mutter/cogl/cogl-material.h

1174 lines
39 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/>.
*
*
*/
#if !defined(__COGL_H_INSIDE__) && !defined(CLUTTER_COMPILATION)
#error "Only <cogl/cogl.h> can be included directly."
#endif
#ifndef __COGL_MATERIAL_H__
#define __COGL_MATERIAL_H__
G_BEGIN_DECLS
#include <cogl/cogl-types.h>
#include <cogl/cogl-matrix.h>
/**
* SECTION:cogl-material
* @short_description: Fuctions for creating and manipulating materials
*
* COGL allows creating and manipulating materials used to fill in
* geometry. Materials may simply be lighting attributes (such as an
* ambient and diffuse colour) or might represent one or more textures
* blended together.
*/
typedef struct _CoglMaterial CoglMaterial;
typedef struct _CoglMaterialLayer CoglMaterialLayer;
#define COGL_MATERIAL(OBJECT) ((CoglMaterial *)OBJECT)
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
/**
* CoglMaterialFilter:
* @COGL_MATERIAL_FILTER_NEAREST: Measuring in manhatten distance from the,
* current pixel center, use the nearest texture texel
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
* @COGL_MATERIAL_FILTER_LINEAR: Use the weighted average of the 4 texels
* nearest the current pixel center
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
* @COGL_MATERIAL_FILTER_NEAREST_MIPMAP_NEAREST: Select the mimap level whose
* texel size most closely matches the current pixel, and use the
* %COGL_MATERIAL_FILTER_NEAREST criterion
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
* @COGL_MATERIAL_FILTER_LINEAR_MIPMAP_NEAREST: Select the mimap level whose
* texel size most closely matches the current pixel, and use the
* %COGL_MATERIAL_FILTER_LINEAR criterion
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
* @COGL_MATERIAL_FILTER_NEAREST_MIPMAP_LINEAR: Select the two mimap levels
* whose texel size most closely matches the current pixel, use
* the %COGL_MATERIAL_FILTER_NEAREST criterion on each one and take
* their weighted average
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
* @COGL_MATERIAL_FILTER_LINEAR_MIPMAP_LINEAR: Select the two mimap levels
* whose texel size most closely matches the current pixel, use
* the %COGL_MATERIAL_FILTER_LINEAR criterion on each one and take
* their weighted average
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
*
* Texture filtering is used whenever the current pixel maps either to more
* than one texture element (texel) or less than one. These filter enums
* correspond to different strategies used to come up with a pixel color, by
* possibly referring to multiple neighbouring texels and taking a weighted
* average or simply using the nearest texel.
*/
typedef enum {
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
COGL_MATERIAL_FILTER_NEAREST = GL_NEAREST,
COGL_MATERIAL_FILTER_LINEAR = GL_LINEAR,
COGL_MATERIAL_FILTER_NEAREST_MIPMAP_NEAREST = GL_NEAREST_MIPMAP_NEAREST,
COGL_MATERIAL_FILTER_LINEAR_MIPMAP_NEAREST = GL_LINEAR_MIPMAP_NEAREST,
COGL_MATERIAL_FILTER_NEAREST_MIPMAP_LINEAR = GL_NEAREST_MIPMAP_LINEAR,
COGL_MATERIAL_FILTER_LINEAR_MIPMAP_LINEAR = GL_LINEAR_MIPMAP_LINEAR
} CoglMaterialFilter;
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
/**
* CoglMaterialWrapMode:
* @COGL_MATERIAL_WRAP_MODE_REPEAT: The texture will be repeated. This
* is useful for example to draw a tiled background.
* @COGL_MATERIAL_WRAP_MODE_CLAMP_TO_EDGE: The coordinates outside the
* range 01 will sample copies of the edge pixels of the
* texture. This is useful to avoid artifacts if only one copy of
* the texture is being rendered.
* @COGL_MATERIAL_WRAP_MODE_AUTOMATIC: Cogl will try to automatically
* decide which of the above two to use. For cogl_rectangle(), it
* will use repeat mode if any of the texture coordinates are
* outside the range 01, otherwise it will use clamp to edge. For
* cogl_polygon() and cogl_vertex_buffer_draw() it will always use
* repeat mode. This is the default value.
*
* The wrap mode specifies what happens when texture coordinates
* outside the range 01 are used. Note that if the filter mode is
* anything but %COGL_MATERIAL_FILTER_NEAREST then texels outside the
* range 01 might be used even when the coordinate is exactly 0 or 1
* because OpenGL will try to sample neighbouring pixels. For example
* if you are trying to render the full texture then you may get
* artifacts around the edges when the pixels from the other side are
* merged in if the wrap mode is set to repeat.
*
* Since: 1.4
*/
/* GL_ALWAYS is just used here as a value that is known not to clash
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
* with any valid GL wrap modes
*
* XXX: keep the values in sync with the CoglMaterialWrapModeInternal
* enum so no conversion is actually needed.
*/
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
typedef enum {
COGL_MATERIAL_WRAP_MODE_REPEAT = GL_REPEAT,
COGL_MATERIAL_WRAP_MODE_CLAMP_TO_EDGE = GL_CLAMP_TO_EDGE,
COGL_MATERIAL_WRAP_MODE_AUTOMATIC = GL_ALWAYS
} CoglMaterialWrapMode;
/**
* cogl_material_new:
*
* Allocates and initializes a blank white material
*
* Return value: a pointer to a new #CoglMaterial
*/
CoglMaterial *
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
cogl_material_new (void);
/**
* cogl_material_copy:
* @source: a #CoglMaterial object to copy
*
* Creates a new material with the configuration copied from the
* source material.
*
* We would strongly advise developers to always aim to use
* cogl_material_copy() instead of cogl_material_new() whenever there will
* be any similarity between two materials. Copying a material helps Cogl
* keep track of a materials ancestry which we may use to help minimize GPU
* state changes.
*
* Returns: a pointer to the newly allocated #CoglMaterial
*
* Since: 1.2
*/
CoglMaterial *
cogl_material_copy (CoglMaterial *source);
#ifndef COGL_DISABLE_DEPRECATED
/**
* cogl_material_ref:
* @material: a #CoglMaterial object.
*
* Increment the reference count for a #CoglMaterial.
*
* Return value: the @material.
*
* Since: 1.0
*
* Deprecated: 1.2: Use cogl_object_ref() instead
*/
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
CoglHandle
cogl_material_ref (CoglHandle handle) G_GNUC_DEPRECATED;
/**
* cogl_material_unref:
* @material: a #CoglMaterial object.
*
* Decrement the reference count for a #CoglMaterial.
*
* Since: 1.0
*
* Deprecated: 1.2: Use cogl_object_unref() instead
*/
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
void
cogl_material_unref (CoglHandle handle) G_GNUC_DEPRECATED;
#endif /* COGL_DISABLE_DEPRECATED */
/**
* cogl_is_material:
* @handle: A CoglHandle
*
* Gets whether the given handle references an existing material object.
*
* Return value: %TRUE if the handle references a #CoglMaterial,
* %FALSE otherwise
*/
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
gboolean
cogl_is_material (CoglHandle handle);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_set_color:
* @material: A #CoglMaterial object
* @color: The components of the color
*
* Sets the basic color of the material, used when no lighting is enabled.
*
* Note that if you don't add any layers to the material then the color
* will be blended unmodified with the destination; the default blend
* expects premultiplied colors: for example, use (0.5, 0.0, 0.0, 0.5) for
* semi-transparent red. See cogl_color_premultiply().
*
* The default value is (1.0, 1.0, 1.0, 1.0)
*
* Since: 1.0
*/
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
void
cogl_material_set_color (CoglMaterial *material,
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
const CoglColor *color);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_set_color4ub:
* @material: A #CoglMaterial object
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
* @red: The red component
* @green: The green component
* @blue: The blue component
* @alpha: The alpha component
*
* Sets the basic color of the material, used when no lighting is enabled.
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* The default value is (0xff, 0xff, 0xff, 0xff)
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* Since: 1.0
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*/
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
void
cogl_material_set_color4ub (CoglMaterial *material,
guint8 red,
guint8 green,
guint8 blue,
guint8 alpha);
/**
* cogl_material_set_color4f:
* @material: A #CoglMaterial object
* @red: The red component
* @green: The green component
* @blue: The blue component
* @alpha: The alpha component
*
* Sets the basic color of the material, used when no lighting is enabled.
*
* The default value is (1.0, 1.0, 1.0, 1.0)
*
* Since: 1.0
*/
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
void
cogl_material_set_color4f (CoglMaterial *material,
float red,
float green,
float blue,
float alpha);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_get_color:
* @material: A #CoglMaterial object
* @color: (out): The location to store the color
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* Retrieves the current material color.
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* Since: 1.0
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*/
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
void
cogl_material_get_color (CoglMaterial *material,
CoglColor *color);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_set_ambient:
* @material: A #CoglMaterial object
* @ambient: The components of the desired ambient color
*
* Sets the material's ambient color, in the standard OpenGL lighting
* model. The ambient color affects the overall color of the object.
*
* Since the diffuse color will be intense when the light hits the surface
* directly, the ambient will be most apparent where the light hits at a
* slant.
*
* The default value is (0.2, 0.2, 0.2, 1.0)
*
* Since: 1.0
*/
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
void
cogl_material_set_ambient (CoglMaterial *material,
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
const CoglColor *ambient);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_get_ambient:
* @material: A #CoglMaterial object
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
* @ambient: The location to store the ambient color
*
* Retrieves the current ambient color for @material
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* Since: 1.0
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*/
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
void
cogl_material_get_ambient (CoglMaterial *material,
CoglColor *ambient);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_set_diffuse:
* @material: A #CoglMaterial object
* @diffuse: The components of the desired diffuse color
*
* Sets the material's diffuse color, in the standard OpenGL lighting
* model. The diffuse color is most intense where the light hits the
* surface directly - perpendicular to the surface.
*
* The default value is (0.8, 0.8, 0.8, 1.0)
*
* Since: 1.0
*/
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
void
cogl_material_set_diffuse (CoglMaterial *material,
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
const CoglColor *diffuse);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_get_diffuse:
* @material: A #CoglMaterial object
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
* @diffuse: The location to store the diffuse color
*
* Retrieves the current diffuse color for @material
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* Since: 1.0
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*/
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
void
cogl_material_get_diffuse (CoglMaterial *material,
CoglColor *diffuse);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_set_ambient_and_diffuse:
* @material: A #CoglMaterial object
* @color: The components of the desired ambient and diffuse colors
*
* Conveniently sets the diffuse and ambient color of @material at the same
* time. See cogl_material_set_ambient() and cogl_material_set_diffuse().
*
* The default ambient color is (0.2, 0.2, 0.2, 1.0)
*
* The default diffuse color is (0.8, 0.8, 0.8, 1.0)
*
* Since: 1.0
*/
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
void
cogl_material_set_ambient_and_diffuse (CoglMaterial *material,
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
const CoglColor *color);
/**
* cogl_material_set_specular:
* @material: A #CoglMaterial object
* @specular: The components of the desired specular color
*
* Sets the material's specular color, in the standard OpenGL lighting
* model. The intensity of the specular color depends on the viewport
* position, and is brightest along the lines of reflection.
*
* The default value is (0.0, 0.0, 0.0, 1.0)
*
* Since: 1.0
*/
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
void
cogl_material_set_specular (CoglMaterial *material,
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
const CoglColor *specular);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_get_specular:
* @material: A #CoglMaterial object
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
* @specular: The location to store the specular color
*
* Retrieves the materials current specular color.
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* Since: 1.0
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*/
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
void
cogl_material_get_specular (CoglMaterial *material,
CoglColor *specular);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_set_shininess:
* @material: A #CoglMaterial object
* @shininess: The desired shininess; range: [0.0, 1.0]
*
* Sets the materials shininess, in the standard OpenGL lighting model,
* which determines how specular highlights are calculated. A higher
* @shininess will produce smaller brigher highlights.
*
* The default value is 0.0
*
* Since: 1.0
*/
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
void
cogl_material_set_shininess (CoglMaterial *material,
float shininess);
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
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_get_shininess:
* @material: A #CoglMaterial object
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* Retrieves the materials current emission color.
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* Return value: The materials current shininess value
*
* Since: 1.0
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*/
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
float
cogl_material_get_shininess (CoglMaterial *material);
/**
* cogl_material_set_emission:
* @material: A #CoglMaterial object
* @emission: The components of the desired emissive color
*
* Sets the material's emissive color, in the standard OpenGL lighting
* model. It will look like the surface is a light source emitting this
* color.
*
* The default value is (0.0, 0.0, 0.0, 1.0)
*
* Since: 1.0
*/
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
void
cogl_material_set_emission (CoglMaterial *material,
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
const CoglColor *emission);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* cogl_material_get_emission:
* @material: A #CoglMaterial object
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
* @emission: The location to store the emission color
*
* Retrieves the materials current emission color.
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*
* Since: 1.0
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
*/
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
void
cogl_material_get_emission (CoglMaterial *material,
CoglColor *emission);
Fully integrates CoglMaterial throughout the rest of Cogl This glues CoglMaterial in as the fundamental way that Cogl describes how to fill in geometry. It adds cogl_set_source (), which is used to set the material which will be used by all subsequent drawing functions It adds cogl_set_source_texture as a convenience for setting up a default material with a single texture layer, and cogl_set_source_color is now also a convenience for setting up a material with a solid fill. "drawing functions" include, cogl_rectangle, cogl_texture_rectangle, cogl_texture_multiple_rectangles, cogl_texture_polygon (though the cogl_texture_* funcs have been renamed; see below for details), cogl_path_fill/stroke and cogl_vertex_buffer_draw*. cogl_texture_rectangle, cogl_texture_multiple_rectangles and cogl_texture_polygon no longer take a texture handle; instead the current source material is referenced. The functions have also been renamed to: cogl_rectangle_with_texture_coords, cogl_rectangles_with_texture_coords and cogl_polygon respectivly. Most code that previously did: cogl_texture_rectangle (tex_handle, x, y,...); needs to be changed to now do: cogl_set_source_texture (tex_handle); cogl_rectangle_with_texture_coords (x, y,....); In the less likely case where you were blending your source texture with a color like: cogl_set_source_color4ub (r,g,b,a); /* where r,g,b,a isn't just white */ cogl_texture_rectangle (tex_handle, x, y,...); you will need your own material to do that: mat = cogl_material_new (); cogl_material_set_color4ub (r,g,b,a); cogl_material_set_layer (mat, 0, tex_handle)); cogl_set_source_material (mat); Code that uses the texture coordinates, 0, 0, 1, 1 don't need to use cog_rectangle_with_texure_coords since these are the coordinates that cogl_rectangle will use. For cogl_texture_polygon; as well as dropping the texture handle, the n_vertices and vertices arguments were transposed for consistency. So code previously written as: cogl_texture_polygon (tex_handle, 3, verts, TRUE); need to be written as: cogl_set_source_texture (tex_handle); cogl_polygon (verts, 3, TRUE); All of the unit tests have been updated to now use the material API and test-cogl-material has been renamed to test-cogl-multitexture since any textured quad is now technically a test of CoglMaterial but this test specifically creates a material with multiple texture layers. Note: The GLES backend has not been updated yet; that will be done in a following commit.
2009-01-23 11:15:40 -05:00
/**
* CoglMaterialAlphaFunc:
* @COGL_MATERIAL_ALPHA_FUNC_NEVER: Never let the fragment through.
* @COGL_MATERIAL_ALPHA_FUNC_LESS: Let the fragment through if the incoming
* alpha value is less than the reference alpha value
* @COGL_MATERIAL_ALPHA_FUNC_EQUAL: Let the fragment through if the incoming
* alpha value equals the reference alpha value
* @COGL_MATERIAL_ALPHA_FUNC_LEQUAL: Let the fragment through if the incoming
* alpha value is less than or equal to the reference alpha value
* @COGL_MATERIAL_ALPHA_FUNC_GREATER: Let the fragment through if the incoming
* alpha value is greater than the reference alpha value
* @COGL_MATERIAL_ALPHA_FUNC_NOTEQUAL: Let the fragment through if the incoming
* alpha value does not equal the reference alpha value
* @COGL_MATERIAL_ALPHA_FUNC_GEQUAL: Let the fragment through if the incoming
* alpha value is greater than or equal to the reference alpha value.
* @COGL_MATERIAL_ALPHA_FUNC_ALWAYS: Always let the fragment through.
*
* Alpha testing happens before blending primitives with the framebuffer and
* gives an opportunity to discard fragments based on a comparison with the
* incoming alpha value and a reference alpha value. The #CoglMaterialAlphaFunc
* determines how the comparison is done.
*/
typedef enum {
COGL_MATERIAL_ALPHA_FUNC_NEVER = GL_NEVER,
COGL_MATERIAL_ALPHA_FUNC_LESS = GL_LESS,
COGL_MATERIAL_ALPHA_FUNC_EQUAL = GL_EQUAL,
COGL_MATERIAL_ALPHA_FUNC_LEQUAL = GL_LEQUAL,
COGL_MATERIAL_ALPHA_FUNC_GREATER = GL_GREATER,
COGL_MATERIAL_ALPHA_FUNC_NOTEQUAL = GL_NOTEQUAL,
COGL_MATERIAL_ALPHA_FUNC_GEQUAL = GL_GEQUAL,
COGL_MATERIAL_ALPHA_FUNC_ALWAYS = GL_ALWAYS
} CoglMaterialAlphaFunc;
/**
* cogl_material_set_alpha_test_function:
* @material: A #CoglMaterial object
* @alpha_func: A @CoglMaterialAlphaFunc constant
* @alpha_reference: A reference point that the chosen alpha function uses
* to compare incoming fragments to.
*
* Before a primitive is blended with the framebuffer, it goes through an
* alpha test stage which lets you discard fragments based on the current
* alpha value. This function lets you change the function used to evaluate
* the alpha channel, and thus determine which fragments are discarded
* and which continue on to the blending stage.
*
* The default is %COGL_MATERIAL_ALPHA_FUNC_ALWAYS
*
* Since: 1.0
*/
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
void
cogl_material_set_alpha_test_function (CoglMaterial *material,
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
CoglMaterialAlphaFunc alpha_func,
float alpha_reference);
/**
* cogl_material_set_blend:
* @material: A #CoglMaterial object
* @blend_string: A <link linkend="cogl-Blend-Strings">Cogl blend string</link>
* describing the desired blend function.
* @error: return location for a #GError that may report lack of driver
* support if you give separate blend string statements for the alpha
* channel and RGB channels since some drivers, or backends such as
* GLES 1.1, don't support this feature. May be %NULL, in which case a
* warning will be printed out using GLib's logging facilities if an
* error is encountered.
*
* If not already familiar; please refer <link linkend="cogl-Blend-Strings">here</link>
* for an overview of what blend strings are, and their syntax.
*
* Blending occurs after the alpha test function, and combines fragments with
* the framebuffer.
* Currently the only blend function Cogl exposes is ADD(). So any valid
* blend statements will be of the form:
*
* |[
* &lt;channel-mask&gt;=ADD(SRC_COLOR*(&lt;factor&gt;), DST_COLOR*(&lt;factor&gt;))
* ]|
*
* <warning>The brackets around blend factors are currently not
* optional!</warning>
*
* This is the list of source-names usable as blend factors:
* <itemizedlist>
* <listitem><para>SRC_COLOR: The color of the in comming fragment</para></listitem>
* <listitem><para>DST_COLOR: The color of the framebuffer</para></listitem>
* <listitem><para>CONSTANT: The constant set via cogl_material_set_blend_constant()</para></listitem>
* </itemizedlist>
*
* The source names can be used according to the
* <link linkend="cogl-Blend-String-syntax">color-source and factor syntax</link>,
* so for example "(1-SRC_COLOR[A])" would be a valid factor, as would
* "(CONSTANT[RGB])"
*
* These can also be used as factors:
* <itemizedlist>
* <listitem>0: (0, 0, 0, 0)</listitem>
* <listitem>1: (1, 1, 1, 1)</listitem>
* <listitem>SRC_ALPHA_SATURATE_FACTOR: (f,f,f,1) where f = MIN(SRC_COLOR[A],1-DST_COLOR[A])</listitem>
* </itemizedlist>
*
* <note>Remember; all color components are normalized to the range [0, 1]
* before computing the result of blending.</note>
*
* <example id="cogl-Blend-Strings-blend-unpremul">
* <title>Blend Strings/1</title>
* <para>Blend a non-premultiplied source over a destination with
* premultiplied alpha:</para>
* <programlisting>
* "RGB = ADD(SRC_COLOR*(SRC_COLOR[A]), DST_COLOR*(1-SRC_COLOR[A]))"
* "A = ADD(SRC_COLOR, DST_COLOR*(1-SRC_COLOR[A]))"
* </programlisting>
* </example>
*
* <example id="cogl-Blend-Strings-blend-premul">
* <title>Blend Strings/2</title>
* <para>Blend a premultiplied source over a destination with
* premultiplied alpha</para>
* <programlisting>
* "RGBA = ADD(SRC_COLOR, DST_COLOR*(1-SRC_COLOR[A]))"
* </programlisting>
* </example>
*
* The default blend string is:
* |[
* RGBA = ADD (SRC_COLOR, DST_COLOR*(1-SRC_COLOR[A]))
* ]|
*
* That gives normal alpha-blending when the calculated color for the material
* is in premultiplied form.
*
* Return value: %TRUE if the blend string was successfully parsed, and the
* described blending is supported by the underlying driver/hardware. If
* there was an error, %FALSE is returned and @error is set accordingly (if
* present).
*
* Since: 1.0
*/
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
gboolean
cogl_material_set_blend (CoglMaterial *material,
const char *blend_string,
GError **error);
/**
* cogl_material_set_blend_constant:
* @material: A #CoglMaterial object
* @constant_color: The constant color you want
*
* When blending is setup to reference a CONSTANT blend factor then
* blending will depend on the constant set with this function.
*
* Since: 1.0
*/
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
void
cogl_material_set_blend_constant (CoglMaterial *material,
const CoglColor *constant_color);
/**
* cogl_material_set_layer:
* @material: A #CoglMaterial object
* @layer_index: the index of the layer
* @texture: a #CoglHandle for the layer object
*
* In addition to the standard OpenGL lighting model a Cogl material may have
* one or more layers comprised of textures that can be blended together in
* order, with a number of different texture combine modes. This function
* defines a new texture layer.
*
* The index values of multiple layers do not have to be consecutive; it is
* only their relative order that is important.
*
* <note>In the future, we may define other types of material layers, such
* as purely GLSL based layers.</note>
*
* Since: 1.0
*/
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
void
cogl_material_set_layer (CoglMaterial *material,
int layer_index,
CoglHandle texture);
/**
* cogl_material_remove_layer:
* @material: A #CoglMaterial object
* @layer_index: Specifies the layer you want to remove
*
* This function removes a layer from your material
*/
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
void
cogl_material_remove_layer (CoglMaterial *material,
int layer_index);
/**
* cogl_material_set_layer_combine:
* @material: A #CoglMaterial object
* @layer_index: Specifies the layer you want define a combine function for
* @blend_string: A <link linkend="cogl-Blend-Strings">Cogl blend string</link>
* describing the desired texture combine function.
* @error: A #GError that may report parse errors or lack of GPU/driver
* support. May be %NULL, in which case a warning will be printed out if an
* error is encountered.
*
* If not already familiar; you can refer
* <link linkend="cogl-Blend-Strings">here</link> for an overview of what blend
* strings are and there syntax.
*
* These are all the functions available for texture combining:
* <itemizedlist>
* <listitem>REPLACE(arg0) = arg0</listitem>
* <listitem>MODULATE(arg0, arg1) = arg0 x arg1</listitem>
* <listitem>ADD(arg0, arg1) = arg0 + arg1</listitem>
* <listitem>ADD_SIGNED(arg0, arg1) = arg0 + arg1 - 0.5</listitem>
* <listitem>INTERPOLATE(arg0, arg1, arg2) = arg0 x arg2 + arg1 x (1 - arg2)</listitem>
* <listitem>SUBTRACT(arg0, arg1) = arg0 - arg1</listitem>
* <listitem>
* <programlisting>
* DOT3_RGB(arg0, arg1) = 4 x ((arg0[R] - 0.5)) * (arg1[R] - 0.5) +
* (arg0[G] - 0.5)) * (arg1[G] - 0.5) +
* (arg0[B] - 0.5)) * (arg1[B] - 0.5))
* </programlisting>
* </listitem>
* <listitem>
* <programlisting>
* DOT3_RGBA(arg0, arg1) = 4 x ((arg0[R] - 0.5)) * (arg1[R] - 0.5) +
* (arg0[G] - 0.5)) * (arg1[G] - 0.5) +
* (arg0[B] - 0.5)) * (arg1[B] - 0.5))
* </programlisting>
* </listitem>
* </itemizedlist>
*
* Refer to the
* <link linkend="cogl-Blend-String-syntax">color-source syntax</link> for
* describing the arguments. The valid source names for texture combining
* are:
* <variablelist>
* <varlistentry>
* <term>TEXTURE</term>
* <listitem>Use the color from the current texture layer</listitem>
* </varlistentry>
* <varlistentry>
* <term>TEXTURE_0, TEXTURE_1, etc</term>
* <listitem>Use the color from the specified texture layer</listitem>
* </varlistentry>
* <varlistentry>
* <term>CONSTANT</term>
* <listitem>Use the color from the constant given with
* cogl_material_set_layer_constant()</listitem>
* </varlistentry>
* <varlistentry>
* <term>PRIMARY</term>
* <listitem>Use the color of the material as set with
* cogl_material_set_color()</listitem>
* </varlistentry>
* <varlistentry>
* <term>PREVIOUS</term>
* <listitem>Either use the texture color from the previous layer, or
* if this is layer 0, use the color of the material as set with
* cogl_material_set_color()</listitem>
* </varlistentry>
* </variablelist>
*
* <refsect2 id="cogl-Layer-Combine-Examples">
* <title>Layer Combine Examples</title>
* <para>This is effectively what the default blending is:</para>
* <informalexample><programlisting>
* RGBA = MODULATE (PREVIOUS, TEXTURE)
* </programlisting></informalexample>
* <para>This could be used to cross-fade between two images, using
* the alpha component of a constant as the interpolator. The constant
* color is given by calling cogl_material_set_layer_constant.</para>
* <informalexample><programlisting>
* RGBA = INTERPOLATE (PREVIOUS, TEXTURE, CONSTANT[A])
* </programlisting></informalexample>
* </refsect2>
*
* <note>You can't give a multiplication factor for arguments as you can
* with blending.</note>
*
* Return value: %TRUE if the blend string was successfully parsed, and the
* described texture combining is supported by the underlying driver and
* or hardware. On failure, %FALSE is returned and @error is set
*
* Since: 1.0
*/
gboolean
cogl_material_set_layer_combine (CoglMaterial *material,
int layer_index,
const char *blend_string,
GError **error);
/**
* cogl_material_set_layer_combine_constant:
* @material: A #CoglMaterial object
* @layer_index: Specifies the layer you want to specify a constant used
* for texture combining
* @constant: The constant color you want
*
* When you are using the 'CONSTANT' color source in a layer combine
* description then you can use this function to define its value.
*
* Since: 1.0
*/
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
void
cogl_material_set_layer_combine_constant (CoglMaterial *material,
int layer_index,
const CoglColor *constant);
/**
* cogl_material_set_layer_matrix:
* @material: A #CoglMaterial object
* @layer_index: the index for the layer inside @material
* @matrix: the transformation matrix for the layer
*
* This function lets you set a matrix that can be used to e.g. translate
* and rotate a single layer of a material used to fill your geometry.
*/
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
void
cogl_material_set_layer_matrix (CoglMaterial *material,
int layer_index,
const CoglMatrix *matrix);
/**
* cogl_material_get_layers:
* @material: A #CoglMaterial object
*
* This function lets you access a material's internal list of layers
* for iteration.
*
* <note>You should avoid using this API if possible since it was only
* made public by mistake and will be deprecated when we have
* suitable alternative.</note>
*
* <note>It's important to understand that the list returned may not
* remain valid if you modify the material or any of the layers in any
* way and so you would have to re-get the list in that
* situation.</note>
*
* Return value: (element-type CoglMaterialLayer) (transfer none): A
* list of #CoglMaterialLayer<!-- -->'s that can be passed to the
* cogl_material_layer_* functions. The list is owned by Cogl and it
* should not be modified or freed
*/
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
G_CONST_RETURN GList *
cogl_material_get_layers (CoglMaterial *material);
[cogl] Improving Cogl journal to minimize driver overheads + GPU state changes Previously the journal was always flushed at the end of _cogl_rectangles_with_multitexture_coords, (i.e. the end of any cogl_rectangle* calls) but now we have broadened the potential for batching geometry. In ideal circumstances we will only flush once per scene. In summary the journal works like this: When you use any of the cogl_rectangle* APIs then nothing is emitted to the GPU at this point, we just log one or more quads into the journal. A journal entry consists of the quad coordinates, an associated material reference, and a modelview matrix. Ideally the journal only gets flushed once at the end of a scene, but in fact there are things to consider that may cause unwanted flushing, including: - modifying materials mid-scene This is because each quad in the journal has an associated material reference (i.e. not copy), so if you try and modify a material that is already referenced in the journal we force a flush first) NOTE: For now this means you should avoid using cogl_set_source_color() since that currently uses a single shared material. Later we should change it to use a pool of materials that is recycled when the journal is flushed. - modifying any state that isn't currently logged, such as depth, fog and backface culling enables. The first thing that happens when flushing, is to upload all the vertex data associated with the journal into a single VBO. We then go through a process of splitting up the journal into batches that have compatible state so they can be emitted to the GPU together. This is currently broken up into 3 levels so we can stagger the state changes: 1) we break the journal up according to changes in the number of material layers associated with logged quads. The number of layers in a material determines the stride of the associated vertices, so we have to update our vertex array offsets at this level. (i.e. calling gl{Vertex,Color},Pointer etc) 2) we further split batches up according to material compatability. (e.g. materials with different textures) We flush material state at this level. 3) Finally we split batches up according to modelview changes. At this level we update the modelview matrix and actually emit the actual draw command. This commit is largely about putting the initial design in-place; this will be followed by other changes that take advantage of the extended batching.
2009-06-17 13:46:42 -04:00
/**
* cogl_material_get_n_layers:
* @material: A #CoglMaterial object
[cogl] Improving Cogl journal to minimize driver overheads + GPU state changes Previously the journal was always flushed at the end of _cogl_rectangles_with_multitexture_coords, (i.e. the end of any cogl_rectangle* calls) but now we have broadened the potential for batching geometry. In ideal circumstances we will only flush once per scene. In summary the journal works like this: When you use any of the cogl_rectangle* APIs then nothing is emitted to the GPU at this point, we just log one or more quads into the journal. A journal entry consists of the quad coordinates, an associated material reference, and a modelview matrix. Ideally the journal only gets flushed once at the end of a scene, but in fact there are things to consider that may cause unwanted flushing, including: - modifying materials mid-scene This is because each quad in the journal has an associated material reference (i.e. not copy), so if you try and modify a material that is already referenced in the journal we force a flush first) NOTE: For now this means you should avoid using cogl_set_source_color() since that currently uses a single shared material. Later we should change it to use a pool of materials that is recycled when the journal is flushed. - modifying any state that isn't currently logged, such as depth, fog and backface culling enables. The first thing that happens when flushing, is to upload all the vertex data associated with the journal into a single VBO. We then go through a process of splitting up the journal into batches that have compatible state so they can be emitted to the GPU together. This is currently broken up into 3 levels so we can stagger the state changes: 1) we break the journal up according to changes in the number of material layers associated with logged quads. The number of layers in a material determines the stride of the associated vertices, so we have to update our vertex array offsets at this level. (i.e. calling gl{Vertex,Color},Pointer etc) 2) we further split batches up according to material compatability. (e.g. materials with different textures) We flush material state at this level. 3) Finally we split batches up according to modelview changes. At this level we update the modelview matrix and actually emit the actual draw command. This commit is largely about putting the initial design in-place; this will be followed by other changes that take advantage of the extended batching.
2009-06-17 13:46:42 -04:00
*
* Retrieves the number of layers defined for the given @material
*
* Return value: the number of layers
[cogl] Improving Cogl journal to minimize driver overheads + GPU state changes Previously the journal was always flushed at the end of _cogl_rectangles_with_multitexture_coords, (i.e. the end of any cogl_rectangle* calls) but now we have broadened the potential for batching geometry. In ideal circumstances we will only flush once per scene. In summary the journal works like this: When you use any of the cogl_rectangle* APIs then nothing is emitted to the GPU at this point, we just log one or more quads into the journal. A journal entry consists of the quad coordinates, an associated material reference, and a modelview matrix. Ideally the journal only gets flushed once at the end of a scene, but in fact there are things to consider that may cause unwanted flushing, including: - modifying materials mid-scene This is because each quad in the journal has an associated material reference (i.e. not copy), so if you try and modify a material that is already referenced in the journal we force a flush first) NOTE: For now this means you should avoid using cogl_set_source_color() since that currently uses a single shared material. Later we should change it to use a pool of materials that is recycled when the journal is flushed. - modifying any state that isn't currently logged, such as depth, fog and backface culling enables. The first thing that happens when flushing, is to upload all the vertex data associated with the journal into a single VBO. We then go through a process of splitting up the journal into batches that have compatible state so they can be emitted to the GPU together. This is currently broken up into 3 levels so we can stagger the state changes: 1) we break the journal up according to changes in the number of material layers associated with logged quads. The number of layers in a material determines the stride of the associated vertices, so we have to update our vertex array offsets at this level. (i.e. calling gl{Vertex,Color},Pointer etc) 2) we further split batches up according to material compatability. (e.g. materials with different textures) We flush material state at this level. 3) Finally we split batches up according to modelview changes. At this level we update the modelview matrix and actually emit the actual draw command. This commit is largely about putting the initial design in-place; this will be followed by other changes that take advantage of the extended batching.
2009-06-17 13:46:42 -04:00
*
* Since: 1.0
*/
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
cogl_material_get_n_layers (CoglMaterial *material);
[cogl] Improving Cogl journal to minimize driver overheads + GPU state changes Previously the journal was always flushed at the end of _cogl_rectangles_with_multitexture_coords, (i.e. the end of any cogl_rectangle* calls) but now we have broadened the potential for batching geometry. In ideal circumstances we will only flush once per scene. In summary the journal works like this: When you use any of the cogl_rectangle* APIs then nothing is emitted to the GPU at this point, we just log one or more quads into the journal. A journal entry consists of the quad coordinates, an associated material reference, and a modelview matrix. Ideally the journal only gets flushed once at the end of a scene, but in fact there are things to consider that may cause unwanted flushing, including: - modifying materials mid-scene This is because each quad in the journal has an associated material reference (i.e. not copy), so if you try and modify a material that is already referenced in the journal we force a flush first) NOTE: For now this means you should avoid using cogl_set_source_color() since that currently uses a single shared material. Later we should change it to use a pool of materials that is recycled when the journal is flushed. - modifying any state that isn't currently logged, such as depth, fog and backface culling enables. The first thing that happens when flushing, is to upload all the vertex data associated with the journal into a single VBO. We then go through a process of splitting up the journal into batches that have compatible state so they can be emitted to the GPU together. This is currently broken up into 3 levels so we can stagger the state changes: 1) we break the journal up according to changes in the number of material layers associated with logged quads. The number of layers in a material determines the stride of the associated vertices, so we have to update our vertex array offsets at this level. (i.e. calling gl{Vertex,Color},Pointer etc) 2) we further split batches up according to material compatability. (e.g. materials with different textures) We flush material state at this level. 3) Finally we split batches up according to modelview changes. At this level we update the modelview matrix and actually emit the actual draw command. This commit is largely about putting the initial design in-place; this will be followed by other changes that take advantage of the extended batching.
2009-06-17 13:46:42 -04:00
/**
* CoglMaterialLayerType:
* @COGL_MATERIAL_LAYER_TYPE_TEXTURE: The layer represents a
* <link linkend="cogl-Textures">texture</link>
*
* Available types of layers for a #CoglMaterial. This enumeration
* might be expanded in later versions.
*
* Since: 1.0
*/
typedef enum {
COGL_MATERIAL_LAYER_TYPE_TEXTURE
} CoglMaterialLayerType;
/**
* cogl_material_layer_get_type:
* @layer: A #CoglMaterialLayer object
*
* Retrieves the type of the layer
*
* Currently there is only one type of layer defined:
* %COGL_MATERIAL_LAYER_TYPE_TEXTURE, but considering we may add purely GLSL
* based layers in the future, you should write code that checks the type
* first.
*
* Return value: the type of the layer
*/
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
CoglMaterialLayerType
cogl_material_layer_get_type (CoglMaterialLayer *layer);
/**
* cogl_material_layer_get_texture:
* @layer: A #CoglMaterialLayer object
*
* Extracts a texture handle for a specific layer.
*
* <note>In the future Cogl may support purely GLSL based layers; for those
* layers this function which will likely return %COGL_INVALID_HANDLE if you
* try to get the texture handle from them. Considering this scenario, you
* should call cogl_material_layer_get_type() first in order check it is of
* type %COGL_MATERIAL_LAYER_TYPE_TEXTURE before calling this function.</note>
*
* Return value: a #CoglHandle for the texture inside the layer
*/
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
CoglHandle
cogl_material_layer_get_texture (CoglMaterialLayer *layer);
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
/**
* cogl_material_layer_get_min_filter:
* @layer: a #CoglHandle for a material layer
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
*
* Queries the currently set downscaling filter for a material layer
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
*
* Return value: the current downscaling filter
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
*/
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
CoglMaterialFilter
cogl_material_layer_get_min_filter (CoglMaterialLayer *layer);
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
/**
* cogl_material_layer_get_mag_filter:
* @layer: A #CoglMaterialLayer object
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
*
* Queries the currently set downscaling filter for a material later
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
*
* Return value: the current downscaling filter
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
*/
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
CoglMaterialFilter
cogl_material_layer_get_mag_filter (CoglMaterialLayer *layer);
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
/**
* cogl_material_set_layer_filters:
* @material: A #CoglMaterial object
[cogl] Move the texture filters to be a property of the material layer The texture filters are now a property of the material layer rather than the texture object. Whenever a texture is painted with a material it sets the filters on all of the GL textures in the Cogl texture. The filter is cached so that it won't be changed unnecessarily. The automatic mipmap generation has changed so that the mipmaps are only generated when the texture is painted instead of every time the data changes. Changing the texture sets a flag to mark that the mipmaps are dirty. This works better if the FBO extension is available because we can use glGenerateMipmap. If the extension is not available it will temporarily enable automatic mipmap generation and reupload the first pixel of each slice. This requires tracking the data for the first pixel. The COGL_TEXTURE_AUTO_MIPMAP flag has been replaced with COGL_TEXTURE_NO_AUTO_MIPMAP so that it will default to auto-mipmapping. The mipmap generation is now effectively free if you are not using a mipmap filter mode so you would only want to disable it if you had some special reason to generate your own mipmaps. ClutterTexture no longer has to store its own copy of the filter mode. Instead it stores it in the material and the property is directly set and read from that. This fixes problems with the filters getting out of sync when a cogl handle is set on the texture directly. It also avoids the mess of having to rerealize the texture if the filter quality changes to HIGH because Cogl will take of generating the mipmaps if needed.
2009-06-04 11:04:57 -04:00
* @layer_index: the layer number to change.
* @min_filter: the filter used when scaling a texture down.
* @mag_filter: the filter used when magnifying a texture.
*
* Changes the decimation and interpolation filters used when a texture is
* drawn at other scales than 100%.
*/
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
void
cogl_material_set_layer_filters (CoglMaterial *material,
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 layer_index,
CoglMaterialFilter min_filter,
CoglMaterialFilter mag_filter);
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
/**
* cogl_material_set_layer_wrap_mode_s:
* @material: A #CoglMaterial object
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
* @layer_index: the layer number to change.
* @mode: the new wrap mode
*
* Sets the wrap mode for the 's' coordinate of texture lookups on this layer.
*
* Since: 1.4
*/
void
cogl_material_set_layer_wrap_mode_s (CoglMaterial *material,
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
int layer_index,
CoglMaterialWrapMode mode);
/**
* cogl_material_set_layer_wrap_mode_t:
* @material: A #CoglMaterial object
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
* @layer_index: the layer number to change.
* @mode: the new wrap mode
*
* Sets the wrap mode for the 't' coordinate of texture lookups on this layer.
*
* Since: 1.4
*/
void
cogl_material_set_layer_wrap_mode_t (CoglMaterial *material,
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
int layer_index,
CoglMaterialWrapMode mode);
/**
* cogl_material_set_layer_wrap_mode:
* @material: A #CoglMaterial object
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
* @layer_index: the layer number to change.
* @mode: the new wrap mode
*
* Sets the wrap mode for both coordinates of texture lookups on this
* layer. This is equivalent to calling
* cogl_material_set_layer_wrap_mode_s() and
* cogl_material_set_layer_wrap_mode_t() separately.
*
* Since: 1.4
*/
void
cogl_material_set_layer_wrap_mode (CoglMaterial *material,
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
int layer_index,
CoglMaterialWrapMode mode);
/**
* cogl_material_layer_get_wrap_mode_s:
* @layer: A #CoglMaterialLayer object
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
*
* Gets the wrap mode for the 's' coordinate of texture lookups on this layer.
*
* Return value: the wrap mode value for the s coordinate.
*
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
* Since: 1.4
*/
CoglMaterialWrapMode
cogl_material_layer_get_wrap_mode_s (CoglMaterialLayer *layer);
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
/**
* cogl_material_layer_get_wrap_mode_t:
* @layer: A #CoglMaterialLayer object
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
*
* Gets the wrap mode for the 't' coordinate of texture lookups on this layer.
*
* Return value: the wrap mode value for the t coordinate.
*
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
* Since: 1.4
*/
CoglMaterialWrapMode
cogl_material_layer_get_wrap_mode_t (CoglMaterialLayer *layer);
cogl-material: Add support for setting the wrap mode for a layer Previously, Cogl's texture coordinate system was effectively always GL_REPEAT so that if an application specifies coordinates outside the range 0→1 it would get repeated copies of the texture. It would however change the mode to GL_CLAMP_TO_EDGE if all of the coordinates are in the range 0→1 so that in the common case that the whole texture is being drawn with linear filtering it will not blend in edge pixels from the opposite sides. This patch adds the option for applications to change the wrap mode per layer. There are now three wrap modes: 'repeat', 'clamp-to-edge' and 'automatic'. The automatic map mode is the default and it implements the previous behaviour. The wrap mode can be changed for the s and t coordinates independently. I've tried to make the internals support setting the r coordinate but as we don't support 3D textures yet I haven't exposed any public API for it. The texture backends still have a set_wrap_mode virtual but this value is intended to be transitory and it will be changed whenever the material is flushed (although the backends are expected to cache it so that it won't use too many GL calls). In my understanding this value was always meant to be transitory and all primitives were meant to set the value before drawing. However there were comments suggesting that this is not the expected behaviour. In particular the vertex buffer drawing code never set a wrap mode so it would end up with whatever the texture was previously used for. These issues are now fixed because the material will always set the wrap modes. There is code to manually implement clamp-to-edge for textures that can't be hardware repeated. However this doesn't fully work because it relies on being able to draw the stretched parts using quads with the same values for tx1 and tx2. The texture iteration code doesn't support this so it breaks. This is a separate bug and it isn't trivially solved. When flushing a material there are now extra options to set wrap mode overrides. The overrides are an array of values for each layer that specifies an override for the s, t or r coordinates. The primitives use this to implement the automatic wrap mode. cogl_polygon also uses it to set GL_CLAMP_TO_BORDER mode for its trick to render sliced textures. Although this code has been added it looks like the sliced trick has been broken for a while and I haven't attempted to fix it here. I've added a constant to represent the maximum number of layers that a material supports so that I can size the overrides array. I've set it to 32 because as far as I can tell we have that limit imposed anyway because the other flush options use a guint32 to store a flag about each layer. The overrides array ends up adding 32 bytes to each flush options struct which may be a concern. http://bugzilla.openedhand.com/show_bug.cgi?id=2063
2010-04-01 06:31:33 -04:00
/* XXX: should this be CoglMaterialDepthTestFunction?
* It makes it very verbose but would be consistent with
* CoglMaterialWrapMode */
/**
* CoglDepthTestFunction:
* @COGL_DEPTH_TEST_FUNCTION_NEVER: Never passes.
* @COGL_DEPTH_TEST_FUNCTION_LESS: Passes if the fragment's depth
* value is less than the value currently in the depth buffer.
* @COGL_DEPTH_TEST_FUNCTION_EQUAL: Passes if the fragment's depth
* value is equal to the value currently in the depth buffer.
* @COGL_DEPTH_TEST_FUNCTION_LEQUAL: Passes if the fragment's depth
* value is less or equal to the value currently in the depth buffer.
* @COGL_DEPTH_TEST_FUNCTION_GREATER: Passes if the fragment's depth
* value is greater than the value currently in the depth buffer.
* @COGL_DEPTH_TEST_FUNCTION_NOTEQUAL: Passes if the fragment's depth
* value is not equal to the value currently in the depth buffer.
* @COGL_DEPTH_TEST_FUNCTION_GEQUAL: Passes if the fragment's depth
* value greater than or equal to the value currently in the depth buffer.
* @COGL_DEPTH_TEST_FUNCTION_ALWAYS: Always passes.
*
* When using depth testing one of these functions is used to compare
* the depth of an incoming fragment against the depth value currently
* stored in the depth buffer. The function is changed using
* cogl_material_set_depth_test_function().
*
* The test is only done when depth testing is explicitly enabled. (See
* cogl_material_set_depth_test_enabled())
*/
typedef enum
{
COGL_DEPTH_TEST_FUNCTION_NEVER = GL_NEVER,
COGL_DEPTH_TEST_FUNCTION_LESS = GL_LESS,
COGL_DEPTH_TEST_FUNCTION_EQUAL = GL_EQUAL,
COGL_DEPTH_TEST_FUNCTION_LEQUAL = GL_LEQUAL,
COGL_DEPTH_TEST_FUNCTION_GREATER = GL_GREATER,
COGL_DEPTH_TEST_FUNCTION_NOTEQUAL = GL_NOTEQUAL,
COGL_DEPTH_TEST_FUNCTION_GEQUAL = GL_GEQUAL,
COGL_DEPTH_TEST_FUNCTION_ALWAYS = GL_ALWAYS
} CoglDepthTestFunction;
/* XXX: to avoid having to split this into a separate include that can
* in #included internally without needing the
* COGL_ENABLE_EXPERIMENTAL_API define this isn't guarded. It's still
* considered experimental but it's guarded instead by the fact that
* there's no corresponding API. */
#ifdef COGL_ENABLE_EXPERIMENTAL_API
/**
* cogl_material_set_depth_test_enabled:
* @material: A #CoglMaterial object
* @enable: The enable state you want
*
* Enables or disables depth testing according to the value of
* @enable.
*
* If depth testing is enable then the #CoglDepthTestFunction set
* using cogl_material_set_depth_test_function() us used to evaluate
* the depth value of incoming fragments against the corresponding
* value stored in the current depth buffer, and if the test passes
* then the fragments depth value is used to update the depth buffer.
* (unless you have disabled depth writing via
* cogl_material_set_depth_writing_enabled ())
*
* By default depth testing is disabled.
*
* Since: 1.4
*/
void
cogl_material_set_depth_test_enabled (CoglMaterial *material,
gboolean enable);
/**
* cogl_material_get_depth_test_enabled:
* @material: A #CoglMaterial object
*
* Gets the current depth test enabled state as previously set by
* cogl_material_set_depth_test_enabled().
*
* Returns: The material's current depth test enabled state.
* Since: 1.4
*/
gboolean
cogl_material_get_depth_test_enabled (CoglMaterial *material);
/**
* cogl_material_set_depth_writing_enabled:
* @material: A #CoglMaterial object
* @enable: The enable state you want
*
* Enables or disables depth buffer writing according to the value of
* @enable. Normally when depth testing is enabled and the comparison
* between a fragment's depth value and the corresponding depth buffer
* value passes then the fragment's depth is written to the depth
* buffer unless writing is disabled here.
*
* By default depth writing is enabled
*
* Since: 1.4
*/
void
cogl_material_set_depth_writing_enabled (CoglMaterial *material,
gboolean enable);
/**
* cogl_material_get_depth_writing_enabled:
* @material: A #CoglMaterial object
*
* Gets the depth writing enable state as set by the corresponding
* cogl_material_set_depth_writing_enabled.
*
* Returns: The current depth writing enable state
* Since: 1.4
*/
gboolean
cogl_material_get_depth_writing_enabled (CoglMaterial *material);
/**
* cogl_material_set_depth_test_function:
* @material: A #CoglMaterial object
* @function: The #CoglDepthTestFunction to set
*
* Sets the #CoglDepthTestFunction used to compare the depth value of
* an incoming fragment against the corresponding value in the current
* depth buffer.
*
* Since: 1.4
*/
void
cogl_material_set_depth_test_function (CoglMaterial *material,
CoglDepthTestFunction function);
/**
* cogl_material_get_depth_test_function:
* @material: A #CoglMaterial object
*
* Gets the current depth test enable state as previously set via
* cogl_material_set_depth_test_enabled().
*
* Returns: The current depth test enable state.
* Since: 1.4
*/
CoglDepthTestFunction
cogl_material_get_depth_test_function (CoglMaterial *material);
/**
* cogl_material_set_depth_range:
* @material: A #CoglMaterial object
* @near_val: The near component of the desired depth range which will be
* clamped to the range [0, 1]
* @far_val: The far component of the desired depth range which will be
* clamped to the range [0, 1]
* @error: location to store an error of type #CoglError
*
* Sets the range to map depth values in normalized device coordinates
* to before writing out to a depth buffer.
*
* After your geometry has be transformed, clipped and had perspective
* division applied placing it in normalized device
* coordinates all depth values between the near and far z clipping
* planes are in the range -1 to 1. Before writing any depth value to
* the depth buffer though the value is mapped into the range [0, 1].
*
* With this function you can change the range which depth values are
* mapped too although the range must still lye within the range [0,
* 1].
*
* If your driver does not support this feature (for example you are
* using GLES 1 drivers) then this will return %FALSE and set an error
* if @error isn't NULL. You can check ahead of time for the
* %COGL_FEATURE_DEPTH_RANGE feature with cogl_features_available() to
* know if this function will succeed.
*
* By default normalized device coordinate depth values are mapped to
* the full range of depth buffer values, [0, 1].
*
* Returns: %TRUE if driver support is available else %FALSE.
*
* Since: 1.4
*/
gboolean
cogl_material_set_depth_range (CoglMaterial *material,
float near_val,
float far_val,
GError **error);
/**
* cogl_material_get_depth_range_mapping:
* @material: A #CoglMaterial object
* @near_val: A pointer to store the near component of the depth range
* @far_val: A pointer to store the far component of the depth range
*
* Gets the current range to which normalized depth values are mapped
* before writing to the depth buffer. This corresponds to the range
* set with cogl_material_set_depth_range().
*
* Since: 1.4
*/
void
cogl_material_get_depth_range (CoglMaterial *material,
float *near_val,
float *far_val);
#endif /* COGL_ENABLE_EXPERIMENTAL_API */
G_END_DECLS
#endif /* __COGL_MATERIAL_H__ */