mutter/cogl/cogl/cogl-matrix-stack.c

1212 lines
35 KiB
C
Raw Normal View History

/*
* Cogl
*
This re-licenses Cogl 1.18 under the MIT license Since the Cogl 1.18 branch is actively maintained in parallel with the master branch; this is a counter part to commit 1b83ef938fc16b which re-licensed the master branch to use the MIT license. This re-licensing is a follow up to the proposal that was sent to the Cogl mailing list: http://lists.freedesktop.org/archives/cogl/2013-December/001465.html Note: there was a copyright assignment policy in place for Clutter (and therefore Cogl which was part of Clutter at the time) until the 11th of June 2010 and so we only checked the details after that point (commit 0bbf50f905) For each file, authors were identified via this Git command: $ git blame -p -C -C -C20 -M -M10 0bbf50f905..HEAD We received blanket approvals for re-licensing all Red Hat and Collabora contributions which reduced how many people needed to be contacted individually: - http://lists.freedesktop.org/archives/cogl/2013-December/001470.html - http://lists.freedesktop.org/archives/cogl/2014-January/001536.html Individual approval requests were sent to all the other identified authors who all confirmed the re-license on the Cogl mailinglist: http://lists.freedesktop.org/archives/cogl/2014-January As well as updating the copyright header in all sources files, the COPYING file has been updated to reflect the license change and also document the other licenses used in Cogl such as the SGI Free Software License B, version 2.0 and the 3-clause BSD license. This patch was not simply cherry-picked from master; but the same methodology was used to check the source files.
2014-02-21 20:28:54 -05:00
* A Low Level GPU Graphics and Utilities API
*
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
* Copyright (C) 2009,2010,2012 Intel Corporation.
*
This re-licenses Cogl 1.18 under the MIT license Since the Cogl 1.18 branch is actively maintained in parallel with the master branch; this is a counter part to commit 1b83ef938fc16b which re-licensed the master branch to use the MIT license. This re-licensing is a follow up to the proposal that was sent to the Cogl mailing list: http://lists.freedesktop.org/archives/cogl/2013-December/001465.html Note: there was a copyright assignment policy in place for Clutter (and therefore Cogl which was part of Clutter at the time) until the 11th of June 2010 and so we only checked the details after that point (commit 0bbf50f905) For each file, authors were identified via this Git command: $ git blame -p -C -C -C20 -M -M10 0bbf50f905..HEAD We received blanket approvals for re-licensing all Red Hat and Collabora contributions which reduced how many people needed to be contacted individually: - http://lists.freedesktop.org/archives/cogl/2013-December/001470.html - http://lists.freedesktop.org/archives/cogl/2014-January/001536.html Individual approval requests were sent to all the other identified authors who all confirmed the re-license on the Cogl mailinglist: http://lists.freedesktop.org/archives/cogl/2014-January As well as updating the copyright header in all sources files, the COPYING file has been updated to reflect the license change and also document the other licenses used in Cogl such as the SGI Free Software License B, version 2.0 and the 3-clause BSD license. This patch was not simply cherry-picked from master; but the same methodology was used to check the source files.
2014-02-21 20:28:54 -05:00
* Permission is hereby granted, free of charge, to any person
* obtaining a copy of this software and associated documentation
* files (the "Software"), to deal in the Software without
* restriction, including without limitation the rights to use, copy,
* modify, merge, publish, distribute, sublicense, and/or sell copies
* of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
This re-licenses Cogl 1.18 under the MIT license Since the Cogl 1.18 branch is actively maintained in parallel with the master branch; this is a counter part to commit 1b83ef938fc16b which re-licensed the master branch to use the MIT license. This re-licensing is a follow up to the proposal that was sent to the Cogl mailing list: http://lists.freedesktop.org/archives/cogl/2013-December/001465.html Note: there was a copyright assignment policy in place for Clutter (and therefore Cogl which was part of Clutter at the time) until the 11th of June 2010 and so we only checked the details after that point (commit 0bbf50f905) For each file, authors were identified via this Git command: $ git blame -p -C -C -C20 -M -M10 0bbf50f905..HEAD We received blanket approvals for re-licensing all Red Hat and Collabora contributions which reduced how many people needed to be contacted individually: - http://lists.freedesktop.org/archives/cogl/2013-December/001470.html - http://lists.freedesktop.org/archives/cogl/2014-January/001536.html Individual approval requests were sent to all the other identified authors who all confirmed the re-license on the Cogl mailinglist: http://lists.freedesktop.org/archives/cogl/2014-January As well as updating the copyright header in all sources files, the COPYING file has been updated to reflect the license change and also document the other licenses used in Cogl such as the SGI Free Software License B, version 2.0 and the 3-clause BSD license. This patch was not simply cherry-picked from master; but the same methodology was used to check the source files.
2014-02-21 20:28:54 -05:00
* The above copyright notice and this permission notice shall be
* included in all copies or substantial portions of the Software.
*
This re-licenses Cogl 1.18 under the MIT license Since the Cogl 1.18 branch is actively maintained in parallel with the master branch; this is a counter part to commit 1b83ef938fc16b which re-licensed the master branch to use the MIT license. This re-licensing is a follow up to the proposal that was sent to the Cogl mailing list: http://lists.freedesktop.org/archives/cogl/2013-December/001465.html Note: there was a copyright assignment policy in place for Clutter (and therefore Cogl which was part of Clutter at the time) until the 11th of June 2010 and so we only checked the details after that point (commit 0bbf50f905) For each file, authors were identified via this Git command: $ git blame -p -C -C -C20 -M -M10 0bbf50f905..HEAD We received blanket approvals for re-licensing all Red Hat and Collabora contributions which reduced how many people needed to be contacted individually: - http://lists.freedesktop.org/archives/cogl/2013-December/001470.html - http://lists.freedesktop.org/archives/cogl/2014-January/001536.html Individual approval requests were sent to all the other identified authors who all confirmed the re-license on the Cogl mailinglist: http://lists.freedesktop.org/archives/cogl/2014-January As well as updating the copyright header in all sources files, the COPYING file has been updated to reflect the license change and also document the other licenses used in Cogl such as the SGI Free Software License B, version 2.0 and the 3-clause BSD license. This patch was not simply cherry-picked from master; but the same methodology was used to check the source files.
2014-02-21 20:28:54 -05:00
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*
* Authors:
* Havoc Pennington <hp@pobox.com> for litl
[cogl] Make sure we draw upside down to offscreen draw buffers First a few notes about Cogl coordinate systems: - Cogl defines the window origin, viewport origin and texture coordinates origin to be top left unlike OpenGL which defines them as bottom left. - Cogl defines the modelview and projection identity matrices in exactly the same way as OpenGL. - I.e. we believe that for 2D centric constructs: windows/framebuffers, viewports and textures developers are more used to dealing with a top left origin, but when modeling objects in 3D; an origin at the center with y going up is quite natural. The way Cogl handles textures is by uploading data upside down in OpenGL terms so that bottom left becomes top left. (Note: This also has the benefit that we don't need to flip the data we get from image decoding libraries since they typically also consider top left to be the image origin.) The viewport and window coords are mostly handled with various y = height - y tweaks before we pass y coordinates to OpenGL. Generally speaking though the handling of coordinate spaces in Cogl is a bit fragile. I guess partly because none of it was design to be, it just evolved from how Clutter defines its coordinates without much consideration or testing. I hope to improve this over a number of commits; starting here. This commit deals with the fact that offscreen draw buffers may be bound to textures but we don't "upload" the texture data upside down, and so if you texture from an offscreen draw buffer you need to manually flip the texture coordinates to get it the right way around. We now force offscreen rendering to be flipped upside down by tweaking the projection matrix right before we submit it to OpenGL to scale y by -1. The tweak is entirely hidden from the user such that if you call cogl_get_projection you will not see this scale.
2009-10-22 11:13:01 -04:00
* Robert Bragg <robert@linux.intel.com>
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
* Neil Roberts <neil@linux.intel.com>
*/
#ifdef HAVE_CONFIG_H
#include "cogl-config.h"
#endif
#include "cogl-context-private.h"
#include "cogl-util-gl-private.h"
#include "cogl-matrix-stack.h"
#include "cogl-framebuffer-private.h"
#include "cogl-object-private.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-private.h"
#include "cogl-magazine-private.h"
#include "cogl-gtype-private.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
static void _cogl_matrix_stack_free (CoglMatrixStack *stack);
COGL_OBJECT_DEFINE (MatrixStack, matrix_stack);
COGL_GTYPE_DEFINE_CLASS (MatrixStack, matrix_stack);
COGL_GTYPE_DEFINE_BOXED (MatrixEntry, matrix_entry,
cogl_matrix_entry_ref,
cogl_matrix_entry_unref);
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 CoglMagazine *cogl_matrix_stack_magazine;
static CoglMagazine *cogl_matrix_stack_matrices_magazine;
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
/* XXX: Note: this leaves entry->parent uninitialized! */
static CoglMatrixEntry *
_cogl_matrix_entry_new (CoglMatrixOp operation)
{
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 *entry =
_cogl_magazine_chunk_alloc (cogl_matrix_stack_magazine);
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->ref_count = 1;
entry->op = operation;
#ifdef COGL_DEBUG_ENABLED
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->composite_gets = 0;
#endif
return entry;
}
static void *
_cogl_matrix_stack_push_entry (CoglMatrixStack *stack,
CoglMatrixEntry *entry)
{
/* NB: The initial reference of the entry is transferred to the
* stack here.
*
* The stack only maintains a reference to the top of the stack (the
* last entry pushed) and each entry in-turn maintains a reference
* to its parent.
*
* We don't need to take a reference to the parent from the entry
* here because the we are stealing the reference that was held by
* the stack while that parent was previously the top of the stack.
*/
entry->parent = stack->last_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
stack->last_entry = entry;
return entry;
}
static void *
_cogl_matrix_stack_push_operation (CoglMatrixStack *stack,
CoglMatrixOp operation)
{
CoglMatrixEntry *entry = _cogl_matrix_entry_new (operation);
_cogl_matrix_stack_push_entry (stack, 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
return entry;
}
static void *
_cogl_matrix_stack_push_replacement_entry (CoglMatrixStack *stack,
CoglMatrixOp operation)
{
CoglMatrixEntry *old_top = stack->last_entry;
CoglMatrixEntry *new_top;
/* This would only be called for operations that completely replace
* the matrix. In that case we don't need to keep a reference to
* anything up to the last save entry. This optimisation could be
* important for applications that aren't using the stack but
* instead just perform their own matrix manipulations and load a
* new stack every frame. If this optimisation isn't done then the
* stack would just grow endlessly. See the comments
* cogl_matrix_stack_pop for a description of how popping works. */
for (new_top = old_top;
new_top->op != COGL_MATRIX_OP_SAVE && new_top->parent;
new_top = new_top->parent)
;
cogl_matrix_entry_ref (new_top);
cogl_matrix_entry_unref (old_top);
stack->last_entry = new_top;
return _cogl_matrix_stack_push_operation (stack, operation);
}
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
void
_cogl_matrix_entry_identity_init (CoglMatrixEntry *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
entry->ref_count = 1;
entry->op = COGL_MATRIX_OP_LOAD_IDENTITY;
entry->parent = NULL;
#ifdef COGL_DEBUG_ENABLED
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->composite_gets = 0;
#endif
}
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
void
cogl_matrix_stack_load_identity (CoglMatrixStack *stack)
{
_cogl_matrix_stack_push_replacement_entry (stack,
COGL_MATRIX_OP_LOAD_IDENTITY);
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
}
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
void
cogl_matrix_stack_translate (CoglMatrixStack *stack,
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 x,
float y,
float z)
{
CoglMatrixEntryTranslate *entry;
entry = _cogl_matrix_stack_push_operation (stack, COGL_MATRIX_OP_TRANSLATE);
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->x = x;
entry->y = y;
entry->z = z;
}
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
void
cogl_matrix_stack_rotate (CoglMatrixStack *stack,
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 angle,
float x,
float y,
float z)
{
CoglMatrixEntryRotate *entry;
entry = _cogl_matrix_stack_push_operation (stack, COGL_MATRIX_OP_ROTATE);
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->angle = angle;
entry->x = x;
entry->y = y;
entry->z = z;
}
void
cogl_matrix_stack_rotate_quaternion (CoglMatrixStack *stack,
const CoglQuaternion *quaternion)
{
CoglMatrixEntryRotateQuaternion *entry;
entry = _cogl_matrix_stack_push_operation (stack,
COGL_MATRIX_OP_ROTATE_QUATERNION);
entry->values[0] = quaternion->w;
entry->values[1] = quaternion->x;
entry->values[2] = quaternion->y;
entry->values[3] = quaternion->z;
}
void
cogl_matrix_stack_rotate_euler (CoglMatrixStack *stack,
const CoglEuler *euler)
{
CoglMatrixEntryRotateEuler *entry;
entry = _cogl_matrix_stack_push_operation (stack,
COGL_MATRIX_OP_ROTATE_EULER);
entry->heading = euler->heading;
entry->pitch = euler->pitch;
entry->roll = euler->roll;
}
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
void
cogl_matrix_stack_scale (CoglMatrixStack *stack,
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 x,
float y,
float z)
{
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
CoglMatrixEntryScale *entry;
entry = _cogl_matrix_stack_push_operation (stack, COGL_MATRIX_OP_SCALE);
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->x = x;
entry->y = y;
entry->z = z;
}
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
void
cogl_matrix_stack_multiply (CoglMatrixStack *stack,
const CoglMatrix *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
{
CoglMatrixEntryMultiply *entry;
entry = _cogl_matrix_stack_push_operation (stack, COGL_MATRIX_OP_MULTIPLY);
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 =
_cogl_magazine_chunk_alloc (cogl_matrix_stack_matrices_magazine);
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_init_from_array (entry->matrix, (float *)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
void
cogl_matrix_stack_set (CoglMatrixStack *stack,
const CoglMatrix *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
CoglMatrixEntryLoad *entry;
entry =
_cogl_matrix_stack_push_replacement_entry (stack,
COGL_MATRIX_OP_LOAD);
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 =
_cogl_magazine_chunk_alloc (cogl_matrix_stack_matrices_magazine);
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_init_from_array (entry->matrix, (float *)matrix);
}
void
cogl_matrix_stack_frustum (CoglMatrixStack *stack,
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 left,
float right,
float bottom,
float top,
float z_near,
float z_far)
{
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
CoglMatrixEntryLoad *entry;
entry =
_cogl_matrix_stack_push_replacement_entry (stack,
COGL_MATRIX_OP_LOAD);
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 =
_cogl_magazine_chunk_alloc (cogl_matrix_stack_matrices_magazine);
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_init_identity (entry->matrix);
cogl_matrix_frustum (entry->matrix,
left, right, bottom, top,
z_near, z_far);
}
void
cogl_matrix_stack_perspective (CoglMatrixStack *stack,
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 fov_y,
float aspect,
float z_near,
float z_far)
{
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
CoglMatrixEntryLoad *entry;
entry =
_cogl_matrix_stack_push_replacement_entry (stack,
COGL_MATRIX_OP_LOAD);
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 =
_cogl_magazine_chunk_alloc (cogl_matrix_stack_matrices_magazine);
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_init_identity (entry->matrix);
cogl_matrix_perspective (entry->matrix,
fov_y, aspect, z_near, z_far);
}
void
cogl_matrix_stack_orthographic (CoglMatrixStack *stack,
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 x_1,
float y_1,
float x_2,
float y_2,
float near,
float far)
{
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
CoglMatrixEntryLoad *entry;
entry =
_cogl_matrix_stack_push_replacement_entry (stack,
COGL_MATRIX_OP_LOAD);
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 =
_cogl_magazine_chunk_alloc (cogl_matrix_stack_matrices_magazine);
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_init_identity (entry->matrix);
cogl_matrix_orthographic (entry->matrix,
x_1, y_1, x_2, y_2, near, far);
}
void
cogl_matrix_stack_push (CoglMatrixStack *stack)
{
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
CoglMatrixEntrySave *entry;
entry = _cogl_matrix_stack_push_operation (stack, COGL_MATRIX_OP_SAVE);
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->cache_valid = 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
CoglMatrixEntry *
cogl_matrix_entry_ref (CoglMatrixEntry *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
/* A NULL pointer is considered a valid stack so we should accept
that as an argument */
if (entry)
entry->ref_count++;
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
return entry;
}
void
cogl_matrix_entry_unref (CoglMatrixEntry *entry)
{
CoglMatrixEntry *parent;
for (; entry && --entry->ref_count <= 0; entry = parent)
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
{
parent = entry->parent;
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
switch (entry->op)
{
case COGL_MATRIX_OP_LOAD_IDENTITY:
case COGL_MATRIX_OP_TRANSLATE:
case COGL_MATRIX_OP_ROTATE:
case COGL_MATRIX_OP_ROTATE_QUATERNION:
case COGL_MATRIX_OP_ROTATE_EULER:
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
case COGL_MATRIX_OP_SCALE:
break;
case COGL_MATRIX_OP_MULTIPLY:
{
CoglMatrixEntryMultiply *multiply =
(CoglMatrixEntryMultiply *)entry;
_cogl_magazine_chunk_free (cogl_matrix_stack_matrices_magazine,
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
multiply->matrix);
break;
}
case COGL_MATRIX_OP_LOAD:
{
CoglMatrixEntryLoad *load = (CoglMatrixEntryLoad *)entry;
_cogl_magazine_chunk_free (cogl_matrix_stack_matrices_magazine,
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
load->matrix);
break;
}
case COGL_MATRIX_OP_SAVE:
{
CoglMatrixEntrySave *save = (CoglMatrixEntrySave *)entry;
if (save->cache_valid)
_cogl_magazine_chunk_free (cogl_matrix_stack_matrices_magazine,
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
save->cache);
break;
}
}
_cogl_magazine_chunk_free (cogl_matrix_stack_magazine, 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
}
}
void
cogl_matrix_stack_pop (CoglMatrixStack *stack)
{
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 *old_top;
CoglMatrixEntry *new_top;
_COGL_RETURN_IF_FAIL (stack != NULL);
old_top = stack->last_entry;
_COGL_RETURN_IF_FAIL (old_top != NULL);
/* 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_matrix_entry_unref will unref the parent.
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
*/
/* Find the last save operation and remove it */
/* XXX: it would be an error to pop to the very beginning of the
* stack so we don't need to check for NULL pointer dereferencing. */
for (new_top = old_top;
new_top->op != COGL_MATRIX_OP_SAVE;
new_top = new_top->parent)
;
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
new_top = new_top->parent;
cogl_matrix_entry_ref (new_top);
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_unref (old_top);
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
stack->last_entry = new_top;
}
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
CoglBool
cogl_matrix_stack_get_inverse (CoglMatrixStack *stack,
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 *inverse)
{
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 matrix;
CoglMatrix *internal = cogl_matrix_stack_get (stack, &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
if (internal)
return cogl_matrix_get_inverse (internal, inverse);
else
return cogl_matrix_get_inverse (&matrix, inverse);
}
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
/* In addition to writing the stack matrix into the give @matrix
* argument this function *may* sometimes also return a pointer
* to a matrix too so if we are querying the inverse matrix we
* should query from the return matrix so that the result can
* be cached within the stack. */
CoglMatrix *
cogl_matrix_entry_get (CoglMatrixEntry *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 *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
int depth;
CoglMatrixEntry *current;
CoglMatrixEntry **children;
int i;
for (depth = 0, current = entry;
current;
current = current->parent, depth++)
{
switch (current->op)
{
case COGL_MATRIX_OP_LOAD_IDENTITY:
cogl_matrix_init_identity (matrix);
goto initialized;
case COGL_MATRIX_OP_LOAD:
{
CoglMatrixEntryLoad *load = (CoglMatrixEntryLoad *)current;
_cogl_matrix_init_from_matrix_without_inverse (matrix,
load->matrix);
goto initialized;
}
case COGL_MATRIX_OP_SAVE:
{
CoglMatrixEntrySave *save = (CoglMatrixEntrySave *)current;
if (!save->cache_valid)
{
CoglMagazine *matrices_magazine =
cogl_matrix_stack_matrices_magazine;
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
save->cache = _cogl_magazine_chunk_alloc (matrices_magazine);
cogl_matrix_entry_get (current->parent, save->cache);
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
save->cache_valid = TRUE;
}
_cogl_matrix_init_from_matrix_without_inverse (matrix, save->cache);
goto initialized;
}
default:
continue;
}
}
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
initialized:
if (depth == 0)
{
switch (entry->op)
{
case COGL_MATRIX_OP_LOAD_IDENTITY:
case COGL_MATRIX_OP_TRANSLATE:
case COGL_MATRIX_OP_ROTATE:
case COGL_MATRIX_OP_ROTATE_QUATERNION:
case COGL_MATRIX_OP_ROTATE_EULER:
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
case COGL_MATRIX_OP_SCALE:
case COGL_MATRIX_OP_MULTIPLY:
return NULL;
case COGL_MATRIX_OP_LOAD:
{
CoglMatrixEntryLoad *load = (CoglMatrixEntryLoad *)entry;
return load->matrix;
}
case COGL_MATRIX_OP_SAVE:
{
CoglMatrixEntrySave *save = (CoglMatrixEntrySave *)entry;
return save->cache;
}
}
g_warn_if_reached ();
return NULL;
}
#ifdef COGL_ENABLE_DEBUG
if (!current)
{
g_warning ("Inconsistent matrix stack");
return NULL;
}
entry->composite_gets++;
#endif
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
children = g_alloca (sizeof (CoglMatrixEntry) * depth);
/* We need walk the list of entries from the init/load/save entry
* back towards the leaf node but the nodes don't link to their
* children so we need to re-walk them here to add to a separate
* array. */
for (i = depth - 1, current = entry;
i >= 0 && current;
i--, current = current->parent)
{
children[i] = current;
}
#ifdef COGL_ENABLE_DEBUG
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 (COGL_DEBUG_ENABLED (COGL_DEBUG_PERFORMANCE) &&
entry->composite_gets >= 2)
{
COGL_NOTE (PERFORMANCE,
"Re-composing a matrix stack entry multiple times");
}
#endif
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
for (i = 0; i < depth; i++)
{
switch (children[i]->op)
{
case COGL_MATRIX_OP_TRANSLATE:
{
CoglMatrixEntryTranslate *translate =
(CoglMatrixEntryTranslate *)children[i];
cogl_matrix_translate (matrix,
translate->x,
translate->y,
translate->z);
continue;
}
case COGL_MATRIX_OP_ROTATE:
{
CoglMatrixEntryRotate *rotate=
(CoglMatrixEntryRotate *)children[i];
cogl_matrix_rotate (matrix,
rotate->angle,
rotate->x,
rotate->y,
rotate->z);
continue;
}
case COGL_MATRIX_OP_ROTATE_EULER:
{
CoglMatrixEntryRotateEuler *rotate =
(CoglMatrixEntryRotateEuler *)children[i];
CoglEuler euler;
cogl_euler_init (&euler,
rotate->heading,
rotate->pitch,
rotate->roll);
cogl_matrix_rotate_euler (matrix,
&euler);
continue;
}
case COGL_MATRIX_OP_ROTATE_QUATERNION:
{
CoglMatrixEntryRotateQuaternion *rotate =
(CoglMatrixEntryRotateQuaternion *)children[i];
CoglQuaternion quaternion;
cogl_quaternion_init_from_array (&quaternion, rotate->values);
cogl_matrix_rotate_quaternion (matrix, &quaternion);
continue;
}
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
case COGL_MATRIX_OP_SCALE:
{
CoglMatrixEntryScale *scale =
(CoglMatrixEntryScale *)children[i];
cogl_matrix_scale (matrix,
scale->x,
scale->y,
scale->z);
continue;
}
case COGL_MATRIX_OP_MULTIPLY:
{
CoglMatrixEntryMultiply *multiply =
(CoglMatrixEntryMultiply *)children[i];
cogl_matrix_multiply (matrix, matrix, multiply->matrix);
continue;
}
case COGL_MATRIX_OP_LOAD_IDENTITY:
case COGL_MATRIX_OP_LOAD:
case COGL_MATRIX_OP_SAVE:
g_warn_if_reached ();
continue;
}
}
return NULL;
}
CoglMatrixEntry *
cogl_matrix_stack_get_entry (CoglMatrixStack *stack)
{
return stack->last_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
/* In addition to writing the stack matrix into the give @matrix
* argument this function *may* sometimes also return a pointer
* to a matrix too so if we are querying the inverse matrix we
* should query from the return matrix so that the result can
* be cached within the stack. */
CoglMatrix *
cogl_matrix_stack_get (CoglMatrixStack *stack,
CoglMatrix *matrix)
{
return cogl_matrix_entry_get (stack->last_entry, 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
}
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
_cogl_matrix_stack_free (CoglMatrixStack *stack)
{
cogl_matrix_entry_unref (stack->last_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
g_slice_free (CoglMatrixStack, stack);
}
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
CoglMatrixStack *
cogl_matrix_stack_new (CoglContext *ctx)
{
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
CoglMatrixStack *stack = g_slice_new (CoglMatrixStack);
if (G_UNLIKELY (cogl_matrix_stack_magazine == NULL))
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_stack_magazine =
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_magazine_new (sizeof (CoglMatrixEntryFull), 20);
cogl_matrix_stack_matrices_magazine =
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_magazine_new (sizeof (CoglMatrix), 20);
}
stack->context = ctx;
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
stack->last_entry = NULL;
cogl_matrix_entry_ref (&ctx->identity_entry);
_cogl_matrix_stack_push_entry (stack, &ctx->identity_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
return _cogl_matrix_stack_object_new (stack);
}
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 CoglMatrixEntry *
_cogl_matrix_entry_skip_saves (CoglMatrixEntry *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
/* We currently assume that every stack starts with an
* _OP_LOAD_IDENTITY so we don't need to worry about
* NULL pointer dereferencing here. */
while (entry->op == COGL_MATRIX_OP_SAVE)
entry = entry->parent;
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
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
CoglBool
cogl_matrix_entry_calculate_translation (CoglMatrixEntry *entry0,
CoglMatrixEntry *entry1,
float *x,
float *y,
float *z)
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
{
GSList *head0 = NULL;
GSList *head1 = NULL;
CoglMatrixEntry *node0;
CoglMatrixEntry *node1;
int len0 = 0;
int len1 = 0;
int count;
GSList *common_ancestor0;
GSList *common_ancestor1;
/* Algorithm:
*
* 1) Ignoring _OP_SAVE entries walk the ancestors of each entry to
* the root node or any non-translation node, adding a pointer to
* each ancestor node to two linked lists.
*
* 2) Compare the lists to find the nodes where they start to
* differ marking the common_ancestor node for each list.
*
* 3) For the list corresponding to entry0, start iterating after
* the common ancestor applying the negative of all translations
* to x, y and z.
*
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
* 4) For the list corresponding to entry1, start iterating after
* the common ancestor applying the positive of all translations
* to x, y and z.
*
* If we come across any non-translation operations during 3) or 4)
* then bail out returning 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
for (node0 = entry0; node0; node0 = node0->parent)
{
GSList *link;
if (node0->op == COGL_MATRIX_OP_SAVE)
continue;
link = alloca (sizeof (GSList));
link->next = head0;
link->data = node0;
head0 = link;
len0++;
if (node0->op != COGL_MATRIX_OP_TRANSLATE)
break;
}
for (node1 = entry1; node1; node1 = node1->parent)
{
GSList *link;
if (node1->op == COGL_MATRIX_OP_SAVE)
continue;
link = alloca (sizeof (GSList));
link->next = head1;
link->data = node1;
head1 = link;
len1++;
if (node1->op != COGL_MATRIX_OP_TRANSLATE)
break;
}
if (head0->data != head1->data)
return FALSE;
common_ancestor0 = head0;
common_ancestor1 = head1;
head0 = head0->next;
head1 = head1->next;
count = MIN (len0, len1) - 1;
while (count--)
{
if (head0->data != head1->data)
break;
common_ancestor0 = head0;
common_ancestor1 = head1;
head0 = head0->next;
head1 = head1->next;
}
*x = 0;
*y = 0;
*z = 0;
for (head0 = common_ancestor0->next; head0; head0 = head0->next)
{
CoglMatrixEntryTranslate *translate;
node0 = head0->data;
if (node0->op != COGL_MATRIX_OP_TRANSLATE)
return FALSE;
translate = (CoglMatrixEntryTranslate *)node0;
*x = *x - translate->x;
*y = *y - translate->y;
*z = *z - translate->z;
}
for (head1 = common_ancestor1->next; head1; head1 = head1->next)
{
CoglMatrixEntryTranslate *translate;
node1 = head1->data;
if (node1->op != COGL_MATRIX_OP_TRANSLATE)
return FALSE;
translate = (CoglMatrixEntryTranslate *)node1;
*x = *x + translate->x;
*y = *y + translate->y;
*z = *z + translate->z;
}
return TRUE;
}
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
CoglBool
cogl_matrix_entry_is_identity (CoglMatrixEntry *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
return entry ? entry->op == COGL_MATRIX_OP_LOAD_IDENTITY : FALSE;
}
static void
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_flush_to_gl_builtin (CoglContext *ctx,
CoglBool is_identity,
CoglMatrix *matrix,
CoglMatrixMode mode)
{
g_assert (_cogl_has_private_feature (ctx, COGL_PRIVATE_FEATURE_GL_FIXED));
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
#if defined (HAVE_COGL_GL) || defined (HAVE_COGL_GLES)
if (ctx->flushed_matrix_mode != mode)
{
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
GLenum gl_mode = 0;
switch (mode)
{
case COGL_MATRIX_MODELVIEW:
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
gl_mode = GL_MODELVIEW;
break;
case COGL_MATRIX_PROJECTION:
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
gl_mode = GL_PROJECTION;
break;
case COGL_MATRIX_TEXTURE:
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
gl_mode = GL_TEXTURE;
break;
}
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
GE (ctx, glMatrixMode (gl_mode));
ctx->flushed_matrix_mode = mode;
}
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
if (is_identity)
GE (ctx, glLoadIdentity ());
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
else
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
GE (ctx, glLoadMatrixf (cogl_matrix_get_array (matrix)));
#endif
}
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
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
void
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 (CoglContext *ctx,
CoglMatrixEntry *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
CoglMatrixMode mode,
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
CoglFramebuffer *framebuffer,
CoglBool disable_flip)
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
{
g_assert (_cogl_has_private_feature (ctx, COGL_PRIVATE_FEATURE_GL_FIXED));
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
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
#if defined (HAVE_COGL_GL) || defined (HAVE_COGL_GLES)
{
CoglBool needs_flip;
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
CoglMatrixEntryCache *cache;
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
if (mode == COGL_MATRIX_PROJECTION)
{
/* Because Cogl defines texture coordinates to have a top left
* origin and because offscreen framebuffers may be used for
* rendering to textures we always render upside down to
* offscreen buffers. Also for some backends we need to render
* onscreen buffers upside-down too.
*/
if (disable_flip)
needs_flip = FALSE;
else
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
needs_flip = cogl_is_offscreen (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
cache = &ctx->builtin_flushed_projection;
}
else
{
needs_flip = FALSE;
if (mode == COGL_MATRIX_MODELVIEW)
cache = &ctx->builtin_flushed_modelview;
else
cache = NULL;
}
/* We don't need to do anything if the state is the same */
if (!cache ||
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_cache_maybe_update (cache, entry, needs_flip))
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
{
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
CoglBool is_identity;
CoglMatrix matrix;
if (entry->op == COGL_MATRIX_OP_LOAD_IDENTITY)
is_identity = TRUE;
else
{
is_identity = FALSE;
cogl_matrix_entry_get (entry, &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
}
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
if (needs_flip)
{
CoglMatrix flipped_matrix;
cogl_matrix_multiply (&flipped_matrix,
&ctx->y_flip_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
is_identity ?
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
&ctx->identity_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
&matrix);
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
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_flush_to_gl_builtin (ctx,
/* not identity */
FALSE,
&flipped_matrix,
mode);
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
}
else
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_flush_to_gl_builtin (ctx,
is_identity,
&matrix,
mode);
}
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
}
}
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
#endif
}
CoglBool
cogl_matrix_entry_equal (CoglMatrixEntry *entry0,
CoglMatrixEntry *entry1)
{
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
for (;
entry0 && entry1;
entry0 = entry0->parent, entry1 = entry1->parent)
{
entry0 = _cogl_matrix_entry_skip_saves (entry0);
entry1 = _cogl_matrix_entry_skip_saves (entry1);
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 (entry0 == entry1)
return TRUE;
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 (entry0->op != entry1->op)
return FALSE;
switch (entry0->op)
{
case COGL_MATRIX_OP_LOAD_IDENTITY:
return TRUE;
case COGL_MATRIX_OP_TRANSLATE:
{
CoglMatrixEntryTranslate *translate0 =
(CoglMatrixEntryTranslate *)entry0;
CoglMatrixEntryTranslate *translate1 =
(CoglMatrixEntryTranslate *)entry1;
/* We could perhaps use an epsilon to compare here?
* I expect the false negatives are probaly never going to
* be a problem and this is a bit cheaper. */
if (translate0->x != translate1->x ||
translate0->y != translate1->y ||
translate0->z != translate1->z)
return FALSE;
}
break;
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
case COGL_MATRIX_OP_ROTATE:
{
CoglMatrixEntryRotate *rotate0 =
(CoglMatrixEntryRotate *)entry0;
CoglMatrixEntryRotate *rotate1 =
(CoglMatrixEntryRotate *)entry1;
if (rotate0->angle != rotate1->angle ||
rotate0->x != rotate1->x ||
rotate0->y != rotate1->y ||
rotate0->z != rotate1->z)
return FALSE;
}
break;
case COGL_MATRIX_OP_ROTATE_QUATERNION:
{
CoglMatrixEntryRotateQuaternion *rotate0 =
(CoglMatrixEntryRotateQuaternion *)entry0;
CoglMatrixEntryRotateQuaternion *rotate1 =
(CoglMatrixEntryRotateQuaternion *)entry1;
int i;
for (i = 0; i < 4; i++)
if (rotate0->values[i] != rotate1->values[i])
return FALSE;
}
break;
case COGL_MATRIX_OP_ROTATE_EULER:
{
CoglMatrixEntryRotateEuler *rotate0 =
(CoglMatrixEntryRotateEuler *)entry0;
CoglMatrixEntryRotateEuler *rotate1 =
(CoglMatrixEntryRotateEuler *)entry1;
if (rotate0->heading != rotate1->heading ||
rotate0->pitch != rotate1->pitch ||
rotate0->roll != rotate1->roll)
return FALSE;
}
break;
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
case COGL_MATRIX_OP_SCALE:
{
CoglMatrixEntryScale *scale0 = (CoglMatrixEntryScale *)entry0;
CoglMatrixEntryScale *scale1 = (CoglMatrixEntryScale *)entry1;
if (scale0->x != scale1->x ||
scale0->y != scale1->y ||
scale0->z != scale1->z)
return FALSE;
}
break;
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
case COGL_MATRIX_OP_MULTIPLY:
{
CoglMatrixEntryMultiply *mult0 = (CoglMatrixEntryMultiply *)entry0;
CoglMatrixEntryMultiply *mult1 = (CoglMatrixEntryMultiply *)entry1;
if (!cogl_matrix_equal (mult0->matrix, mult1->matrix))
return FALSE;
}
break;
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
case COGL_MATRIX_OP_LOAD:
{
CoglMatrixEntryLoad *load0 = (CoglMatrixEntryLoad *)entry0;
CoglMatrixEntryLoad *load1 = (CoglMatrixEntryLoad *)entry1;
/* There's no need to check any further since an
* _OP_LOAD makes all the ancestors redundant as far as
* the final matrix value is concerned. */
return cogl_matrix_equal (load0->matrix, load1->matrix);
}
case COGL_MATRIX_OP_SAVE:
/* We skip over saves above so we shouldn't see save entries */
g_warn_if_reached ();
}
}
return FALSE;
}
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
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
void
cogl_debug_matrix_entry_print (CoglMatrixEntry *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
{
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
int depth;
CoglMatrixEntry *e;
CoglMatrixEntry **children;
int i;
for (depth = 0, e = entry; e; e = e->parent)
depth++;
children = g_alloca (sizeof (CoglMatrixEntry) * depth);
for (i = depth - 1, e = entry;
i >= 0 && e;
i--, e = e->parent)
{
children[i] = e;
}
g_print ("MatrixEntry %p =\n", entry);
for (i = 0; i < depth; i++)
{
entry = children[i];
switch (entry->op)
{
case COGL_MATRIX_OP_LOAD_IDENTITY:
g_print (" LOAD IDENTITY\n");
continue;
case COGL_MATRIX_OP_TRANSLATE:
{
CoglMatrixEntryTranslate *translate =
(CoglMatrixEntryTranslate *)entry;
g_print (" TRANSLATE X=%f Y=%f Z=%f\n",
translate->x,
translate->y,
translate->z);
continue;
}
case COGL_MATRIX_OP_ROTATE:
{
CoglMatrixEntryRotate *rotate =
(CoglMatrixEntryRotate *)entry;
g_print (" ROTATE ANGLE=%f X=%f Y=%f Z=%f\n",
rotate->angle,
rotate->x,
rotate->y,
rotate->z);
continue;
}
case COGL_MATRIX_OP_ROTATE_QUATERNION:
{
CoglMatrixEntryRotateQuaternion *rotate =
(CoglMatrixEntryRotateQuaternion *)entry;
g_print (" ROTATE QUATERNION w=%f x=%f y=%f z=%f\n",
rotate->values[0],
rotate->values[1],
rotate->values[2],
rotate->values[3]);
continue;
}
case COGL_MATRIX_OP_ROTATE_EULER:
{
CoglMatrixEntryRotateEuler *rotate =
(CoglMatrixEntryRotateEuler *)entry;
g_print (" ROTATE EULER heading=%f pitch=%f roll=%f\n",
rotate->heading,
rotate->pitch,
rotate->roll);
continue;
}
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
case COGL_MATRIX_OP_SCALE:
{
CoglMatrixEntryScale *scale = (CoglMatrixEntryScale *)entry;
g_print (" SCALE X=%f Y=%f Z=%f\n",
scale->x,
scale->y,
scale->z);
continue;
}
case COGL_MATRIX_OP_MULTIPLY:
{
CoglMatrixEntryMultiply *mult = (CoglMatrixEntryMultiply *)entry;
g_print (" MULT:\n");
_cogl_matrix_prefix_print (" ", mult->matrix);
continue;
}
case COGL_MATRIX_OP_LOAD:
{
CoglMatrixEntryLoad *load = (CoglMatrixEntryLoad *)entry;
g_print (" LOAD:\n");
_cogl_matrix_prefix_print (" ", load->matrix);
continue;
}
case COGL_MATRIX_OP_SAVE:
g_print (" SAVE\n");
}
}
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
}
void
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_cache_init (CoglMatrixEntryCache *cache)
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
{
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
cache->entry = NULL;
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
cache->flushed_identity = FALSE;
cache->flipped = FALSE;
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
}
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 function can report false negatives since it never does a
* deep comparison of the stack matrices. */
CoglBool
_cogl_matrix_entry_cache_maybe_update (CoglMatrixEntryCache *cache,
CoglMatrixEntry *entry,
CoglBool flip)
{
CoglBool is_identity;
CoglBool updated = FALSE;
if (cache->flipped != flip)
{
cache->flipped = flip;
updated = TRUE;
}
is_identity = (entry->op == COGL_MATRIX_OP_LOAD_IDENTITY);
if (cache->flushed_identity != is_identity)
{
cache->flushed_identity = is_identity;
updated = TRUE;
}
if (cache->entry != entry)
{
cogl_matrix_entry_ref (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
if (cache->entry)
cogl_matrix_entry_unref (cache->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
cache->entry = entry;
/* We want to make sure here that if the cache->entry and the
* given @entry are both identity matrices then even though they
* are different entries we don't want to consider this an
* update...
*/
updated |= !is_identity;
}
return updated;
}
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
void
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_cache_destroy (CoglMatrixEntryCache *cache)
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
{
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 (cache->entry)
cogl_matrix_entry_unref (cache->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
}