mutter/cogl/cogl-journal.c

952 lines
33 KiB
C
Raw Normal View History

/*
* Cogl
*
* An object oriented GL/GLES Abstraction/Utility Layer
*
* Copyright (C) 2007,2008,2009 Intel Corporation.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library. If not, see <http://www.gnu.org/licenses/>.
*
*
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "cogl.h"
#include "cogl-debug.h"
#include "cogl-internal.h"
#include "cogl-context.h"
#include "cogl-journal-private.h"
#include "cogl-texture-private.h"
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
#include "cogl-pipeline-private.h"
#include "cogl-pipeline-opengl-private.h"
#include "cogl-vertex-buffer-private.h"
#include "cogl-framebuffer-private.h"
#include "cogl-profile.h"
#include "cogl-vertex-attribute-private.h"
#include <string.h>
#include <gmodule.h>
#include <math.h>
/* XXX NB:
* Our journal's vertex data is arranged as follows:
* 4 vertices per quad:
* 2 or 3 GLfloats per position (3 when doing software transforms)
* 4 RGBA GLubytes,
* 2 GLfloats per tex coord * n_layers
*
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
* Where n_layers corresponds to the number of pipeline layers enabled
*
* To avoid frequent changes in the stride of our vertex data we always pad
* n_layers to be >= 2
*
* When we are transforming quads in software we need to also track the z
* coordinate of transformed vertices.
*
* So for a given number of layers this gets the stride in 32bit words:
*/
#define SW_TRANSFORM (!(cogl_debug_flags & \
COGL_DEBUG_DISABLE_SOFTWARE_TRANSFORM))
#define POS_STRIDE (SW_TRANSFORM ? 3 : 2) /* number of 32bit words */
#define N_POS_COMPONENTS POS_STRIDE
#define COLOR_STRIDE 1 /* number of 32bit words */
#define TEX_STRIDE 2 /* number of 32bit words */
#define MIN_LAYER_PADING 2
#define GET_JOURNAL_VB_STRIDE_FOR_N_LAYERS(N_LAYERS) \
(POS_STRIDE + COLOR_STRIDE + \
TEX_STRIDE * (N_LAYERS < MIN_LAYER_PADING ? MIN_LAYER_PADING : N_LAYERS))
typedef struct _CoglJournalFlushState
{
CoglVertexArray *vertex_array;
GArray *attributes;
int current_attribute;
gsize stride;
size_t array_offset;
GLuint current_vertex;
#ifndef HAVE_COGL_GL
CoglIndices *indices;
gsize indices_type_size;
#endif
CoglMatrixStack *modelview_stack;
CoglMatrixStack *projection_stack;
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
CoglPipeline *source;
} CoglJournalFlushState;
typedef void (*CoglJournalBatchCallback) (CoglJournalEntry *start,
int n_entries,
void *data);
typedef gboolean (*CoglJournalBatchTest) (CoglJournalEntry *entry0,
CoglJournalEntry *entry1);
void
_cogl_journal_dump_quad_vertices (guint8 *data, int n_layers)
{
cogl: improves header and coding style consistency We've had complaints that our Cogl code/headers are a bit "special" so this is a first pass at tidying things up by giving them some consistency. These changes are all consistent with how new code in Cogl is being written, but the style isn't consistently applied across all code yet. There are two parts to this patch; but since each one required a large amount of effort to maintain tidy indenting it made sense to combine the changes to reduce the time spent re indenting the same lines. The first change is to use a consistent style for declaring function prototypes in headers. Cogl headers now consistently use this style for prototypes: return_type cogl_function_name (CoglType arg0, CoglType arg1); Not everyone likes this style, but it seems that most of the currently active Cogl developers agree on it. The second change is to constrain the use of redundant glib data types in Cogl. Uses of gint, guint, gfloat, glong, gulong and gchar have all been replaced with int, unsigned int, float, long, unsigned long and char respectively. When talking about pixel data; use of guchar has been replaced with guint8, otherwise unsigned char can be used. The glib types that we continue to use for portability are gboolean, gint{8,16,32,64}, guint{8,16,32,64} and gsize. The general intention is that Cogl should look palatable to the widest range of C programmers including those outside the Gnome community so - especially for the public API - we want to minimize the number of foreign looking typedefs.
2010-02-09 20:57:32 -05:00
gsize stride = GET_JOURNAL_VB_STRIDE_FOR_N_LAYERS (n_layers);
int i;
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
g_print ("n_layers = %d; stride = %d; pos stride = %d; color stride = %d; "
"tex stride = %d; stride in bytes = %d\n",
n_layers, (int)stride, POS_STRIDE, COLOR_STRIDE,
TEX_STRIDE, (int)stride * 4);
for (i = 0; i < 4; i++)
{
float *v = (float *)data + (i * stride);
guint8 *c = data + (POS_STRIDE * 4) + (i * stride * 4);
int j;
if (G_UNLIKELY (cogl_debug_flags &
COGL_DEBUG_DISABLE_SOFTWARE_TRANSFORM))
g_print ("v%d: x = %f, y = %f, rgba=0x%02X%02X%02X%02X",
i, v[0], v[1], c[0], c[1], c[2], c[3]);
else
g_print ("v%d: x = %f, y = %f, z = %f, rgba=0x%02X%02X%02X%02X",
i, v[0], v[1], v[2], c[0], c[1], c[2], c[3]);
for (j = 0; j < n_layers; j++)
{
float *t = v + POS_STRIDE + COLOR_STRIDE + TEX_STRIDE * j;
g_print (", tx%d = %f, ty%d = %f", j, t[0], j, t[1]);
}
g_print ("\n");
}
}
void
_cogl_journal_dump_quad_batch (guint8 *data, int n_layers, int n_quads)
{
cogl: improves header and coding style consistency We've had complaints that our Cogl code/headers are a bit "special" so this is a first pass at tidying things up by giving them some consistency. These changes are all consistent with how new code in Cogl is being written, but the style isn't consistently applied across all code yet. There are two parts to this patch; but since each one required a large amount of effort to maintain tidy indenting it made sense to combine the changes to reduce the time spent re indenting the same lines. The first change is to use a consistent style for declaring function prototypes in headers. Cogl headers now consistently use this style for prototypes: return_type cogl_function_name (CoglType arg0, CoglType arg1); Not everyone likes this style, but it seems that most of the currently active Cogl developers agree on it. The second change is to constrain the use of redundant glib data types in Cogl. Uses of gint, guint, gfloat, glong, gulong and gchar have all been replaced with int, unsigned int, float, long, unsigned long and char respectively. When talking about pixel data; use of guchar has been replaced with guint8, otherwise unsigned char can be used. The glib types that we continue to use for portability are gboolean, gint{8,16,32,64}, guint{8,16,32,64} and gsize. The general intention is that Cogl should look palatable to the widest range of C programmers including those outside the Gnome community so - especially for the public API - we want to minimize the number of foreign looking typedefs.
2010-02-09 20:57:32 -05:00
gsize byte_stride = GET_JOURNAL_VB_STRIDE_FOR_N_LAYERS (n_layers) * 4;
int i;
g_print ("_cogl_journal_dump_quad_batch: n_layers = %d, n_quads = %d\n",
n_layers, n_quads);
for (i = 0; i < n_quads; i++)
_cogl_journal_dump_quad_vertices (data + byte_stride * 4 * i, n_layers);
}
static void
batch_and_call (CoglJournalEntry *entries,
int n_entries,
CoglJournalBatchTest can_batch_callback,
CoglJournalBatchCallback batch_callback,
void *data)
{
int i;
int batch_len = 1;
CoglJournalEntry *batch_start = entries;
if (n_entries < 1)
return;
for (i = 1; i < n_entries; i++)
{
CoglJournalEntry *entry0 = &entries[i - 1];
CoglJournalEntry *entry1 = entry0 + 1;
if (can_batch_callback (entry0, entry1))
{
batch_len++;
continue;
}
batch_callback (batch_start, batch_len, data);
batch_start = entry1;
batch_len = 1;
}
/* The last batch... */
batch_callback (batch_start, batch_len, data);
}
static void
_cogl_journal_flush_modelview_and_entries (CoglJournalEntry *batch_start,
int batch_len,
void *data)
{
CoglJournalFlushState *state = data;
CoglVertexAttribute **attributes;
COGL_STATIC_TIMER (time_flush_modelview_and_entries,
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
"flush: pipeline+entries", /* parent */
"flush: modelview+entries",
"The time spent flushing modelview + entries",
0 /* no application private data */);
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
COGL_TIMER_START (_cogl_uprof_context, time_flush_modelview_and_entries);
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_BATCHING))
g_print ("BATCHING: modelview batch len = %d\n", batch_len);
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_DISABLE_SOFTWARE_TRANSFORM))
{
_cogl_matrix_stack_set (state->modelview_stack,
&batch_start->model_view);
_cogl_matrix_stack_flush_to_gl (state->modelview_stack,
COGL_MATRIX_MODELVIEW);
}
attributes = (CoglVertexAttribute **)state->attributes->data;
cogl_push_source (state->source);
#ifdef HAVE_COGL_GL
/* XXX: it's rather evil that we sneak in the GL_QUADS enum here... */
_cogl_draw_vertex_attributes_array (GL_QUADS,
state->current_vertex, batch_len * 4,
attributes);
#else /* HAVE_COGL_GL */
if (batch_len > 1)
{
_cogl_draw_indexed_vertex_attributes_array (COGL_VERTICES_MODE_TRIANGLES,
state->current_vertex * 6 / 4,
batch_len * 6,
state->indices,
attributes);
}
else
{
_cogl_draw_vertex_attributes_array (COGL_VERTICES_MODE_TRIANGLE_FAN,
state->current_vertex, 4,
attributes);
}
#endif
/* DEBUGGING CODE XXX: This path will cause all rectangles to be
* drawn with a coloured outline. Each batch will be rendered with
* the same color. This may e.g. help with debugging texture slicing
* issues, visually seeing what is batched and debugging blending
* issues, plus it looks quite cool.
*/
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_RECTANGLES))
{
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
static CoglPipeline *outline = NULL;
guint8 color_intensity;
int i;
CoglVertexAttribute *loop_attributes[2];
_COGL_GET_CONTEXT (ctxt, NO_RETVAL);
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
if (outline == NULL)
outline = cogl_pipeline_new ();
/* The least significant three bits represent the three
components so that the order of colours goes red, green,
yellow, blue, magenta, cyan. Black and white are skipped. The
next two bits give four scales of intensity for those colours
in the order 0xff, 0xcc, 0x99, and 0x66. This gives a total
of 24 colours. If there are more than 24 batches on the stage
then it will wrap around */
color_intensity = 0xff - 0x33 * (ctxt->journal_rectangles_color >> 3);
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
cogl_pipeline_set_color4ub (outline,
(ctxt->journal_rectangles_color & 1) ?
color_intensity : 0,
(ctxt->journal_rectangles_color & 2) ?
color_intensity : 0,
(ctxt->journal_rectangles_color & 4) ?
color_intensity : 0,
0xff);
cogl_set_source (outline);
loop_attributes[0] = attributes[0]; /* we just want the position */
loop_attributes[1] = NULL;
for (i = 0; i < batch_len; i++)
_cogl_draw_vertex_attributes_array (COGL_VERTICES_MODE_LINE_LOOP,
4 * i + state->current_vertex, 4,
loop_attributes);
/* Go to the next color */
do
ctxt->journal_rectangles_color = ((ctxt->journal_rectangles_color + 1) &
((1 << 5) - 1));
/* We don't want to use black or white */
while ((ctxt->journal_rectangles_color & 0x07) == 0
|| (ctxt->journal_rectangles_color & 0x07) == 0x07);
}
state->current_vertex += (4 * batch_len);
cogl_pop_source ();
COGL_TIMER_STOP (_cogl_uprof_context, time_flush_modelview_and_entries);
}
static gboolean
compare_entry_modelviews (CoglJournalEntry *entry0,
CoglJournalEntry *entry1)
{
/* Batch together quads with the same model view matrix */
/* FIXME: this is nasty, there are much nicer ways to track this
* (at the add_quad_vertices level) without resorting to a memcmp!
*
* E.g. If the cogl-current-matrix code maintained an "age" for
* the modelview matrix we could simply check in add_quad_vertices
* if the age has increased, and if so record the change as a
* boolean in the journal.
*/
if (memcmp (&entry0->model_view, &entry1->model_view,
sizeof (GLfloat) * 16) == 0)
return TRUE;
else
return FALSE;
}
/* At this point we have a run of quads that we know have compatible
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
* pipelines, but they may not all have the same modelview matrix */
static void
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
_cogl_journal_flush_pipeline_and_entries (CoglJournalEntry *batch_start,
cogl: improves header and coding style consistency We've had complaints that our Cogl code/headers are a bit "special" so this is a first pass at tidying things up by giving them some consistency. These changes are all consistent with how new code in Cogl is being written, but the style isn't consistently applied across all code yet. There are two parts to this patch; but since each one required a large amount of effort to maintain tidy indenting it made sense to combine the changes to reduce the time spent re indenting the same lines. The first change is to use a consistent style for declaring function prototypes in headers. Cogl headers now consistently use this style for prototypes: return_type cogl_function_name (CoglType arg0, CoglType arg1); Not everyone likes this style, but it seems that most of the currently active Cogl developers agree on it. The second change is to constrain the use of redundant glib data types in Cogl. Uses of gint, guint, gfloat, glong, gulong and gchar have all been replaced with int, unsigned int, float, long, unsigned long and char respectively. When talking about pixel data; use of guchar has been replaced with guint8, otherwise unsigned char can be used. The glib types that we continue to use for portability are gboolean, gint{8,16,32,64}, guint{8,16,32,64} and gsize. The general intention is that Cogl should look palatable to the widest range of C programmers including those outside the Gnome community so - especially for the public API - we want to minimize the number of foreign looking typedefs.
2010-02-09 20:57:32 -05:00
int batch_len,
void *data)
{
CoglJournalFlushState *state = data;
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
COGL_STATIC_TIMER (time_flush_pipeline_entries,
"flush: texcoords+pipeline+entries", /* parent */
"flush: pipeline+entries",
"The time spent flushing pipeline + entries",
0 /* no application private data */);
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
COGL_TIMER_START (_cogl_uprof_context, time_flush_pipeline_entries);
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_BATCHING))
g_print ("BATCHING: pipeline batch len = %d\n", batch_len);
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
state->source = batch_start->pipeline;
/* If we haven't transformed the quads in software then we need to also break
* up batches according to changes in the modelview matrix... */
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_DISABLE_SOFTWARE_TRANSFORM))
{
batch_and_call (batch_start,
batch_len,
compare_entry_modelviews,
_cogl_journal_flush_modelview_and_entries,
data);
}
else
_cogl_journal_flush_modelview_and_entries (batch_start, batch_len, data);
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
COGL_TIMER_STOP (_cogl_uprof_context, time_flush_pipeline_entries);
}
static gboolean
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
compare_entry_pipelines (CoglJournalEntry *entry0, CoglJournalEntry *entry1)
{
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
/* batch rectangles using compatible pipelines */
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
/* XXX: _cogl_pipeline_equal may give false negatives since it avoids
* deep comparisons as an optimization. It aims to compare enough so
* that we that we are able to batch the 90% common cases, but may not
* look at less common differences. */
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
if (_cogl_pipeline_equal (entry0->pipeline,
entry1->pipeline,
CoglMaterial: Implements sparse materials design This is a complete overhaul of the data structures used to manage CoglMaterial state. We have these requirements that were aiming to meet: (Note: the references to "renderlists" correspond to the effort to support scenegraph level shuffling of Clutter actor primitives so we can minimize GPU state changes) Sparse State: We wanted a design that allows sparse descriptions of state so it scales well as we make CoglMaterial responsible for more and more state. It needs to scale well in terms of memory usage and the cost of operations we need to apply to materials such as comparing, copying and flushing their state. I.e. we would rather have these things scale by the number of real changes a material represents not by how much overall state CoglMaterial becomes responsible for. Cheap Copies: As we add support for renderlists in Clutter we will need to be able to get an immutable handle for a given material's current state so that we can retain a record of a primitive with its associated material without worrying that changes to the original material will invalidate that record. No more flush override options: We want to get rid of the flush overrides mechanism we currently use to deal with texture fallbacks, wrap mode changes and to handle the use of highlevel CoglTextures that need to be resolved into lowlevel textures before flushing the material state. The flush options structure has been expanding in size and the structure is logged with every journal entry so it is not an approach that scales well at all. It also makes flushing material state that much more complex. Weak Materials: Again for renderlists we need a way to create materials derived from other materials but without the strict requirement that modifications to the original material wont affect the derived ("weak") material. The only requirement is that its possible to later check if the original material has been changed. A summary of the new design: A CoglMaterial now basically represents a diff against its parent. Each material has a single parent and a mask of state that it changes. Each group of state (such as the blending state) has an "authority" which is found by walking up from a given material through its ancestors checking the difference mask until a match for that group is found. There is only one root node to the graph of all materials, which is the default material first created when Cogl is being initialized. All the groups of state are divided into two types, such that infrequently changed state belongs in a separate "BigState" structure that is only allocated and attached to a material when necessary. CoglMaterialLayers are another sparse structure. Like CoglMaterials they represent a diff against their parent and all the layers are part of another graph with the "default_layer_0" layer being the root node that Cogl creates during initialization. Copying a material is now basically just a case of slice allocating a CoglMaterial, setting the parent to be the source being copied and zeroing the mask of changes. Flush overrides should now be handled by simply relying on the cheapness of copying a material and making changes to it. (This will be done in a follow on commit) Weak material support will be added in a follow on commit.
2010-04-08 07:21:04 -04:00
TRUE))
return TRUE;
else
return FALSE;
}
/* Since the stride may not reflect the number of texture layers in use
* (due to padding) we deal with texture coordinate offsets separately
* from vertex and color offsets... */
static void
_cogl_journal_flush_texcoord_vbo_offsets_and_entries (
CoglJournalEntry *batch_start,
cogl: improves header and coding style consistency We've had complaints that our Cogl code/headers are a bit "special" so this is a first pass at tidying things up by giving them some consistency. These changes are all consistent with how new code in Cogl is being written, but the style isn't consistently applied across all code yet. There are two parts to this patch; but since each one required a large amount of effort to maintain tidy indenting it made sense to combine the changes to reduce the time spent re indenting the same lines. The first change is to use a consistent style for declaring function prototypes in headers. Cogl headers now consistently use this style for prototypes: return_type cogl_function_name (CoglType arg0, CoglType arg1); Not everyone likes this style, but it seems that most of the currently active Cogl developers agree on it. The second change is to constrain the use of redundant glib data types in Cogl. Uses of gint, guint, gfloat, glong, gulong and gchar have all been replaced with int, unsigned int, float, long, unsigned long and char respectively. When talking about pixel data; use of guchar has been replaced with guint8, otherwise unsigned char can be used. The glib types that we continue to use for portability are gboolean, gint{8,16,32,64}, guint{8,16,32,64} and gsize. The general intention is that Cogl should look palatable to the widest range of C programmers including those outside the Gnome community so - especially for the public API - we want to minimize the number of foreign looking typedefs.
2010-02-09 20:57:32 -05:00
int batch_len,
void *data)
{
CoglJournalFlushState *state = data;
int i;
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
COGL_STATIC_TIMER (time_flush_texcoord_pipeline_entries,
"flush: vbo+texcoords+pipeline+entries", /* parent */
"flush: texcoords+pipeline+entries",
"The time spent flushing texcoord offsets + pipeline "
"+ entries",
0 /* no application private data */);
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
COGL_TIMER_START (_cogl_uprof_context, time_flush_texcoord_pipeline_entries);
/* NB: attributes 0 and 1 are position and color */
for (i = 2; i < state->attributes->len; i++)
cogl_object_unref (g_array_index (state->attributes,
CoglVertexAttribute *, i));
g_array_set_size (state->attributes, batch_start->n_layers + 2);
for (i = 0; i < batch_start->n_layers; i++)
{
CoglVertexAttribute **attribute_entry =
&g_array_index (state->attributes, CoglVertexAttribute *, i + 2);
const char *names[] = {
"cogl_tex_coord0_in",
"cogl_tex_coord1_in",
"cogl_tex_coord2_in",
"cogl_tex_coord3_in",
"cogl_tex_coord4_in",
"cogl_tex_coord5_in",
"cogl_tex_coord6_in",
"cogl_tex_coord7_in"
};
char *name;
/* XXX NB:
* Our journal's vertex data is arranged as follows:
* 4 vertices per quad:
* 2 or 3 floats per position (3 when doing software transforms)
* 4 RGBA bytes,
* 2 floats per tex coord * n_layers
* (though n_layers may be padded; see definition of
* GET_JOURNAL_VB_STRIDE_FOR_N_LAYERS for details)
*/
name = i < 8 ? (char *)names[i] :
g_strdup_printf ("cogl_tex_coord%d_in", i);
/* XXX: it may be worth having some form of static initializer for
* attributes... */
*attribute_entry =
cogl_vertex_attribute_new (state->vertex_array,
name,
state->stride,
state->array_offset +
(POS_STRIDE + COLOR_STRIDE) * 4 +
TEX_STRIDE * 4 * i,
2,
COGL_VERTEX_ATTRIBUTE_TYPE_FLOAT);
if (i >= 8)
g_free (name);
}
batch_and_call (batch_start,
batch_len,
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
compare_entry_pipelines,
_cogl_journal_flush_pipeline_and_entries,
data);
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
COGL_TIMER_STOP (_cogl_uprof_context, time_flush_texcoord_pipeline_entries);
}
static gboolean
compare_entry_n_layers (CoglJournalEntry *entry0, CoglJournalEntry *entry1)
{
if (entry0->n_layers == entry1->n_layers)
return TRUE;
else
return FALSE;
}
/* At this point we know the stride has changed from the previous batch
* of journal entries */
static void
_cogl_journal_flush_vbo_offsets_and_entries (CoglJournalEntry *batch_start,
cogl: improves header and coding style consistency We've had complaints that our Cogl code/headers are a bit "special" so this is a first pass at tidying things up by giving them some consistency. These changes are all consistent with how new code in Cogl is being written, but the style isn't consistently applied across all code yet. There are two parts to this patch; but since each one required a large amount of effort to maintain tidy indenting it made sense to combine the changes to reduce the time spent re indenting the same lines. The first change is to use a consistent style for declaring function prototypes in headers. Cogl headers now consistently use this style for prototypes: return_type cogl_function_name (CoglType arg0, CoglType arg1); Not everyone likes this style, but it seems that most of the currently active Cogl developers agree on it. The second change is to constrain the use of redundant glib data types in Cogl. Uses of gint, guint, gfloat, glong, gulong and gchar have all been replaced with int, unsigned int, float, long, unsigned long and char respectively. When talking about pixel data; use of guchar has been replaced with guint8, otherwise unsigned char can be used. The glib types that we continue to use for portability are gboolean, gint{8,16,32,64}, guint{8,16,32,64} and gsize. The general intention is that Cogl should look palatable to the widest range of C programmers including those outside the Gnome community so - especially for the public API - we want to minimize the number of foreign looking typedefs.
2010-02-09 20:57:32 -05:00
int batch_len,
void *data)
{
CoglJournalFlushState *state = data;
gsize stride;
int i;
CoglVertexAttribute **attribute_entry;
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
COGL_STATIC_TIMER (time_flush_vbo_texcoord_pipeline_entries,
"flush: clip+vbo+texcoords+pipeline+entries", /* parent */
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
"flush: vbo+texcoords+pipeline+entries",
"The time spent flushing vbo + texcoord offsets + "
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
"pipeline + entries",
0 /* no application private data */);
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
COGL_TIMER_START (_cogl_uprof_context,
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
time_flush_vbo_texcoord_pipeline_entries);
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_BATCHING))
g_print ("BATCHING: vbo offset batch len = %d\n", batch_len);
/* XXX NB:
* Our journal's vertex data is arranged as follows:
* 4 vertices per quad:
* 2 or 3 GLfloats per position (3 when doing software transforms)
* 4 RGBA GLubytes,
* 2 GLfloats per tex coord * n_layers
* (though n_layers may be padded; see definition of
* GET_JOURNAL_VB_STRIDE_FOR_N_LAYERS for details)
*/
stride = GET_JOURNAL_VB_STRIDE_FOR_N_LAYERS (batch_start->n_layers);
stride *= sizeof (float);
state->stride = stride;
for (i = 0; i < state->attributes->len; i++)
cogl_object_unref (g_array_index (state->attributes,
CoglVertexAttribute *, i));
g_array_set_size (state->attributes, 2);
attribute_entry =
&g_array_index (state->attributes, CoglVertexAttribute *, 0);
*attribute_entry =
cogl_vertex_attribute_new (state->vertex_array,
"cogl_position_in",
stride,
state->array_offset,
N_POS_COMPONENTS,
COGL_VERTEX_ATTRIBUTE_TYPE_FLOAT);
attribute_entry =
&g_array_index (state->attributes, CoglVertexAttribute *, 1);
*attribute_entry =
cogl_vertex_attribute_new (state->vertex_array,
"cogl_color_in",
stride,
state->array_offset + (POS_STRIDE * 4),
4,
COGL_VERTEX_ATTRIBUTE_TYPE_UNSIGNED_BYTE);
#ifndef HAVE_COGL_GL
state->indices = cogl_get_rectangle_indices (batch_len);
#endif
/* We only create new VertexAttributes when the stride within the
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
* VertexArray changes. (due to a change in the number of pipeline layers)
* While the stride remains constant we walk forward through the above
* VertexArray using a vertex offset passed to cogl_draw_vertex_attributes
*/
state->current_vertex = 0;
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_JOURNAL))
{
guint8 *verts;
if (cogl_get_features () & COGL_FEATURE_VBOS)
verts = ((guint8 *)ctx->logged_vertices->data) +
(size_t)state->array_offset;
else
verts = (guint8 *)state->array_offset;
_cogl_journal_dump_quad_batch (verts,
batch_start->n_layers,
batch_len);
}
batch_and_call (batch_start,
batch_len,
compare_entry_n_layers,
_cogl_journal_flush_texcoord_vbo_offsets_and_entries,
data);
/* progress forward through the VBO containing all our vertices */
state->array_offset += (stride * 4 * batch_len);
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_JOURNAL))
g_print ("new vbo offset = %lu\n", (unsigned long)state->array_offset);
COGL_TIMER_STOP (_cogl_uprof_context,
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
time_flush_vbo_texcoord_pipeline_entries);
}
static gboolean
compare_entry_strides (CoglJournalEntry *entry0, CoglJournalEntry *entry1)
{
/* Currently the only thing that affects the stride for our vertex arrays
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
* is the number of pipeline layers. We need to update our VBO offsets
* whenever the stride changes. */
/* TODO: We should be padding the n_layers == 1 case as if it were
* n_layers == 2 so we can reduce the need to split batches. */
if (entry0->n_layers == entry1->n_layers ||
(entry0->n_layers <= MIN_LAYER_PADING &&
entry1->n_layers <= MIN_LAYER_PADING))
return TRUE;
else
return FALSE;
}
/* At this point we know the batch has a unique clip stack */
static void
_cogl_journal_flush_clip_stacks_and_entries (CoglJournalEntry *batch_start,
int batch_len,
void *data)
{
CoglJournalFlushState *state = data;
COGL_STATIC_TIMER (time_flush_clip_stack_pipeline_entries,
"Journal Flush", /* parent */
"flush: clip+vbo+texcoords+pipeline+entries",
"The time spent flushing clip + vbo + texcoord offsets + "
"pipeline + entries",
0 /* no application private data */);
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
COGL_TIMER_START (_cogl_uprof_context,
time_flush_clip_stack_pipeline_entries);
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_BATCHING))
g_print ("BATCHING: clip stack batch len = %d\n", batch_len);
_cogl_clip_stack_flush (batch_start->clip_stack);
_cogl_matrix_stack_push (state->modelview_stack);
/* If we have transformed all our quads at log time then we ensure
* no further model transform is applied by loading the identity
* matrix here. We need to do this after flushing the clip stack
* because the clip stack flushing code can modify the matrix */
if (G_LIKELY (!(cogl_debug_flags & COGL_DEBUG_DISABLE_SOFTWARE_TRANSFORM)))
{
_cogl_matrix_stack_load_identity (state->modelview_stack);
_cogl_matrix_stack_flush_to_gl (state->modelview_stack,
COGL_MATRIX_MODELVIEW);
}
/* Setting up the clip state can sometimes also flush the projection
matrix so we should flush it again. This will be a no-op if the
clip code didn't modify the projection */
_cogl_matrix_stack_flush_to_gl (state->projection_stack,
COGL_MATRIX_PROJECTION);
batch_and_call (batch_start,
batch_len,
compare_entry_strides,
_cogl_journal_flush_vbo_offsets_and_entries, /* callback */
data);
_cogl_matrix_stack_pop (state->modelview_stack);
COGL_TIMER_STOP (_cogl_uprof_context,
time_flush_clip_stack_pipeline_entries);
}
static gboolean
compare_entry_clip_stacks (CoglJournalEntry *entry0, CoglJournalEntry *entry1)
{
return entry0->clip_stack == entry1->clip_stack;
}
static CoglVertexArray *
upload_vertices (GArray *vertices, CoglJournalFlushState *state)
{
cogl: improves header and coding style consistency We've had complaints that our Cogl code/headers are a bit "special" so this is a first pass at tidying things up by giving them some consistency. These changes are all consistent with how new code in Cogl is being written, but the style isn't consistently applied across all code yet. There are two parts to this patch; but since each one required a large amount of effort to maintain tidy indenting it made sense to combine the changes to reduce the time spent re indenting the same lines. The first change is to use a consistent style for declaring function prototypes in headers. Cogl headers now consistently use this style for prototypes: return_type cogl_function_name (CoglType arg0, CoglType arg1); Not everyone likes this style, but it seems that most of the currently active Cogl developers agree on it. The second change is to constrain the use of redundant glib data types in Cogl. Uses of gint, guint, gfloat, glong, gulong and gchar have all been replaced with int, unsigned int, float, long, unsigned long and char respectively. When talking about pixel data; use of guchar has been replaced with guint8, otherwise unsigned char can be used. The glib types that we continue to use for portability are gboolean, gint{8,16,32,64}, guint{8,16,32,64} and gsize. The general intention is that Cogl should look palatable to the widest range of C programmers including those outside the Gnome community so - especially for the public API - we want to minimize the number of foreign looking typedefs.
2010-02-09 20:57:32 -05:00
gsize needed_vbo_len;
CoglVertexArray *array;
CoglBuffer *buffer;
_COGL_GET_CONTEXT (ctx, 0);
needed_vbo_len = vertices->len * sizeof (float);
g_assert (needed_vbo_len);
array = cogl_vertex_array_new (needed_vbo_len, NULL);
buffer = COGL_BUFFER (array);
cogl_buffer_set_update_hint (buffer, COGL_BUFFER_UPDATE_HINT_STATIC);
cogl_buffer_set_data (buffer, 0, vertices->data, needed_vbo_len);
/* As we flush the journal entries in batches we walk forward through the
* above VBO starting at offset 0... */
state->array_offset = 0;
return array;
}
/* XXX NB: When _cogl_journal_flush() returns all state relating
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
* to pipelines, all glEnable flags and current matrix state
* is undefined.
*/
void
_cogl_journal_flush (void)
{
CoglJournalFlushState state;
int i;
CoglFramebuffer *framebuffer;
[draw-buffers] First pass at overhauling Cogl's framebuffer management Cogl's support for offscreen rendering was originally written just to support the clutter_texture_new_from_actor API and due to lack of documentation and several confusing - non orthogonal - side effects of using the API it wasn't really possible to use directly. This commit does a number of things: - It removes {gl,gles}/cogl-fbo.{c,h} and adds shared cogl-draw-buffer.{c,h} files instead which should be easier to maintain. - internally CoglFbo objects are now called CoglDrawBuffers. A CoglDrawBuffer is an abstract base class that is inherited from to implement CoglOnscreen and CoglOffscreen draw buffers. CoglOffscreen draw buffers will initially be used to support the cogl_offscreen_new_to_texture API, and CoglOnscreen draw buffers will start to be used internally to represent windows as we aim to migrate some of Clutter's backend code to Cogl. - It makes draw buffer objects the owners of the following state: - viewport - projection matrix stack - modelview matrix stack - clip state (This means when you switch between draw buffers you will automatically be switching to their associated viewport, matrix and clip state) Aside from hopefully making cogl_offscreen_new_to_texture be more useful short term by having simpler and well defined semantics for cogl_set_draw_buffer, as mentioned above this is the first step for a couple of other things: - Its a step toward moving ownership for windows down from Clutter backends into Cogl, by (internally at least) introducing the CoglOnscreen draw buffer. Note: the plan is that cogl_set_draw_buffer will accept on or offscreen draw buffer handles, and the "target" argument will become redundant since we will instead query the type of the given draw buffer handle. - Because we have a common type for on and offscreen framebuffers we can provide a unified API for framebuffer management. Things like: - blitting between buffers - managing ancillary buffers (e.g. attaching depth and stencil buffers) - size requisition - clearing
2009-09-25 09:34:34 -04:00
CoglMatrixStack *modelview_stack;
COGL_STATIC_TIMER (flush_timer,
"Mainloop", /* parent */
"Journal Flush",
"The time spent flushing the Cogl journal",
0 /* no application private data */);
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
if (ctx->journal->len == 0)
return;
COGL_TIMER_START (_cogl_uprof_context, flush_timer);
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_BATCHING))
g_print ("BATCHING: journal len = %d\n", ctx->journal->len);
state.vertex_array = upload_vertices (ctx->logged_vertices, &state);
state.attributes = ctx->journal_flush_attributes_array;
framebuffer = _cogl_get_framebuffer ();
modelview_stack = _cogl_framebuffer_get_modelview_stack (framebuffer);
[draw-buffers] First pass at overhauling Cogl's framebuffer management Cogl's support for offscreen rendering was originally written just to support the clutter_texture_new_from_actor API and due to lack of documentation and several confusing - non orthogonal - side effects of using the API it wasn't really possible to use directly. This commit does a number of things: - It removes {gl,gles}/cogl-fbo.{c,h} and adds shared cogl-draw-buffer.{c,h} files instead which should be easier to maintain. - internally CoglFbo objects are now called CoglDrawBuffers. A CoglDrawBuffer is an abstract base class that is inherited from to implement CoglOnscreen and CoglOffscreen draw buffers. CoglOffscreen draw buffers will initially be used to support the cogl_offscreen_new_to_texture API, and CoglOnscreen draw buffers will start to be used internally to represent windows as we aim to migrate some of Clutter's backend code to Cogl. - It makes draw buffer objects the owners of the following state: - viewport - projection matrix stack - modelview matrix stack - clip state (This means when you switch between draw buffers you will automatically be switching to their associated viewport, matrix and clip state) Aside from hopefully making cogl_offscreen_new_to_texture be more useful short term by having simpler and well defined semantics for cogl_set_draw_buffer, as mentioned above this is the first step for a couple of other things: - Its a step toward moving ownership for windows down from Clutter backends into Cogl, by (internally at least) introducing the CoglOnscreen draw buffer. Note: the plan is that cogl_set_draw_buffer will accept on or offscreen draw buffer handles, and the "target" argument will become redundant since we will instead query the type of the given draw buffer handle. - Because we have a common type for on and offscreen framebuffers we can provide a unified API for framebuffer management. Things like: - blitting between buffers - managing ancillary buffers (e.g. attaching depth and stencil buffers) - size requisition - clearing
2009-09-25 09:34:34 -04:00
state.modelview_stack = modelview_stack;
state.projection_stack = _cogl_framebuffer_get_projection_stack (framebuffer);
/* batch_and_call() batches a list of journal entries according to some
* given criteria and calls a callback once for each determined batch.
*
* The process of flushing the journal is staggered to reduce the amount
* of driver/GPU state changes necessary:
* 1) We split the entries according to the clip state.
* 2) We split the entries according to the stride of the vertices:
* Each time the stride of our vertex data changes we need to call
* gl{Vertex,Color}Pointer to inform GL of new VBO offsets.
* Currently the only thing that affects the stride of our vertex data
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
* is the number of pipeline layers.
* 3) We split the entries explicitly by the number of pipeline layers:
* We pad our vertex data when the number of layers is < 2 so that we
* can minimize changes in stride. Each time the number of layers
* changes we need to call glTexCoordPointer to inform GL of new VBO
* offsets.
* 4) We then split according to compatible Cogl pipelines:
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
* This is where we flush pipeline state
* 5) Finally we split according to modelview matrix changes:
* This is when we finally tell GL to draw something.
* Note: Splitting by modelview changes is skipped when are doing the
* vertex transformation in software at log time.
*/
batch_and_call ((CoglJournalEntry *)ctx->journal->data, /* first entry */
ctx->journal->len, /* max number of entries to consider */
compare_entry_clip_stacks,
_cogl_journal_flush_clip_stacks_and_entries, /* callback */
&state); /* data */
for (i = 0; i < state.attributes->len; i++)
cogl_object_unref (g_array_index (state.attributes,
CoglVertexAttribute *, i));
g_array_set_size (state.attributes, 0);
cogl_object_unref (state.vertex_array);
for (i = 0; i < ctx->journal->len; i++)
{
CoglJournalEntry *entry =
&g_array_index (ctx->journal, CoglJournalEntry, i);
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
_cogl_pipeline_journal_unref (entry->pipeline);
_cogl_clip_stack_unref (entry->clip_stack);
}
g_array_set_size (ctx->journal, 0);
g_array_set_size (ctx->logged_vertices, 0);
COGL_TIMER_STOP (_cogl_uprof_context, flush_timer);
}
[draw-buffers] First pass at overhauling Cogl's framebuffer management Cogl's support for offscreen rendering was originally written just to support the clutter_texture_new_from_actor API and due to lack of documentation and several confusing - non orthogonal - side effects of using the API it wasn't really possible to use directly. This commit does a number of things: - It removes {gl,gles}/cogl-fbo.{c,h} and adds shared cogl-draw-buffer.{c,h} files instead which should be easier to maintain. - internally CoglFbo objects are now called CoglDrawBuffers. A CoglDrawBuffer is an abstract base class that is inherited from to implement CoglOnscreen and CoglOffscreen draw buffers. CoglOffscreen draw buffers will initially be used to support the cogl_offscreen_new_to_texture API, and CoglOnscreen draw buffers will start to be used internally to represent windows as we aim to migrate some of Clutter's backend code to Cogl. - It makes draw buffer objects the owners of the following state: - viewport - projection matrix stack - modelview matrix stack - clip state (This means when you switch between draw buffers you will automatically be switching to their associated viewport, matrix and clip state) Aside from hopefully making cogl_offscreen_new_to_texture be more useful short term by having simpler and well defined semantics for cogl_set_draw_buffer, as mentioned above this is the first step for a couple of other things: - Its a step toward moving ownership for windows down from Clutter backends into Cogl, by (internally at least) introducing the CoglOnscreen draw buffer. Note: the plan is that cogl_set_draw_buffer will accept on or offscreen draw buffer handles, and the "target" argument will become redundant since we will instead query the type of the given draw buffer handle. - Because we have a common type for on and offscreen framebuffers we can provide a unified API for framebuffer management. Things like: - blitting between buffers - managing ancillary buffers (e.g. attaching depth and stencil buffers) - size requisition - clearing
2009-09-25 09:34:34 -04:00
static void
_cogl_journal_init (void)
{
/* Here we flush anything that we know must remain constant until the
* next the the journal is flushed. Note: This lets up flush things
* that themselves depend on the journal, such as clip state. */
/* NB: the journal deals with flushing the modelview stack and clip
state manually */
_cogl_framebuffer_flush_state (_cogl_get_framebuffer (),
COGL_FRAMEBUFFER_FLUSH_SKIP_MODELVIEW |
COGL_FRAMEBUFFER_FLUSH_SKIP_CLIP_STATE);
[draw-buffers] First pass at overhauling Cogl's framebuffer management Cogl's support for offscreen rendering was originally written just to support the clutter_texture_new_from_actor API and due to lack of documentation and several confusing - non orthogonal - side effects of using the API it wasn't really possible to use directly. This commit does a number of things: - It removes {gl,gles}/cogl-fbo.{c,h} and adds shared cogl-draw-buffer.{c,h} files instead which should be easier to maintain. - internally CoglFbo objects are now called CoglDrawBuffers. A CoglDrawBuffer is an abstract base class that is inherited from to implement CoglOnscreen and CoglOffscreen draw buffers. CoglOffscreen draw buffers will initially be used to support the cogl_offscreen_new_to_texture API, and CoglOnscreen draw buffers will start to be used internally to represent windows as we aim to migrate some of Clutter's backend code to Cogl. - It makes draw buffer objects the owners of the following state: - viewport - projection matrix stack - modelview matrix stack - clip state (This means when you switch between draw buffers you will automatically be switching to their associated viewport, matrix and clip state) Aside from hopefully making cogl_offscreen_new_to_texture be more useful short term by having simpler and well defined semantics for cogl_set_draw_buffer, as mentioned above this is the first step for a couple of other things: - Its a step toward moving ownership for windows down from Clutter backends into Cogl, by (internally at least) introducing the CoglOnscreen draw buffer. Note: the plan is that cogl_set_draw_buffer will accept on or offscreen draw buffer handles, and the "target" argument will become redundant since we will instead query the type of the given draw buffer handle. - Because we have a common type for on and offscreen framebuffers we can provide a unified API for framebuffer management. Things like: - blitting between buffers - managing ancillary buffers (e.g. attaching depth and stencil buffers) - size requisition - clearing
2009-09-25 09:34:34 -04:00
}
void
_cogl_journal_log_quad (const float *position,
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
CoglPipeline *pipeline,
int n_layers,
CoglHandle layer0_override_texture,
const float *tex_coords,
unsigned int tex_coords_len)
{
cogl: improves header and coding style consistency We've had complaints that our Cogl code/headers are a bit "special" so this is a first pass at tidying things up by giving them some consistency. These changes are all consistent with how new code in Cogl is being written, but the style isn't consistently applied across all code yet. There are two parts to this patch; but since each one required a large amount of effort to maintain tidy indenting it made sense to combine the changes to reduce the time spent re indenting the same lines. The first change is to use a consistent style for declaring function prototypes in headers. Cogl headers now consistently use this style for prototypes: return_type cogl_function_name (CoglType arg0, CoglType arg1); Not everyone likes this style, but it seems that most of the currently active Cogl developers agree on it. The second change is to constrain the use of redundant glib data types in Cogl. Uses of gint, guint, gfloat, glong, gulong and gchar have all been replaced with int, unsigned int, float, long, unsigned long and char respectively. When talking about pixel data; use of guchar has been replaced with guint8, otherwise unsigned char can be used. The glib types that we continue to use for portability are gboolean, gint{8,16,32,64}, guint{8,16,32,64} and gsize. The general intention is that Cogl should look palatable to the widest range of C programmers including those outside the Gnome community so - especially for the public API - we want to minimize the number of foreign looking typedefs.
2010-02-09 20:57:32 -05:00
gsize stride;
gsize byte_stride;
int next_vert;
GLfloat *v;
GLubyte *c;
GLubyte *src_c;
int i;
int next_entry;
guint32 disable_layers;
CoglJournalEntry *entry;
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
CoglPipeline *source;
CoglPipelineFlushOptions flush_options;
COGL_STATIC_TIMER (log_timer,
"Mainloop", /* parent */
"Journal Log",
"The time spent logging in the Cogl journal",
0 /* no application private data */);
_COGL_GET_CONTEXT (ctx, NO_RETVAL);
COGL_TIMER_START (_cogl_uprof_context, log_timer);
[draw-buffers] First pass at overhauling Cogl's framebuffer management Cogl's support for offscreen rendering was originally written just to support the clutter_texture_new_from_actor API and due to lack of documentation and several confusing - non orthogonal - side effects of using the API it wasn't really possible to use directly. This commit does a number of things: - It removes {gl,gles}/cogl-fbo.{c,h} and adds shared cogl-draw-buffer.{c,h} files instead which should be easier to maintain. - internally CoglFbo objects are now called CoglDrawBuffers. A CoglDrawBuffer is an abstract base class that is inherited from to implement CoglOnscreen and CoglOffscreen draw buffers. CoglOffscreen draw buffers will initially be used to support the cogl_offscreen_new_to_texture API, and CoglOnscreen draw buffers will start to be used internally to represent windows as we aim to migrate some of Clutter's backend code to Cogl. - It makes draw buffer objects the owners of the following state: - viewport - projection matrix stack - modelview matrix stack - clip state (This means when you switch between draw buffers you will automatically be switching to their associated viewport, matrix and clip state) Aside from hopefully making cogl_offscreen_new_to_texture be more useful short term by having simpler and well defined semantics for cogl_set_draw_buffer, as mentioned above this is the first step for a couple of other things: - Its a step toward moving ownership for windows down from Clutter backends into Cogl, by (internally at least) introducing the CoglOnscreen draw buffer. Note: the plan is that cogl_set_draw_buffer will accept on or offscreen draw buffer handles, and the "target" argument will become redundant since we will instead query the type of the given draw buffer handle. - Because we have a common type for on and offscreen framebuffers we can provide a unified API for framebuffer management. Things like: - blitting between buffers - managing ancillary buffers (e.g. attaching depth and stencil buffers) - size requisition - clearing
2009-09-25 09:34:34 -04:00
if (ctx->logged_vertices->len == 0)
_cogl_journal_init ();
/* The vertex data is logged into a separate array in a layout that can be
* directly passed to OpenGL
*/
/* XXX: See definition of GET_JOURNAL_VB_STRIDE_FOR_N_LAYERS for details
* about how we pack our vertex data */
stride = GET_JOURNAL_VB_STRIDE_FOR_N_LAYERS (n_layers);
/* NB: stride is in 32bit words */
byte_stride = stride * 4;
next_vert = ctx->logged_vertices->len;
g_array_set_size (ctx->logged_vertices, next_vert + 4 * stride);
v = &g_array_index (ctx->logged_vertices, GLfloat, next_vert);
c = (GLubyte *)(v + POS_STRIDE);
/* XXX: All the jumping around to fill in this strided buffer doesn't
* seem ideal. */
/* XXX: we could defer expanding the vertex data for GL until we come
* to flushing the journal. */
/* FIXME: This is a hacky optimization, since it will break if we
* change the definition of CoglColor: */
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
_cogl_pipeline_get_colorubv (pipeline, c);
src_c = c;
for (i = 0; i < 3; i++)
{
c += byte_stride;
memcpy (c, src_c, 4);
}
#define X0 0
#define Y0 1
#define X1 2
#define Y1 3
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_DISABLE_SOFTWARE_TRANSFORM))
{
v[0] = position[X0]; v[1] = position[Y0];
v += stride;
v[0] = position[X0]; v[1] = position[Y1];
v += stride;
v[0] = position[X1]; v[1] = position[Y1];
v += stride;
v[0] = position[X1]; v[1] = position[Y0];
}
else
{
CoglMatrix mv;
float x, y, z, w;
cogl_get_modelview_matrix (&mv);
x = position[X0], y = position[Y0], z = 0; w = 1;
cogl_matrix_transform_point (&mv, &x, &y, &z, &w);
v[0] = x; v[1] = y; v[2] = z;
v += stride;
x = position[X0], y = position[Y1], z = 0; w = 1;
cogl_matrix_transform_point (&mv, &x, &y, &z, &w);
v[0] = x; v[1] = y; v[2] = z;
v += stride;
x = position[X1], y = position[Y1], z = 0; w = 1;
cogl_matrix_transform_point (&mv, &x, &y, &z, &w);
v[0] = x; v[1] = y; v[2] = z;
v += stride;
x = position[X1], y = position[Y0], z = 0; w = 1;
cogl_matrix_transform_point (&mv, &x, &y, &z, &w);
v[0] = x; v[1] = y; v[2] = z;
}
#undef X0
#undef Y0
#undef X1
#undef Y1
for (i = 0; i < n_layers; i++)
{
/* XXX: See definition of GET_JOURNAL_VB_STRIDE_FOR_N_LAYERS for details
* about how we pack our vertex data */
GLfloat *t = &g_array_index (ctx->logged_vertices, GLfloat,
next_vert + POS_STRIDE +
COLOR_STRIDE + TEX_STRIDE * i);
t[0] = tex_coords[i * 4 + 0]; t[1] = tex_coords[i * 4 + 1];
t += stride;
t[0] = tex_coords[i * 4 + 0]; t[1] = tex_coords[i * 4 + 3];
t += stride;
t[0] = tex_coords[i * 4 + 2]; t[1] = tex_coords[i * 4 + 3];
t += stride;
t[0] = tex_coords[i * 4 + 2]; t[1] = tex_coords[i * 4 + 1];
}
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_JOURNAL))
{
g_print ("Logged new quad:\n");
v = &g_array_index (ctx->logged_vertices, GLfloat, next_vert);
_cogl_journal_dump_quad_vertices ((guint8 *)v, n_layers);
}
next_entry = ctx->journal->len;
g_array_set_size (ctx->journal, next_entry + 1);
entry = &g_array_index (ctx->journal, CoglJournalEntry, next_entry);
entry->n_layers = n_layers;
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
source = pipeline;
if (G_UNLIKELY (ctx->legacy_state_set))
{
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
source = cogl_pipeline_copy (pipeline);
_cogl_pipeline_apply_legacy_state (source);
}
flush_options.flags = 0;
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
if (G_UNLIKELY (cogl_pipeline_get_n_layers (pipeline) != n_layers))
{
disable_layers = (1 << n_layers) - 1;
disable_layers = ~disable_layers;
flush_options.disable_layers = disable_layers;
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
flush_options.flags |= COGL_PIPELINE_FLUSH_DISABLE_MASK;
}
if (G_UNLIKELY (layer0_override_texture))
{
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
flush_options.flags |= COGL_PIPELINE_FLUSH_LAYER0_OVERRIDE;
flush_options.layer0_override_texture = layer0_override_texture;
}
if (G_UNLIKELY (flush_options.flags))
{
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
/* If we haven't already created a derived pipeline... */
if (source == pipeline)
source = cogl_pipeline_copy (pipeline);
_cogl_pipeline_apply_overrides (source, &flush_options);
}
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
entry->pipeline = _cogl_pipeline_journal_ref (source);
entry->clip_stack = _cogl_clip_stack_ref (_cogl_get_clip_stack ());
cogl: rename CoglMaterial -> CoglPipeline This applies an API naming change that's been deliberated over for a while now which is to rename CoglMaterial to CoglPipeline. For now the new pipeline API is marked as experimental and public headers continue to talk about materials not pipelines. The CoglMaterial API is now maintained in terms of the cogl_pipeline API internally. Currently this API is targeting Cogl 2.0 so we will have time to integrate it properly with other upcoming Cogl 2.0 work. The basic reasons for the rename are: - That the term "material" implies to many people that they are constrained to fragment processing; perhaps as some kind of high-level texture abstraction. - In Clutter they get exposed by ClutterTexture actors which may be re-inforcing this misconception. - When comparing how other frameworks use the term material, a material sometimes describes a multi-pass fragment processing technique which isn't the case in Cogl. - In code, "CoglPipeline" will hopefully be a much more self documenting summary of what these objects represent; a full GPU pipeline configuration including, for example, vertex processing, fragment processing and blending. - When considering the API documentation story, at some point we need a document introducing developers to how the "GPU pipeline" works so it should become intuitive that CoglPipeline maps back to that description of the GPU pipeline. - This is consistent in terminology and concept to OpenGL 4's new pipeline object which is a container for program objects. Note: The cogl-material.[ch] files have been renamed to cogl-material-compat.[ch] because otherwise git doesn't seem to treat the change as a moving the old cogl-material.c->cogl-pipeline.c and so we loose all our git-blame history.
2010-10-27 13:54:57 -04:00
if (G_UNLIKELY (source != pipeline))
cogl_handle_unref (source);
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_DISABLE_SOFTWARE_TRANSFORM))
cogl_get_modelview_matrix (&entry->model_view);
if (G_UNLIKELY (cogl_debug_flags & COGL_DEBUG_DISABLE_BATCHING))
_cogl_journal_flush ();
COGL_TIMER_STOP (_cogl_uprof_context, log_timer);
}