Commit Graph

87 Commits

Author SHA1 Message Date
Robert Bragg
b025a4a020 material: Add cogl_material_{get,set}_user_program
Previously we had an internal only _cogl_material_set_user_program to
redirect legacy usage of cogl_program_use() through CoglMaterial. This
instead makes the API public because until we implement our planned
"snippet" framework we need a stop-gap solution for using shaders in
Cogl.

The plan is to also support ARBfp with the cogl_program/shader API so
this API will also allow clutter-gst to stop using direct OpenGL calls
that conflict with Cogl's state tracking.
2010-08-09 17:26:57 +01:00
Robert Bragg
937fc0dbec material: chain up in layer_pre_change_notify
A change to a layer is also going to be a change to its owning material
so we have to chain up in _cogl_material_layer_pre_change_notify and
call _cogl_material_pre_change_notify. Previously we were only
considering if the owning material was referenced in the journal but
that ignores that it might also have dependants. We no longer need to
flush the journal directly in layer_pre_change_notify.
2010-08-09 17:12:26 +01:00
Robert Bragg
b493b30257 material: unlink old layer from owner when replacing
In _cogl_material_layer_pre_change_notify when we see that a layer has
dependants and it can't be modified directly then we allocate a new
layer. In this case we also have to link the new layer to its required
owner. If the immutable layer we copied had the same owner though we
weren't unlinking that old layer.
2010-08-09 17:12:26 +01:00
Robert Bragg
021265e634 material: fix initialization of sparse properties
In _cogl_material_pre_change_notify we need to identify if it's a sparse
property being changed and if so initialize the state group if the given
material isn't currently the authority for it.

Previously we were unconditionally calling
_cogl_material_initialize_state which would e.g. NULL the layer
differences list of a material each time a layer change was notified.

It would also call _cogl_material_initialize_state for non-sparse
properties which should always be valid at this point so the function
has been renamed to _cogl_material_initialize_sparse_state to make this
clearer with a corresponding g_return_if_fail check.
2010-08-09 17:12:26 +01:00
Robert Bragg
02ce77dc08 material: fix copying of material layer differences
This fixes how we copy layer differences in
_cogl_material_copy_layer_differences.

We were making a redundant g_list_copy of the src differences and then
iterating the src list calling _cogl_material_add_layer_difference for
each entry which would double the list length, but the initial copy
directly referenced the original layers which wasn't correct.

Also we were initializing dest->n_layers before copying the layer
differences but the act of copying the differences will re-initialize
n_layers to 0 when adding the first layer_difference since it will
trigger a layer_pre_change_notify and since the dest material isn't yet
a STATE_LAYERS authority the state group is initialized before allowing
the change.
2010-08-09 17:12:26 +01:00
Robert Bragg
8caaf5a8b2 material: Improve the dot file debugging code
This improve the dot file output available when calling
_cogl_debug_dump_materials_dot_file. The material graph now directly
points into the layer graph and the layers now show the texture unit
index.
2010-08-09 17:12:25 +01:00
Neil Roberts
abea3d7ead cogl-material: Unref the texture when reverting authority
When the texture is set on a layer so that it is back to the parent's
texture it would clear the texture change flag but it wouldn't unref
the texture. The free function for a material layer does not unref the
texture if the change flag is cleared so the texture would end up
leaking. This happens for ClutterTexture because it disposes the
texture by setting layer 0 of the material to COGL_INVALID_HANDLE
which ends up the same as the default material.
2010-08-06 17:09:54 +01:00
Robert Bragg
eda852e78c material: only deference layer->texture for authority
In _cogl_material_layer_pre_paint we were mistakenly dereferencing the
layer->texture member for the passed layer instead of dereferencing the
texture state authority which was causing crashes in some cases.
2010-08-06 14:16:35 +01:00
Robert Bragg
8640f527cb cogl: don't include cogl-debug.h in cogl.h or install
cogl-debug.h is an internal header so it shouldn't have been included by
cogl.h and the header shouldn't be installed either.
2010-08-02 17:41:42 +01:00
Neil Roberts
0ab6dc9db1 cogl-material: Don't map the shininess value to [0,1]
In OpenGL the 'shininess' lighting parameter is floating point value
limited to the range 0.0→128.0. This number is used to affect the size
of the specular highlight. Cogl materials used to only accept a number
between 0.0 and 1.0 which then gets multiplied by 128.0 before sending
to GL. I think the assumption was that this is just a weird GL quirk
so we don't expose it. However the value is used as an exponent to
raise the attenuation to a power so there is no conceptual limit to
the value.

This removes the mapping and changes some of the documentation.

http://bugzilla.clutter-project.org/show_bug.cgi?id=2222
2010-07-22 17:52:51 +01:00
Neil Roberts
5b6298db10 cogl-material: Consider the shader when deciding if materials equate
_cogl_material_equal was ignoring the user shader state so rectangles
with different shaders would get batched together.

http://bugzilla.clutter-project.org/show_bug.cgi?id=2220
2010-07-20 17:21:12 +01:00
Robert Bragg
96b0e6c304 material: splits out all the state flushing code
This moves the code supporting _cogl_material_flush_gl_state into
cogl-material-opengl.c as part of an effort to reduce the size of
cogl-material.c to keep it manageable.
2010-07-13 19:26:58 +01:00
Robert Bragg
5442e429ba material: split the texture unit management out
In general cogl-material.c has become far to large to manage in one
source file. As one of the ways to try and break it down this patch
starts to move some of lower level texture unit state management out
into cogl-material-opengl.c. The naming is such because the plan is to
follow up and migrate the very GL specific state flushing code into the
same file.
2010-07-13 19:26:58 +01:00
Robert Bragg
be9564b3db material: copy_differences: handle copying fog state
When the support for redirecting the legacy fog state through cogl
material was added in 9b9e764dc, the code to handle copying the fog
state in _cogl_material_copy_differences was missed.
2010-07-13 19:26:57 +01:00
Neil Roberts
671abec8b7 Make the material functions for setting the p wrap mode public
Now that we have 3D texture support it makes sense to expose the wrap
mode for the p coordinate.
2010-07-13 14:29:07 +01:00
Neil Roberts
ec718d4ca4 Rename the third texure coordinate from 'r' to 'p'
Using 'r' to name the third component is problematic because that is
commonly used to represent the red component of a vector representing
a color. Under GLSL this is awkward because the texture swizzling for
a vector uses a single letter for each component and the names for
colors, textures and positions are synonymous. GLSL works around this
by naming the components of the texture s, t, p and q. Cogl already
effectively already exposes this naming because it exposes GLSL so it
makes sense to use that naming consistently. Another alternative could
be u, v and w. This is what Blender and Direct3D use. However the w
component conflicts with the w component of a position vertex.
2010-07-13 14:29:07 +01:00
Neil Roberts
396ef40d43 Rename COGL_ERROR_MISSING_FEATURE to COGL_ERROR_UNSUPPORTED
We can use this error in more unsupported situations than just when we
have a Cogl feature flag for the error. For example if a non-sliced
texture is created with dimensions that are too large then we could
throw this error. Therefore it seems good to rename to something more
general.
2010-07-13 14:28:36 +01:00
Neil Roberts
2c8bf00995 Don't define public cogl_is_* functions for internal types
This adds a COGL_OBJECT_INTERNAL_DEFINE macro and friends that are the
same as COGL_OBJECT_DEFINE except that they prefix the cogl_is_*
function with an underscore so that it doesn't get exported in the
shared library.
2010-07-09 18:57:54 +01:00
Neil Roberts
b25035ad2d cogl: Don't define the deprecated ref/unref accessors for new types
Previously COGL_OBJECT_DEFINE would always define deprecated
cogl_$type_{ref,unref} functions even if the type is new or if the
type is entirely internal. An application would still find it
difficult to use these because they wouldn't be in the headers, but it
still looks bad that they are exported from the shared library. This
patch changes it so that the deprecated ref counting functions are
defined using a separate macro and only the types that have these
functions in the headers call this macro.
2010-07-09 18:57:54 +01:00
Robert Bragg
ea708094c5 material: use common node type for materials and layers
Since 365605cf42, materials and layers are represented in a tree
structure that allows traversing up through parents and iterating down
through children.  This re-works the related typedefs and reparenting
code so that they can be shared.
2010-07-09 18:04:27 +01:00
Neil Roberts
475a72fd94 cogl-material: Fix some problems with flushing texture overrides
There were a few problems flushing texture overrides so that sliced
textures would not work:

* In _cogl_material_set_layer_texture it ignored the 'overriden'
  parameter and always set texture_overridden to FALSE.

* cogl_texture_get_gl_texture wasn't being called correctly in
  override_layer_texture_cb. It returns a gboolean to indicate the
  error status but this boolean was being assigned to gl_target.

* _cogl_material_layer_texture_equal did not take into account the
  override.

* _cogl_material_layer_get_texture_info did not return the overridden
  texture so it would always use the first texture slice.
2010-07-09 11:14:15 +01:00
Neil Roberts
da3be3df6b cogl-material: Add support for point sprites
This adds a new API call to enable point sprite coordinate generation
for a material layer:

void
cogl_material_set_layer_point_sprite_coords_enabled (CoglHandle material,
                                                     int layer_index,
                                                     gboolean enable);

There is also a corresponding get function.

Enabling point sprite coords simply sets the GL_COORD_REPLACE of the
GL_POINT_SPRITE glTexEnv when flusing the material. There is no
separate application control for glEnable(GL_POINT_SPRITE). Instead it
is left permanently enabled under the assumption that it has no affect
unless GL_COORD_REPLACE is enabled for a texture unit.

http://bugzilla.openedhand.com/show_bug.cgi?id=2047
2010-07-08 16:34:30 +01:00
Neil Roberts
06b58baa10 cogl-material: Add a property for setting the point size
This adds cogl_material_{get,set}_point_size. If the point size is not
1.0f then glPointSize will be called when the material is flushed.

http://bugzilla.openedhand.com/show_bug.cgi?id=2047
2010-07-08 16:34:30 +01:00
Robert Bragg
dcc8c66fdd material: support dumping layer state graph to dot file
Recently I added a _cogl_debug_dump_materials_dot_file function for
debugging the sparse material state. This extends the state dumped to
include the graph of layer state also.
2010-07-07 20:46:50 +01:00
Robert Bragg
093251776f material: set ->layer_index on new layers for unit0
We were mistakenly only initializing layer->layer_index for new layers
associated with texture units > 0. This had gone unnoticed because
normally layers associated with texture unit0 have a layer index of 0
too. Mutter was hitting this issue because it was initializing layer 1
before layer 0 for one of its materials so layer 1 was temporarily
associated with texture unit 0.
2010-07-07 20:46:04 +01:00
Robert Bragg
8098052d56 material: route fogging state through CoglMaterial
Previously cogl_set_fog would cause a flush of the Cogl journal and
would directly bang the GL state machine to setup fogging. As part of
the ongoing effort to track most state in CoglMaterial to support
renderlists this now adds an indirection so that cogl_set_fog now just
updates ctx->legacy_fog_state. The fogging state then gets enabled as a
legacy override similar to how the old depth testing API is handled.
2010-07-07 14:12:15 +01:00
Robert Bragg
12a76ca4aa cogl-material: Adds debug api to dump sparse material graph
This adds a _cogl_debug_dump_materials_dot_file function that can be
used to dump all the descendants of the default material to a file using
the dot format which can then be converted to an image to visualize.
2010-07-05 15:04:46 +01:00
Robert Bragg
cc0b76b405 material: don't keep extra ref on copy-on-write materials
In _cogl_material_pre_change_notify if a material with descendants is
modified then we create a new material that is a copy of the one being
modified and reparent those descendants to the new material.

This patch ensures we drop the reference we get from cogl_material_copy
since we can rely on the descendants to keep the new material alive.
2010-07-05 14:51:05 +01:00
Neil Roberts
7961b1c93d cogl-material: Don't doubly ref the parent in cogl_material_copy
cogl_material_copy was taking a reference on the original texture when
making a copy. However it then calls _cogl_material_set_parent on the
material which also takes a reference on the parent. The second
reference is cleaned up whenever _cogl_material_unparent is called and
this is also called by _cogl_material_free. However, it seems that
nothing was cleaning up the first reference. I think the reference is
entirely unnecessary so this patch removes it.
2010-07-01 15:32:08 +01:00
Robert Bragg
be46b872e9 material: Replace CoglHandle with CoglMaterial *
As part of the ongoing effort to remove CoglHandle from the API this
switches the cogl_material API to use a strongly typed CoglMaterial
pointer instead of CoglHandle.
2010-06-30 15:04:18 +01:00
Robert Bragg
17731b3af1 material: Split the fragment processing backends out
This splits the fragment processing backends (glsl, arbfp and fixed) out
from cogl-material.c into their own cogl-material-{glsl,arbfp,fixed}.c
files in an effort to help and keep cogl-material.c maintainable.
2010-06-30 15:04:10 +01:00
Neil Roberts
d99a8bf694 cogl: Add const to some pointer arguments
Some of the arguments to the material and path functions were taking a
pointer to a CoglColor or an array of floats that was not intended to
be written to but were not marked with const.
2010-06-28 15:25:19 +01:00
Robert Bragg
de1f4885dd material: Avoid possibly reading invalid memory
in _cogl_material_prune_empty_layer_difference we sometimes unref the
given layer before dereferencing it to get a pointer to its parent. This
defers the unref until after we have fetched the parent pointer.
2010-06-24 17:05:26 +01:00
Emmanuele Bassi
0244eccfe3 Silence some compiler warnings 2010-06-22 13:48:53 +01:00
Neil Roberts
82e63a4753 cogl-texture: Replace the ensure_mipmaps virtual with pre_paint
Instead of the ensure_mipmaps virtual that is only called whenever the
texture is about to be rendered with a min filter that needs the
mipmap, there is now a pre_paint virtual that is always called when
the texture is about to be painted in any way. It has a flags
parameter which is used to specify whether the mipmap will be needed.

This is useful for CoglTexturePixmapX11 because it needs to do stuff
before painting that is unrelated to mipmapping.
2010-06-22 11:47:33 +01:00
Robert Bragg
9234757c35 material: fix a typo in _cogl_material_journal_unref
The function had a line like:

CoglMaterial *material =
    material = _cogl_material_pointer_from_handle (material_handle);

where the duplicate "material =" wasn't intended, so this patch removes
it.
2010-06-22 11:21:09 +01:00
Neil Roberts
f95cdfc11e cogl-material: Don't use the symbol 'near'
The window headers contain the line

 #define near

so it's not possible to use the symbol 'near' in code that's portable
to Windows. This replaces it with 'near_val'.

I think the define is meant to improve compatibility with code written
for Windows 3.1 where near would be a keyword to make it a smaller
pointer size.
2010-06-18 15:34:48 +01:00
Robert Bragg
c506b58ecf material: Adds arbfp compile counter
This adds a uprof counter to track how many ARBfp programs get compiled
by an application.
2010-06-15 15:26:28 +01:00
Robert Bragg
ba067c3da8 material: Support sparse ARBfp program generation
We don't need to generate a new ARBfp program for every material created
if we can find an ancestor whos state will result in the same program
being generated.

The more code we can have adopt the coding pattern of deriving their
materials from other similar materials using cogl_material_copy() the
more likely this metric will be good enough on its own to minimize the
set of arbfp programs necessary to support a given application.
2010-06-15 15:26:28 +01:00
Robert Bragg
a38e597aaa material: free any layer cache in material_set_parent
Previously in _cogl_material_pre_change_notify we manually freed the
layer caches of a material if we caused a reparent, but it makes more
sense to have _cogl_material_set_parent do this directly instead.
2010-06-15 15:26:28 +01:00
Robert Bragg
92e8fbfdc6 Adds an internal weak material mechanism
This adds a _cogl_material_weak_copy() function that can be used to
create materials that don't count as strong dependants on their parents.
This means the parent can be modified without worrying about how it will
affect weak materials. The material age of the parent can potentially be
queried to determine if a weak material might need to be re-created.
2010-06-15 15:26:28 +01:00
Robert Bragg
9689ec5aad material: Adds a simple material age getter
When we add support for weak materials it's expected that Clutter will
want to attach them as private data to other materials and it needs a
mechanism to determine when a weak material should be re-created because
its parent has changed somehow.

This adds the concept of a material age (internal only currently) which
increments whenever a material is modified. Clutter can then save the
age of the material which its weak materials are derived from and later
determine when the weak material may be invalid.
2010-06-15 15:26:28 +01:00
Robert Bragg
20910a1215 material: generalize how we compare simple material properties
In _cogl_material_equal we were repeating the same code pattern to
compare several of the state groups so this just adds
simple_property_equal function that's now used instead.
2010-06-15 15:26:28 +01:00
Robert Bragg
dab4ddc718 material: Make CoglMaterial responsible for depth state
This redirects the legacy depth testing APIs through CoglMaterial and
adds a new experimental cogl_material_ API for handling the depth
testing state.

This adds the following new functions:
cogl_material_set_depth_test_enabled
cogl_material_get_depth_test_enabled
cogl_material_set_depth_writing_enabled
cogl_material_get_depth_writing_enabled
cogl_material_set_depth_test_function
cogl_material_get_depth_test_function
cogl_material_set_depth_range
cogl_material_get_depth_range

As with other experimental Cogl API you need to define
COGL_ENABLE_EXPERIMENTAL_API to access them and their stability isn't
yet guaranteed.
2010-06-15 15:26:28 +01:00
Robert Bragg
b57f68245c material: Adds simple breadcrumb debugging mechanism
Since it can sometimes be awkward to figure out where a particular
material came from when debugging, this adds a breadcrumb mechanism that
lets you associate a const string with a material that may give a clue
about its origin.
2010-06-15 15:26:27 +01:00
Robert Bragg
6b5281401c material: remove override options from _cogl_material_equal
As a follow on to using cogl_material_copy instead of flush options this
patch now removes the ability to pass flush options to
_cogl_material_equal which is the final reference to the
CoglMaterialFlushOptions mechanism.
2010-06-15 15:26:27 +01:00
Robert Bragg
3979de6982 cogl: remove _cogl_material_flush_gl_state flush options
Since cogl_material_copy should now be cheap to use we can simplify
how we handle fallbacks and wrap mode overrides etc by simply copying
the original material and making our override changes on the new
material. This avoids the need for a sideband state structure that has
been growing in size and makes flushing material state more complex.

Note the plan is to eventually use weak materials for these override
materials and attach these as private data to the original materials so
we aren't making so many one-shot materials.
2010-06-15 15:26:27 +01:00
Robert Bragg
1cc3ae6944 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.
2010-06-15 15:26:27 +01:00
Robert Bragg
6c8b8cef9e material: fix #ifdef guard around use of _TEXTURE_RECTANGLE
We were incorrectly guarding the use of GL_TEXTURE_RECTANGLE_ARB with
ifdef ARB_texture_rectangle instead of ifdef GL_ARB_texture_rectangle
which broke test-cogl-texture-rectangle.
2010-06-15 15:26:27 +01:00
Robert Bragg
26628553ed blend-strings: removes the AUTO_COMPOSITE function enum
This was mistakenly added some time ago because at some point when we
were discussing how to handle premultiplied alpha in Clutter/Cogl we
were considering having a magic "just do the right thing" option which
was later abandoned.
2010-06-15 15:26:27 +01:00