
|
|
2181 |
if (self_intersecting && intersection_of_small_bars(q, 0, 1, 2, 3, &ry, &ey)) {
|
|
|
2182 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[3].y, swap_axes, &dc, orient)) < 0)
|
|
|
2183 |
return code;
|
|
|
2184 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 3, 2, q[3].y, ry + ey, swap_axes, &dc, orient)) < 0)
|
|
|
2185 |
return code;
|
|
|
2186 |
if ((code = gx_shade_trapezoid(pfs, q, 3, 2, 0, 1, ry, q[2].y, swap_axes, &dc, orient)) < 0)
|
|
|
2187 |
return code;
|
|
|
2188 |
return gx_shade_trapezoid(pfs, q, 2, 1, 0, 1, q[2].y, q[1].y, swap_axes, &dc, orient);
|
|
|
2189 |
} else {
|
|
|
2190 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 0, 3, q[0].y, q[3].y, swap_axes, &dc, orient)) < 0)
|
|
|
2191 |
return code;
|
|
|
2192 |
if ((code = gx_shade_trapezoid(pfs, q, 0, 1, 3, 2, q[3].y, q[2].y, swap_axes, &dc, orient)) < 0)
|
|
|
2193 |
return code;
|
|
|
2194 |
return gx_shade_trapezoid(pfs, q, 0, 1, 2, 1, q[2].y, q[1].y, swap_axes, &dc, orient);
|
|
|
2195 |
}
|
|
|
2196 |
} else {
|
|
|
2197 |
/* Impossible. */
|
|
|
2198 |
return_error(gs_error_unregistered);
|
|
|
2199 |
}
|
|
|
2200 |
}
|
|
|
2201 |
|
|
|
2202 |
private inline void
|
|
|
2203 |
divide_quadrangle_by_v(patch_fill_state_t *pfs, quadrangle_patch *s0, quadrangle_patch *s1,
|
|
|
2204 |
shading_vertex_t q[2], const quadrangle_patch *p)
|
|
|
2205 |
{
|
|
|
2206 |
q[0].p.x = (p->p[0][0]->p.x + p->p[1][0]->p.x) / 2;
|
|
|
2207 |
q[1].p.x = (p->p[0][1]->p.x + p->p[1][1]->p.x) / 2;
|
|
|
2208 |
q[0].p.y = (p->p[0][0]->p.y + p->p[1][0]->p.y) / 2;
|
|
|
2209 |
q[1].p.y = (p->p[0][1]->p.y + p->p[1][1]->p.y) / 2;
|
|
|
2210 |
patch_interpolate_color(&q[0].c, &p->p[0][0]->c, &p->p[1][0]->c, pfs, 0.5);
|
|
|
2211 |
patch_interpolate_color(&q[1].c, &p->p[0][1]->c, &p->p[1][1]->c, pfs, 0.5);
|
|
|
2212 |
s0->p[0][0] = p->p[0][0];
|
|
|
2213 |
s0->p[0][1] = p->p[0][1];
|
|
|
2214 |
s0->p[1][0] = s1->p[0][0] = &q[0];
|
|
|
2215 |
s0->p[1][1] = s1->p[0][1] = &q[1];
|
|
|
2216 |
s1->p[1][0] = p->p[1][0];
|
|
|
2217 |
s1->p[1][1] = p->p[1][1];
|
|
|
2218 |
}
|
|
|
2219 |
|
|
|
2220 |
private inline void
|
|
|
2221 |
divide_quadrangle_by_u(patch_fill_state_t *pfs, quadrangle_patch *s0, quadrangle_patch *s1,
|
|
|
2222 |
shading_vertex_t q[2], const quadrangle_patch *p)
|
|
|
2223 |
{
|
|
|
2224 |
q[0].p.x = (p->p[0][0]->p.x + p->p[0][1]->p.x) / 2;
|
|
|
2225 |
q[1].p.x = (p->p[1][0]->p.x + p->p[1][1]->p.x) / 2;
|
|
|
2226 |
q[0].p.y = (p->p[0][0]->p.y + p->p[0][1]->p.y) / 2;
|
|
|
2227 |
q[1].p.y = (p->p[1][0]->p.y + p->p[1][1]->p.y) / 2;
|
|
|
2228 |
patch_interpolate_color(&q[0].c, &p->p[0][0]->c, &p->p[0][1]->c, pfs, 0.5);
|
|
|
2229 |
patch_interpolate_color(&q[1].c, &p->p[1][0]->c, &p->p[1][1]->c, pfs, 0.5);
|
|
|
2230 |
s0->p[0][0] = p->p[0][0];
|
|
|
2231 |
s0->p[1][0] = p->p[1][0];
|
|
|
2232 |
s0->p[0][1] = s1->p[0][0] = &q[0];
|
|
|
2233 |
s0->p[1][1] = s1->p[1][0] = &q[1];
|
|
|
2234 |
s1->p[0][1] = p->p[0][1];
|
|
|
2235 |
s1->p[1][1] = p->p[1][1];
|
|
|
2236 |
}
|
|
|
2237 |
|
|
|
2238 |
private inline int
|
|
|
2239 |
is_quadrangle_color_monotonic(const patch_fill_state_t *pfs, const quadrangle_patch *p,
|
|
|
2240 |
bool *not_monotonic_by_u, bool *not_monotonic_by_v)
|
|
|
2241 |
{ /* returns : 1 = monotonic, 0 = don't know, <0 = error. */
|
|
|
2242 |
int code;
|
|
|
2243 |
|
|
|
2244 |
code = isnt_color_monotonic(pfs, &p->p[0][0]->c, &p->p[1][1]->c);
|
|
|
2245 |
if (code <= 0)
|
|
|
2246 |
return code;
|
|
|
2247 |
if (code & 1)
|
|
|
2248 |
*not_monotonic_by_v = true;
|
|
|
2249 |
if (code & 2)
|
|
|
2250 |
*not_monotonic_by_u = true;
|
|
|
2251 |
return !code;
|
|
|
2252 |
}
|
|
|
2253 |
|
|
|
2254 |
private inline bool
|
|
|
2255 |
quadrangle_bbox_covers_pixel_centers(const quadrangle_patch *p)
|
|
|
2256 |
{
|
|
|
2257 |
fixed xbot, xtop, ybot, ytop;
|
|
|
2258 |
|
|
|
2259 |
xbot = min(min(p->p[0][0]->p.x, p->p[0][1]->p.x),
|
|
|
2260 |
min(p->p[1][0]->p.x, p->p[1][1]->p.x));
|
|
|
2261 |
xtop = max(max(p->p[0][0]->p.x, p->p[0][1]->p.x),
|
|
|
2262 |
max(p->p[1][0]->p.x, p->p[1][1]->p.x));
|
|
|
2263 |
if (covers_pixel_centers(xbot, xtop))
|
|
|
2264 |
return true;
|
|
|
2265 |
ybot = min(min(p->p[0][0]->p.y, p->p[0][1]->p.y),
|
|
|
2266 |
min(p->p[1][0]->p.y, p->p[1][1]->p.y));
|
|
|
2267 |
ytop = max(max(p->p[0][0]->p.y, p->p[0][1]->p.y),
|
|
|
2268 |
max(p->p[1][0]->p.y, p->p[1][1]->p.y));
|
|
|
2269 |
if (covers_pixel_centers(ybot, ytop))
|
|
|
2270 |
return true;
|
|
|
2271 |
return false;
|
|
|
2272 |
}
|
|
|
2273 |
|
|
|
2274 |
private inline void
|
|
|
2275 |
divide_bar(patch_fill_state_t *pfs,
|
|
|
2276 |
const shading_vertex_t *p0, const shading_vertex_t *p1, int radix, shading_vertex_t *p)
|
|
|
2277 |
{
|
|
|
2278 |
p->p.x = (fixed)((int64_t)p0->p.x * (radix - 1) + p1->p.x) / radix;
|
|
|
2279 |
p->p.y = (fixed)((int64_t)p0->p.y * (radix - 1) + p1->p.y) / radix;
|
|
|
2280 |
patch_interpolate_color(&p->c, &p0->c, &p1->c, pfs, (double)(radix - 1) / radix);
|
|
|
2281 |
}
|
|
|
2282 |
|
|
|
2283 |
private inline void
|
|
|
2284 |
bbox_of_points(gs_fixed_rect *r,
|
|
|
2285 |
const gs_fixed_point *p0, const gs_fixed_point *p1,
|
|
|
2286 |
const gs_fixed_point *p2, const gs_fixed_point *p3)
|
|
|
2287 |
{
|
|
|
2288 |
r->p.x = r->q.x = p0->x;
|
|
|
2289 |
r->p.y = r->q.y = p0->y;
|
|
|
2290 |
|
|
|
2291 |
if (r->p.x > p1->x)
|
|
|
2292 |
r->p.x = p1->x;
|
|
|
2293 |
if (r->q.x < p1->x)
|
|
|
2294 |
r->q.x = p1->x;
|
|
|
2295 |
if (r->p.y > p1->y)
|
|
|
2296 |
r->p.y = p1->y;
|
|
|
2297 |
if (r->q.y < p1->y)
|
|
|
2298 |
r->q.y = p1->y;
|
|
|
2299 |
|
|
|
2300 |
if (r->p.x > p2->x)
|
|
|
2301 |
r->p.x = p2->x;
|
|
|
2302 |
if (r->q.x < p2->x)
|
|
|
2303 |
r->q.x = p2->x;
|
|
|
2304 |
if (r->p.y > p2->y)
|
|
|
2305 |
r->p.y = p2->y;
|
|
|
2306 |
if (r->q.y < p2->y)
|
|
|
2307 |
r->q.y = p2->y;
|
|
|
2308 |
|
|
|
2309 |
if (p3 == NULL)
|
|
|
2310 |
return;
|
|
|
2311 |
|
|
|
2312 |
if (r->p.x > p3->x)
|
|
|
2313 |
r->p.x = p3->x;
|
|
|
2314 |
if (r->q.x < p3->x)
|
|
|
2315 |
r->q.x = p3->x;
|
|
|
2316 |
if (r->p.y > p3->y)
|
|
|
2317 |
r->p.y = p3->y;
|
|
|
2318 |
if (r->q.y < p3->y)
|
|
|
2319 |
r->q.y = p3->y;
|
|
|
2320 |
}
|
|
|
2321 |
|
|
|
2322 |
private int
|
|
|
2323 |
triangle_by_4(patch_fill_state_t *pfs,
|
|
|
2324 |
const shading_vertex_t *p0, const shading_vertex_t *p1, const shading_vertex_t *p2,
|
|
|
2325 |
wedge_vertex_list_t *l01, wedge_vertex_list_t *l12, wedge_vertex_list_t *l20,
|
|
|
2326 |
double cd, fixed sd)
|
|
|
2327 |
{
|
|
|
2328 |
shading_vertex_t p01, p12, p20;
|
|
|
2329 |
wedge_vertex_list_t L01, L12, L20, L[3];
|
|
|
2330 |
bool inside_save = pfs->inside;
|
|
|
2331 |
gs_fixed_rect r, r1;
|
|
|
2332 |
int code;
|
|
|
2333 |
|
|
|
2334 |
if (!pfs->inside) {
|
|
|
2335 |
bbox_of_points(&r, &p0->p, &p1->p, &p2->p, NULL);
|
|
|
2336 |
r1 = r;
|
|
|
2337 |
rect_intersect(r, pfs->rect);
|
|
|
2338 |
if (r.q.x <= r.p.x || r.q.y <= r.p.y)
|
|
|
2339 |
return 0; /* Outside. */
|
|
|
2340 |
}
|
|
|
2341 |
code = try_device_linear_color(pfs, false, p0, p1, p2);
|
|
|
2342 |
switch(code) {
|
|
|
2343 |
case 0: /* The area is filled. */
|
|
|
2344 |
return 0;
|
|
|
2345 |
case 2: /* decompose to constant color areas */
|
|
|
2346 |
if (sd < fixed_1 * 4)
|
|
|
2347 |
return constant_color_triangle(pfs, p2, p0, p1);
|
|
|
2348 |
if (pfs->Function != NULL) {
|
|
|
2349 |
double d01 = color_span(pfs, &p1->c, &p0->c);
|
|
|
2350 |
double d12 = color_span(pfs, &p2->c, &p1->c);
|
|
|
2351 |
double d20 = color_span(pfs, &p0->c, &p2->c);
|
|
|
2352 |
|
|
|
2353 |
if (d01 <= pfs->smoothness / COLOR_CONTIGUITY &&
|
|
|
2354 |
d12 <= pfs->smoothness / COLOR_CONTIGUITY &&
|
|
|
2355 |
d20 <= pfs->smoothness / COLOR_CONTIGUITY)
|
|
|
2356 |
return constant_color_triangle(pfs, p2, p0, p1);
|
|
|
2357 |
} else if (cd <= pfs->smoothness / COLOR_CONTIGUITY)
|
|
|
2358 |
return constant_color_triangle(pfs, p2, p0, p1);
|
|
|
2359 |
break;
|
|
|
2360 |
case 1: /* decompose to linear color areas */
|
|
|
2361 |
if (sd < fixed_1)
|
|
|
2362 |
return constant_color_triangle(pfs, p2, p0, p1);
|
|
|
2363 |
break;
|
|
|
2364 |
default: /* Error. */
|
|
|
2365 |
return code;
|
|
|
2366 |
}
|
|
|
2367 |
if (!pfs->inside) {
|
|
|
2368 |
if (r.p.x == r1.p.x && r.p.y == r1.p.y &&
|
|
|
2369 |
r.q.x == r1.q.x && r.q.y == r1.q.y)
|
|
|
2370 |
pfs->inside = true;
|
|
|
2371 |
}
|
|
|
2372 |
divide_bar(pfs, p0, p1, 2, &p01);
|
|
|
2373 |
divide_bar(pfs, p1, p2, 2, &p12);
|
|
|
2374 |
divide_bar(pfs, p2, p0, 2, &p20);
|
|
|
2375 |
if (LAZY_WEDGES) {
|
|
|
2376 |
init_wedge_vertex_list(L, count_of(L));
|
|
|
2377 |
make_wedge_median(pfs, &L01, l01, true, &p0->p, &p1->p, &p01.p);
|
|
|
2378 |
make_wedge_median(pfs, &L12, l12, true, &p1->p, &p2->p, &p12.p);
|
|
|
2379 |
make_wedge_median(pfs, &L20, l20, false, &p2->p, &p0->p, &p20.p);
|
|
|
2380 |
} else {
|
|
|
2381 |
code = fill_triangle_wedge(pfs, p0, p1, &p01);
|
|
|
2382 |
if (code < 0)
|
|
|
2383 |
return code;
|
|
|
2384 |
code = fill_triangle_wedge(pfs, p1, p2, &p12);
|
|
|
2385 |
if (code < 0)
|
|
|
2386 |
return code;
|
|
|
2387 |
code = fill_triangle_wedge(pfs, p2, p0, &p20);
|
|
|
2388 |
if (code < 0)
|
|
|
2389 |
return code;
|
|
|
2390 |
}
|
|
|
2391 |
code = triangle_by_4(pfs, p0, &p01, &p20, &L01, &L[0], &L20, cd / 2, sd / 2);
|
|
|
2392 |
if (code < 0)
|
|
|
2393 |
return code;
|
|
|
2394 |
if (LAZY_WEDGES) {
|
|
|
2395 |
move_wedge(&L01, l01, true);
|
|
|
2396 |
move_wedge(&L20, l20, false);
|
|
|
2397 |
}
|
|
|
2398 |
code = triangle_by_4(pfs, p1, &p12, &p01, &L12, &L[1], &L01, cd / 2, sd / 2);
|
|
|
2399 |
if (code < 0)
|
|
|
2400 |
return code;
|
|
|
2401 |
if (LAZY_WEDGES)
|
|
|
2402 |
move_wedge(&L12, l12, true);
|
|
|
2403 |
code = triangle_by_4(pfs, p2, &p20, &p12, &L20, &L[2], &L12, cd / 2, sd / 2);
|
|
|
2404 |
if (code < 0)
|
|
|
2405 |
return code;
|
|
|
2406 |
L[0].last_side = L[1].last_side = L[2].last_side = true;
|
|
|
2407 |
code = triangle_by_4(pfs, &p01, &p12, &p20, &L[1], &L[2], &L[0], cd / 2, sd / 2);
|
|
|
2408 |
if (code < 0)
|
|
|
2409 |
return code;
|
|
|
2410 |
if (LAZY_WEDGES) {
|
|
|
2411 |
code = close_wedge_median(pfs, l01, &p0->c, &p1->c);
|
|
|
2412 |
if (code < 0)
|
|
|
2413 |
return code;
|
|
|
2414 |
code = close_wedge_median(pfs, l12, &p1->c, &p2->c);
|
|
|
2415 |
if (code < 0)
|
|
|
2416 |
return code;
|
|
|
2417 |
code = close_wedge_median(pfs, l20, &p2->c, &p0->c);
|
|
|
2418 |
if (code < 0)
|
|
|
2419 |
return code;
|
|
|
2420 |
code = terminate_wedge_vertex_list(pfs, &L[0], &p01.c, &p20.c);
|
|
|
2421 |
if (code < 0)
|
|
|
2422 |
return code;
|
|
|
2423 |
code = terminate_wedge_vertex_list(pfs, &L[1], &p12.c, &p01.c);
|
|
|
2424 |
if (code < 0)
|
|
|
2425 |
return code;
|
|
|
2426 |
code = terminate_wedge_vertex_list(pfs, &L[2], &p20.c, &p12.c);
|
|
|
2427 |
if (code < 0)
|
|
|
2428 |
return code;
|
|
|
2429 |
}
|
|
|
2430 |
pfs->inside = inside_save;
|
|
|
2431 |
return 0;
|
|
|
2432 |
}
|
|
|
2433 |
|
|
|
2434 |
private inline int
|
|
|
2435 |
fill_triangle(patch_fill_state_t *pfs,
|
|
|
2436 |
const shading_vertex_t *p0, const shading_vertex_t *p1, const shading_vertex_t *p2,
|
|
|
2437 |
wedge_vertex_list_t *l01, wedge_vertex_list_t *l12, wedge_vertex_list_t *l20)
|
|
|
2438 |
{
|
|
|
2439 |
fixed sd01 = max(any_abs(p1->p.x - p0->p.x), any_abs(p1->p.y - p0->p.y));
|
|
|
2440 |
fixed sd12 = max(any_abs(p2->p.x - p1->p.x), any_abs(p2->p.y - p1->p.y));
|
|
|
2441 |
fixed sd20 = max(any_abs(p0->p.x - p2->p.x), any_abs(p0->p.y - p2->p.y));
|
|
|
2442 |
fixed sd1 = max(sd01, sd12);
|
|
|
2443 |
fixed sd = max(sd1, sd20);
|
|
|
2444 |
double cd = 0;
|
|
|
2445 |
|
|
|
2446 |
# if SKIP_TEST
|
|
|
2447 |
dbg_triangle_cnt++;
|
|
|
2448 |
# endif
|
|
|
2449 |
if (pfs->Function == NULL) {
|
|
|
2450 |
double d01 = color_span(pfs, &p1->c, &p0->c);
|
|
|
2451 |
double d12 = color_span(pfs, &p2->c, &p1->c);
|
|
|
2452 |
double d20 = color_span(pfs, &p0->c, &p2->c);
|
|
|
2453 |
double cd1 = max(d01, d12);
|
|
|
2454 |
|
|
|
2455 |
cd = max(cd1, d20);
|
|
|
2456 |
}
|
|
|
2457 |
return triangle_by_4(pfs, p0, p1, p2, l01, l12, l20, cd, sd);
|
|
|
2458 |
}
|
|
|
2459 |
|
|
|
2460 |
private int
|
|
|
2461 |
small_mesh_triangle(patch_fill_state_t *pfs,
|
|
|
2462 |
const shading_vertex_t *p0, const shading_vertex_t *p1, const shading_vertex_t *p2)
|
|
|
2463 |
{
|
|
|
2464 |
int code;
|
|
|
2465 |
wedge_vertex_list_t l[3];
|
|
|
2466 |
|
|
|
2467 |
init_wedge_vertex_list(l, count_of(l));
|
|
|
2468 |
code = fill_triangle(pfs, p0, p1, p2, &l[0], &l[1], &l[2]);
|
|
|
2469 |
if (code < 0)
|
|
|
2470 |
return code;
|
|
|
2471 |
code = terminate_wedge_vertex_list(pfs, &l[0], &p0->c, &p1->c);
|
|
|
2472 |
if (code < 0)
|
|
|
2473 |
return code;
|
|
|
2474 |
code = terminate_wedge_vertex_list(pfs, &l[1], &p1->c, &p2->c);
|
|
|
2475 |
if (code < 0)
|
|
|
2476 |
return code;
|
|
|
2477 |
return terminate_wedge_vertex_list(pfs, &l[2], &p2->c, &p0->c);
|
|
|
2478 |
}
|
|
|
2479 |
|
|
|
2480 |
private int
|
|
|
2481 |
mesh_triangle_rec(patch_fill_state_t *pfs,
|
|
|
2482 |
const shading_vertex_t *p0, const shading_vertex_t *p1, const shading_vertex_t *p2)
|
|
|
2483 |
{
|
|
|
2484 |
pfs->unlinear = !is_linear_color_applicable(pfs);
|
|
|
2485 |
if (manhattan_dist(&p0->p, &p1->p) < pfs->max_small_coord &&
|
|
|
2486 |
manhattan_dist(&p1->p, &p2->p) < pfs->max_small_coord &&
|
|
|
2487 |
manhattan_dist(&p2->p, &p0->p) < pfs->max_small_coord)
|
|
|
2488 |
return small_mesh_triangle(pfs, p0, p1, p2);
|
|
|
2489 |
else {
|
|
|
2490 |
/* Subdivide into 4 triangles with 3 triangle non-lazy wedges.
|
|
|
2491 |
Doing so against the wedge_vertex_list_elem_buffer overflow.
|
|
|
2492 |
We could apply a smarter method, dividing long sides
|
|
|
2493 |
with no wedges and short sides with lazy wedges.
|
|
|
2494 |
This needs to start wedges dynamically when
|
|
|
2495 |
a side becomes short. We don't do so because the
|
|
|
2496 |
number of checks per call significantly increases
|
|
|
2497 |
and the logics is complicated, but the performance
|
|
|
2498 |
advantage appears small due to big meshes are rare.
|
|
|
2499 |
*/
|
|
|
2500 |
shading_vertex_t p01, p12, p20;
|
|
|
2501 |
int code;
|
|
|
2502 |
|
|
|
2503 |
divide_bar(pfs, p0, p1, 2, &p01);
|
|
|
2504 |
divide_bar(pfs, p1, p2, 2, &p12);
|
|
|
2505 |
divide_bar(pfs, p2, p0, 2, &p20);
|
|
|
2506 |
code = fill_triangle_wedge(pfs, p0, p1, &p01);
|
|
|
2507 |
if (code < 0)
|
|
|
2508 |
return code;
|
|
|
2509 |
code = fill_triangle_wedge(pfs, p1, p2, &p12);
|
|
|
2510 |
if (code < 0)
|
|
|
2511 |
return code;
|
|
|
2512 |
code = fill_triangle_wedge(pfs, p2, p0, &p20);
|
|
|
2513 |
if (code < 0)
|
|
|
2514 |
return code;
|
|
|
2515 |
code = mesh_triangle_rec(pfs, p0, &p01, &p20);
|
|
|
2516 |
if (code < 0)
|
|
|
2517 |
return code;
|
|
|
2518 |
code = mesh_triangle_rec(pfs, p1, &p12, &p01);
|
|
|
2519 |
if (code < 0)
|
|
|
2520 |
return code;
|
|
|
2521 |
code = mesh_triangle_rec(pfs, p2, &p20, &p12);
|
|
|
2522 |
if (code < 0)
|
|
|
2523 |
return code;
|
|
|
2524 |
return mesh_triangle_rec(pfs, &p01, &p12, &p20);
|
|
|
2525 |
}
|
|
|
2526 |
}
|
|
|
2527 |
|
|
|
2528 |
int
|
|
|
2529 |
mesh_triangle(patch_fill_state_t *pfs,
|
|
|
2530 |
const shading_vertex_t *p0, const shading_vertex_t *p1, const shading_vertex_t *p2)
|
|
|
2531 |
{
|
|
|
2532 |
if ((*dev_proc(pfs->dev, pattern_manage))(pfs->dev,
|
|
|
2533 |
gs_no_id, NULL, pattern_manage__shading_area) > 0) {
|
|
|
2534 |
/* Inform the device with the shading coverage area.
|
|
|
2535 |
First compute the sign of the area, because
|
|
|
2536 |
all areas to be clipped in same direction. */
|
|
|
2537 |
gx_device *pdev = pfs->dev;
|
|
|
2538 |
gx_path path;
|
|
|
2539 |
int code;
|
|
|
2540 |
fixed d01x = p1->p.x - p0->p.x, d01y = p1->p.y - p0->p.y;
|
|
|
2541 |
fixed d12x = p2->p.x - p1->p.x, d12y = p2->p.y - p1->p.y;
|
|
|
2542 |
int64_t s1 = (int64_t)d01x * d12y - (int64_t)d01y * d12x;
|
|
|
2543 |
|
|
|
2544 |
gx_path_init_local(&path, pdev->memory);
|
|
|
2545 |
code = gx_path_add_point(&path, p0->p.x, p0->p.y);
|
|
|
2546 |
if (code >= 0 && s1 >= 0)
|
|
|
2547 |
code = gx_path_add_line(&path, p1->p.x, p1->p.y);
|
|
|
2548 |
if (code >= 0)
|
|
|
2549 |
code = gx_path_add_line(&path, p2->p.x, p2->p.y);
|
|
|
2550 |
if (code >= 0 && s1 < 0)
|
|
|
2551 |
code = gx_path_add_line(&path, p1->p.x, p1->p.y);
|
|
|
2552 |
if (code >= 0)
|
|
|
2553 |
code = gx_path_close_subpath(&path);
|
|
|
2554 |
if (code >= 0)
|
|
|
2555 |
code = (*dev_proc(pfs->dev, fill_path))(pdev, NULL, &path, NULL, NULL, NULL);
|
|
|
2556 |
gx_path_free(&path, "mesh_triangle");
|
|
|
2557 |
if (code < 0)
|
|
|
2558 |
return code;
|
|
|
2559 |
}
|
|
|
2560 |
return mesh_triangle_rec(pfs, p0, p1, p2);
|
|
|
2561 |
}
|
|
|
2562 |
|
|
|
2563 |
private inline int
|
|
|
2564 |
triangles4(patch_fill_state_t *pfs, const quadrangle_patch *p, bool dummy_argument)
|
|
|
2565 |
{
|
|
|
2566 |
shading_vertex_t p0001, p1011, q;
|
|
|
2567 |
wedge_vertex_list_t l[4];
|
|
|
2568 |
int code;
|
|
|
2569 |
|
|
|
2570 |
init_wedge_vertex_list(l, count_of(l));
|
|
|
2571 |
divide_bar(pfs, p->p[0][0], p->p[0][1], 2, &p0001);
|
|
|
2572 |
divide_bar(pfs, p->p[1][0], p->p[1][1], 2, &p1011);
|
|
|
2573 |
divide_bar(pfs, &p0001, &p1011, 2, &q);
|
|
|
2574 |
code = fill_triangle(pfs, p->p[0][0], p->p[0][1], &q, p->l0001, &l[0], &l[3]);
|
|
|
2575 |
if (code < 0)
|
|
|
2576 |
return code;
|
|
|
2577 |
l[0].last_side = true;
|
|
|
2578 |
l[3].last_side = true;
|
|
|
2579 |
code = fill_triangle(pfs, p->p[0][1], p->p[1][1], &q, p->l0111, &l[1], &l[0]);
|
|
|
2580 |
if (code < 0)
|
|
|
2581 |
return code;
|
|
|
2582 |
l[1].last_side = true;
|
|
|
2583 |
code = fill_triangle(pfs, p->p[1][1], p->p[1][0], &q, p->l1110, &l[2], &l[1]);
|
|
|
2584 |
if (code < 0)
|
|
|
2585 |
return code;
|
|
|
2586 |
l[2].last_side = true;
|
|
|
2587 |
code = fill_triangle(pfs, p->p[1][0], p->p[0][0], &q, p->l1000, &l[3], &l[2]);
|
|
|
2588 |
if (code < 0)
|
|
|
2589 |
return code;
|
|
|
2590 |
code = terminate_wedge_vertex_list(pfs, &l[0], &p->p[0][1]->c, &q.c);
|
|
|
2591 |
if (code < 0)
|
|
|
2592 |
return code;
|
|
|
2593 |
code = terminate_wedge_vertex_list(pfs, &l[1], &p->p[1][1]->c, &q.c);
|
|
|
2594 |
if (code < 0)
|
|
|
2595 |
return code;
|
|
|
2596 |
code = terminate_wedge_vertex_list(pfs, &l[2], &p->p[1][0]->c, &q.c);
|
|
|
2597 |
if (code < 0)
|
|
|
2598 |
return code;
|
|
|
2599 |
code = terminate_wedge_vertex_list(pfs, &l[3], &q.c, &p->p[0][0]->c);
|
|
|
2600 |
if (code < 0)
|
|
|
2601 |
return code;
|
|
|
2602 |
return 0;
|
|
|
2603 |
}
|
|
|
2604 |
|
|
|
2605 |
private inline int
|
|
|
2606 |
triangles2(patch_fill_state_t *pfs, const quadrangle_patch *p, bool dummy_argument)
|
|
|
2607 |
{
|
|
|
2608 |
wedge_vertex_list_t l;
|
|
|
2609 |
int code;
|
|
|
2610 |
|
|
|
2611 |
init_wedge_vertex_list(&l, 1);
|
|
|
2612 |
code = fill_triangle(pfs, p->p[0][0], p->p[0][1], p->p[1][1], p->l0001, p->l0111, &l);
|
|
|
2613 |
if (code < 0)
|
|
|
2614 |
return code;
|
|
|
2615 |
l.last_side = true;
|
|
|
2616 |
code = fill_triangle(pfs, p->p[1][1], p->p[1][0], p->p[0][0], p->l1110, p->l1000, &l);
|
|
|
2617 |
if (code < 0)
|
|
|
2618 |
return code;
|
|
|
2619 |
code = terminate_wedge_vertex_list(pfs, &l, &p->p[1][1]->c, &p->p[0][0]->c);
|
|
|
2620 |
if (code < 0)
|
|
|
2621 |
return code;
|
|
|
2622 |
return 0;
|
|
|
2623 |
}
|
|
|
2624 |
|
|
|
2625 |
private inline void
|
|
|
2626 |
make_quadrangle(const tensor_patch *p, shading_vertex_t qq[2][2],
|
|
|
2627 |
wedge_vertex_list_t l[4], quadrangle_patch *q)
|
|
|
2628 |
{
|
|
|
2629 |
qq[0][0].p = p->pole[0][0];
|
|
|
2630 |
qq[0][1].p = p->pole[0][3];
|
|
|
2631 |
qq[1][0].p = p->pole[3][0];
|
|
|
2632 |
qq[1][1].p = p->pole[3][3];
|
|
|
2633 |
qq[0][0].c = p->c[0][0];
|
|
|
2634 |
qq[0][1].c = p->c[0][1];
|
|
|
2635 |
qq[1][0].c = p->c[1][0];
|
|
|
2636 |
qq[1][1].c = p->c[1][1];
|
|
|
2637 |
q->p[0][0] = &qq[0][0];
|
|
|
2638 |
q->p[0][1] = &qq[0][1];
|
|
|
2639 |
q->p[1][0] = &qq[1][0];
|
|
|
2640 |
q->p[1][1] = &qq[1][1];
|
|
|
2641 |
q->l0001 = &l[0];
|
|
|
2642 |
q->l0111 = &l[1];
|
|
|
2643 |
q->l1110 = &l[2];
|
|
|
2644 |
q->l1000 = &l[3];
|
|
|
2645 |
}
|
|
|
2646 |
|
|
|
2647 |
private inline int
|
|
|
2648 |
is_quadrangle_color_linear_by_u(const patch_fill_state_t *pfs, const quadrangle_patch *p)
|
|
|
2649 |
{ /* returns : 1 = linear, 0 = unlinear, <0 = error. */
|
|
|
2650 |
int code;
|
|
|
2651 |
|
|
|
2652 |
code = is_color_linear(pfs, &p->p[0][0]->c, &p->p[0][1]->c);
|
|
|
2653 |
if (code <= 0)
|
|
|
2654 |
return code;
|
|
|
2655 |
return is_color_linear(pfs, &p->p[1][0]->c, &p->p[1][1]->c);
|
|
|
2656 |
}
|
|
|
2657 |
|
|
|
2658 |
private inline int
|
|
|
2659 |
is_quadrangle_color_linear_by_v(const patch_fill_state_t *pfs, const quadrangle_patch *p)
|
|
|
2660 |
{ /* returns : 1 = linear, 0 = unlinear, <0 = error. */
|
|
|
2661 |
int code;
|
|
|
2662 |
|
|
|
2663 |
code = is_color_linear(pfs, &p->p[0][0]->c, &p->p[1][0]->c);
|
|
|
2664 |
if (code <= 0)
|
|
|
2665 |
return code;
|
|
|
2666 |
return is_color_linear(pfs, &p->p[0][1]->c, &p->p[1][1]->c);
|
|
|
2667 |
}
|
|
|
2668 |
|
|
|
2669 |
private inline int
|
|
|
2670 |
is_quadrangle_color_linear_by_diagonals(const patch_fill_state_t *pfs, const quadrangle_patch *p)
|
|
|
2671 |
{ /* returns : 1 = linear, 0 = unlinear, <0 = error. */
|
|
|
2672 |
int code;
|
|
|
2673 |
|
|
|
2674 |
code = is_color_linear(pfs, &p->p[0][0]->c, &p->p[1][1]->c);
|
|
|
2675 |
if (code <= 0)
|
|
|
2676 |
return code;
|
|
|
2677 |
return is_color_linear(pfs, &p->p[0][1]->c, &p->p[1][0]->c);
|
|
|
2678 |
}
|
|
|
2679 |
|
|
|
2680 |
typedef enum {
|
|
|
2681 |
color_change_small,
|
|
|
2682 |
color_change_gradient,
|
|
|
2683 |
color_change_linear,
|
|
|
2684 |
color_change_bilinear,
|
|
|
2685 |
color_change_general
|
|
|
2686 |
} color_change_type_t;
|
|
|
2687 |
|
|
|
2688 |
private inline color_change_type_t
|
|
|
2689 |
quadrangle_color_change(const patch_fill_state_t *pfs, const quadrangle_patch *p,
|
|
|
2690 |
bool is_big_u, bool is_big_v, bool *divide_u, bool *divide_v)
|
|
|
2691 |
{
|
|
|
2692 |
patch_color_t d0001, d1011, d;
|
|
|
2693 |
double D, D0001, D1011, D0010, D0111, D0011, D0110;
|
|
|
2694 |
double Du, Dv;
|
|
|
2695 |
|
|
|
2696 |
color_diff(pfs, &p->p[0][0]->c, &p->p[0][1]->c, &d0001);
|
|
|
2697 |
color_diff(pfs, &p->p[1][0]->c, &p->p[1][1]->c, &d1011);
|
|
|
2698 |
D0001 = color_norm(pfs, &d0001);
|
|
|
2699 |
D1011 = color_norm(pfs, &d1011);
|
|
|
2700 |
D0010 = color_span(pfs, &p->p[0][0]->c, &p->p[1][0]->c);
|
|
|
2701 |
D0111 = color_span(pfs, &p->p[0][1]->c, &p->p[1][1]->c);
|
|
|
2702 |
D0011 = color_span(pfs, &p->p[0][0]->c, &p->p[1][1]->c);
|
|
|
2703 |
D0110 = color_span(pfs, &p->p[0][1]->c, &p->p[1][0]->c);
|
|
|
2704 |
if (pfs->unlinear) {
|
|
|
2705 |
if (D0001 <= pfs->smoothness && D1011 <= pfs->smoothness &&
|
|
|
2706 |
D0010 <= pfs->smoothness && D0111 <= pfs->smoothness &&
|
|
|
2707 |
D0011 <= pfs->smoothness && D0110 <= pfs->smoothness)
|
|
|
2708 |
return color_change_small;
|
|
|
2709 |
if (D0001 <= pfs->smoothness && D1011 <= pfs->smoothness) {
|
|
|
2710 |
if (!is_big_v) {
|
|
|
2711 |
/* The color function looks uncontiguous. */
|
|
|
2712 |
return color_change_small;
|
|
|
2713 |
}
|
|
|
2714 |
*divide_v = true;
|
|
|
2715 |
return color_change_gradient;
|
|
|
2716 |
}
|
|
|
2717 |
if (D0010 <= pfs->smoothness && D0111 <= pfs->smoothness) {
|
|
|
2718 |
if (!is_big_u) {
|
|
|
2719 |
/* The color function looks uncontiguous. */
|
|
|
2720 |
return color_change_small;
|
|
|
2721 |
}
|
|
|
2722 |
*divide_u = true;
|
|
|
2723 |
return color_change_gradient;
|
|
|
2724 |
}
|
|
|
2725 |
}
|
|
|
2726 |
color_diff(pfs, &d0001, &d1011, &d);
|
|
|
2727 |
Du = max(D0001, D1011);
|
|
|
2728 |
Dv = max(D0010, D0111);
|
|
|
2729 |
if (Du <= pfs->smoothness / 8 && Dv <= pfs->smoothness / 8)
|
|
|
2730 |
return color_change_small;
|
|
|
2731 |
if (Du <= pfs->smoothness / 8)
|
|
|
2732 |
return color_change_linear;
|
|
|
2733 |
if (Dv <= pfs->smoothness / 8)
|
|
|
2734 |
return color_change_linear;
|
|
|
2735 |
D = color_norm(pfs, &d);
|
|
|
2736 |
if (D <= pfs->smoothness)
|
|
|
2737 |
return color_change_bilinear;
|
|
|
2738 |
#if 0 /* Disabled due to a 0.5% slowdown with the test file of the Bug 687948. */
|
|
|
2739 |
if (Du > Dv && is_big_u)
|
|
|
2740 |
*divide_u = true;
|
|
|
2741 |
else if (Du < Dv && is_big_v)
|
|
|
2742 |
*divide_v = true;
|
|
|
2743 |
else if (is_big_u)
|
|
|
2744 |
*divide_u = true;
|
|
|
2745 |
else if (is_big_v)
|
|
|
2746 |
*divide_v = true;
|
|
|
2747 |
else {
|
|
|
2748 |
/* The color function looks uncontiguous. */
|
|
|
2749 |
return color_change_small;
|
|
|
2750 |
}
|
|
|
2751 |
#else
|
|
|
2752 |
if (Du > Dv)
|
|
|
2753 |
*divide_u = true;
|
|
|
2754 |
else
|
|
|
2755 |
*divide_v = true;
|
|
|
2756 |
#endif
|
|
|
2757 |
return color_change_general;
|
|
|
2758 |
}
|
|
|
2759 |
|
|
|
2760 |
private int
|
|
|
2761 |
fill_quadrangle(patch_fill_state_t *pfs, const quadrangle_patch *p, bool big, int level)
|
|
|
2762 |
{
|
|
|
2763 |
/* The quadrangle is flattened enough by V and U, so ignore inner poles. */
|
|
|
2764 |
/* Assuming the XY span is restricted with curve_samples.
|
|
|
2765 |
It is important for intersection_of_small_bars to compute faster. */
|
|
|
2766 |
quadrangle_patch s0, s1;
|
|
|
2767 |
wedge_vertex_list_t l0, l1, l2;
|
|
|
2768 |
int code;
|
|
|
2769 |
bool divide_u = false, divide_v = false, big1 = big;
|
|
|
2770 |
shading_vertex_t q[2];
|
|
|
2771 |
bool monotonic_color_save = pfs->monotonic_color;
|
|
|
2772 |
bool linear_color_save = pfs->linear_color;
|
|
|
2773 |
bool inside_save = pfs->inside;
|
|
|
2774 |
gs_fixed_rect r, r1;
|
|
|
2775 |
/* Warning : pfs->monotonic_color is not restored on error. */
|
|
|
2776 |
|
|
|
2777 |
if (level > 100)
|
|
|
2778 |
return_error(gs_error_unregistered); /* Safety. */
|
|
|
2779 |
if (!pfs->inside) {
|
|
|
2780 |
bbox_of_points(&r, &p->p[0][0]->p, &p->p[0][1]->p, &p->p[1][0]->p, &p->p[1][1]->p);
|
|
|
2781 |
r1 = r;
|
|
|
2782 |
rect_intersect(r, pfs->rect);
|
|
|
2783 |
if (r.q.x <= r.p.x || r.q.y <= r.p.y)
|
|
|
2784 |
return 0; /* Outside. */
|
|
|
2785 |
}
|
|
|
2786 |
if (big) {
|
|
|
2787 |
/* Likely 'big' is an unuseful rudiment due to curve_samples
|
|
|
2788 |
restricts lengthes. We keep it for a while because its implementation
|
|
|
2789 |
isn't obvious and its time consumption is invisibly small.
|
|
|
2790 |
*/
|
|
|
2791 |
fixed size_u = max(max(any_abs(p->p[0][0]->p.x - p->p[0][1]->p.x),
|
|
|
2792 |
any_abs(p->p[1][0]->p.x - p->p[1][1]->p.x)),
|
|
|
2793 |
max(any_abs(p->p[0][0]->p.y - p->p[0][1]->p.y),
|
|
|
2794 |
any_abs(p->p[1][0]->p.y - p->p[1][1]->p.y)));
|
|
|
2795 |
fixed size_v = max(max(any_abs(p->p[0][0]->p.x - p->p[1][0]->p.x),
|
|
|
2796 |
any_abs(p->p[0][1]->p.x - p->p[1][1]->p.x)),
|
|
|
2797 |
max(any_abs(p->p[0][0]->p.y - p->p[1][0]->p.y),
|
|
|
2798 |
any_abs(p->p[0][1]->p.y - p->p[1][1]->p.y)));
|
|
|
2799 |
|
|
|
2800 |
if (QUADRANGLES && pfs->maybe_self_intersecting) {
|
|
|
2801 |
if (size_v > pfs->max_small_coord) {
|
|
|
2802 |
/* constant_color_quadrangle can't handle big self-intersecting areas
|
|
|
2803 |
because we don't want int64_t in it. */
|
|
|
2804 |
divide_v = true;
|
|
|
2805 |
} else if (size_u > pfs->max_small_coord) {
|
|
|
2806 |
/* constant_color_quadrangle can't handle big self-intersecting areas,
|
|
|
2807 |
because we don't want int64_t in it. */
|
|
|
2808 |
divide_u = true;
|
|
|
2809 |
} else
|
|
|
2810 |
big1 = false;
|
|
|
2811 |
} else
|
|
|
2812 |
big1 = false;
|
|
|
2813 |
}
|
|
|
2814 |
if (!big1) {
|
|
|
2815 |
bool is_big_u = false, is_big_v = false;
|
|
|
2816 |
double d0001x = any_abs(p->p[0][0]->p.x - p->p[0][1]->p.x);
|
|
|
2817 |
double d1011x = any_abs(p->p[1][0]->p.x - p->p[1][1]->p.x);
|
|
|
2818 |
double d0001y = any_abs(p->p[0][0]->p.y - p->p[0][1]->p.y);
|
|
|
2819 |
double d1011y = any_abs(p->p[1][0]->p.y - p->p[1][1]->p.y);
|
|
|
2820 |
double d0010x = any_abs(p->p[0][0]->p.x - p->p[1][0]->p.x);
|
|
|
2821 |
double d0111x = any_abs(p->p[0][1]->p.x - p->p[1][1]->p.x);
|
|
|
2822 |
double d0010y = any_abs(p->p[0][0]->p.y - p->p[1][0]->p.y);
|
|
|
2823 |
double d0111y = any_abs(p->p[0][1]->p.y - p->p[1][1]->p.y);
|
|
|
2824 |
|
|
|
2825 |
if (d0001x > fixed_1 || d1011x > fixed_1 || d0001y > fixed_1 || d1011y > fixed_1)
|
|
|
2826 |
is_big_u = true;
|
|
|
2827 |
if (d0010x > fixed_1 || d0111x > fixed_1 || d0010y > fixed_1 || d0111y > fixed_1)
|
|
|
2828 |
is_big_v = true;
|
|
|
2829 |
else if (!is_big_u)
|
|
|
2830 |
return (QUADRANGLES || !pfs->maybe_self_intersecting ?
|
|
|
2831 |
constant_color_quadrangle : triangles4)(pfs, p,
|
|
|
2832 |
pfs->maybe_self_intersecting);
|
|
|
2833 |
if (!pfs->monotonic_color) {
|
|
|
2834 |
bool not_monotonic_by_u = false, not_monotonic_by_v = false;
|
|
|
2835 |
|
|
|
2836 |
code = is_quadrangle_color_monotonic(pfs, p, ¬_monotonic_by_u, ¬_monotonic_by_v);
|
|
|
2837 |
if (code < 0)
|
|
|
2838 |
return code;
|
|
|
2839 |
if (is_big_u)
|
|
|
2840 |
divide_u = not_monotonic_by_u;
|
|
|
2841 |
if (is_big_v)
|
|
|
2842 |
divide_v = not_monotonic_by_v;
|
|
|
2843 |
if (!divide_u && !divide_v)
|
|
|
2844 |
pfs->monotonic_color = true;
|
|
|
2845 |
}
|
|
|
2846 |
if (pfs->monotonic_color && !pfs->linear_color) {
|
|
|
2847 |
if (divide_v && divide_u) {
|
|
|
2848 |
if (d0001x + d1011x + d0001y + d1011y > d0010x + d0111x + d0010y + d0111y)
|
|
|
2849 |
divide_v = false;
|
|
|
2850 |
else
|
|
|
2851 |
divide_u = false;
|
|
|
2852 |
} else if (!divide_u && !divide_v && !pfs->unlinear) {
|
|
|
2853 |
if (is_big_u) {
|
|
|
2854 |
code = is_quadrangle_color_linear_by_u(pfs, p);
|
|
|
2855 |
if (code < 0)
|
|
|
2856 |
return code;
|
|
|
2857 |
divide_u = !code;
|
|
|
2858 |
}
|
|
|
2859 |
if (is_big_v) {
|
|
|
2860 |
code = is_quadrangle_color_linear_by_v(pfs, p);
|
|
|
2861 |
if (code < 0)
|
|
|
2862 |
return code;
|
|
|
2863 |
divide_v = !code;
|
|
|
2864 |
}
|
|
|
2865 |
if (is_big_u && is_big_v) {
|
|
|
2866 |
code = is_quadrangle_color_linear_by_diagonals(pfs, p);
|
|
|
2867 |
if (code < 0)
|
|
|
2868 |
return code;
|
|
|
2869 |
if (!code) {
|
|
|
2870 |
if (d0001x + d1011x + d0001y + d1011y > d0010x + d0111x + d0010y + d0111y) {
|
|
|
2871 |
divide_u = true;
|
|
|
2872 |
divide_v = false;
|
|
|
2873 |
} else {
|
|
|
2874 |
divide_v = true;
|
|
|
2875 |
divide_u = false;
|
|
|
2876 |
}
|
|
|
2877 |
}
|
|
|
2878 |
}
|
|
|
2879 |
}
|
|
|
2880 |
if (!divide_u && !divide_v)
|
|
|
2881 |
pfs->linear_color = true;
|
|
|
2882 |
}
|
|
|
2883 |
if (!pfs->linear_color) {
|
|
|
2884 |
/* go to divide. */
|
|
|
2885 |
} else switch(quadrangle_color_change(pfs, p, is_big_u, is_big_v, ÷_u, ÷_v)) {
|
|
|
2886 |
case color_change_small:
|
|
|
2887 |
code = (QUADRANGLES || !pfs->maybe_self_intersecting ?
|
|
|
2888 |
constant_color_quadrangle : triangles4)(pfs, p,
|
|
|
2889 |
pfs->maybe_self_intersecting);
|
|
|
2890 |
pfs->monotonic_color = monotonic_color_save;
|
|
|
2891 |
pfs->linear_color = linear_color_save;
|
|
|
2892 |
return code;
|
|
|
2893 |
case color_change_bilinear:
|
|
|
2894 |
if (!QUADRANGLES) {
|
|
|
2895 |
code = triangles4(pfs, p, true);
|
|
|
2896 |
pfs->monotonic_color = monotonic_color_save;
|
|
|
2897 |
pfs->linear_color = linear_color_save;
|
|
|
2898 |
return code;
|
|
|
2899 |
}
|
|
|
2900 |
case color_change_linear:
|
|
|
2901 |
if (!QUADRANGLES) {
|
|
|
2902 |
code = triangles2(pfs, p, true);
|
|
|
2903 |
pfs->monotonic_color = monotonic_color_save;
|
|
|
2904 |
pfs->linear_color = linear_color_save;
|
|
|
2905 |
return code;
|
|
|
2906 |
}
|
|
|
2907 |
case color_change_gradient:
|
|
|
2908 |
case color_change_general:
|
|
|
2909 |
; /* goto divide. */
|
|
|
2910 |
}
|
|
|
2911 |
}
|
|
|
2912 |
if (!pfs->inside) {
|
|
|
2913 |
if (r.p.x == r1.p.x && r.p.y == r1.p.y &&
|
|
|
2914 |
r.q.x == r1.q.x && r.q.y == r1.q.y)
|
|
|
2915 |
pfs->inside = true;
|
|
|
2916 |
}
|
|
|
2917 |
if (LAZY_WEDGES)
|
|
|
2918 |
init_wedge_vertex_list(&l0, 1);
|
|
|
2919 |
if (divide_v) {
|
|
|
2920 |
divide_quadrangle_by_v(pfs, &s0, &s1, q, p);
|
|
|
2921 |
if (LAZY_WEDGES) {
|
|
|
2922 |
make_wedge_median(pfs, &l1, p->l0111, true, &p->p[0][1]->p, &p->p[1][1]->p, &s0.p[1][1]->p);
|
|
|
2923 |
make_wedge_median(pfs, &l2, p->l1000, false, &p->p[1][0]->p, &p->p[0][0]->p, &s0.p[1][0]->p);
|
|
|
2924 |
s0.l1110 = s1.l0001 = &l0;
|
|
|
2925 |
s0.l0111 = s1.l0111 = &l1;
|
|
|
2926 |
s0.l1000 = s1.l1000 = &l2;
|
|
|
2927 |
s0.l0001 = p->l0001;
|
|
|
2928 |
s1.l1110 = p->l1110;
|
|
|
2929 |
} else {
|
|
|
2930 |
code = fill_triangle_wedge(pfs, s0.p[0][0], s1.p[1][0], s0.p[1][0]);
|
|
|
2931 |
if (code < 0)
|
|
|
2932 |
return code;
|
|
|
2933 |
code = fill_triangle_wedge(pfs, s0.p[0][1], s1.p[1][1], s0.p[1][1]);
|
|
|
2934 |
if (code < 0)
|
|
|
2935 |
return code;
|
|
|
2936 |
}
|
|
|
2937 |
code = fill_quadrangle(pfs, &s0, big, level + 1);
|
|
|
2938 |
if (code < 0)
|
|
|
2939 |
return code;
|
|
|
2940 |
if (LAZY_WEDGES) {
|
|
|
2941 |
l0.last_side = true;
|
|
|
2942 |
move_wedge(&l1, p->l0111, true);
|
|
|
2943 |
move_wedge(&l2, p->l1000, false);
|
|
|
2944 |
}
|
|
|
2945 |
code = fill_quadrangle(pfs, &s1, big1, level + 1);
|
|
|
2946 |
if (LAZY_WEDGES) {
|
|
|
2947 |
if (code < 0)
|
|
|
2948 |
return code;
|
|
|
2949 |
code = close_wedge_median(pfs, p->l0111, &p->p[0][1]->c, &p->p[1][1]->c);
|
|
|
2950 |
if (code < 0)
|
|
|
2951 |
return code;
|
|
|
2952 |
code = close_wedge_median(pfs, p->l1000, &p->p[1][0]->c, &p->p[0][0]->c);
|
|
|
2953 |
if (code < 0)
|
|
|
2954 |
return code;
|
|
|
2955 |
code = terminate_wedge_vertex_list(pfs, &l0, &s0.p[1][0]->c, &s0.p[1][1]->c);
|
|
|
2956 |
}
|
|
|
2957 |
} else if (divide_u) {
|
|
|
2958 |
divide_quadrangle_by_u(pfs, &s0, &s1, q, p);
|
|
|
2959 |
if (LAZY_WEDGES) {
|
|
|
2960 |
make_wedge_median(pfs, &l1, p->l0001, true, &p->p[0][0]->p, &p->p[0][1]->p, &s0.p[0][1]->p);
|
|
|
2961 |
make_wedge_median(pfs, &l2, p->l1110, false, &p->p[1][1]->p, &p->p[1][0]->p, &s0.p[1][1]->p);
|
|
|
2962 |
s0.l0111 = s1.l1000 = &l0;
|
|
|
2963 |
s0.l0001 = s1.l0001 = &l1;
|
|
|
2964 |
s0.l1110 = s1.l1110 = &l2;
|
|
|
2965 |
s0.l1000 = p->l1000;
|
|
|
2966 |
s1.l0111 = p->l0111;
|
|
|
2967 |
} else {
|
|
|
2968 |
code = fill_triangle_wedge(pfs, s0.p[0][0], s1.p[0][1], s0.p[0][1]);
|
|
|
2969 |
if (code < 0)
|
|
|
2970 |
return code;
|
|
|
2971 |
code = fill_triangle_wedge(pfs, s0.p[1][0], s1.p[1][1], s0.p[1][1]);
|
|
|
2972 |
if (code < 0)
|
|
|
2973 |
return code;
|
|
|
2974 |
}
|
|
|
2975 |
code = fill_quadrangle(pfs, &s0, big1, level + 1);
|
|
|
2976 |
if (code < 0)
|
|
|
2977 |
return code;
|
|
|
2978 |
if (LAZY_WEDGES) {
|
|
|
2979 |
l0.last_side = true;
|
|
|
2980 |
move_wedge(&l1, p->l0001, true);
|
|
|
2981 |
move_wedge(&l2, p->l1110, false);
|
|
|
2982 |
}
|
|
|
2983 |
code = fill_quadrangle(pfs, &s1, big1, level + 1);
|
|
|
2984 |
if (LAZY_WEDGES) {
|
|
|
2985 |
if (code < 0)
|
|
|
2986 |
return code;
|
|
|
2987 |
code = close_wedge_median(pfs, p->l0001, &p->p[0][0]->c, &p->p[0][1]->c);
|
|
|
2988 |
if (code < 0)
|
|
|
2989 |
return code;
|
|
|
2990 |
code = close_wedge_median(pfs, p->l1110, &p->p[1][1]->c, &p->p[1][0]->c);
|
|
|
2991 |
if (code < 0)
|
|
|
2992 |
return code;
|
|
|
2993 |
code = terminate_wedge_vertex_list(pfs, &l0, &s0.p[0][1]->c, &s0.p[1][1]->c);
|
|
|
2994 |
}
|
|
|
2995 |
} else
|
|
|
2996 |
code = (QUADRANGLES || !pfs->maybe_self_intersecting ?
|
|
|
2997 |
constant_color_quadrangle : triangles4)(pfs, p,
|
|
|
2998 |
pfs->maybe_self_intersecting);
|
|
|
2999 |
pfs->monotonic_color = monotonic_color_save;
|
|
|
3000 |
pfs->linear_color = linear_color_save;
|
|
|
3001 |
pfs->inside = inside_save;
|
|
|
3002 |
return code;
|
|
|
3003 |
}
|
|
|
3004 |
|
|
|
3005 |
|
|
|
3006 |
|
|
|
3007 |
private inline void
|
|
|
3008 |
split_stripe(patch_fill_state_t *pfs, tensor_patch *s0, tensor_patch *s1, const tensor_patch *p)
|
|
|
3009 |
{
|
|
|
3010 |
split_curve_s(p->pole[0], s0->pole[0], s1->pole[0], 1);
|
|
|
3011 |
split_curve_s(p->pole[1], s0->pole[1], s1->pole[1], 1);
|
|
|
3012 |
split_curve_s(p->pole[2], s0->pole[2], s1->pole[2], 1);
|
|
|
3013 |
split_curve_s(p->pole[3], s0->pole[3], s1->pole[3], 1);
|
|
|
3014 |
s0->c[0][0] = p->c[0][0];
|
|
|
3015 |
s0->c[1][0] = p->c[1][0];
|
|
|
3016 |
patch_interpolate_color(&s0->c[0][1], &p->c[0][0], &p->c[0][1], pfs, 0.5);
|
|
|
3017 |
patch_interpolate_color(&s0->c[1][1], &p->c[1][0], &p->c[1][1], pfs, 0.5);
|
|
|
3018 |
s1->c[0][0] = s0->c[0][1];
|
|
|
3019 |
s1->c[1][0] = s0->c[1][1];
|
|
|
3020 |
s1->c[0][1] = p->c[0][1];
|
|
|
3021 |
s1->c[1][1] = p->c[1][1];
|
|
|
3022 |
}
|
|
|
3023 |
|
|
|
3024 |
private inline void
|
|
|
3025 |
split_patch(patch_fill_state_t *pfs, tensor_patch *s0, tensor_patch *s1, const tensor_patch *p)
|
|
|
3026 |
{
|
|
|
3027 |
split_curve_s(&p->pole[0][0], &s0->pole[0][0], &s1->pole[0][0], 4);
|
|
|
3028 |
split_curve_s(&p->pole[0][1], &s0->pole[0][1], &s1->pole[0][1], 4);
|
|
|
3029 |
split_curve_s(&p->pole[0][2], &s0->pole[0][2], &s1->pole[0][2], 4);
|
|
|
3030 |
split_curve_s(&p->pole[0][3], &s0->pole[0][3], &s1->pole[0][3], 4);
|
|
|
3031 |
s0->c[0][0] = p->c[0][0];
|
|
|
3032 |
s0->c[0][1] = p->c[0][1];
|
|
|
3033 |
patch_interpolate_color(&s0->c[1][0], &p->c[0][0], &p->c[1][0], pfs, 0.5);
|
|
|
3034 |
patch_interpolate_color(&s0->c[1][1], &p->c[0][1], &p->c[1][1], pfs, 0.5);
|
|
|
3035 |
s1->c[0][0] = s0->c[1][0];
|
|
|
3036 |
s1->c[0][1] = s0->c[1][1];
|
|
|
3037 |
s1->c[1][0] = p->c[1][0];
|
|
|
3038 |
s1->c[1][1] = p->c[1][1];
|
|
|
3039 |
}
|
|
|
3040 |
|
|
|
3041 |
private int
|
|
|
3042 |
decompose_stripe(patch_fill_state_t *pfs, const tensor_patch *p, int ku)
|
|
|
3043 |
{
|
|
|
3044 |
if (ku > 1) {
|
|
|
3045 |
tensor_patch s0, s1;
|
|
|
3046 |
int code;
|
|
|
3047 |
|
|
|
3048 |
split_stripe(pfs, &s0, &s1, p);
|
|
|
3049 |
if (0) { /* Debug purpose only. */
|
|
|
3050 |
draw_patch(&s0, true, RGB(0, 128, 128));
|
|
|
3051 |
draw_patch(&s1, true, RGB(0, 128, 128));
|
|
|
3052 |
}
|
|
|
3053 |
code = decompose_stripe(pfs, &s0, ku / 2);
|
|
|
3054 |
if (code < 0)
|
|
|
3055 |
return code;
|
|
|
3056 |
return decompose_stripe(pfs, &s1, ku / 2);
|
|
|
3057 |
} else {
|
|
|
3058 |
quadrangle_patch q;
|
|
|
3059 |
shading_vertex_t qq[2][2];
|
|
|
3060 |
wedge_vertex_list_t l[4];
|
|
|
3061 |
int code;
|
|
|
3062 |
|
|
|
3063 |
init_wedge_vertex_list(l, count_of(l));
|
|
|
3064 |
make_quadrangle(p, qq, l, &q);
|
|
|
3065 |
# if SKIP_TEST
|
|
|
3066 |
dbg_quad_cnt++;
|
|
|
3067 |
# endif
|
|
|
3068 |
code = fill_quadrangle(pfs, &q, true, 0);
|
|
|
3069 |
if (LAZY_WEDGES) {
|
|
|
3070 |
code = terminate_wedge_vertex_list(pfs, &l[0], &q.p[0][0]->c, &q.p[0][1]->c);
|
|
|
3071 |
if (code < 0)
|
|
|
3072 |
return code;
|
|
|
3073 |
code = terminate_wedge_vertex_list(pfs, &l[1], &q.p[0][1]->c, &q.p[1][1]->c);
|
|
|
3074 |
if (code < 0)
|
|
|
3075 |
return code;
|
|
|
3076 |
code = terminate_wedge_vertex_list(pfs, &l[2], &q.p[1][1]->c, &q.p[1][0]->c);
|
|
|
3077 |
if (code < 0)
|
|
|
3078 |
return code;
|
|
|
3079 |
code = terminate_wedge_vertex_list(pfs, &l[3], &q.p[1][0]->c, &q.p[0][1]->c);
|
|
|
3080 |
if (code < 0)
|
|
|
3081 |
return code;
|
|
|
3082 |
}
|
|
|
3083 |
return code;
|
|
|
3084 |
}
|
|
|
3085 |
}
|
|
|
3086 |
|
|
|
3087 |
private int
|
|
|
3088 |
fill_stripe(patch_fill_state_t *pfs, const tensor_patch *p)
|
|
|
3089 |
{
|
|
|
3090 |
/* The stripe is flattened enough by V, so ignore inner poles. */
|
|
|
3091 |
int ku[4], kum, code;
|
|
|
3092 |
|
|
|
3093 |
/* We would like to apply iterations for enumerating the kum curve parts,
|
|
|
3094 |
but the roundinmg errors would be too complicated due to
|
|
|
3095 |
the dependence on the direction. Note that neigbour
|
|
|
3096 |
patches may use the opposite direction for same bounding curve.
|
|
|
3097 |
We apply the recursive dichotomy, in which
|
|
|
3098 |
the rounding errors do not depend on the direction. */
|
|
|
3099 |
ku[0] = curve_samples(pfs, p->pole[0], 1, pfs->fixed_flat);
|
|
|
3100 |
ku[3] = curve_samples(pfs, p->pole[3], 1, pfs->fixed_flat);
|
|
|
3101 |
kum = max(ku[0], ku[3]);
|
|
|
3102 |
code = fill_wedges(pfs, ku[0], kum, p->pole[0], 1, &p->c[0][0], &p->c[0][1], inpatch_wedge);
|
|
|
3103 |
if (code < 0)
|
|
|
3104 |
return code;
|
|
|
3105 |
if (INTERPATCH_PADDING) {
|
|
|
3106 |
vd_bar(p->pole[0][0].x, p->pole[0][0].y, p->pole[3][0].x, p->pole[3][0].y, 0, RGB(255, 0, 0));
|
|
|
3107 |
code = mesh_padding(pfs, &p->pole[0][0], &p->pole[3][0], &p->c[0][0], &p->c[1][0]);
|
|
|
3108 |
if (code < 0)
|
|
|
3109 |
return code;
|
|
|
3110 |
vd_bar(p->pole[0][3].x, p->pole[0][3].y, p->pole[3][3].x, p->pole[3][3].y, 0, RGB(255, 0, 0));
|
|
|
3111 |
code = mesh_padding(pfs, &p->pole[0][3], &p->pole[3][3], &p->c[0][1], &p->c[1][1]);
|
|
|
3112 |
if (code < 0)
|
|
|
3113 |
return code;
|
|
|
3114 |
}
|
|
|
3115 |
code = decompose_stripe(pfs, p, kum);
|
|
|
3116 |
if (code < 0)
|
|
|
3117 |
return code;
|
|
|
3118 |
return fill_wedges(pfs, ku[3], kum, p->pole[3], 1, &p->c[1][0], &p->c[1][1], inpatch_wedge);
|
|
|
3119 |
}
|
|
|
3120 |
|
|
|
3121 |
private inline bool
|
|
|
3122 |
is_curve_x_monotonic(const gs_fixed_point *pole, int pole_step)
|
|
|
3123 |
{ /* true = monotonic, false = don't know. */
|
|
|
3124 |
return (pole[0 * pole_step].x <= pole[1 * pole_step].x &&
|
|
|
3125 |
pole[1 * pole_step].x <= pole[2 * pole_step].x &&
|
|
|
3126 |
pole[2 * pole_step].x <= pole[3 * pole_step].x) ||
|
|
|
3127 |
(pole[0 * pole_step].x >= pole[1 * pole_step].x &&
|
|
|
3128 |
pole[1 * pole_step].x >= pole[2 * pole_step].x &&
|
|
|
3129 |
pole[2 * pole_step].x >= pole[3 * pole_step].x);
|
|
|
3130 |
}
|
|
|
3131 |
|
|
|
3132 |
private inline bool
|
|
|
3133 |
is_curve_y_monotonic(const gs_fixed_point *pole, int pole_step)
|
|
|
3134 |
{ /* true = monotonic, false = don't know. */
|
|
|
3135 |
return (pole[0 * pole_step].y <= pole[1 * pole_step].y &&
|
|
|
3136 |
pole[1 * pole_step].y <= pole[2 * pole_step].y &&
|
|
|
3137 |
pole[2 * pole_step].y <= pole[3 * pole_step].y) ||
|
|
|
3138 |
(pole[0 * pole_step].y >= pole[1 * pole_step].y &&
|
|
|
3139 |
pole[1 * pole_step].y >= pole[2 * pole_step].y &&
|
|
|
3140 |
pole[2 * pole_step].y >= pole[3 * pole_step].y);
|
|
|
3141 |
}
|
|
|
3142 |
|
|
|
3143 |
private inline bool neqs(int *a, int b)
|
|
|
3144 |
{ /* Unequal signs. Assuming -1, 0, 1 only. */
|
|
|
3145 |
if (*a * b < 0)
|
|
|
3146 |
return true;
|
|
|
3147 |
if (!*a)
|
|
|
3148 |
*a = b;
|
|
|
3149 |
return false;
|
|
|
3150 |
}
|
|
|
3151 |
|
|
|
3152 |
private inline int
|
|
|
3153 |
vector_pair_orientation(const gs_fixed_point *p0, const gs_fixed_point *p1, const gs_fixed_point *p2)
|
|
|
3154 |
{ fixed dx1 = p1->x - p0->x, dy1 = p1->y - p0->y;
|
|
|
3155 |
fixed dx2 = p2->x - p0->x, dy2 = p2->y - p0->y;
|
|
|
3156 |
int64_t vp = (int64_t)dx1 * dy2 - (int64_t)dy1 * dx2;
|
|
|
3157 |
|
|
|
3158 |
return (vp > 0 ? 1 : vp < 0 ? -1 : 0);
|
|
|
3159 |
}
|
|
|
3160 |
|
|
|
3161 |
private inline bool
|
|
|
3162 |
is_x_bended(const tensor_patch *p)
|
|
|
3163 |
{
|
|
|
3164 |
int sign = vector_pair_orientation(&p->pole[0][0], &p->pole[0][1], &p->pole[1][0]);
|
|
|
3165 |
|
|
|
3166 |
if (neqs(&sign, vector_pair_orientation(&p->pole[0][1], &p->pole[0][2], &p->pole[1][1])))
|
|
|
3167 |
return true;
|
|
|
3168 |
if (neqs(&sign, vector_pair_orientation(&p->pole[0][2], &p->pole[0][3], &p->pole[1][2])))
|
|
|
3169 |
return true;
|
|
|
3170 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[0][3], &p->pole[0][2], &p->pole[1][3])))
|
|
|
3171 |
return true;
|
|
|
3172 |
|
|
|
3173 |
if (neqs(&sign, vector_pair_orientation(&p->pole[1][1], &p->pole[1][2], &p->pole[2][1])))
|
|
|
3174 |
return true;
|
|
|
3175 |
if (neqs(&sign, vector_pair_orientation(&p->pole[1][1], &p->pole[1][2], &p->pole[2][1])))
|
|
|
3176 |
return true;
|
|
|
3177 |
if (neqs(&sign, vector_pair_orientation(&p->pole[1][2], &p->pole[1][3], &p->pole[2][2])))
|
|
|
3178 |
return true;
|
|
|
3179 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[1][3], &p->pole[1][2], &p->pole[2][3])))
|
|
|
3180 |
return true;
|
|
|
3181 |
|
|
|
3182 |
if (neqs(&sign, vector_pair_orientation(&p->pole[2][1], &p->pole[2][2], &p->pole[3][1])))
|
|
|
3183 |
return true;
|
|
|
3184 |
if (neqs(&sign, vector_pair_orientation(&p->pole[2][1], &p->pole[2][2], &p->pole[3][1])))
|
|
|
3185 |
return true;
|
|
|
3186 |
if (neqs(&sign, vector_pair_orientation(&p->pole[2][2], &p->pole[2][3], &p->pole[3][2])))
|
|
|
3187 |
return true;
|
|
|
3188 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[2][3], &p->pole[2][2], &p->pole[3][3])))
|
|
|
3189 |
return true;
|
|
|
3190 |
|
|
|
3191 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[3][1], &p->pole[3][2], &p->pole[2][1])))
|
|
|
3192 |
return true;
|
|
|
3193 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[3][1], &p->pole[3][2], &p->pole[2][1])))
|
|
|
3194 |
return true;
|
|
|
3195 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[3][2], &p->pole[3][3], &p->pole[2][2])))
|
|
|
3196 |
return true;
|
|
|
3197 |
if (neqs(&sign, vector_pair_orientation(&p->pole[3][3], &p->pole[3][2], &p->pole[2][3])))
|
|
|
3198 |
return true;
|
|
|
3199 |
return false;
|
|
|
3200 |
}
|
|
|
3201 |
|
|
|
3202 |
private inline bool
|
|
|
3203 |
is_y_bended(const tensor_patch *p)
|
|
|
3204 |
{
|
|
|
3205 |
int sign = vector_pair_orientation(&p->pole[0][0], &p->pole[1][0], &p->pole[0][1]);
|
|
|
3206 |
|
|
|
3207 |
if (neqs(&sign, vector_pair_orientation(&p->pole[1][0], &p->pole[2][0], &p->pole[1][1])))
|
|
|
3208 |
return true;
|
|
|
3209 |
if (neqs(&sign, vector_pair_orientation(&p->pole[2][0], &p->pole[3][0], &p->pole[2][1])))
|
|
|
3210 |
return true;
|
|
|
3211 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[3][0], &p->pole[2][0], &p->pole[3][1])))
|
|
|
3212 |
return true;
|
|
|
3213 |
|
|
|
3214 |
if (neqs(&sign, vector_pair_orientation(&p->pole[1][1], &p->pole[2][1], &p->pole[1][2])))
|
|
|
3215 |
return true;
|
|
|
3216 |
if (neqs(&sign, vector_pair_orientation(&p->pole[1][1], &p->pole[2][1], &p->pole[1][2])))
|
|
|
3217 |
return true;
|
|
|
3218 |
if (neqs(&sign, vector_pair_orientation(&p->pole[2][1], &p->pole[3][1], &p->pole[2][2])))
|
|
|
3219 |
return true;
|
|
|
3220 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[3][1], &p->pole[2][1], &p->pole[3][2])))
|
|
|
3221 |
return true;
|
|
|
3222 |
|
|
|
3223 |
if (neqs(&sign, vector_pair_orientation(&p->pole[1][2], &p->pole[2][2], &p->pole[1][3])))
|
|
|
3224 |
return true;
|
|
|
3225 |
if (neqs(&sign, vector_pair_orientation(&p->pole[1][2], &p->pole[2][2], &p->pole[1][3])))
|
|
|
3226 |
return true;
|
|
|
3227 |
if (neqs(&sign, vector_pair_orientation(&p->pole[2][2], &p->pole[3][2], &p->pole[2][3])))
|
|
|
3228 |
return true;
|
|
|
3229 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[3][2], &p->pole[2][2], &p->pole[3][3])))
|
|
|
3230 |
return true;
|
|
|
3231 |
|
|
|
3232 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[1][3], &p->pole[2][3], &p->pole[1][2])))
|
|
|
3233 |
return true;
|
|
|
3234 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[1][3], &p->pole[2][3], &p->pole[1][2])))
|
|
|
3235 |
return true;
|
|
|
3236 |
if (neqs(&sign, -vector_pair_orientation(&p->pole[2][3], &p->pole[3][3], &p->pole[2][2])))
|
|
|
3237 |
return true;
|
|
|
3238 |
if (neqs(&sign, vector_pair_orientation(&p->pole[3][3], &p->pole[2][3], &p->pole[3][2])))
|
|
|
3239 |
return true;
|
|
|
3240 |
return false;
|
|
|
3241 |
}
|
|
|
3242 |
|
|
|
3243 |
private inline bool
|
|
|
3244 |
is_curve_x_small(const gs_fixed_point *pole, int pole_step, fixed fixed_flat)
|
|
|
3245 |
{ /* Is curve within a single pixel, or smaller than half pixel ? */
|
|
|
3246 |
fixed xmin0 = min(pole[0 * pole_step].x, pole[1 * pole_step].x);
|
|
|
3247 |
fixed xmin1 = min(pole[2 * pole_step].x, pole[3 * pole_step].x);
|
|
|
3248 |
fixed xmin = min(xmin0, xmin1);
|
|
|
3249 |
fixed xmax0 = max(pole[0 * pole_step].x, pole[1 * pole_step].x);
|
|
|
3250 |
fixed xmax1 = max(pole[2 * pole_step].x, pole[3 * pole_step].x);
|
|
|
3251 |
fixed xmax = max(xmax0, xmax1);
|
|
|
3252 |
|
|
|
3253 |
if(xmax - xmin <= fixed_1)
|
|
|
3254 |
return true;
|
|
|
3255 |
return false;
|
|
|
3256 |
}
|
|
|
3257 |
|
|
|
3258 |
private inline bool
|
|
|
3259 |
is_curve_y_small(const gs_fixed_point *pole, int pole_step, fixed fixed_flat)
|
|
|
3260 |
{ /* Is curve within a single pixel, or smaller than half pixel ? */
|
|
|
3261 |
fixed ymin0 = min(pole[0 * pole_step].y, pole[1 * pole_step].y);
|
|
|
3262 |
fixed ymin1 = min(pole[2 * pole_step].y, pole[3 * pole_step].y);
|
|
|
3263 |
fixed ymin = min(ymin0, ymin1);
|
|
|
3264 |
fixed ymax0 = max(pole[0 * pole_step].y, pole[1 * pole_step].y);
|
|
|
3265 |
fixed ymax1 = max(pole[2 * pole_step].y, pole[3 * pole_step].y);
|
|
|
3266 |
fixed ymax = max(ymax0, ymax1);
|
|
|
3267 |
|
|
|
3268 |
if (ymax - ymin <= fixed_1)
|
|
|
3269 |
return true;
|
|
|
3270 |
return false;
|
|
|
3271 |
}
|
|
|
3272 |
|
|
|
3273 |
private inline bool
|
|
|
3274 |
is_patch_narrow(const patch_fill_state_t *pfs, const tensor_patch *p)
|
|
|
3275 |
{
|
|
|
3276 |
if (!is_curve_x_small(&p->pole[0][0], 4, pfs->fixed_flat))
|
|
|
3277 |
return false;
|
|
|
3278 |
if (!is_curve_x_small(&p->pole[0][1], 4, pfs->fixed_flat))
|
|
|
3279 |
return false;
|
|
|
3280 |
if (!is_curve_x_small(&p->pole[0][2], 4, pfs->fixed_flat))
|
|
|
3281 |
return false;
|
|
|
3282 |
if (!is_curve_x_small(&p->pole[0][3], 4, pfs->fixed_flat))
|
|
|
3283 |
return false;
|
|
|
3284 |
if (!is_curve_y_small(&p->pole[0][0], 4, pfs->fixed_flat))
|
|
|
3285 |
return false;
|
|
|
3286 |
if (!is_curve_y_small(&p->pole[0][1], 4, pfs->fixed_flat))
|
|
|
3287 |
return false;
|
|
|
3288 |
if (!is_curve_y_small(&p->pole[0][2], 4, pfs->fixed_flat))
|
|
|
3289 |
return false;
|
|
|
3290 |
if (!is_curve_y_small(&p->pole[0][3], 4, pfs->fixed_flat))
|
|
|
3291 |
return false;
|
|
|
3292 |
return true;
|
|
|
3293 |
}
|
|
|
3294 |
|
|
|
3295 |
private int
|
|
|
3296 |
fill_patch(patch_fill_state_t *pfs, const tensor_patch *p, int kv, int kv0, int kv1)
|
|
|
3297 |
{
|
|
|
3298 |
if (kv <= 1) {
|
|
|
3299 |
if (is_patch_narrow(pfs, p))
|
|
|
3300 |
return fill_stripe(pfs, p);
|
|
|
3301 |
if (!is_x_bended(p))
|
|
|
3302 |
return fill_stripe(pfs, p);
|
|
|
3303 |
}
|
|
|
3304 |
{ tensor_patch s0, s1;
|
|
|
3305 |
shading_vertex_t q0, q1, q2;
|
|
|
3306 |
int code;
|
|
|
3307 |
|
|
|
3308 |
split_patch(pfs, &s0, &s1, p);
|
|
|
3309 |
if (kv0 <= 1) {
|
|
|
3310 |
q0.p = s0.pole[0][0];
|
|
|
3311 |
q0.c = s0.c[0][0];
|
|
|
3312 |
q1.p = s1.pole[3][0];
|
|
|
3313 |
q1.c = s1.c[1][0];
|
|
|
3314 |
q2.p = s0.pole[3][0];
|
|
|
3315 |
q2.c = s0.c[1][0];
|
|
|
3316 |
code = fill_triangle_wedge(pfs, &q0, &q1, &q2);
|
|
|
3317 |
if (code < 0)
|
|
|
3318 |
return code;
|
|
|
3319 |
}
|
|
|
3320 |
if (kv1 <= 1) {
|
|
|
3321 |
q0.p = s0.pole[0][3];
|
|
|
3322 |
q0.c = s0.c[0][1];
|
|
|
3323 |
q1.p = s1.pole[3][3];
|
|
|
3324 |
q1.c = s1.c[1][1];
|
|
|
3325 |
q2.p = s0.pole[3][3];
|
|
|
3326 |
q2.c = s0.c[1][1];
|
|
|
3327 |
code = fill_triangle_wedge(pfs, &q0, &q1, &q2);
|
|
|
3328 |
if (code < 0)
|
|
|
3329 |
return code;
|
|
|
3330 |
}
|
|
|
3331 |
code = fill_patch(pfs, &s0, kv / 2, kv0 / 2, kv1 / 2);
|
|
|
3332 |
if (code < 0)
|
|
|
3333 |
return code;
|
|
|
3334 |
return fill_patch(pfs, &s1, kv / 2, kv0 / 2, kv1 / 2);
|
|
|
3335 |
/* fixme : To privide the precise filling order, we must
|
|
|
3336 |
decompose left and right wedges into pieces by intersections
|
|
|
3337 |
with stripes, and fill each piece with its stripe.
|
|
|
3338 |
A lazy wedge list would be fine for storing
|
|
|
3339 |
the necessary information.
|
|
|
3340 |
|
|
|
3341 |
If the patch is created from a radial shading,
|
|
|
3342 |
the wedge color appears a constant, so the filling order
|
|
|
3343 |
isn't important. The order is important for other
|
|
|
3344 |
self-overlapping patches, but the visible effect is
|
|
|
3345 |
just a slight norrowing the patch (as its lower layer appears
|
|
|
3346 |
visible through the upper layer near the side).
|
|
|
3347 |
This kind of dropout isn't harmful, because
|
|
|
3348 |
contacring self-overlapping patches are painted
|
|
|
3349 |
one after one by definition, so that a side coverage break
|
|
|
3350 |
appears unavoidable by definition.
|
|
|
3351 |
|
|
|
3352 |
Delaying this improvement because it is low important.
|
|
|
3353 |
*/
|
|
|
3354 |
}
|
|
|
3355 |
}
|
|
|
3356 |
|
|
|
3357 |
private inline fixed
|
|
|
3358 |
lcp1(fixed p0, fixed p3)
|
|
|
3359 |
{ /* Computing the 1st pole of a 3d order besier, which applears a line. */
|
|
|
3360 |
return (p0 + p0 + p3) / 3;
|
|
|
3361 |
}
|
|
|
3362 |
private inline fixed
|
|
|
3363 |
lcp2(fixed p0, fixed p3)
|
|
|
3364 |
{ /* Computing the 2nd pole of a 3d order besier, which applears a line. */
|
|
|
3365 |
return (p0 + p3 + p3) / 3;
|
|
|
3366 |
}
|
|
|
3367 |
|
|
|
3368 |
private void
|
|
|
3369 |
patch_set_color(const patch_fill_state_t *pfs, patch_color_t *c, const float *cc)
|
|
|
3370 |
{
|
|
|
3371 |
if (pfs->Function) {
|
|
|
3372 |
c->t[0] = cc[0];
|
|
|
3373 |
c->t[1] = cc[1];
|
|
|
3374 |
} else
|
|
|
3375 |
memcpy(c->cc.paint.values, cc, sizeof(c->cc.paint.values[0]) * pfs->num_components);
|
|
|
3376 |
}
|
|
|
3377 |
|
|
|
3378 |
private void
|
|
|
3379 |
make_tensor_patch(const patch_fill_state_t *pfs, tensor_patch *p, const patch_curve_t curve[4],
|
|
|
3380 |
const gs_fixed_point interior[4])
|
|
|
3381 |
{
|
|
|
3382 |
const gs_color_space *pcs = pfs->direct_space;
|
|
|
3383 |
|
|
|
3384 |
p->pole[0][0] = curve[0].vertex.p;
|
|
|
3385 |
p->pole[1][0] = curve[0].control[0];
|
|
|
3386 |
p->pole[2][0] = curve[0].control[1];
|
|
|
3387 |
p->pole[3][0] = curve[1].vertex.p;
|
|
|
3388 |
p->pole[3][1] = curve[1].control[0];
|
|
|
3389 |
p->pole[3][2] = curve[1].control[1];
|
|
|
3390 |
p->pole[3][3] = curve[2].vertex.p;
|
|
|
3391 |
p->pole[2][3] = curve[2].control[0];
|
|
|
3392 |
p->pole[1][3] = curve[2].control[1];
|
|
|
3393 |
p->pole[0][3] = curve[3].vertex.p;
|
|
|
3394 |
p->pole[0][2] = curve[3].control[0];
|
|
|
3395 |
p->pole[0][1] = curve[3].control[1];
|
|
|
3396 |
if (interior != NULL) {
|
|
|
3397 |
p->pole[1][1] = interior[0];
|
|
|
3398 |
p->pole[1][2] = interior[1];
|
|
|
3399 |
p->pole[2][2] = interior[2];
|
|
|
3400 |
p->pole[2][1] = interior[3];
|
|
|
3401 |
} else {
|
|
|
3402 |
p->pole[1][1].x = lcp1(p->pole[0][1].x, p->pole[3][1].x) +
|
|
|
3403 |
lcp1(p->pole[1][0].x, p->pole[1][3].x) -
|
|
|
3404 |
lcp1(lcp1(p->pole[0][0].x, p->pole[0][3].x),
|
|
|
3405 |
lcp1(p->pole[3][0].x, p->pole[3][3].x));
|
|
|
3406 |
p->pole[1][2].x = lcp1(p->pole[0][2].x, p->pole[3][2].x) +
|
|
|
3407 |
lcp2(p->pole[1][0].x, p->pole[1][3].x) -
|
|
|
3408 |
lcp1(lcp2(p->pole[0][0].x, p->pole[0][3].x),
|
|
|
3409 |
lcp2(p->pole[3][0].x, p->pole[3][3].x));
|
|
|
3410 |
p->pole[2][1].x = lcp2(p->pole[0][1].x, p->pole[3][1].x) +
|
|
|
3411 |
lcp1(p->pole[2][0].x, p->pole[2][3].x) -
|
|
|
3412 |
lcp2(lcp1(p->pole[0][0].x, p->pole[0][3].x),
|
|
|
3413 |
lcp1(p->pole[3][0].x, p->pole[3][3].x));
|
|
|
3414 |
p->pole[2][2].x = lcp2(p->pole[0][2].x, p->pole[3][2].x) +
|
|
|
3415 |
lcp2(p->pole[2][0].x, p->pole[2][3].x) -
|
|
|
3416 |
lcp2(lcp2(p->pole[0][0].x, p->pole[0][3].x),
|
|
|
3417 |
lcp2(p->pole[3][0].x, p->pole[3][3].x));
|
|
|
3418 |
|
|
|
3419 |
p->pole[1][1].y = lcp1(p->pole[0][1].y, p->pole[3][1].y) +
|
|
|
3420 |
lcp1(p->pole[1][0].y, p->pole[1][3].y) -
|
|
|
3421 |
lcp1(lcp1(p->pole[0][0].y, p->pole[0][3].y),
|
|
|
3422 |
lcp1(p->pole[3][0].y, p->pole[3][3].y));
|
|
|
3423 |
p->pole[1][2].y = lcp1(p->pole[0][2].y, p->pole[3][2].y) +
|
|
|
3424 |
lcp2(p->pole[1][0].y, p->pole[1][3].y) -
|
|
|
3425 |
lcp1(lcp2(p->pole[0][0].y, p->pole[0][3].y),
|
|
|
3426 |
lcp2(p->pole[3][0].y, p->pole[3][3].y));
|
|
|
3427 |
p->pole[2][1].y = lcp2(p->pole[0][1].y, p->pole[3][1].y) +
|
|
|
3428 |
lcp1(p->pole[2][0].y, p->pole[2][3].y) -
|
|
|
3429 |
lcp2(lcp1(p->pole[0][0].y, p->pole[0][3].y),
|
|
|
3430 |
lcp1(p->pole[3][0].y, p->pole[3][3].y));
|
|
|
3431 |
p->pole[2][2].y = lcp2(p->pole[0][2].y, p->pole[3][2].y) +
|
|
|
3432 |
lcp2(p->pole[2][0].y, p->pole[2][3].y) -
|
|
|
3433 |
lcp2(lcp2(p->pole[0][0].y, p->pole[0][3].y),
|
|
|
3434 |
lcp2(p->pole[3][0].y, p->pole[3][3].y));
|
|
|
3435 |
}
|
|
|
3436 |
patch_set_color(pfs, &p->c[0][0], curve[0].vertex.cc);
|
|
|
3437 |
patch_set_color(pfs, &p->c[1][0], curve[1].vertex.cc);
|
|
|
3438 |
patch_set_color(pfs, &p->c[1][1], curve[2].vertex.cc);
|
|
|
3439 |
patch_set_color(pfs, &p->c[0][1], curve[3].vertex.cc);
|
|
|
3440 |
patch_resolve_color_inline(&p->c[0][0], pfs);
|
|
|
3441 |
patch_resolve_color_inline(&p->c[0][1], pfs);
|
|
|
3442 |
patch_resolve_color_inline(&p->c[1][0], pfs);
|
|
|
3443 |
patch_resolve_color_inline(&p->c[1][1], pfs);
|
|
|
3444 |
if (!pfs->Function) {
|
|
|
3445 |
pcs->type->restrict_color(&p->c[0][0].cc, pcs);
|
|
|
3446 |
pcs->type->restrict_color(&p->c[0][1].cc, pcs);
|
|
|
3447 |
pcs->type->restrict_color(&p->c[1][0].cc, pcs);
|
|
|
3448 |
pcs->type->restrict_color(&p->c[1][1].cc, pcs);
|
|
|
3449 |
}
|
|
|
3450 |
}
|
|
|
3451 |
|
|
|
3452 |
int
|
|
|
3453 |
gx_shade_background(gx_device *pdev, const gs_fixed_rect *rect,
|
|
|
3454 |
const gx_device_color *pdevc, gs_logical_operation_t log_op)
|
|
|
3455 |
{
|
|
|
3456 |
gs_fixed_edge le, re;
|
|
|
3457 |
|
|
|
3458 |
le.start.x = rect->p.x - INTERPATCH_PADDING;
|
|
|
3459 |
le.start.y = rect->p.y - INTERPATCH_PADDING;
|
|
|
3460 |
le.end.x = rect->p.x - INTERPATCH_PADDING;
|
|
|
3461 |
le.end.y = rect->q.y + INTERPATCH_PADDING;
|
|
|
3462 |
re.start.x = rect->q.x + INTERPATCH_PADDING;
|
|
|
3463 |
re.start.y = rect->p.y - INTERPATCH_PADDING;
|
|
|
3464 |
re.end.x = rect->q.x + INTERPATCH_PADDING;
|
|
|
3465 |
re.end.y = rect->q.y + INTERPATCH_PADDING;
|
|
|
3466 |
return dev_proc(pdev, fill_trapezoid)(pdev,
|
|
|
3467 |
&le, &re, le.start.y, le.end.y, false, pdevc, log_op);
|
|
|
3468 |
}
|
|
|
3469 |
|
|
|
3470 |
|
|
|
3471 |
int
|
|
|
3472 |
patch_fill(patch_fill_state_t *pfs, const patch_curve_t curve[4],
|
|
|
3473 |
const gs_fixed_point interior[4],
|
|
|
3474 |
void (*transform) (gs_fixed_point *, const patch_curve_t[4],
|
|
|
3475 |
const gs_fixed_point[4], floatp, floatp))
|
|
|
3476 |
{
|
|
|
3477 |
tensor_patch p;
|
|
|
3478 |
int kv[4], kvm, ku[4], kum, km;
|
|
|
3479 |
int code = 0;
|
|
|
3480 |
|
|
|
3481 |
#if SKIP_TEST
|
|
|
3482 |
dbg_patch_cnt++;
|
|
|
3483 |
/*if (dbg_patch_cnt != 67 && dbg_patch_cnt != 78)
|
|
|
3484 |
return 0;*/
|
|
|
3485 |
#endif
|
|
|
3486 |
/* We decompose the patch into tiny quadrangles,
|
|
|
3487 |
possibly inserting wedges between them against a dropout. */
|
|
|
3488 |
make_tensor_patch(pfs, &p, curve, interior);
|
|
|
3489 |
pfs->unlinear = !is_linear_color_applicable(pfs);
|
|
|
3490 |
pfs->linear_color = false;
|
|
|
3491 |
if ((*dev_proc(pfs->dev, pattern_manage))(pfs->dev,
|
|
|
3492 |
gs_no_id, NULL, pattern_manage__shading_area) > 0) {
|
|
|
3493 |
/* Inform the device with the shading coverage area.
|
|
|
3494 |
First compute the sign of the area, because
|
|
|
3495 |
all areas to be clipped in same direction. */
|
|
|
3496 |
gx_device *pdev = pfs->dev;
|
|
|
3497 |
gx_path path;
|
|
|
3498 |
fixed d01x = (curve[1].vertex.p.x - curve[0].vertex.p.x) >> 1;
|
|
|
3499 |
fixed d01y = (curve[1].vertex.p.y - curve[0].vertex.p.y) >> 1;
|
|
|
3500 |
fixed d12x = (curve[2].vertex.p.x - curve[1].vertex.p.x) >> 1;
|
|
|
3501 |
fixed d12y = (curve[2].vertex.p.y - curve[1].vertex.p.y) >> 1;
|
|
|
3502 |
fixed d23x = (curve[3].vertex.p.x - curve[2].vertex.p.x) >> 1;
|
|
|
3503 |
fixed d23y = (curve[3].vertex.p.y - curve[2].vertex.p.y) >> 1;
|
|
|
3504 |
fixed d30x = (curve[0].vertex.p.x - curve[3].vertex.p.x) >> 1;
|
|
|
3505 |
fixed d30y = (curve[0].vertex.p.y - curve[3].vertex.p.y) >> 1;
|
|
|
3506 |
int64_t s1 = (int64_t)d01x * d12y - (int64_t)d01y * d12x;
|
|
|
3507 |
int64_t s2 = (int64_t)d23x * d30y - (int64_t)d23y * d30x;
|
|
|
3508 |
int s = (s1 + s2 > 0 ? 1 : 3), i, j, k, jj, l = (s == 1 ? 0 : 1);
|
|
|
3509 |
|
|
|
3510 |
gx_path_init_local(&path, pdev->memory);
|
|
|
3511 |
if (is_x_bended(&p) || is_y_bended(&p)) {
|
|
|
3512 |
/* The patch possibly is self-overlapping,
|
|
|
3513 |
so the patch coverage may fall outside the patch outline.
|
|
|
3514 |
In this case we pass an empty path,
|
|
|
3515 |
and the device must use a bitmap mask instead clipping. */
|
|
|
3516 |
} else {
|
|
|
3517 |
code = gx_path_add_point(&path, curve[0].vertex.p.x, curve[0].vertex.p.y);
|
|
|
3518 |
for (i = k = 0; k < 4 && code >= 0; i = j, k++) {
|
|
|
3519 |
j = (i + s) % 4, jj = (s == 1 ? i : j);
|
|
|
3520 |
if (curve[jj].straight)
|
|
|
3521 |
code = gx_path_add_line(&path, curve[j].vertex.p.x,
|
|
|
3522 |
curve[j].vertex.p.y);
|
|
|
3523 |
else
|
|
|
3524 |
code = gx_path_add_curve(&path, curve[jj].control[l].x, curve[jj].control[l].y,
|
|
|
3525 |
curve[jj].control[(l + 1) & 1].x, curve[jj].control[(l + 1) & 1].y,
|
|
|
3526 |
curve[j].vertex.p.x,
|
|
|
3527 |
curve[j].vertex.p.y);
|
|
|
3528 |
}
|
|
|
3529 |
if (code >= 0)
|
|
|
3530 |
code = gx_path_close_subpath(&path);
|
|
|
3531 |
}
|
|
|
3532 |
if (code >= 0)
|
|
|
3533 |
code = (*dev_proc(pfs->dev, fill_path))(pdev, NULL, &path, NULL, NULL, NULL);
|
|
|
3534 |
gx_path_free(&path, "patch_fill");
|
|
|
3535 |
if (code < 0)
|
|
|
3536 |
return code;
|
|
|
3537 |
}
|
|
|
3538 |
/* draw_patch(&p, true, RGB(0, 0, 0)); */
|
|
|
3539 |
kv[0] = curve_samples(pfs, &p.pole[0][0], 4, pfs->fixed_flat);
|
|
|
3540 |
kv[1] = curve_samples(pfs, &p.pole[0][1], 4, pfs->fixed_flat);
|
|
|
3541 |
kv[2] = curve_samples(pfs, &p.pole[0][2], 4, pfs->fixed_flat);
|
|
|
3542 |
kv[3] = curve_samples(pfs, &p.pole[0][3], 4, pfs->fixed_flat);
|
|
|
3543 |
kvm = max(max(kv[0], kv[1]), max(kv[2], kv[3]));
|
|
|
3544 |
ku[0] = curve_samples(pfs, p.pole[0], 1, pfs->fixed_flat);
|
|
|
3545 |
ku[3] = curve_samples(pfs, p.pole[3], 1, pfs->fixed_flat);
|
|
|
3546 |
kum = max(ku[0], ku[3]);
|
|
|
3547 |
km = max(kvm, kum);
|
|
|
3548 |
# if NOFILL_TEST
|
|
|
3549 |
dbg_nofill = false;
|
|
|
3550 |
# endif
|
|
|
3551 |
code = fill_wedges(pfs, ku[0], kum, p.pole[0], 1, &p.c[0][0], &p.c[0][1],
|
|
|
3552 |
interpatch_padding | inpatch_wedge);
|
|
|
3553 |
if (code >= 0) {
|
|
|
3554 |
/* We would like to apply iterations for enumerating the kvm curve parts,
|
|
|
3555 |
but the roundinmg errors would be too complicated due to
|
|
|
3556 |
the dependence on the direction. Note that neigbour
|
|
|
3557 |
patches may use the opposite direction for same bounding curve.
|
|
|
3558 |
We apply the recursive dichotomy, in which
|
|
|
3559 |
the rounding errors do not depend on the direction. */
|
|
|
3560 |
# if NOFILL_TEST
|
|
|
3561 |
dbg_nofill = false;
|
|
|
3562 |
code = fill_patch(pfs, &p, kvm, kv[0], kv[3]);
|
|
|
3563 |
dbg_nofill = true;
|
|
|
3564 |
# endif
|
|
|
3565 |
code = fill_patch(pfs, &p, kvm, kv[0], kv[3]);
|
|
|
3566 |
}
|
|
|
3567 |
if (code >= 0)
|
|
|
3568 |
code = fill_wedges(pfs, ku[3], kum, p.pole[3], 1, &p.c[1][0], &p.c[1][1],
|
|
|
3569 |
interpatch_padding | inpatch_wedge);
|
|
|
3570 |
return code;
|
|
|
3571 |
}
|