cairo/src/cairo-traps.c

1124 lines
33 KiB
C
Raw Normal View History

/* -*- Mode: c; tab-width: 8; c-basic-offset: 4; indent-tabs-mode: t; -*- */
2002-08-07 19:48:49 +00:00
/*
* Copyright © 2002 Keith Packard
* Copyright © 2007 Red Hat, Inc.
2002-08-07 19:48:49 +00:00
*
* This library is free software; you can redistribute it and/or
* modify it either under the terms of the GNU Lesser General Public
* License version 2.1 as published by the Free Software Foundation
* (the "LGPL") or, at your option, under the terms of the Mozilla
* Public License Version 1.1 (the "MPL"). If you do not alter this
* notice, a recipient may use your version of this file under either
* the MPL or the LGPL.
2002-08-07 19:48:49 +00:00
*
* You should have received a copy of the LGPL along with this library
* in the file COPYING-LGPL-2.1; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Suite 500, Boston, MA 02110-1335, USA
* You should have received a copy of the MPL along with this library
* in the file COPYING-MPL-1.1
*
* The contents of this file are subject to the Mozilla Public License
* Version 1.1 (the "License"); you may not use this file except in
* compliance with the License. You may obtain a copy of the License at
* http://www.mozilla.org/MPL/
*
* This software is distributed on an "AS IS" basis, WITHOUT WARRANTY
* OF ANY KIND, either express or implied. See the LGPL or the MPL for
* the specific language governing rights and limitations.
*
* The Original Code is the cairo graphics library.
*
* The Initial Developer of the Original Code is Keith Packard
*
* Contributor(s):
* Keith R. Packard <keithp@keithp.com>
* Carl D. Worth <cworth@cworth.org>
2002-08-07 19:48:49 +00:00
*
* 2002-07-15: Converted from XRenderCompositeDoublePoly to #cairo_trap_t. Carl D. Worth
2002-08-07 19:48:49 +00:00
*/
2003-07-18 11:34:19 +00:00
#include "cairoint.h"
2002-08-07 19:48:49 +00:00
#include "cairo-box-inline.h"
#include "cairo-boxes-private.h"
#include "cairo-error-private.h"
#include "cairo-line-private.h"
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
#include "cairo-region-private.h"
#include "cairo-slope-private.h"
#include "cairo-traps-private.h"
#include "cairo-spans-private.h"
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
2002-08-07 19:48:49 +00:00
/* private functions */
void
_cairo_traps_init (cairo_traps_t *traps)
2002-08-07 19:48:49 +00:00
{
VG (VALGRIND_MAKE_MEM_UNDEFINED (traps, sizeof (cairo_traps_t)));
traps->status = CAIRO_STATUS_SUCCESS;
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
traps->maybe_region = 1;
traps->is_rectilinear = 0;
traps->is_rectangular = 0;
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
traps->num_traps = 0;
2002-08-07 19:48:49 +00:00
traps->traps_size = ARRAY_LENGTH (traps->traps_embedded);
traps->traps = traps->traps_embedded;
traps->num_limits = 0;
traps->has_intersections = FALSE;
}
void
_cairo_traps_limit (cairo_traps_t *traps,
const cairo_box_t *limits,
int num_limits)
{
int i;
traps->limits = limits;
traps->num_limits = num_limits;
traps->bounds = limits[0];
for (i = 1; i < num_limits; i++)
_cairo_box_add_box (&traps->bounds, &limits[i]);
2002-08-07 19:48:49 +00:00
}
void
_cairo_traps_init_with_clip (cairo_traps_t *traps,
const cairo_clip_t *clip)
{
_cairo_traps_init (traps);
if (clip)
_cairo_traps_limit (traps, clip->boxes, clip->num_boxes);
}
void
_cairo_traps_clear (cairo_traps_t *traps)
{
traps->status = CAIRO_STATUS_SUCCESS;
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
traps->maybe_region = 1;
traps->is_rectilinear = 0;
traps->is_rectangular = 0;
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
traps->num_traps = 0;
traps->has_intersections = FALSE;
}
2002-08-07 19:48:49 +00:00
void
_cairo_traps_fini (cairo_traps_t *traps)
2002-08-07 19:48:49 +00:00
{
if (traps->traps != traps->traps_embedded)
free (traps->traps);
VG (VALGRIND_MAKE_MEM_UNDEFINED (traps, sizeof (cairo_traps_t)));
2002-08-07 19:48:49 +00:00
}
/* make room for at least one more trap */
static cairo_bool_t
_cairo_traps_grow (cairo_traps_t *traps)
{
cairo_trapezoid_t *new_traps;
int new_size = 4 * traps->traps_size;
if (CAIRO_INJECT_FAULT ()) {
traps->status = _cairo_error (CAIRO_STATUS_NO_MEMORY);
return FALSE;
}
if (traps->traps == traps->traps_embedded) {
new_traps = _cairo_malloc_ab (new_size, sizeof (cairo_trapezoid_t));
if (new_traps != NULL)
memcpy (new_traps, traps->traps, sizeof (traps->traps_embedded));
} else {
new_traps = _cairo_realloc_ab (traps->traps,
new_size, sizeof (cairo_trapezoid_t));
}
if (unlikely (new_traps == NULL)) {
traps->status = _cairo_error (CAIRO_STATUS_NO_MEMORY);
return FALSE;
}
traps->traps = new_traps;
traps->traps_size = new_size;
return TRUE;
}
void
_cairo_traps_add_trap (cairo_traps_t *traps,
cairo_fixed_t top, cairo_fixed_t bottom,
const cairo_line_t *left,
const cairo_line_t *right)
2002-08-07 19:48:49 +00:00
{
cairo_trapezoid_t *trap;
2002-08-07 19:48:49 +00:00
assert (left->p1.y != left->p2.y);
assert (right->p1.y != right->p2.y);
assert (bottom > top);
if (unlikely (traps->num_traps == traps->traps_size)) {
if (unlikely (! _cairo_traps_grow (traps)))
2009-07-24 13:51:23 +01:00
return;
}
trap = &traps->traps[traps->num_traps++];
trap->top = top;
trap->bottom = bottom;
trap->left = *left;
trap->right = *right;
}
static void
_cairo_traps_add_clipped_trap (cairo_traps_t *traps,
cairo_fixed_t _top, cairo_fixed_t _bottom,
const cairo_line_t *_left,
const cairo_line_t *_right)
{
/* Note: With the goofy trapezoid specification, (where an
* arbitrary two points on the lines can specified for the left
* and right edges), these limit checks would not work in
* general. For example, one can imagine a trapezoid entirely
* within the limits, but with two points used to specify the left
* edge entirely to the right of the limits. Fortunately, for our
* purposes, cairo will never generate such a crazy
* trapezoid. Instead, cairo always uses for its points the
* extreme positions of the edge that are visible on at least some
* trapezoid. With this constraint, it's impossible for both
* points to be outside the limits while the relevant edge is
* entirely inside the limits.
*/
if (traps->num_limits) {
const cairo_box_t *b = &traps->bounds;
cairo_fixed_t top = _top, bottom = _bottom;
cairo_line_t left = *_left, right = *_right;
/* Trivially reject if trapezoid is entirely to the right or
* to the left of the limits. */
if (left.p1.x >= b->p2.x && left.p2.x >= b->p2.x)
return;
if (right.p1.x <= b->p1.x && right.p2.x <= b->p1.x)
return;
/* And reject if the trapezoid is entirely above or below */
if (top >= b->p2.y || bottom <= b->p1.y)
return;
/* Otherwise, clip the trapezoid to the limits. We only clip
* where an edge is entirely outside the limits. If we wanted
* to be more clever, we could handle cases where a trapezoid
* edge intersects the edge of the limits, but that would
* require slicing this trapezoid into multiple trapezoids,
* and I'm not sure the effort would be worth it. */
if (top < b->p1.y)
top = b->p1.y;
if (bottom > b->p2.y)
bottom = b->p2.y;
if (left.p1.x <= b->p1.x && left.p2.x <= b->p1.x)
left.p1.x = left.p2.x = b->p1.x;
if (right.p1.x >= b->p2.x && right.p2.x >= b->p2.x)
right.p1.x = right.p2.x = b->p2.x;
/* Trivial discards for empty trapezoids that are likely to
* be produced by our tessellators (most notably convex_quad
* when given a simple rectangle).
*/
if (top >= bottom)
return;
/* cheap colinearity check */
if (right.p1.x <= left.p1.x && right.p1.y == left.p1.y &&
right.p2.x <= left.p2.x && right.p2.y == left.p2.y)
return;
_cairo_traps_add_trap (traps, top, bottom, &left, &right);
} else
_cairo_traps_add_trap (traps, _top, _bottom, _left, _right);
}
static int
_compare_point_fixed_by_y (const void *av, const void *bv)
{
const cairo_point_t *a = av, *b = bv;
int ret = a->y - b->y;
if (ret == 0)
ret = a->x - b->x;
return ret;
}
void
_cairo_traps_tessellate_convex_quad (cairo_traps_t *traps,
const cairo_point_t q[4])
{
int a, b, c, d;
int i;
cairo_slope_t ab, ad;
cairo_bool_t b_left_of_d;
cairo_line_t left;
cairo_line_t right;
/* Choose a as a point with minimal y */
a = 0;
for (i = 1; i < 4; i++)
if (_compare_point_fixed_by_y (&q[i], &q[a]) < 0)
a = i;
/* b and d are adjacent to a, while c is opposite */
b = (a + 1) % 4;
c = (a + 2) % 4;
d = (a + 3) % 4;
/* Choose between b and d so that b.y is less than d.y */
if (_compare_point_fixed_by_y (&q[d], &q[b]) < 0) {
b = (a + 3) % 4;
d = (a + 1) % 4;
}
/* Without freedom left to choose anything else, we have four
* cases to tessellate.
*
* First, we have to determine the Y-axis sort of the four
* vertices, (either abcd or abdc). After that we need to determine
* which edges will be "left" and which will be "right" in the
* resulting trapezoids. This can be determined by computing a
* slope comparison of ab and ad to determine if b is left of d or
* not.
*
* Note that "left of" here is in the sense of which edges should
* be the left vs. right edges of the trapezoid. In particular, b
* left of d does *not* mean that b.x is less than d.x.
*
* This should hopefully be made clear in the lame ASCII art
* below. Since the same slope comparison is used in all cases, we
* compute it before testing for the Y-value sort. */
/* Note: If a == b then the ab slope doesn't give us any
* information. In that case, we can replace it with the ac (or
* equivalenly the bc) slope which gives us exactly the same
* information we need. At worst the names of the identifiers ab
* and b_left_of_d are inaccurate in this case, (would be ac, and
* c_left_of_d). */
if (q[a].x == q[b].x && q[a].y == q[b].y)
_cairo_slope_init (&ab, &q[a], &q[c]);
else
_cairo_slope_init (&ab, &q[a], &q[b]);
_cairo_slope_init (&ad, &q[a], &q[d]);
b_left_of_d = _cairo_slope_compare (&ab, &ad) > 0;
if (q[c].y <= q[d].y) {
if (b_left_of_d) {
/* Y-sort is abcd and b is left of d, (slope(ab) > slope (ad))
*
* top bot left right
* _a a a
* / / /| |\ a.y b.y ab ad
* b / b | b \
* / / | | \ \ b.y c.y bc ad
* c / c | c \
* | / \| \ \ c.y d.y cd ad
* d d d
*/
left.p1 = q[a]; left.p2 = q[b];
right.p1 = q[a]; right.p2 = q[d];
_cairo_traps_add_clipped_trap (traps, q[a].y, q[b].y, &left, &right);
left.p1 = q[b]; left.p2 = q[c];
_cairo_traps_add_clipped_trap (traps, q[b].y, q[c].y, &left, &right);
left.p1 = q[c]; left.p2 = q[d];
_cairo_traps_add_clipped_trap (traps, q[c].y, q[d].y, &left, &right);
} else {
/* Y-sort is abcd and b is right of d, (slope(ab) <= slope (ad))
*
* a a a_
* /| |\ \ \ a.y b.y ad ab
* / b | b \ b
* / / | | \ \ b.y c.y ad bc
* / c | c \ c
* / / |/ \ | c.y d.y ad cd
* d d d
*/
left.p1 = q[a]; left.p2 = q[d];
right.p1 = q[a]; right.p2 = q[b];
_cairo_traps_add_clipped_trap (traps, q[a].y, q[b].y, &left, &right);
right.p1 = q[b]; right.p2 = q[c];
_cairo_traps_add_clipped_trap (traps, q[b].y, q[c].y, &left, &right);
right.p1 = q[c]; right.p2 = q[d];
_cairo_traps_add_clipped_trap (traps, q[c].y, q[d].y, &left, &right);
}
} else {
if (b_left_of_d) {
/* Y-sort is abdc and b is left of d, (slope (ab) > slope (ad))
*
* a a a
* // / \ |\ a.y b.y ab ad
* /b/ b \ b \
* / / \ \ \ \ b.y d.y bc ad
* /d/ \ d \ d
* // \ / \| d.y c.y bc dc
* c c c
*/
left.p1 = q[a]; left.p2 = q[b];
right.p1 = q[a]; right.p2 = q[d];
_cairo_traps_add_clipped_trap (traps, q[a].y, q[b].y, &left, &right);
left.p1 = q[b]; left.p2 = q[c];
_cairo_traps_add_clipped_trap (traps, q[b].y, q[d].y, &left, &right);
right.p1 = q[d]; right.p2 = q[c];
_cairo_traps_add_clipped_trap (traps, q[d].y, q[c].y, &left, &right);
} else {
/* Y-sort is abdc and b is right of d, (slope (ab) <= slope (ad))
*
* a a a
* /| / \ \\ a.y b.y ad ab
* / b / b \b\
* / / / / \ \ b.y d.y ad bc
* d / d / \d\
* |/ \ / \\ d.y c.y dc bc
* c c c
*/
left.p1 = q[a]; left.p2 = q[d];
right.p1 = q[a]; right.p2 = q[b];
_cairo_traps_add_clipped_trap (traps, q[a].y, q[b].y, &left, &right);
right.p1 = q[b]; right.p2 = q[c];
_cairo_traps_add_clipped_trap (traps, q[b].y, q[d].y, &left, &right);
left.p1 = q[d]; left.p2 = q[c];
_cairo_traps_add_clipped_trap (traps, q[d].y, q[c].y, &left, &right);
}
}
}
static void add_tri (cairo_traps_t *traps,
int y1, int y2,
const cairo_line_t *left,
const cairo_line_t *right)
{
if (y2 < y1) {
int tmp = y1;
y1 = y2;
y2 = tmp;
}
if (cairo_lines_compare_at_y (left, right, y1) > 0) {
const cairo_line_t *tmp = left;
left = right;
right = tmp;
}
_cairo_traps_add_clipped_trap (traps, y1, y2, left, right);
}
void
_cairo_traps_tessellate_triangle_with_edges (cairo_traps_t *traps,
const cairo_point_t t[3],
const cairo_point_t edges[4])
{
cairo_line_t lines[3];
if (edges[0].y <= edges[1].y) {
lines[0].p1 = edges[0];
lines[0].p2 = edges[1];
} else {
lines[0].p1 = edges[1];
lines[0].p2 = edges[0];
}
if (edges[2].y <= edges[3].y) {
lines[1].p1 = edges[2];
lines[1].p2 = edges[3];
} else {
lines[1].p1 = edges[3];
lines[1].p2 = edges[2];
}
if (t[1].y == t[2].y) {
add_tri (traps, t[0].y, t[1].y, &lines[0], &lines[1]);
return;
}
if (t[1].y <= t[2].y) {
lines[2].p1 = t[1];
lines[2].p2 = t[2];
} else {
lines[2].p1 = t[2];
lines[2].p2 = t[1];
}
if (((t[1].y - t[0].y) < 0) ^ ((t[2].y - t[0].y) < 0)) {
add_tri (traps, t[0].y, t[1].y, &lines[0], &lines[2]);
add_tri (traps, t[0].y, t[2].y, &lines[1], &lines[2]);
} else if (abs(t[1].y - t[0].y) < abs(t[2].y - t[0].y)) {
add_tri (traps, t[0].y, t[1].y, &lines[0], &lines[1]);
add_tri (traps, t[1].y, t[2].y, &lines[2], &lines[1]);
} else {
add_tri (traps, t[0].y, t[2].y, &lines[1], &lines[0]);
add_tri (traps, t[1].y, t[2].y, &lines[2], &lines[0]);
}
}
/**
* _cairo_traps_init_boxes:
* @traps: a #cairo_traps_t
* @box: an array box that will each be converted to a single trapezoid
* to store in @traps.
*
* Initializes a #cairo_traps_t to contain an array of rectangular
* trapezoids.
**/
cairo_status_t
_cairo_traps_init_boxes (cairo_traps_t *traps,
const cairo_boxes_t *boxes)
{
cairo_trapezoid_t *trap;
const struct _cairo_boxes_chunk *chunk;
_cairo_traps_init (traps);
while (traps->traps_size < boxes->num_boxes) {
if (unlikely (! _cairo_traps_grow (traps))) {
_cairo_traps_fini (traps);
return _cairo_error (CAIRO_STATUS_NO_MEMORY);
}
}
traps->num_traps = boxes->num_boxes;
traps->is_rectilinear = TRUE;
traps->is_rectangular = TRUE;
traps->maybe_region = boxes->is_pixel_aligned;
trap = &traps->traps[0];
for (chunk = &boxes->chunks; chunk != NULL; chunk = chunk->next) {
const cairo_box_t *box;
int i;
box = chunk->base;
for (i = 0; i < chunk->count; i++) {
trap->top = box->p1.y;
trap->bottom = box->p2.y;
trap->left.p1 = box->p1;
trap->left.p2.x = box->p1.x;
trap->left.p2.y = box->p2.y;
trap->right.p1.x = box->p2.x;
trap->right.p1.y = box->p1.y;
trap->right.p2 = box->p2;
box++, trap++;
}
}
return CAIRO_STATUS_SUCCESS;
}
2009-07-24 13:51:23 +01:00
cairo_status_t
_cairo_traps_tessellate_rectangle (cairo_traps_t *traps,
const cairo_point_t *top_left,
const cairo_point_t *bottom_right)
{
cairo_line_t left;
cairo_line_t right;
cairo_fixed_t top, bottom;
if (top_left->y == bottom_right->y)
return CAIRO_STATUS_SUCCESS;
if (top_left->x == bottom_right->x)
return CAIRO_STATUS_SUCCESS;
2009-07-24 13:51:23 +01:00
left.p1.x = left.p2.x = top_left->x;
left.p1.y = right.p1.y = top_left->y;
right.p1.x = right.p2.x = bottom_right->x;
left.p2.y = right.p2.y = bottom_right->y;
top = top_left->y;
bottom = bottom_right->y;
if (traps->num_limits) {
cairo_bool_t reversed;
int n;
if (top >= traps->bounds.p2.y || bottom <= traps->bounds.p1.y)
return CAIRO_STATUS_SUCCESS;
/* support counter-clockwise winding for rectangular tessellation */
reversed = top_left->x > bottom_right->x;
if (reversed) {
right.p1.x = right.p2.x = top_left->x;
left.p1.x = left.p2.x = bottom_right->x;
}
if (left.p1.x >= traps->bounds.p2.x || right.p1.x <= traps->bounds.p1.x)
return CAIRO_STATUS_SUCCESS;
for (n = 0; n < traps->num_limits; n++) {
const cairo_box_t *limits = &traps->limits[n];
cairo_line_t _left, _right;
cairo_fixed_t _top, _bottom;
if (top >= limits->p2.y)
continue;
if (bottom <= limits->p1.y)
continue;
/* Trivially reject if trapezoid is entirely to the right or
* to the left of the limits. */
if (left.p1.x >= limits->p2.x)
continue;
if (right.p1.x <= limits->p1.x)
continue;
/* Otherwise, clip the trapezoid to the limits. */
_top = top;
if (_top < limits->p1.y)
_top = limits->p1.y;
_bottom = bottom;
if (_bottom > limits->p2.y)
_bottom = limits->p2.y;
if (_bottom <= _top)
continue;
_left = left;
if (_left.p1.x < limits->p1.x) {
_left.p1.x = limits->p1.x;
_left.p1.y = limits->p1.y;
_left.p2.x = limits->p1.x;
_left.p2.y = limits->p2.y;
}
_right = right;
if (_right.p1.x > limits->p2.x) {
_right.p1.x = limits->p2.x;
_right.p1.y = limits->p1.y;
_right.p2.x = limits->p2.x;
_right.p2.y = limits->p2.y;
}
if (left.p1.x >= right.p1.x)
continue;
if (reversed)
_cairo_traps_add_trap (traps, _top, _bottom, &_right, &_left);
else
_cairo_traps_add_trap (traps, _top, _bottom, &_left, &_right);
}
} else {
_cairo_traps_add_trap (traps, top, bottom, &left, &right);
}
2002-08-07 19:48:49 +00:00
2009-07-24 13:51:23 +01:00
return traps->status;
2002-08-07 19:48:49 +00:00
}
void
_cairo_traps_translate (cairo_traps_t *traps, int x, int y)
{
cairo_fixed_t xoff, yoff;
cairo_trapezoid_t *t;
int i;
/* Ugh. The cairo_composite/(Render) interface doesn't allow
an offset for the trapezoids. Need to manually shift all
the coordinates to align with the offset origin of the
intermediate surface. */
xoff = _cairo_fixed_from_int (x);
yoff = _cairo_fixed_from_int (y);
for (i = 0, t = traps->traps; i < traps->num_traps; i++, t++) {
t->top += yoff;
t->bottom += yoff;
t->left.p1.x += xoff;
t->left.p1.y += yoff;
t->left.p2.x += xoff;
t->left.p2.y += yoff;
t->right.p1.x += xoff;
t->right.p1.y += yoff;
t->right.p2.x += xoff;
t->right.p2.y += yoff;
}
}
void
_cairo_trapezoid_array_translate_and_scale (cairo_trapezoid_t *offset_traps,
cairo_trapezoid_t *src_traps,
int num_traps,
double tx, double ty,
double sx, double sy)
{
int i;
cairo_fixed_t xoff = _cairo_fixed_from_double (tx);
cairo_fixed_t yoff = _cairo_fixed_from_double (ty);
if (sx == 1.0 && sy == 1.0) {
for (i = 0; i < num_traps; i++) {
offset_traps[i].top = src_traps[i].top + yoff;
offset_traps[i].bottom = src_traps[i].bottom + yoff;
offset_traps[i].left.p1.x = src_traps[i].left.p1.x + xoff;
offset_traps[i].left.p1.y = src_traps[i].left.p1.y + yoff;
offset_traps[i].left.p2.x = src_traps[i].left.p2.x + xoff;
offset_traps[i].left.p2.y = src_traps[i].left.p2.y + yoff;
offset_traps[i].right.p1.x = src_traps[i].right.p1.x + xoff;
offset_traps[i].right.p1.y = src_traps[i].right.p1.y + yoff;
offset_traps[i].right.p2.x = src_traps[i].right.p2.x + xoff;
offset_traps[i].right.p2.y = src_traps[i].right.p2.y + yoff;
}
} else {
cairo_fixed_t xsc = _cairo_fixed_from_double (sx);
cairo_fixed_t ysc = _cairo_fixed_from_double (sy);
for (i = 0; i < num_traps; i++) {
offset_traps[i].top = _cairo_fixed_mul (src_traps[i].top + yoff, ysc);
offset_traps[i].bottom = _cairo_fixed_mul (src_traps[i].bottom + yoff, ysc);
offset_traps[i].left.p1.x = _cairo_fixed_mul (src_traps[i].left.p1.x + xoff, xsc);
offset_traps[i].left.p1.y = _cairo_fixed_mul (src_traps[i].left.p1.y + yoff, ysc);
offset_traps[i].left.p2.x = _cairo_fixed_mul (src_traps[i].left.p2.x + xoff, xsc);
offset_traps[i].left.p2.y = _cairo_fixed_mul (src_traps[i].left.p2.y + yoff, ysc);
offset_traps[i].right.p1.x = _cairo_fixed_mul (src_traps[i].right.p1.x + xoff, xsc);
offset_traps[i].right.p1.y = _cairo_fixed_mul (src_traps[i].right.p1.y + yoff, ysc);
offset_traps[i].right.p2.x = _cairo_fixed_mul (src_traps[i].right.p2.x + xoff, xsc);
offset_traps[i].right.p2.y = _cairo_fixed_mul (src_traps[i].right.p2.y + yoff, ysc);
}
}
}
static cairo_bool_t
_cairo_trap_contains (cairo_trapezoid_t *t, cairo_point_t *pt)
{
cairo_slope_t slope_left, slope_pt, slope_right;
if (t->top > pt->y)
return FALSE;
if (t->bottom < pt->y)
return FALSE;
_cairo_slope_init (&slope_left, &t->left.p1, &t->left.p2);
_cairo_slope_init (&slope_pt, &t->left.p1, pt);
if (_cairo_slope_compare (&slope_left, &slope_pt) < 0)
return FALSE;
_cairo_slope_init (&slope_right, &t->right.p1, &t->right.p2);
_cairo_slope_init (&slope_pt, &t->right.p1, pt);
if (_cairo_slope_compare (&slope_pt, &slope_right) < 0)
return FALSE;
return TRUE;
}
cairo_bool_t
_cairo_traps_contain (const cairo_traps_t *traps,
double x, double y)
{
int i;
cairo_point_t point;
point.x = _cairo_fixed_from_double (x);
point.y = _cairo_fixed_from_double (y);
for (i = 0; i < traps->num_traps; i++) {
if (_cairo_trap_contains (&traps->traps[i], &point))
return TRUE;
}
return FALSE;
}
2004-04-06 09:36:12 +00:00
2009-07-24 13:51:23 +01:00
static cairo_fixed_t
_line_compute_intersection_x_for_y (const cairo_line_t *line,
cairo_fixed_t y)
{
return _cairo_edge_compute_intersection_x_for_y (&line->p1, &line->p2, y);
}
2004-04-06 09:36:12 +00:00
void
_cairo_traps_extents (const cairo_traps_t *traps,
2009-07-24 13:51:23 +01:00
cairo_box_t *extents)
2004-04-06 09:36:12 +00:00
{
2009-07-24 13:51:23 +01:00
int i;
if (traps->num_traps == 0) {
2009-07-24 13:51:23 +01:00
extents->p1.x = extents->p1.y = 0;
extents->p2.x = extents->p2.y = 0;
return;
}
extents->p1.x = extents->p1.y = INT32_MAX;
extents->p2.x = extents->p2.y = INT32_MIN;
for (i = 0; i < traps->num_traps; i++) {
const cairo_trapezoid_t *trap = &traps->traps[i];
if (trap->top < extents->p1.y)
extents->p1.y = trap->top;
if (trap->bottom > extents->p2.y)
extents->p2.y = trap->bottom;
if (trap->left.p1.x < extents->p1.x) {
cairo_fixed_t x = trap->left.p1.x;
if (trap->top != trap->left.p1.y) {
x = _line_compute_intersection_x_for_y (&trap->left,
trap->top);
if (x < extents->p1.x)
extents->p1.x = x;
} else
extents->p1.x = x;
}
if (trap->left.p2.x < extents->p1.x) {
cairo_fixed_t x = trap->left.p2.x;
if (trap->bottom != trap->left.p2.y) {
x = _line_compute_intersection_x_for_y (&trap->left,
trap->bottom);
if (x < extents->p1.x)
extents->p1.x = x;
} else
extents->p1.x = x;
}
if (trap->right.p1.x > extents->p2.x) {
cairo_fixed_t x = trap->right.p1.x;
if (trap->top != trap->right.p1.y) {
x = _line_compute_intersection_x_for_y (&trap->right,
trap->top);
if (x > extents->p2.x)
extents->p2.x = x;
} else
extents->p2.x = x;
}
if (trap->right.p2.x > extents->p2.x) {
cairo_fixed_t x = trap->right.p2.x;
if (trap->bottom != trap->right.p2.y) {
x = _line_compute_intersection_x_for_y (&trap->right,
trap->bottom);
if (x > extents->p2.x)
extents->p2.x = x;
} else
extents->p2.x = x;
}
}
2004-04-06 09:36:12 +00:00
}
static cairo_bool_t
_mono_edge_is_vertical (const cairo_line_t *line)
{
return _cairo_fixed_integer_round_down (line->p1.x) == _cairo_fixed_integer_round_down (line->p2.x);
}
static cairo_bool_t
_traps_are_pixel_aligned (cairo_traps_t *traps,
cairo_antialias_t antialias)
{
int i;
if (antialias == CAIRO_ANTIALIAS_NONE) {
for (i = 0; i < traps->num_traps; i++) {
if (! _mono_edge_is_vertical (&traps->traps[i].left) ||
! _mono_edge_is_vertical (&traps->traps[i].right))
{
traps->maybe_region = FALSE;
return FALSE;
}
}
} else {
for (i = 0; i < traps->num_traps; i++) {
if (traps->traps[i].left.p1.x != traps->traps[i].left.p2.x ||
traps->traps[i].right.p1.x != traps->traps[i].right.p2.x ||
! _cairo_fixed_is_integer (traps->traps[i].top) ||
! _cairo_fixed_is_integer (traps->traps[i].bottom) ||
! _cairo_fixed_is_integer (traps->traps[i].left.p1.x) ||
! _cairo_fixed_is_integer (traps->traps[i].right.p1.x))
{
traps->maybe_region = FALSE;
return FALSE;
}
}
}
return TRUE;
}
2009-07-24 13:51:23 +01:00
/**
* _cairo_traps_extract_region:
* @traps: a #cairo_traps_t
* @region: a #cairo_region_t
*
* Determines if a set of trapezoids are exactly representable as a
* cairo region. If so, the passed-in region is initialized to
* the area representing the given traps. It should be finalized
* with cairo_region_fini(). If not, %CAIRO_INT_STATUS_UNSUPPORTED
* is returned.
*
* Return value: %CAIRO_STATUS_SUCCESS, %CAIRO_INT_STATUS_UNSUPPORTED
* or %CAIRO_STATUS_NO_MEMORY
**/
cairo_int_status_t
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
_cairo_traps_extract_region (cairo_traps_t *traps,
cairo_antialias_t antialias,
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
cairo_region_t **region)
{
cairo_rectangle_int_t stack_rects[CAIRO_STACK_ARRAY_LENGTH (cairo_rectangle_int_t)];
cairo_rectangle_int_t *rects = stack_rects;
cairo_int_status_t status;
int i, rect_count;
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
/* we only treat this a hint... */
if (antialias != CAIRO_ANTIALIAS_NONE && ! traps->maybe_region)
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
return CAIRO_INT_STATUS_UNSUPPORTED;
if (! _traps_are_pixel_aligned (traps, antialias)) {
traps->maybe_region = FALSE;
return CAIRO_INT_STATUS_UNSUPPORTED;
}
if (traps->num_traps > ARRAY_LENGTH (stack_rects)) {
rects = _cairo_malloc_ab (traps->num_traps, sizeof (cairo_rectangle_int_t));
if (unlikely (rects == NULL))
return _cairo_error (CAIRO_STATUS_NO_MEMORY);
}
rect_count = 0;
for (i = 0; i < traps->num_traps; i++) {
int x1, y1, x2, y2;
if (antialias == CAIRO_ANTIALIAS_NONE) {
x1 = _cairo_fixed_integer_round_down (traps->traps[i].left.p1.x);
y1 = _cairo_fixed_integer_round_down (traps->traps[i].top);
x2 = _cairo_fixed_integer_round_down (traps->traps[i].right.p1.x);
y2 = _cairo_fixed_integer_round_down (traps->traps[i].bottom);
} else {
x1 = _cairo_fixed_integer_part (traps->traps[i].left.p1.x);
y1 = _cairo_fixed_integer_part (traps->traps[i].top);
x2 = _cairo_fixed_integer_part (traps->traps[i].right.p1.x);
y2 = _cairo_fixed_integer_part (traps->traps[i].bottom);
}
if (x2 > x1 && y2 > y1) {
rects[rect_count].x = x1;
rects[rect_count].y = y1;
rects[rect_count].width = x2 - x1;
rects[rect_count].height = y2 - y1;
rect_count++;
}
}
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
*region = cairo_region_create_rectangles (rects, rect_count);
Remove clip handling from generic surface layer. Handling clip as part of the surface state, as opposed to being part of the operation state, is cumbersome and a hindrance to providing true proxy surface support. For example, the clip must be copied from the surface onto the fallback image, but this was forgotten causing undue hassle in each backend. Another example is the contortion the meta surface endures to ensure the clip is correctly recorded. By contrast passing the clip along with the operation is quite simple and enables us to write generic handlers for providing surface wrappers. (And in the future, we should be able to write more esoteric wrappers, e.g. automatic 2x FSAA, trivially.) In brief, instead of the surface automatically applying the clip before calling the backend, the backend can call into a generic helper to apply clipping. For raster surfaces, clip regions are handled automatically as part of the composite interface. For vector surfaces, a clip helper is introduced to replay and callback into an intersect_clip_path() function as necessary. Whilst this is not primarily a performance related change (the change should just move the computation of the clip from the moment it is applied by the user to the moment it is required by the backend), it is important to track any potential regression: ppc: Speedups ======== image-rgba evolution-20090607-0 1026085.22 0.18% -> 672972.07 0.77%: 1.52x speedup ▌ image-rgba evolution-20090618-0 680579.98 0.12% -> 573237.66 0.16%: 1.19x speedup ▎ image-rgba swfdec-fill-rate-4xaa-0 460296.92 0.36% -> 407464.63 0.42%: 1.13x speedup ▏ image-rgba swfdec-fill-rate-2xaa-0 128431.95 0.47% -> 115051.86 0.42%: 1.12x speedup ▏ Slowdowns ========= image-rgba firefox-periodic-table-0 56837.61 0.78% -> 66055.17 3.20%: 1.09x slowdown ▏
2009-07-23 15:32:13 +01:00
status = (*region)->status;
if (rects != stack_rects)
free (rects);
return status;
}
cairo_bool_t
_cairo_traps_to_boxes (cairo_traps_t *traps,
cairo_antialias_t antialias,
cairo_boxes_t *boxes)
{
int i;
for (i = 0; i < traps->num_traps; i++) {
if (traps->traps[i].left.p1.x != traps->traps[i].left.p2.x ||
traps->traps[i].right.p1.x != traps->traps[i].right.p2.x)
return FALSE;
}
_cairo_boxes_init (boxes);
boxes->num_boxes = traps->num_traps;
boxes->chunks.base = (cairo_box_t *) traps->traps;
boxes->chunks.count = traps->num_traps;
boxes->chunks.size = traps->num_traps;
if (antialias != CAIRO_ANTIALIAS_NONE) {
for (i = 0; i < traps->num_traps; i++) {
/* Note the traps and boxes alias so we need to take the local copies first. */
cairo_fixed_t x1 = traps->traps[i].left.p1.x;
cairo_fixed_t x2 = traps->traps[i].right.p1.x;
cairo_fixed_t y1 = traps->traps[i].top;
cairo_fixed_t y2 = traps->traps[i].bottom;
boxes->chunks.base[i].p1.x = x1;
boxes->chunks.base[i].p1.y = y1;
boxes->chunks.base[i].p2.x = x2;
boxes->chunks.base[i].p2.y = y2;
if (boxes->is_pixel_aligned) {
boxes->is_pixel_aligned =
_cairo_fixed_is_integer (x1) && _cairo_fixed_is_integer (y1) &&
_cairo_fixed_is_integer (x2) && _cairo_fixed_is_integer (y2);
}
}
} else {
boxes->is_pixel_aligned = TRUE;
for (i = 0; i < traps->num_traps; i++) {
/* Note the traps and boxes alias so we need to take the local copies first. */
cairo_fixed_t x1 = traps->traps[i].left.p1.x;
cairo_fixed_t x2 = traps->traps[i].right.p1.x;
cairo_fixed_t y1 = traps->traps[i].top;
cairo_fixed_t y2 = traps->traps[i].bottom;
/* round down here to match Pixman's behavior when using traps. */
boxes->chunks.base[i].p1.x = _cairo_fixed_round_down (x1);
boxes->chunks.base[i].p1.y = _cairo_fixed_round_down (y1);
boxes->chunks.base[i].p2.x = _cairo_fixed_round_down (x2);
boxes->chunks.base[i].p2.y = _cairo_fixed_round_down (y2);
}
}
return TRUE;
}
/* moves trap points such that they become the actual corners of the trapezoid */
static void
_sanitize_trap (cairo_trapezoid_t *t)
{
cairo_trapezoid_t s = *t;
#define FIX(lr, tb, p) \
if (t->lr.p.y != t->tb) { \
2009-07-24 13:51:23 +01:00
t->lr.p.x = s.lr.p2.x + _cairo_fixed_mul_div_floor (s.lr.p1.x - s.lr.p2.x, s.tb - s.lr.p2.y, s.lr.p1.y - s.lr.p2.y); \
t->lr.p.y = s.tb; \
}
FIX (left, top, p1);
FIX (left, bottom, p2);
FIX (right, top, p1);
FIX (right, bottom, p2);
}
cairo_private cairo_status_t
_cairo_traps_path (const cairo_traps_t *traps,
cairo_path_fixed_t *path)
{
int i;
for (i = 0; i < traps->num_traps; i++) {
cairo_status_t status;
cairo_trapezoid_t trap = traps->traps[i];
if (trap.top == trap.bottom)
continue;
_sanitize_trap (&trap);
status = _cairo_path_fixed_move_to (path, trap.left.p1.x, trap.top);
if (unlikely (status)) return status;
status = _cairo_path_fixed_line_to (path, trap.right.p1.x, trap.top);
if (unlikely (status)) return status;
status = _cairo_path_fixed_line_to (path, trap.right.p2.x, trap.bottom);
if (unlikely (status)) return status;
status = _cairo_path_fixed_line_to (path, trap.left.p2.x, trap.bottom);
if (unlikely (status)) return status;
status = _cairo_path_fixed_close_path (path);
if (unlikely (status)) return status;
}
return CAIRO_STATUS_SUCCESS;
}
2011-10-09 19:53:56 +01:00
void
_cairo_debug_print_traps (FILE *file, const cairo_traps_t *traps)
{
cairo_box_t extents;
int n;
#if 0
if (traps->has_limits) {
printf ("%s: limits=(%d, %d, %d, %d)\n",
filename,
traps->limits.p1.x, traps->limits.p1.y,
traps->limits.p2.x, traps->limits.p2.y);
}
#endif
_cairo_traps_extents (traps, &extents);
fprintf (file, "extents=(%d, %d, %d, %d)\n",
extents.p1.x, extents.p1.y,
extents.p2.x, extents.p2.y);
for (n = 0; n < traps->num_traps; n++) {
fprintf (file, "%d %d L:(%d, %d), (%d, %d) R:(%d, %d), (%d, %d)\n",
traps->traps[n].top,
traps->traps[n].bottom,
traps->traps[n].left.p1.x,
traps->traps[n].left.p1.y,
traps->traps[n].left.p2.x,
traps->traps[n].left.p2.y,
traps->traps[n].right.p1.x,
traps->traps[n].right.p1.y,
traps->traps[n].right.p2.x,
traps->traps[n].right.p2.y);
}
}
struct cairo_trap_renderer {
cairo_span_renderer_t base;
cairo_traps_t *traps;
};
static cairo_status_t
span_to_traps (void *abstract_renderer, int y, int h,
const cairo_half_open_span_t *spans, unsigned num_spans)
{
struct cairo_trap_renderer *r = abstract_renderer;
cairo_fixed_t top, bot;
if (num_spans == 0)
return CAIRO_STATUS_SUCCESS;
top = _cairo_fixed_from_int (y);
bot = _cairo_fixed_from_int (y + h);
do {
if (spans[0].coverage) {
cairo_fixed_t x0 = _cairo_fixed_from_int(spans[0].x);
cairo_fixed_t x1 = _cairo_fixed_from_int(spans[1].x);
cairo_line_t left = { { x0, top }, { x0, bot } },
right = { { x1, top }, { x1, bot } };
_cairo_traps_add_trap (r->traps, top, bot, &left, &right);
}
spans++;
} while (--num_spans > 1);
return CAIRO_STATUS_SUCCESS;
}
cairo_int_status_t
_cairo_rasterise_polygon_to_traps (cairo_polygon_t *polygon,
cairo_fill_rule_t fill_rule,
cairo_antialias_t antialias,
cairo_traps_t *traps)
{
struct cairo_trap_renderer renderer;
cairo_scan_converter_t *converter;
cairo_int_status_t status;
cairo_rectangle_int_t r;
TRACE ((stderr, "%s: fill_rule=%d, antialias=%d\n",
__FUNCTION__, fill_rule, antialias));
assert(antialias == CAIRO_ANTIALIAS_NONE);
renderer.traps = traps;
renderer.base.render_rows = span_to_traps;
_cairo_box_round_to_rectangle (&polygon->extents, &r);
converter = _cairo_mono_scan_converter_create (r.x, r.y,
r.x + r.width,
r.y + r.height,
fill_rule);
status = _cairo_mono_scan_converter_add_polygon (converter, polygon);
if (likely (status == CAIRO_INT_STATUS_SUCCESS))
status = converter->generate (converter, &renderer.base);
converter->destroy (converter);
return status;
}