mutter/cogl/cogl-clip-stack.c

1064 lines
35 KiB
C
Raw Normal View History

/*
* Cogl
*
* An object oriented GL/GLES Abstraction/Utility Layer
*
* Copyright (C) 2007,2008,2009,2010 Intel Corporation.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library. If not, see <http://www.gnu.org/licenses/>.
*
*
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
#include <string.h>
#include <math.h>
#include <glib.h>
#include "cogl-clip-stack.h"
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
#include "cogl-primitives.h"
#include "cogl-context-private.h"
#include "cogl-internal.h"
#include "cogl-framebuffer-private.h"
#include "cogl-journal-private.h"
#include "cogl-util.h"
#include "cogl-path-private.h"
#include "cogl-matrix-private.h"
#include "cogl-primitives-private.h"
Add internal _cogl_push_source to optionally disable legacy state Some code in Cogl such as when flushing a stencil clip assumes that it can push a temporary simple pipeline to reset to a known state for internal drawing operations. However this breaks down if the application has set any legacy state because that is set globally so it will also get applied to the internal pipeline. _cogl_draw_attributes already had an internal flag to disable applying the legacy state but I think this is quite awkward to use because not all places that push a pipeline draw the attribute buffers directly so it is difficult to pass the flag down through the layers. Conceptually the legacy state is meant to be like a layer on top of the purely pipeline-based state API so I think ideally we should have an internal function to push the source without the applying the legacy state. The legacy state can't be applied as the pipeline is pushed because the global state can be modified even after it is pushed. This patch adds a _cogl_push_source() function which takes an extra boolean flag to mark whether to enable the legacy state. The value of this flag is stored alongside the pipeline in the pipeline stack. Another new internal function called _cogl_get_enable_legacy_state queries whether the top entry in the pipeline stack has legacy state enabled. cogl-primitives and the vertex array drawing code now use this to determine whether to apply the legacy state when drawing. The COGL_DRAW_SKIP_LEGACY_STATE flag is now removed. Reviewed-by: Robert Bragg <robert@linux.intel.com>
2011-09-14 07:17:09 -04:00
#include "cogl-private.h"
#include "cogl-pipeline-opengl-private.h"
#include "cogl-attribute-private.h"
#include "cogl-primitive-private.h"
#include "cogl1-context.h"
#include "cogl-offscreen.h"
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
#include "cogl-matrix-stack.h"
#ifndef GL_CLIP_PLANE0
#define GL_CLIP_PLANE0 0x3000
#define GL_CLIP_PLANE1 0x3001
#define GL_CLIP_PLANE2 0x3002
#define GL_CLIP_PLANE3 0x3003
#define GL_CLIP_PLANE4 0x3004
#define GL_CLIP_PLANE5 0x3005
#endif
static void
project_vertex (const CoglMatrix *modelview_projection,
float *vertex)
{
int i;
cogl_matrix_transform_point (modelview_projection,
&vertex[0], &vertex[1],
&vertex[2], &vertex[3]);
/* Convert from homogenized coordinates */
for (i = 0; i < 4; i++)
vertex[i] /= vertex[3];
}
static void
set_clip_plane (CoglFramebuffer *framebuffer,
GLint plane_num,
const float *vertex_a,
const float *vertex_b)
{
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
GLfloat planef[4];
double planed[4];
GLfloat angle;
CoglMatrixStack *modelview_stack =
_cogl_framebuffer_get_modelview_stack (framebuffer);
CoglMatrixStack *projection_stack =
_cogl_framebuffer_get_projection_stack (framebuffer);
CoglMatrix inverse_projection;
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
_cogl_matrix_stack_get_inverse (projection_stack, &inverse_projection);
/* Calculate the angle between the axes and the line crossing the
two points */
angle = atan2f (vertex_b[1] - vertex_a[1],
vertex_b[0] - vertex_a[0]) * (180.0/G_PI);
_cogl_matrix_stack_push (modelview_stack);
/* Load the inverse of the projection matrix so we can specify the plane
* in screen coordinates */
_cogl_matrix_stack_set (modelview_stack, &inverse_projection);
/* Rotate about point a */
_cogl_matrix_stack_translate (modelview_stack,
vertex_a[0], vertex_a[1], vertex_a[2]);
/* Rotate the plane by the calculated angle so that it will connect
the two points */
_cogl_matrix_stack_rotate (modelview_stack, angle, 0.0f, 0.0f, 1.0f);
_cogl_matrix_stack_translate (modelview_stack,
-vertex_a[0], -vertex_a[1], -vertex_a[2]);
Flush matrices in the progend and flip with a vector Previously flushing the matrices was performed as part of the framebuffer state. When on GLES2 this matrix flushing is actually diverted so that it only keeps a reference to the intended matrix stack. This is necessary because on GLES2 there are no builtin uniforms so it can't actually flush the matrices until the program for the pipeline is generated. When the matrices are flushed it would store the age of modifications on the matrix stack so that it could detect when the matrix hasn't changed and avoid flushing it. This patch changes it so that the pipeline is responsible for flushing the matrices even when we are using the GL builtins. The same mechanism for detecting unmodified matrix stacks is used in all cases. There is a new CoglMatrixStackCache type which is used to store a reference to the intended matrix stack along with its last flushed age. There are now two of these attached to the CoglContext to track the flushed state for the global matrix builtins and also two for each glsl progend program state to track the flushed state for a program. The framebuffer matrix flush now just updates the intended matrix stacks without actually trying to flush. When a vertex snippet is attached to the pipeline, the GLSL vertend will now avoid using the projection matrix to flip the rendering. This is necessary because any vertex snippet may cause the projection matrix not to be used. Instead the flip is done as a forced final step by multiplying cogl_position_out by a vec4 uniform. This uniform is updated as part of the progend pre_paint depending on whether the framebuffer is offscreen or not. Reviewed-by: Robert Bragg <robert@linux.intel.com>
2011-11-29 09:21:07 -05:00
/* Clip planes can only be used when a fixed function backend is in
use so we know we can directly push this matrix to the builtin
state */
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
_cogl_matrix_entry_flush_to_gl_builtins (ctx,
modelview_stack->last_entry,
Flush matrices in the progend and flip with a vector Previously flushing the matrices was performed as part of the framebuffer state. When on GLES2 this matrix flushing is actually diverted so that it only keeps a reference to the intended matrix stack. This is necessary because on GLES2 there are no builtin uniforms so it can't actually flush the matrices until the program for the pipeline is generated. When the matrices are flushed it would store the age of modifications on the matrix stack so that it could detect when the matrix hasn't changed and avoid flushing it. This patch changes it so that the pipeline is responsible for flushing the matrices even when we are using the GL builtins. The same mechanism for detecting unmodified matrix stacks is used in all cases. There is a new CoglMatrixStackCache type which is used to store a reference to the intended matrix stack along with its last flushed age. There are now two of these attached to the CoglContext to track the flushed state for the global matrix builtins and also two for each glsl progend program state to track the flushed state for a program. The framebuffer matrix flush now just updates the intended matrix stacks without actually trying to flush. When a vertex snippet is attached to the pipeline, the GLSL vertend will now avoid using the projection matrix to flip the rendering. This is necessary because any vertex snippet may cause the projection matrix not to be used. Instead the flip is done as a forced final step by multiplying cogl_position_out by a vec4 uniform. This uniform is updated as part of the progend pre_paint depending on whether the framebuffer is offscreen or not. Reviewed-by: Robert Bragg <robert@linux.intel.com>
2011-11-29 09:21:07 -05:00
COGL_MATRIX_MODELVIEW,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
framebuffer,
Flush matrices in the progend and flip with a vector Previously flushing the matrices was performed as part of the framebuffer state. When on GLES2 this matrix flushing is actually diverted so that it only keeps a reference to the intended matrix stack. This is necessary because on GLES2 there are no builtin uniforms so it can't actually flush the matrices until the program for the pipeline is generated. When the matrices are flushed it would store the age of modifications on the matrix stack so that it could detect when the matrix hasn't changed and avoid flushing it. This patch changes it so that the pipeline is responsible for flushing the matrices even when we are using the GL builtins. The same mechanism for detecting unmodified matrix stacks is used in all cases. There is a new CoglMatrixStackCache type which is used to store a reference to the intended matrix stack along with its last flushed age. There are now two of these attached to the CoglContext to track the flushed state for the global matrix builtins and also two for each glsl progend program state to track the flushed state for a program. The framebuffer matrix flush now just updates the intended matrix stacks without actually trying to flush. When a vertex snippet is attached to the pipeline, the GLSL vertend will now avoid using the projection matrix to flip the rendering. This is necessary because any vertex snippet may cause the projection matrix not to be used. Instead the flip is done as a forced final step by multiplying cogl_position_out by a vec4 uniform. This uniform is updated as part of the progend pre_paint depending on whether the framebuffer is offscreen or not. Reviewed-by: Robert Bragg <robert@linux.intel.com>
2011-11-29 09:21:07 -05:00
FALSE /* don't disable flip */);
Dynamically load the GL or GLES library The GL or GLES library is now dynamically loaded by the CoglRenderer so that it can choose between GL, GLES1 and GLES2 at runtime. The library is loaded by the renderer because it needs to be done before calling eglInitialize. There is a new environment variable called COGL_DRIVER to choose between gl, gles1 or gles2. The #ifdefs for HAVE_COGL_GL, HAVE_COGL_GLES and HAVE_COGL_GLES2 have been changed so that they don't assume the ifdefs are mutually exclusive. They haven't been removed entirely so that it's possible to compile the GLES backends without the the enums from the GL headers. When using GLX the winsys additionally dynamically loads libGL because that also contains the GLX API. It can't be linked in directly because that would probably conflict with the GLES API if the EGL is selected. When compiling with EGL support the library links directly to libEGL because it doesn't contain any GL API so it shouldn't have any conflicts. When building for WGL or OSX Cogl still directly links against the GL API so there is a #define in config.h so that Cogl won't try to dlopen the library. Cogl-pango previously had a #ifdef to detect when the GL backend is used so that it can sneakily pass GL_QUADS to cogl_vertex_buffer_draw. This is now changed so that it queries the CoglContext for the backend. However to get this to work Cogl now needs to export the _cogl_context_get_default symbol and cogl-pango needs some extra -I flags to so that it can include cogl-context-private.h
2011-07-07 15:44:56 -04:00
planef[0] = 0;
planef[1] = -1.0;
planef[2] = 0;
planef[3] = vertex_a[1];
switch (ctx->driver)
{
default:
g_assert_not_reached ();
break;
case COGL_DRIVER_GLES1:
GE( ctx, glClipPlanef (plane_num, planef) );
break;
case COGL_DRIVER_GL:
planed[0] = planef[0];
planed[1] = planef[1];
planed[2] = planef[2];
planed[3] = planef[3];
GE( ctx, glClipPlane (plane_num, planed) );
break;
}
_cogl_matrix_stack_pop (modelview_stack);
}
static void
set_clip_planes (CoglFramebuffer *framebuffer,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrixEntry *modelview_entry,
float x_1,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
float y_1,
float x_2,
float y_2)
{
CoglMatrix modelview_matrix;
CoglMatrixStack *projection_stack =
_cogl_framebuffer_get_projection_stack (framebuffer);
CoglMatrix projection_matrix;
CoglMatrix modelview_projection;
float signed_area;
float vertex_tl[4] = { x_1, y_1, 0, 1.0 };
float vertex_tr[4] = { x_2, y_1, 0, 1.0 };
float vertex_bl[4] = { x_1, y_2, 0, 1.0 };
float vertex_br[4] = { x_2, y_2, 0, 1.0 };
_cogl_matrix_stack_get (projection_stack, &projection_matrix);
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
_cogl_matrix_entry_get (modelview_entry, &modelview_matrix);
cogl_matrix_multiply (&modelview_projection,
&projection_matrix,
&modelview_matrix);
project_vertex (&modelview_projection, vertex_tl);
project_vertex (&modelview_projection, vertex_tr);
project_vertex (&modelview_projection, vertex_bl);
project_vertex (&modelview_projection, vertex_br);
/* Calculate the signed area of the polygon formed by the four
vertices so that we can know its orientation */
signed_area = (vertex_tl[0] * (vertex_tr[1] - vertex_bl[1])
+ vertex_tr[0] * (vertex_br[1] - vertex_tl[1])
+ vertex_br[0] * (vertex_bl[1] - vertex_tr[1])
+ vertex_bl[0] * (vertex_tl[1] - vertex_br[1]));
/* Set the clip planes to form lines between all of the vertices
using the same orientation as we calculated */
if (signed_area > 0.0f)
{
/* counter-clockwise */
set_clip_plane (framebuffer, GL_CLIP_PLANE0, vertex_tl, vertex_bl);
set_clip_plane (framebuffer, GL_CLIP_PLANE1, vertex_bl, vertex_br);
set_clip_plane (framebuffer, GL_CLIP_PLANE2, vertex_br, vertex_tr);
set_clip_plane (framebuffer, GL_CLIP_PLANE3, vertex_tr, vertex_tl);
}
else
{
/* clockwise */
set_clip_plane (framebuffer, GL_CLIP_PLANE0, vertex_tl, vertex_tr);
set_clip_plane (framebuffer, GL_CLIP_PLANE1, vertex_tr, vertex_br);
set_clip_plane (framebuffer, GL_CLIP_PLANE2, vertex_br, vertex_bl);
set_clip_plane (framebuffer, GL_CLIP_PLANE3, vertex_bl, vertex_tl);
}
}
static void
add_stencil_clip_rectangle (CoglFramebuffer *framebuffer,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrixEntry *modelview_entry,
float x_1,
float y_1,
float x_2,
float y_2,
CoglBool first)
{
CoglMatrixStack *projection_stack =
_cogl_framebuffer_get_projection_stack (framebuffer);
CoglContext *ctx = cogl_framebuffer_get_context (framebuffer);
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
/* NB: This can be called while flushing the journal so we need
* to be very conservative with what state we change.
*/
_cogl_context_set_current_projection_entry (ctx,
projection_stack->last_entry);
_cogl_context_set_current_modelview_entry (ctx, modelview_entry);
if (first)
{
GE( ctx, glEnable (GL_STENCIL_TEST) );
/* Initially disallow everything */
GE( ctx, glClearStencil (0) );
GE( ctx, glClear (GL_STENCIL_BUFFER_BIT) );
/* Punch out a hole to allow the rectangle */
GE( ctx, glStencilFunc (GL_NEVER, 0x1, 0x1) );
GE( ctx, glStencilOp (GL_REPLACE, GL_REPLACE, GL_REPLACE) );
_cogl_rectangle_immediate (framebuffer,
ctx->stencil_pipeline,
x_1, y_1, x_2, y_2);
}
else
{
/* Add one to every pixel of the stencil buffer in the
rectangle */
GE( ctx, glStencilFunc (GL_NEVER, 0x1, 0x3) );
GE( ctx, glStencilOp (GL_INCR, GL_INCR, GL_INCR) );
_cogl_rectangle_immediate (framebuffer,
ctx->stencil_pipeline,
x_1, y_1, x_2, y_2);
/* Subtract one from all pixels in the stencil buffer so that
only pixels where both the original stencil buffer and the
rectangle are set will be valid */
GE( ctx, glStencilOp (GL_DECR, GL_DECR, GL_DECR) );
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
_cogl_context_set_current_projection_entry (ctx, &ctx->identity_entry);
_cogl_context_set_current_modelview_entry (ctx, &ctx->identity_entry);
_cogl_rectangle_immediate (framebuffer,
ctx->stencil_pipeline,
-1.0, -1.0, 1.0, 1.0);
}
/* Restore the stencil mode */
GE( ctx, glStencilFunc (GL_EQUAL, 0x1, 0x1) );
GE( ctx, glStencilOp (GL_KEEP, GL_KEEP, GL_KEEP) );
}
typedef void (*SilhouettePaintCallback) (CoglFramebuffer *framebuffer,
CoglPipeline *pipeline,
void *user_data);
static void
add_stencil_clip_silhouette (CoglFramebuffer *framebuffer,
SilhouettePaintCallback silhouette_callback,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrixEntry *modelview_entry,
float bounds_x1,
float bounds_y1,
float bounds_x2,
float bounds_y2,
CoglBool merge,
CoglBool need_clear,
void *user_data)
{
CoglMatrixStack *projection_stack =
_cogl_framebuffer_get_projection_stack (framebuffer);
CoglContext *ctx = cogl_framebuffer_get_context (framebuffer);
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
/* NB: This can be called while flushing the journal so we need
* to be very conservative with what state we change.
*/
_cogl_context_set_current_projection_entry (ctx,
projection_stack->last_entry);
_cogl_context_set_current_modelview_entry (ctx, modelview_entry);
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
_cogl_pipeline_flush_gl_state (ctx->stencil_pipeline, framebuffer, FALSE, 0);
GE( ctx, glEnable (GL_STENCIL_TEST) );
GE( ctx, glColorMask (FALSE, FALSE, FALSE, FALSE) );
GE( ctx, glDepthMask (FALSE) );
if (merge)
{
GE (ctx, glStencilMask (2));
GE (ctx, glStencilFunc (GL_LEQUAL, 0x2, 0x6));
}
else
{
/* If we're not using the stencil buffer for clipping then we
don't need to clear the whole stencil buffer, just the area
that will be drawn */
if (need_clear)
/* If this is being called from the clip stack code then it
will have set up a scissor for the minimum bounding box of
all of the clips. That box will likely mean that this
_cogl_clear won't need to clear the entire
buffer. _cogl_framebuffer_clear_without_flush4f is used instead
of cogl_clear because it won't try to flush the journal */
_cogl_framebuffer_clear_without_flush4f (framebuffer,
COGL_BUFFER_BIT_STENCIL,
0, 0, 0, 0);
else
{
/* Just clear the bounding box */
GE( ctx, glStencilMask (~(GLuint) 0) );
GE( ctx, glStencilOp (GL_ZERO, GL_ZERO, GL_ZERO) );
_cogl_rectangle_immediate (framebuffer,
ctx->stencil_pipeline,
bounds_x1, bounds_y1,
bounds_x2, bounds_y2);
}
GE (ctx, glStencilMask (1));
GE (ctx, glStencilFunc (GL_LEQUAL, 0x1, 0x3));
}
GE (ctx, glStencilOp (GL_INVERT, GL_INVERT, GL_INVERT));
silhouette_callback (framebuffer, ctx->stencil_pipeline, user_data);
if (merge)
{
/* Now we have the new stencil buffer in bit 1 and the old
stencil buffer in bit 0 so we need to intersect them */
GE (ctx, glStencilMask (3));
GE (ctx, glStencilFunc (GL_NEVER, 0x2, 0x3));
GE (ctx, glStencilOp (GL_DECR, GL_DECR, GL_DECR));
/* Decrement all of the bits twice so that only pixels where the
value is 3 will remain */
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
_cogl_context_set_current_projection_entry (ctx, &ctx->identity_entry);
_cogl_context_set_current_modelview_entry (ctx, &ctx->identity_entry);
_cogl_rectangle_immediate (framebuffer, ctx->stencil_pipeline,
-1.0, -1.0, 1.0, 1.0);
_cogl_rectangle_immediate (framebuffer, ctx->stencil_pipeline,
-1.0, -1.0, 1.0, 1.0);
}
GE (ctx, glStencilMask (~(GLuint) 0));
GE (ctx, glDepthMask (TRUE));
GE (ctx, glColorMask (TRUE, TRUE, TRUE, TRUE));
GE (ctx, glStencilFunc (GL_EQUAL, 0x1, 0x1));
GE (ctx, glStencilOp (GL_KEEP, GL_KEEP, GL_KEEP));
}
static void
paint_path_silhouette (CoglFramebuffer *framebuffer,
CoglPipeline *pipeline,
void *user_data)
{
CoglPath *path = user_data;
if (path->data->path_nodes->len >= 3)
_cogl_path_fill_nodes (path,
framebuffer,
pipeline,
COGL_DRAW_SKIP_JOURNAL_FLUSH |
COGL_DRAW_SKIP_PIPELINE_VALIDATION |
COGL_DRAW_SKIP_FRAMEBUFFER_FLUSH);
}
static void
add_stencil_clip_path (CoglFramebuffer *framebuffer,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrixEntry *modelview_entry,
CoglPath *path,
CoglBool merge,
CoglBool need_clear)
{
CoglPathData *data = path->data;
add_stencil_clip_silhouette (framebuffer,
paint_path_silhouette,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
modelview_entry,
data->path_nodes_min.x,
data->path_nodes_min.y,
data->path_nodes_max.x,
data->path_nodes_max.y,
merge,
need_clear,
path);
}
static void
paint_primitive_silhouette (CoglFramebuffer *framebuffer,
CoglPipeline *pipeline,
void *user_data)
{
_cogl_framebuffer_draw_primitive (framebuffer,
pipeline,
user_data,
COGL_DRAW_SKIP_JOURNAL_FLUSH |
COGL_DRAW_SKIP_PIPELINE_VALIDATION |
COGL_DRAW_SKIP_FRAMEBUFFER_FLUSH |
COGL_DRAW_SKIP_LEGACY_STATE);
}
Add -Wmissing-declarations to maintainer flags and fix problems This option to GCC makes it give a warning whenever a global function is defined without a declaration. This should catch cases were we've defined a function but forgot to put it in a header. In that case it is either only used within one file so we should make it static or we should declare it in a header. The following changes where made to fix problems: • Some functions were made static • cogl-path.h (the one containing the 1.0 API) was split into two files, one defining the functions and one defining the enums so that cogl-path.c can include the enum and function declarations from the 2.0 API as well as the function declarations from the 1.0 API. • cogl2-clip-state has been removed. This only had one experimental function called cogl_clip_push_from_path but as this is unstable we might as well remove it favour of the equivalent cogl_framebuffer_* API. • The GLX, SDL and WGL winsys's now have a private header to define their get_vtable function instead of directly declaring in the C file where it is called. • All places that were calling COGL_OBJECT_DEFINE need to have the cogl_is_whatever function declared so these have been added either as a public function or in a private header. • Some files that were not including the header containing their function declarations have been fixed to do so. • Any unused error quark functions have been removed. If we later want them we should add them back one by one and add a declaration for them in a header. • _cogl_is_framebuffer has been renamed to cogl_is_framebuffer and made a public function with a declaration in cogl-framebuffer.h • Similarly for CoglOnscreen. • cogl_vdraw_indexed_attributes is called cogl_framebuffer_vdraw_indexed_attributes in the header. The definition has been changed to match the header. • cogl_index_buffer_allocate has been removed. This had no declaration and I'm not sure what it's supposed to do. • CoglJournal has been changed to use the internal CoglObject macro so that it won't define an exported cogl_is_journal symbol. • The _cogl_blah_pointer_from_handle functions have been removed. CoglHandle isn't used much anymore anyway and in the few places where it is used I think it's safe to just use the implicit cast from void* to the right type. • The test-utils.h header for the conformance tests explicitly disables the -Wmissing-declaration option using a pragma because all of the tests declare their main function without a header. Any mistakes relating to missing declarations aren't really important for the tests. • cogl_quaternion_init_from_quaternion and init_from_matrix have been given declarations in cogl-quaternion.h Reviewed-by: Robert Bragg <robert@linux.intel.com>
2012-03-06 13:21:28 -05:00
static void
add_stencil_clip_primitive (CoglFramebuffer *framebuffer,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrixEntry *modelview_entry,
CoglPrimitive *primitive,
float bounds_x1,
float bounds_y1,
float bounds_x2,
float bounds_y2,
CoglBool merge,
CoglBool need_clear)
{
add_stencil_clip_silhouette (framebuffer,
paint_primitive_silhouette,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
modelview_entry,
bounds_x1,
bounds_y1,
bounds_x2,
bounds_y2,
merge,
need_clear,
primitive);
}
static void
disable_stencil_buffer (void)
{
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
GE( ctx, glDisable (GL_STENCIL_TEST) );
}
static void
enable_clip_planes (void)
{
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
GE( ctx, glEnable (GL_CLIP_PLANE0) );
GE( ctx, glEnable (GL_CLIP_PLANE1) );
GE( ctx, glEnable (GL_CLIP_PLANE2) );
GE( ctx, glEnable (GL_CLIP_PLANE3) );
}
static void
disable_clip_planes (void)
{
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
GE( ctx, glDisable (GL_CLIP_PLANE3) );
GE( ctx, glDisable (GL_CLIP_PLANE2) );
GE( ctx, glDisable (GL_CLIP_PLANE1) );
GE( ctx, glDisable (GL_CLIP_PLANE0) );
}
static void *
_cogl_clip_stack_push_entry (CoglClipStack *clip_stack,
size_t size,
CoglClipStackType type)
{
CoglClipStack *entry = g_slice_alloc (size);
/* The new entry starts with a ref count of 1 because the stack
holds a reference to it as it is the top entry */
entry->ref_count = 1;
entry->type = type;
entry->parent = clip_stack;
/* We don't need to take a reference to the parent from the entry
because the we are stealing the ref in the new stack top */
return entry;
}
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
static void
get_transformed_corners (float x_1,
float y_1,
float x_2,
float y_2,
CoglMatrix *modelview,
CoglMatrix *projection,
const float *viewport,
float *transformed_corners)
{
int i;
transformed_corners[0] = x_1;
transformed_corners[1] = y_1;
transformed_corners[2] = x_2;
transformed_corners[3] = y_1;
transformed_corners[4] = x_2;
transformed_corners[5] = y_2;
transformed_corners[6] = x_1;
transformed_corners[7] = y_2;
/* Project the coordinates to window space coordinates */
for (i = 0; i < 4; i++)
{
float *v = transformed_corners + i * 2;
_cogl_transform_point (modelview, projection, viewport, v, v + 1);
}
}
/* Sets the window-space bounds of the entry based on the projected
coordinates of the given rectangle */
static void
_cogl_clip_stack_entry_set_bounds (CoglClipStack *entry,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
float *transformed_corners)
{
float min_x = G_MAXFLOAT, min_y = G_MAXFLOAT;
float max_x = -G_MAXFLOAT, max_y = -G_MAXFLOAT;
int i;
for (i = 0; i < 4; i++)
{
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
float *v = transformed_corners + i * 2;
if (v[0] > max_x)
max_x = v[0];
if (v[0] < min_x)
min_x = v[0];
if (v[1] > max_y)
max_y = v[1];
if (v[1] < min_y)
min_y = v[1];
}
entry->bounds_x0 = floorf (min_x);
entry->bounds_x1 = ceilf (max_x);
entry->bounds_y0 = floorf (min_y);
entry->bounds_y1 = ceilf (max_y);
}
CoglClipStack *
_cogl_clip_stack_push_window_rectangle (CoglClipStack *stack,
int x_offset,
int y_offset,
int width,
int height)
{
CoglClipStack *entry;
entry = _cogl_clip_stack_push_entry (stack,
sizeof (CoglClipStackWindowRect),
COGL_CLIP_STACK_WINDOW_RECT);
entry->bounds_x0 = x_offset;
entry->bounds_x1 = x_offset + width;
entry->bounds_y0 = y_offset;
entry->bounds_y1 = y_offset + height;
return entry;
}
CoglClipStack *
_cogl_clip_stack_push_rectangle (CoglClipStack *stack,
float x_1,
float y_1,
float x_2,
float y_2,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrixEntry *modelview_entry,
CoglMatrixEntry *projection_entry,
const float *viewport)
{
CoglClipStackRect *entry;
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrix modelview;
CoglMatrix projection;
CoglMatrix modelview_projection;
/* Corners of the given rectangle in an clockwise order:
* (0, 1) (2, 3)
*
*
*
* (6, 7) (4, 5)
*/
float rect[] = {
x_1, y_1,
x_2, y_1,
x_2, y_2,
x_1, y_2
};
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
/* Make a new entry */
entry = _cogl_clip_stack_push_entry (stack,
sizeof (CoglClipStackRect),
COGL_CLIP_STACK_RECT);
entry->x0 = x_1;
entry->y0 = y_1;
entry->x1 = x_2;
entry->y1 = y_2;
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
entry->matrix_entry = _cogl_matrix_entry_ref (modelview_entry);
_cogl_matrix_entry_get (modelview_entry, &modelview);
_cogl_matrix_entry_get (projection_entry, &projection);
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
cogl_matrix_multiply (&modelview_projection,
&projection,
&modelview);
/* Technically we could avoid the viewport transform at this point
* if we want to make this a bit faster. */
_cogl_transform_point (&modelview, &projection, viewport, &rect[0], &rect[1]);
_cogl_transform_point (&modelview, &projection, viewport, &rect[2], &rect[3]);
_cogl_transform_point (&modelview, &projection, viewport, &rect[4], &rect[5]);
_cogl_transform_point (&modelview, &projection, viewport, &rect[6], &rect[7]);
/* If the fully transformed rectangle isn't still axis aligned we
* can't handle it using a scissor.
*
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
* We don't use an epsilon here since we only really aim to catch
* simple cases where the transform doesn't leave the rectangle screen
* aligned and don't mind some false positives.
*/
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
if (rect[0] != rect[6] ||
rect[1] != rect[3] ||
rect[2] != rect[4] ||
rect[7] != rect[5])
{
entry->can_be_scissor = FALSE;
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
_cogl_clip_stack_entry_set_bounds ((CoglClipStack *) entry,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
rect);
}
else
{
CoglClipStack *base_entry = (CoglClipStack *) entry;
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
x_1 = rect[0];
y_1 = rect[1];
x_2 = rect[4];
y_2 = rect[5];
/* Consider that the modelview matrix may flip the rectangle
* along the x or y axis... */
#define SWAP(A,B) do { float tmp = B; B = A; A = tmp; } while (0)
if (x_1 > x_2)
SWAP (x_1, x_2);
if (y_1 > y_2)
SWAP (y_1, y_2);
#undef SWAP
base_entry->bounds_x0 = COGL_UTIL_NEARBYINT (x_1);
base_entry->bounds_y0 = COGL_UTIL_NEARBYINT (y_1);
base_entry->bounds_x1 = COGL_UTIL_NEARBYINT (x_2);
base_entry->bounds_y1 = COGL_UTIL_NEARBYINT (y_2);
entry->can_be_scissor = TRUE;
}
return (CoglClipStack *) entry;
}
CoglClipStack *
_cogl_clip_stack_push_from_path (CoglClipStack *stack,
CoglPath *path,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrixEntry *modelview_entry,
CoglMatrixEntry *projection_entry,
const float *viewport)
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
{
float x_1, y_1, x_2, y_2;
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
_cogl_path_get_bounds (path, &x_1, &y_1, &x_2, &y_2);
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
/* If the path is a simple rectangle then we can divert to pushing a
rectangle clip instead which usually won't involve the stencil
buffer */
if (_cogl_path_is_rectangle (path))
return _cogl_clip_stack_push_rectangle (stack,
x_1, y_1,
x_2, y_2,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
modelview_entry,
projection_entry,
viewport);
else
{
CoglClipStackPath *entry;
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrix modelview;
CoglMatrix projection;
float transformed_corners[8];
entry = _cogl_clip_stack_push_entry (stack,
sizeof (CoglClipStackPath),
COGL_CLIP_STACK_PATH);
entry->path = cogl_path_copy (path);
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
entry->matrix_entry = _cogl_matrix_entry_ref (modelview_entry);
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
_cogl_matrix_entry_get (modelview_entry, &modelview);
_cogl_matrix_entry_get (projection_entry, &projection);
get_transformed_corners (x_1, y_1, x_2, y_2,
&modelview,
&projection,
viewport,
transformed_corners);
_cogl_clip_stack_entry_set_bounds ((CoglClipStack *) entry,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
transformed_corners);
return (CoglClipStack *) entry;
}
}
CoglClipStack *
_cogl_clip_stack_push_primitive (CoglClipStack *stack,
CoglPrimitive *primitive,
float bounds_x1,
float bounds_y1,
float bounds_x2,
float bounds_y2,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrixEntry *modelview_entry,
CoglMatrixEntry *projection_entry,
const float *viewport)
{
CoglClipStackPrimitive *entry;
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglMatrix modelview;
CoglMatrix projection;
float transformed_corners[8];
entry = _cogl_clip_stack_push_entry (stack,
sizeof (CoglClipStackPrimitive),
COGL_CLIP_STACK_PRIMITIVE);
entry->primitive = cogl_object_ref (primitive);
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
entry->matrix_entry = _cogl_matrix_entry_ref (modelview_entry);
entry->bounds_x1 = bounds_x1;
entry->bounds_y1 = bounds_y1;
entry->bounds_x2 = bounds_x2;
entry->bounds_y2 = bounds_y2;
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
_cogl_matrix_entry_get (modelview_entry, &modelview);
_cogl_matrix_entry_get (modelview_entry, &projection);
get_transformed_corners (bounds_x1, bounds_y1, bounds_x2, bounds_y2,
&modelview,
&projection,
viewport,
transformed_corners);
/* NB: this is referring to the bounds in window coordinates as opposed
* to the bounds above in primitive local coordinates. */
_cogl_clip_stack_entry_set_bounds ((CoglClipStack *) entry,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
transformed_corners);
return (CoglClipStack *) entry;
}
CoglClipStack *
_cogl_clip_stack_ref (CoglClipStack *entry)
{
/* A NULL pointer is considered a valid stack so we should accept
that as an argument */
if (entry)
entry->ref_count++;
return entry;
}
void
_cogl_clip_stack_unref (CoglClipStack *entry)
{
/* Unref all of the entries until we hit the root of the list or the
entry still has a remaining reference */
while (entry && --entry->ref_count <= 0)
{
CoglClipStack *parent = entry->parent;
switch (entry->type)
{
case COGL_CLIP_STACK_RECT:
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
{
CoglClipStackRect *rect = (CoglClipStackRect *) entry;
_cogl_matrix_entry_unref (rect->matrix_entry);
g_slice_free1 (sizeof (CoglClipStackRect), entry);
break;
}
case COGL_CLIP_STACK_WINDOW_RECT:
g_slice_free1 (sizeof (CoglClipStackWindowRect), entry);
break;
case COGL_CLIP_STACK_PATH:
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
{
CoglClipStackPath *path_entry = (CoglClipStackPath *) entry;
_cogl_matrix_entry_unref (path_entry->matrix_entry);
cogl_object_unref (path_entry->path);
g_slice_free1 (sizeof (CoglClipStackPath), entry);
break;
}
case COGL_CLIP_STACK_PRIMITIVE:
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
{
CoglClipStackPrimitive *primitive_entry =
(CoglClipStackPrimitive *) entry;
_cogl_matrix_entry_unref (primitive_entry->matrix_entry);
cogl_object_unref (primitive_entry->primitive);
g_slice_free1 (sizeof (CoglClipStackPrimitive), entry);
break;
}
default:
g_assert_not_reached ();
}
entry = parent;
}
}
CoglClipStack *
_cogl_clip_stack_pop (CoglClipStack *stack)
{
CoglClipStack *new_top;
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
_COGL_RETURN_VAL_IF_FAIL (stack != NULL, NULL);
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
/* To pop we are moving the top of the stack to the old top's parent
node. The stack always needs to have a reference to the top entry
so we must take a reference to the new top. The stack would have
previously had a reference to the old top so we need to decrease
the ref count on that. We need to ref the new head first in case
this stack was the only thing referencing the old top. In that
case the call to _cogl_clip_stack_entry_unref will unref the
parent. */
new_top = stack->parent;
_cogl_clip_stack_ref (new_top);
_cogl_clip_stack_unref (stack);
return new_top;
}
void
_cogl_clip_stack_get_bounds (CoglClipStack *stack,
int *scissor_x0,
int *scissor_y0,
int *scissor_x1,
int *scissor_y1)
{
CoglClipStack *entry;
*scissor_x0 = 0;
*scissor_y0 = 0;
*scissor_x1 = G_MAXINT;
*scissor_y1 = G_MAXINT;
for (entry = stack; entry; entry = entry->parent)
{
/* Get the intersection of the current scissor and the bounding
box of this clip */
*scissor_x0 = MAX (*scissor_x0, entry->bounds_x0);
*scissor_y0 = MAX (*scissor_y0, entry->bounds_y0);
*scissor_x1 = MIN (*scissor_x1, entry->bounds_x1);
*scissor_y1 = MIN (*scissor_y1, entry->bounds_y1);
}
}
void
_cogl_clip_stack_flush (CoglClipStack *stack,
CoglFramebuffer *framebuffer)
{
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
CoglContext *ctx = framebuffer->context;
[draw-buffers] First pass at overhauling Cogl's framebuffer management Cogl's support for offscreen rendering was originally written just to support the clutter_texture_new_from_actor API and due to lack of documentation and several confusing - non orthogonal - side effects of using the API it wasn't really possible to use directly. This commit does a number of things: - It removes {gl,gles}/cogl-fbo.{c,h} and adds shared cogl-draw-buffer.{c,h} files instead which should be easier to maintain. - internally CoglFbo objects are now called CoglDrawBuffers. A CoglDrawBuffer is an abstract base class that is inherited from to implement CoglOnscreen and CoglOffscreen draw buffers. CoglOffscreen draw buffers will initially be used to support the cogl_offscreen_new_to_texture API, and CoglOnscreen draw buffers will start to be used internally to represent windows as we aim to migrate some of Clutter's backend code to Cogl. - It makes draw buffer objects the owners of the following state: - viewport - projection matrix stack - modelview matrix stack - clip state (This means when you switch between draw buffers you will automatically be switching to their associated viewport, matrix and clip state) Aside from hopefully making cogl_offscreen_new_to_texture be more useful short term by having simpler and well defined semantics for cogl_set_draw_buffer, as mentioned above this is the first step for a couple of other things: - Its a step toward moving ownership for windows down from Clutter backends into Cogl, by (internally at least) introducing the CoglOnscreen draw buffer. Note: the plan is that cogl_set_draw_buffer will accept on or offscreen draw buffer handles, and the "target" argument will become redundant since we will instead query the type of the given draw buffer handle. - Because we have a common type for on and offscreen framebuffers we can provide a unified API for framebuffer management. Things like: - blitting between buffers - managing ancillary buffers (e.g. attaching depth and stencil buffers) - size requisition - clearing
2009-09-25 09:34:34 -04:00
int has_clip_planes;
CoglBool using_clip_planes = FALSE;
CoglBool using_stencil_buffer = FALSE;
int scissor_x0;
int scissor_y0;
int scissor_x1;
int scissor_y1;
CoglClipStack *entry;
int scissor_y_start;
/* If we have already flushed this state then we don't need to do
anything */
if (ctx->current_clip_stack_valid)
{
if (ctx->current_clip_stack == stack)
return;
_cogl_clip_stack_unref (ctx->current_clip_stack);
}
ctx->current_clip_stack_valid = TRUE;
ctx->current_clip_stack = _cogl_clip_stack_ref (stack);
has_clip_planes =
ctx->private_feature_flags & COGL_PRIVATE_FEATURE_FOUR_CLIP_PLANES;
[draw-buffers] First pass at overhauling Cogl's framebuffer management Cogl's support for offscreen rendering was originally written just to support the clutter_texture_new_from_actor API and due to lack of documentation and several confusing - non orthogonal - side effects of using the API it wasn't really possible to use directly. This commit does a number of things: - It removes {gl,gles}/cogl-fbo.{c,h} and adds shared cogl-draw-buffer.{c,h} files instead which should be easier to maintain. - internally CoglFbo objects are now called CoglDrawBuffers. A CoglDrawBuffer is an abstract base class that is inherited from to implement CoglOnscreen and CoglOffscreen draw buffers. CoglOffscreen draw buffers will initially be used to support the cogl_offscreen_new_to_texture API, and CoglOnscreen draw buffers will start to be used internally to represent windows as we aim to migrate some of Clutter's backend code to Cogl. - It makes draw buffer objects the owners of the following state: - viewport - projection matrix stack - modelview matrix stack - clip state (This means when you switch between draw buffers you will automatically be switching to their associated viewport, matrix and clip state) Aside from hopefully making cogl_offscreen_new_to_texture be more useful short term by having simpler and well defined semantics for cogl_set_draw_buffer, as mentioned above this is the first step for a couple of other things: - Its a step toward moving ownership for windows down from Clutter backends into Cogl, by (internally at least) introducing the CoglOnscreen draw buffer. Note: the plan is that cogl_set_draw_buffer will accept on or offscreen draw buffer handles, and the "target" argument will become redundant since we will instead query the type of the given draw buffer handle. - Because we have a common type for on and offscreen framebuffers we can provide a unified API for framebuffer management. Things like: - blitting between buffers - managing ancillary buffers (e.g. attaching depth and stencil buffers) - size requisition - clearing
2009-09-25 09:34:34 -04:00
if (has_clip_planes)
disable_clip_planes ();
disable_stencil_buffer ();
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
/* If the stack is empty then there's nothing else to do */
if (stack == NULL)
{
COGL_NOTE (CLIPPING, "Flushed empty clip stack");
ctx->current_clip_stack_uses_stencil = FALSE;
GE (ctx, glDisable (GL_SCISSOR_TEST));
return;
}
/* Calculate the scissor rect first so that if we eventually have to
clear the stencil buffer then the clear will be clipped to the
intersection of all of the bounding boxes. This saves having to
clear the whole stencil buffer */
_cogl_clip_stack_get_bounds (stack,
&scissor_x0, &scissor_y0,
&scissor_x1, &scissor_y1);
/* Enable scissoring as soon as possible */
if (scissor_x0 >= scissor_x1 || scissor_y0 >= scissor_y1)
scissor_x0 = scissor_y0 = scissor_x1 = scissor_y1 = scissor_y_start = 0;
else
{
/* We store the entry coordinates in Cogl coordinate space
* but OpenGL requires the window origin to be the bottom
* left so we may need to convert the incoming coordinates.
*
* NB: Cogl forces all offscreen rendering to be done upside
* down so in this case no conversion is needed.
*/
if (cogl_is_offscreen (framebuffer))
scissor_y_start = scissor_y0;
else
{
int framebuffer_height =
cogl_framebuffer_get_height (framebuffer);
scissor_y_start = framebuffer_height - scissor_y1;
}
}
COGL_NOTE (CLIPPING, "Flushing scissor to (%i, %i, %i, %i)",
scissor_x0, scissor_y0,
scissor_x1, scissor_y1);
GE (ctx, glEnable (GL_SCISSOR_TEST));
GE (ctx, glScissor (scissor_x0, scissor_y_start,
scissor_x1 - scissor_x0,
scissor_y1 - scissor_y0));
/* Add all of the entries. This will end up adding them in the
reverse order that they were specified but as all of the clips
are intersecting it should work out the same regardless of the
order */
for (entry = stack; entry; entry = entry->parent)
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
{
switch (entry->type)
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
{
case COGL_CLIP_STACK_PATH:
{
CoglClipStackPath *path_entry = (CoglClipStackPath *) entry;
COGL_NOTE (CLIPPING, "Adding stencil clip for path");
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
add_stencil_clip_path (framebuffer,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
path_entry->matrix_entry,
path_entry->path,
using_stencil_buffer,
TRUE);
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
using_stencil_buffer = TRUE;
break;
}
case COGL_CLIP_STACK_PRIMITIVE:
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
{
CoglClipStackPrimitive *primitive_entry =
(CoglClipStackPrimitive *) entry;
COGL_NOTE (CLIPPING, "Adding stencil clip for primitive");
add_stencil_clip_primitive (framebuffer,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
primitive_entry->matrix_entry,
primitive_entry->primitive,
primitive_entry->bounds_x1,
primitive_entry->bounds_y1,
primitive_entry->bounds_x2,
primitive_entry->bounds_y2,
using_stencil_buffer,
TRUE);
using_stencil_buffer = TRUE;
break;
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
}
case COGL_CLIP_STACK_RECT:
{
CoglClipStackRect *rect = (CoglClipStackRect *) entry;
/* We don't need to do anything extra if the clip for this
rectangle was entirely described by its scissor bounds */
if (!rect->can_be_scissor)
{
/* If we support clip planes and we haven't already used
them then use that instead */
if (has_clip_planes)
{
COGL_NOTE (CLIPPING,
"Adding clip planes clip for rectangle");
set_clip_planes (framebuffer,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
rect->matrix_entry,
rect->x0,
rect->y0,
rect->x1,
rect->y1);
using_clip_planes = TRUE;
/* We can't use clip planes a second time */
has_clip_planes = FALSE;
}
else
{
COGL_NOTE (CLIPPING, "Adding stencil clip for rectangle");
add_stencil_clip_rectangle (framebuffer,
Re-design the matrix stack using a graph of ops This re-designs the matrix stack so we now keep track of each separate operation such as rotating, scaling, translating and multiplying as immutable, ref-counted nodes in a graph. Being a "graph" here means that different transformations composed of a sequence of linked operation nodes may share nodes. The first node in a matrix-stack is always a LOAD_IDENTITY operation. As an example consider if an application where to draw three rectangles A, B and C something like this: cogl_framebuffer_scale (fb, 2, 2, 2); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_translate (fb, 10, 0, 0); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_rotate (fb, 45, 0, 0, 1); cogl_framebuffer_draw_rectangle (...); /* A */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_draw_rectangle (...); /* B */ cogl_framebuffer_pop_matrix(fb); cogl_framebuffer_push_matrix(fb); cogl_framebuffer_set_modelview_matrix (fb, &mv); cogl_framebuffer_draw_rectangle (...); /* C */ cogl_framebuffer_pop_matrix(fb); That would result in a graph of nodes like this: LOAD_IDENTITY | SCALE / \ SAVE LOAD | | TRANSLATE RECTANGLE(C) | \ SAVE RECTANGLE(B) | ROTATE | RECTANGLE(A) Each push adds a SAVE operation which serves as a marker to rewind too when a corresponding pop is issued and also each SAVE node may also store a cached matrix representing the composition of all its ancestor nodes. This means if we repeatedly need to resolve a real CoglMatrix for a given node then we don't need to repeat the composition. Some advantages of this design are: - A single pointer to any node in the graph can now represent a complete, immutable transformation that can be logged for example into a journal. Previously we were storing a full CoglMatrix in each journal entry which is 16 floats for the matrix itself as well as space for flags and another 16 floats for possibly storing a cache of the inverse. This means that we significantly reduce the size of the journal when drawing lots of primitives and we also avoid copying over 128 bytes per entry. - It becomes much cheaper to check for equality. In cases where some (unlikely) false negatives are allowed simply comparing the pointers of two matrix stack graph entries is enough. Previously we would use memcmp() to compare matrices. - It becomes easier to do comparisons of transformations. By looking for the common ancestry between nodes we can determine the operations that differentiate the transforms and use those to gain a high level understanding of the differences. For example we use this in the journal to be able to efficiently determine when two rectangle transforms only differ by some translation so that we can perform software clipping. Reviewed-by: Neil Roberts <neil@linux.intel.com> (cherry picked from commit f75aee93f6b293ca7a7babbd8fcc326ee6bf7aef)
2012-02-20 10:59:48 -05:00
rect->matrix_entry,
rect->x0,
rect->y0,
rect->x1,
rect->y1,
!using_stencil_buffer);
using_stencil_buffer = TRUE;
}
}
break;
}
case COGL_CLIP_STACK_WINDOW_RECT:
break;
/* We don't need to do anything for window space rectangles because
* their functionality is entirely implemented by the entry bounding
* box */
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
}
}
/* Enabling clip planes is delayed to now so that they won't affect
setting up the stencil buffer */
if (using_clip_planes)
enable_clip_planes ();
Bug 1172 - Disjoint paths and clip to path * clutter/cogl/cogl-path.h: * clutter/cogl/common/cogl-primitives.c: * clutter/cogl/common/cogl-primitives.h: * clutter/cogl/gl/cogl-primitives.c: * clutter/cogl/gles/cogl-primitives.c: Changed the semantics of cogl_path_move_to. Previously this always started a new path but now it instead starts a new disjoint sub path. The path isn't cleared until you call either cogl_path_stroke, cogl_path_fill or cogl_path_new. There are also cogl_path_stroke_preserve and cogl_path_fill_preserve functions. * clutter/cogl/gl/cogl-context.c: * clutter/cogl/gl/cogl-context.h: * clutter/cogl/gles/cogl-context.c: * clutter/cogl/gles/cogl-context.h: Convert the path nodes array to a GArray. * clutter/cogl/gl/cogl-texture.c: * clutter/cogl/gles/cogl-texture.c: Call cogl_clip_ensure * clutter/cogl/common/cogl-clip-stack.c: * clutter/cogl/common/cogl-clip-stack.h: Simplified the clip stack code quite a bit to make it more maintainable. Previously whenever you added a new clip it would go through a separate route to immediately intersect with the current clip and when you removed it again it would immediately rebuild the entire clip. Now when you add or remove a clip it doesn't do anything immediately but just sets a dirty flag instead. * clutter/cogl/gl/cogl.c: * clutter/cogl/gles/cogl.c: Taken away the code to intersect stencil clips when there is exactly one stencil bit. It won't work with path clips and I don't know of any platform that doesn't have eight or zero stencil bits. It needs at least three bits to intersect a path with an existing clip. cogl_features_init now just decides you don't have a stencil buffer at all if you have less than three bits. * clutter/cogl/cogl.h.in: New functions and documentation. * tests/interactive/test-clip.c: Replaced with a different test that lets you add and remove clips. The three different mouse buttons add clips in different shapes. This makes it easier to test multiple levels of clipping. * tests/interactive/test-cogl-primitives.c: Use cogl_path_stroke_preserve when using the same path again. * doc/reference/cogl/cogl-sections.txt: Document the new functions.
2008-12-04 08:45:09 -05:00
ctx->current_clip_stack_uses_stencil = using_stencil_buffer;
}