| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of maintaining optional support for using deps/eglib instead of
glib we now always use this code. This way the code is always being
tested so we are less likely to break support for building Cogl without
glib, and because the api no longer needs to be swappable we are free to
make Cogl specific changes to the api.
So that we can still support glib integration the eglib library has been
renamed to ulib and symbols are prefixed with 'u_' or 'U_' so that they
don't clash with glib.
GLib integration for GError and GType continues to be supported.
|
|
|
|
|
|
|
|
|
|
|
| |
This removes a number of ./configure time compiler checks in favour of
compile time checks and also assumes you are building with a modern
compiler. For example we assume <stdint.h> and <inttypes.h> headers are
available.
These changes will make it easier to maintain the ability to build Cogl
on Android via an Android.mk file where it's not usual to use the build
system to check compiler features.
|
|
|
|
|
| |
This simply removes a couple of eglib apis that aren't needed by Cogl
and the corresponding tests.
|
|
|
|
|
|
| |
The eglib as found in the mono git repository can be built separately
from the rest of mono, but to avoid maintaining parallel build scripts
for eglib this strips away the autotool related files from deps/eglib.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds a deps/eglib api which is a small MIT licensed library that is
api compatible with enough of the glib api to run Cogl. This replaces
deps/glib and deps/gmodule for use when Cogl is built with
--enable-standalone.
This code is based on eglib from the mono project, with a few updates
and additions to handle apis that Cogl uses that mono doesn't.
As a reference point for monitoring eglib changes in the mono project,
this code is based on commit d50c836721f22df74a45185d51d0b9d37c370e33
|
|
|
|
|
| |
This fixes a few build issues with compiling Cogl against our internal
deps/glib and deps/gmodule libraries.
|
|
|
|
|
|
| |
This now simply defines G_STATIC_ASSERT in terms of _Static_assert if __GNUC__ is
defined, and disables the assertion if not. This stops lots of verbose
compiler warnings caused by the previous typedef based approach.
|
|
|
|
|
| |
This upgrades stb_image to 1.33 which notably fixes some problems with
inconsistently using #if and #ifdef to check the STBI_SIMD define.
|
|
|
|
|
|
| |
This fixes a warning by making the _COGL_RETURN_VAL_IF_FAIL check in
_cogl_offscreen_gl_allocate return FALSE not NULL, since the api returns
a CoglBool not a pointer.
|
|
|
|
|
|
|
|
|
| |
This makes sure video textures being uploaded via
video_texture_new_from_data are allocated before the function returns.
This function create a CoglBitmap to wrap the data from gstreamer and by
allowing cogl to allocate the texture lazily it's possible that the data
being pointed to by the bitmap won't remain valid until we actually come
to allocate the texture.
|
|
|
|
|
|
|
|
|
|
|
|
| |
Not doing so leads to the following error, if stddef.h is not included
indirectly through EGL headers:
| libdrm/drm.h:132:2: error: unknown type name 'size_t'
| size_t name_len; /**< Length of name buffer */
Signed-off-by: Andreas Oberritter <obi@saftware.de>
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The _cogl_egl_texture_2d_new_from_image function has a CoglError
argument which implies that it is unlike the other texture
constructors and returns errors immediately rather than having a
delayed-allocation mechanism. cogl_wayland_texture_2d_new_from_buffer
which calls it is also like this. We can't rely on delayed-allocation
semantics for this without changing the applications because the
texture needs to be allocated before the corresponding EGLImage is
destroyed. This patch just makes it immediately allocate.
A better patch might be to remove the error argument to make it
obvious that there are delayed-allocation semantics and then fix all
of the applications.
This was breaking Cogland and Mutter.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The previous implementation was dereferencing the sample pointer in
order to get the offset to subtract from the member pointer. The
resulting value is then only used to get a pointer to the member in
order to calculate the offset so it doesn't actually read from the
memory location and shouldn't cause any problems. However this is
probably technically invalid and could have undefined behaviour. It
looks like clang takes advantage of this undefined behaviour and
doesn't actually offset the pointer. It also generates a warning when
it does this.
This patch splits the _cogl_container_of macro into two
implementations. Previously the macro was always used in the list
iterator macros like this:
SomeType *sample = _cogl_container_of(list_node, sample, link)
Instead of doing that there is now a new macro called
_cogl_list_set_iterator which explicitly assigns to the sample pointer
with an initial value before assigning to it again with the real
offset. This redundant initialisation gets optimised out by compiler.
The second macro is still called _cogl_container_of but instead of
taking a sample pointer it just directly takes the type name. That way
it can use the standard offsetof macro.
https://bugzilla.gnome.org/show_bug.cgi?id=723530
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If the URI argument given on the command line doesn't look like a URI
then the basic video player will now try to parse it as a GStreamer
pipeline description. The pipeline must contain a coglsink element
which the player will look for. This can be used to test Cogl GST. For
example, to test NV12 this can be used:
./cogl-basic-video-player 'videotestsrc !
video/x-raw,format=(string)NV12 !
coglsink'
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If no context is set on the CoglGstVideoSink then it would previously
call gst_caps_ref with a NULL pointer. This patch makes it just return
NULL instead. I think that is a valid thing to do because that is what
gst_base_sink_default_get_caps does. If we don't do this then it's not
possible to use CoglGstVideoSink with GstParse because that tries to
link the pipeline after parsing the string. That was previously
causing a critical error because the freshly parsed sink doesn't have
a CoglContext.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
Reviewed-by: Lionel Landwerlin <llandwerlin@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds a cogl-gst renderer to decode NV12 data. NV12 is split into
two buffers, one for the luma component and another for the two
chrominance components at a quarter of the resolution. The second
buffer is uploaded to a two-component RG texture. RG-component
textures are only supported if COGL_FEATURE_ID_TEXTURE_RG is
advertised by Cogl so the NV12 cap is also only advertised when that
is available.
Based on a patch by Lionel Landwerlin which was in turn based on a
patch from Edward Hervey and Matthieu Bouron for Clutter-Gst:
https://bugzilla.gnome.org/show_bug.cgi?id=712832
Reviewed-by: Robert Bragg <robert@linux.intel.com>
Reviewed-by: Lionel Landwerlin <llandwerlin@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 99a53c82e9ab0a1e5 removed the internal format argument when
uploading a video frame to a texture so that the format will just be
determined automatically from the image format. However this also
leaves the premultiplied state at the default and the default is TRUE.
That means that when we upload RGBA data Cogl will do a premultiplied
conversion on the CPU. We probably don't want to be putting a CPU
conversion in the way of video frames so this patch changes it to set
the premultiplied state to FALSE on the textures and then do the
premultiplied conversion in the shader.
This is particularly important for AYUV which uses the alpha channel
for the V component so doing a premultiplied conversion on the CPU
just creates garbage and messes up the image.
The RGB and RGBA renderers have each been split into two; one that
uses GLSL and one that uses a regular pipeline. The RGBA pipeline
without GLSL is then changed to use 2 layers so we that we can do the
premultiplied conversion in the second layer with a special layer
combine string.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
| |
|
|
|
|
| |
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
| |
The -L option makes curl follow redirections. This is needed for
downloading glext.h because khronos.org is using a redirect.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
| |
Since 248a76f5eac7e5ae4fb45208577f9a55360812a7 cogl.h can no longer be
included in internal source files so the WGL winsys was no longer
compiling.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
|
| |
The texture formats which have 10-bit components are necessarily not
byte-aligned so they ought to have the COGL_BITWISE_BIT flag. This
flag is enough to make them unique so they no longer need a new enum
value. The last enum value is therefore reset back to 1.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds COGL_PIXEL_FORMAT_RG_88 and COGL_TEXTURE_COMPONENTS_RG in
order to support two-component textures. The RG components for a
texture is only supported if COGL_FEATURE_ID_TEXTURE_RG is advertised.
This is only available on GL 3, GL 2 with the GL_ARB_texture_rg
extension or GLES with the GL_EXT_texture_rg extension. The RG pixel
format is always supported for images because Cogl can easily do the
conversion if an application uses this format to upload to a texture
with a different format.
If an application tries to create an RG texture when the feature isn't
supported then it will raise an error when the texture is allocated.
https://bugzilla.gnome.org/show_bug.cgi?id=712830
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The following changes are made to the documentation for CoglTexture:
• The description of the default value for the components property is
changed to say that it is always RGBA for textures created by the
‘_with_size’ textures. Previously it said that the default is based
on the pixel format used the first time data is set on the texture,
but this is only true if the data is set using a constructor.
• Added documentation for the CoglTextureComponents enum.
• Changed it to say that it _specifies_ what components are required
for sampling rather than determinging [sic] them.
• Added ‘Since: 1.18’ to
cogl_texture_{set,get}_{components,premultiplied}
• Changed the since tag for CoglTextureError from 2.0 to 1.8.
• Added documentation for COGL_TEXTURE_ERROR_{FORMAT,TYPE}.
• Added the following to the cogl2-sections.txt file:
COGL_TEXTURE_ERROR
CoglTextureError
cogl_texture_allocate
cogl_texture_set_components
cogl_texture_get_components
cogl_texture_set_premultiplied
cogl_texture_get_premultiplied
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This follows up on the proposal that was sent to the Cogl mailing list
to re-license from the LGPL to the MIT license:
http://lists.freedesktop.org/archives/cogl/2013-December/001465.html
Note: there was a copyright assignment policy in place for Clutter (and
therefore Cogl which was part of Clutter at the time) until the 11th of
June 2010 and so we only checked the details after that point (commit
0bbf50f905)
For each file, authors were identified via this Git command:
$ git blame -p -C -C -C20 -M -M10 0bbf50f905..HEAD
We received blanket approvals for re-licensing all Red Hat and Collabora
contributions which reduced how many people needed to be contacted
individually:
- http://lists.freedesktop.org/archives/cogl/2013-December/001470.html
- http://lists.freedesktop.org/archives/cogl/2014-January/001536.html
Individual approval requests were sent to all the other identified authors
who all confirmed the re-license on the Cogl mailinglist:
http://lists.freedesktop.org/archives/cogl/2014-January
As well as updating the copyright header in all sources files, the
COPYING file has been updated to reflect the license change and also
document the other licenses used in Cogl such as the SGI Free Software
License B, version 2.0 and the 3-clause BSD license.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously when a pipeline is added to the cache it would never be
removed. If the application is generating a lot of unique pipelines
this can end up effectively leaking a large number of resources
including the GL program objects. Arguably this isn't really a problem
because if the application is generating that many unique pipelines
then it is doing something wrong anyway. It also implies that it will
be recompiling shaders very often so the cache leaking will likely be
the least of the problems.
This patch makes it keep track of which pipelines in the cache are in
use. The cache now returns a struct representing the entry instead of
directly returning the pipeline. This entry contains a usage counter
which the pipeline backends can use to mark when there is a pipeline
alive that is using the cache entry. When the hash table decides that
it's a good time to prune some entries, it will make a list of all of
the pipelines that are not in use and then remove the least recently
used half of the pipelines. That way it is less likely to remove
pipelines that the application is actually regenerating often even if
they aren't in use all of the time.
When the cache is pruned the hash table makes a note of how small the
cache could be if it removed all of the unused pipelines. The hash
table starts pruning when there are more entries than twice this
minimum expected size. The idea is that if that case it hit then the
hash table is more than half full of useless pipelines so the
application is generating lots of redundant pipelines and it is a good
time to remove them.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
| |
Commit a750f80c6aaa was supposed to enable it for newer mesa but was wrong,
fix that.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The bug that prevented MESA_copy_sub_buffer to work for swrast /
llvmpipe got fixed in mesa 10.1 git so enable it for mesa 10.1+.
https://bugzilla.gnome.org/show_bug.cgi?id=721450
When landing the patch, it was tweaked to #include "cogl-version.h" to
avoid a compiler warning about COGL_VERSION_ENCODE being implicitly
defined. -- Robert Bragg
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
| |
This fixes the cogl_texture_get_components() prototype to have a return
type of CoglTextureComponents instead of CoglBool which was probably a
copy and paste error.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Texture allocation is now consistently handled lazily such that the
internal format can now be controlled using
cogl_texture_set_components() and cogl_texture_set_premultiplied()
before allocating the texture with cogl_texture_allocate(). This means
that the internal_format arguments to texture constructors are now
redundant and since most of the texture constructors now can't ever fail
the error arguments are also redundant. This now means we no longer
use CoglPixelFormat in the public api for describing the internal format
of textures which had been bad solution originally due to how specific
CoglPixelFormat is which is missleading when we don't support such
explicit control over the internal format.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This introduces the internal idea of texture loaders that track the
state for loading and allocating a texture. This defers a lot more work
until the texture is allocated.
There are several intentions to this change:
- provides a means for extending how textures are allocated without
requiring all the parameters to be supplied in a single _texture_new()
function call.
- allow us to remove the internal_format argument from all
_texture_new() apis since using CoglPixelFormat is bad way of
expressing the internal format constraints because it is too specific.
For now the internal_format arguments haven't actually been removed
but this patch does introduce replacement apis for controlling the
internal format:
cogl_texture_set_components() lets you specify what components your
texture needs when it is allocated.
cogl_texture_set_premultiplied() lets you specify whether a texture
data should be interpreted as premultiplied or not.
- Enable us to support asynchronous texture loading + allocation in the
future.
Of note, the _new_from_data() texture constructors all continue to
allocate textures immediately so that existing code doesn't need to be
adapted to manage the lifetime of the data being uploaded.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
When reading a texture back by first wrapping it as an offscreen
framebuffer and using _read_pixels_into_bitmap() we now make sure the
offscreen framebuffer has an internal format that matches the
meta-texture being read not that of the current sub-texture being
iterated. In the case of atlas textures the subtexture is a shared
texture whose format doesn't reflect the premultipled alpha status of
individual atlas-textures, nor whether the alpha component is valid.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
| |
This changes cogl_texture_get_width/height to return signed integers
just because unsigned integers in C often cause un-obvious arithmetic
issues due to implicit casting.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
| |
CoglPixelFormat is not a good way of describing the internal
format of a texture because it's too specific given that we don't
actually have exact knowledge of the internal format used by the driver.
This makes cogl_texture_get_format private and in the future we'll
provide a better way of querying the channels and their precision.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
| |
This defers checking the internal format and whether accelerated
migration is supported until allocating the texture.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
| |
Instead of throwing a CoglError exception if an application tries to
allocate a zero size atlas texture this make that a programmer error
instead.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
| |
The plan is to defer a lot more work in creating a texture until
allocation time. This means that for some texture backends we might not
know until after allocation whether the texture is sliced or can support
hardware repeating. This makes sure we trigger an allocation if either
of these are queried.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The plan is to defer a lot more work in creating a texture until
allocation time. This means we wont be able to assume that all textures
being used to render must have already been allocated when data was
specified.
The latest point at which we will generally require a texture to be
allocated will be when we need to know the underlying GL handle for a
texture and so this updates cogl_texture_get_gl_texture() to ensure the
texture is allocated.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The plan is to defer more of the work for creating a texture until
allocation time, but that means we won't be able to always assume
we can query the size of a texture when creating an offscreen
framebuffer from a texture (consider for example using
_texture_new_from_file() where the size isn't known until the file has
been loaded). This defers needing to know the size of the texture
underlying an offscreen framebuffer until calling
cogl_framebuffer_allocate().
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This ensures framebuffers are implicitly allocated when querying the
width, height or viewport width/height if the framebuffer's size is
currently unknown. The plan is to allow texture backends to defer
calculating the size of textures until they are allocated which in turn
means we won't know the size of offscreen framebuffers until the texture
has been allocated. Potentially we could be more specific about this in
the future and only ensure the texture is allocated, but for now it will
be simplest to just ensure the framebuffer is allocated.
Note: in the case of onscreen buffers which are always initialized with
a requested size we are careful to avoid triggering an allocation when
this is queried otherwise we will see recursion when the winsys code
queries the requested size during allocation.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
| |
Since we are planning on deferring more texture allocation work this
makes sure we don't query whether a texture is sliced until we know it
has been allocated.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This removes cogl_framebuffer_get_format() since the actual internal
format isn't strictly controlled by us. CoglFramebuffer::format has been
renamed to ::internal_format to make it clearer that it only really
represents the premultiplication status.
The plan is to make most of the work involved in creating a texture
happen lazily when allocating so this patch also changes
_cogl_framebuffer_init() to not take a format argument anymore since we
won't know the format of offscreen framebuffers until the framebuffer is
allocated, after the corresponding texture has been allocated. In the
case of offscreen framebuffers we now update the framebuffer
internal_format during allocation.
Reviewed-by: Neil Roberts <neil@linux.intel.com>
|
|
|
|
|
|
|
| |
This fixes a mistake in commit 637728dd89d51bc28 that was meant to fix
yv12/i420 support but when the patch was updated based on review to use
COGL_PIXEL_FORMAT_A instead of COGL_PIXEL_FORMAT_G the corresponding
glsl code wasn't also updated to sample the .a component.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
This means that we can't cache the journal read_pixels optimization.
https://bugzilla.gnome.org/show_bug.cgi?id=719582
Reviewed-by: Neil Roberts <neil@linux.intel.com>
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
| |
When projecting the bounding rectangle of a primitive it was using the
modelview matrix twice instead of the modelview and projection
matrices so it was coming out with garbage.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
| |
The test makes an L-shaped path that fills the whole framebuffer
except for the top right quadrant. It then clips to that and tries to
fill the framebuffer with a rectangle. Then it verifies that all of
the quadrants have the expected colour.
This is currently failing due to a bug in the primitive clipping.
Reviewed-by: Robert Bragg <robert@linux.intel.com>
|