Commit Graph

2707 Commits

Author SHA1 Message Date
Milo Casagrande
4a4d8153f8 [l10n] Updated Italian translation. 2012-09-29 18:02:19 +02:00
Neil Roberts
9563799655 Fix the terminator in one of the extension lists
The list of extension names in COGL_EXT_BEGIN should be a zero
separated list of strings which is terminated by an empty string. The
name for the GL_ARB_shader_objects extension was missing the zero
separator so presumably it was relying on the following byte to happen
to be a zero in order not to crash.

Reviewed-by: Robert Bragg <robert@linux.intel.com>

(cherry picked from commit f63381f23fa8b0b17e030561940b8a38efff221f)
2012-09-28 17:15:11 +01:00
Neil Roberts
f9e7f8df94 winsys-glx: Remove the vblank counter feature when indirect rendering
Previously when Cogl detects that the GLX context is indirect it
resets the function pointers for the VBLANK_COUNTER feature to NULL.
However it wasn't removing the VBLANK_COUNTER feature flag. Some other
parts of the winsys check for that feature flag rather than checking
whether the pointer is NULL so it would end up calling an invalid
function pointer and crashing. This just fixes it to also clear the
feature flag.

https://bugzilla.gnome.org/show_bug.cgi?id=684917

Reviewed-by: Robert Bragg <robert@linux.intel.com>

(cherry picked from commit e947c713a541086f80a308d22774229f0720196a)
2012-09-28 17:15:11 +01:00
Chun-wei Fan
cfdc1c95eb Clean up Visual C++ projects
Since commit 7253c5ca (Bug 682071-cogl/cogl-sdl.h: MSVC: Link to SDL when
apps are built), on Visual C++ we link to SDL.lib and SDLmain.lib using
a #pragma comment directive in cogl/cogl-sdl.h, so we no longer need
specific project configs for Cogl programs which build against a Cogl built
with SDL.  This removes those unneeded configs.

Also "install" cogl/cogl-sdl.h when Cogl is built with the SDL winsys, as
it is a needed header
2012-09-28 18:45:41 +08:00
Chun-wei Fan
c0e9d9235a Update Visual C++ projects
-Make "install" parts for VS9 and VS10 more consistent with each other
-Create the .def files in the respective intermediate/.obj folders, so that
 it is cleaner duringg the build
-Make up for missed files to "install"
2012-09-28 17:14:28 +08:00
Aurimas Černius
244ee67284 Updated Lithuanian translation 2012-09-27 22:06:21 +03:00
Мирослав Николић
3dd87cbf54 Updated Serbian translation 2012-09-27 05:45:39 +02:00
Ihar Hrachyshka
e2ad3cb86b Updated Belarusian translation. 2012-09-26 15:09:36 +03:00
Rūdolfs Mazurs
248182d3d1 Updated Latvian translation 2012-09-26 12:02:54 +03:00
Tom Tryfonidis
e3373e7cb3 Updated Greek translation 2012-09-26 11:21:43 +03:00
Andika Triwidada
30d3f1a7d3 Update French translation 2012-09-26 08:48:27 +02:00
Andika Triwidada
441e4cf8f6 Updated Indonesian translation 2012-09-26 09:00:39 +07:00
Robert Bragg
a8fa6c7c4d Post-release version bump to 1.12.1 2012-09-24 14:19:41 +01:00
Robert Bragg
fe198cf424 Merge branch 'cogl-1.12-release' into cogl-1.12 2012-09-24 14:08:22 +01:00
Robert Bragg
b1aafcc15f Release 1.12.0 (release) 2012-09-24 13:46:40 +01:00
Robert Bragg
d32bbaf0e0 Updates NEWS for the 1.12.0 release 2012-09-24 13:46:01 +01:00
Yuri Myasoedov
95ad4ce2ef Updated Russian translation
(cherry picked from commit c34e8bb0ca44c90d05753e6790727a462d7fd732)
2012-09-24 13:38:18 +01:00
Mario Blättermann
8a533157b7 [l10n] Updated German translation
(cherry picked from commit b9afd31baabf1e139f51e6042104ab0904e77488)
2012-09-24 13:38:09 +01:00
YunQiang Su
b6ffa02d2f Update zh_CN translation
(cherry picked from commit 7003bd0bf092ea38eab4491ecb56cb8db155190a)
2012-09-24 13:37:55 +01:00
Ani Peter
eb18e471ef Updated Malayalam file
(cherry picked from commit 9cdaf2e8e13c9b255343357e95038e1881578bf7)
2012-09-24 13:37:38 +01:00
Nishio Futoshi
6e9ae288e1 [l10n] Update Japanese translation
(cherry picked from commit 857f9d336ba1940211becfdd637dcfe08b52e079)
2012-09-24 13:37:28 +01:00
Dr.T.Vasudevan
59e6078454 updated Tamil translation
(cherry picked from commit af265ab5d2e2f2dab8294c549e872ef3b2fdf1d3)
2012-09-24 13:37:18 +01:00
Dr.T.Vasudevan
19fa0efb72 updated Tamil translation
(cherry picked from commit 1b7743f0c29981b4d91ca9faa16fa4543662c2a7)
2012-09-24 13:37:07 +01:00
Alexander Shopov
700e7df85d Updated Bulgarian translation
(cherry picked from commit 93d7defa7fbb1a9ccb56beef882699008b7b9d44)
2012-09-24 13:36:53 +01:00
Rajesh Ranjan
fc0c3c8be7 hindi update
(cherry picked from commit 38201324e9543fa4c4b613829b6c1199778f02d5)
2012-09-24 13:36:42 +01:00
Chun-wei Fan
7253c5ca29 Bug 682071-cogl/cogl-sdl.h: MSVC: Link to SDL when apps are built
Link to SDL.lib and SDLmain.lib if Cogl was built with the SDL winsys.

Recent changes to the SDL winsys introduced a direct dependency to
SDLmain.lib (and hence SDL.lib) when programs are built, causing linker
errors to appear when any programs using cogl (with the SDL winsys built
in) are built.

Since we cannot determine whether a Cogl build is built with the SDL winsys
at build time easily, we could use #pragma comment (lib, ...) whenever
cogl-sdl.h is included by cogl.h so that SDLmain.lib and SDL.lib is linked
into the resulting binary, so that the program can link and run correctly.

This does not add any external dependencies as the Cogl DLL already depends
on SDL.dll when it is built with the SDL winsys.

https://bugzilla.gnome.org/show_bug.cgi?id=682071

(cherry picked from commit 2921d2a4d9c79f1ca7530171e0dfa8c945607bc7)
2012-09-24 13:35:18 +01:00
Carles Ferrando
e3e53f2c70 [l10n]Updated Catalan (Valencian) translation
(cherry picked from commit 829e090deb5decbeb60519c26199a403ae8174b0)
2012-09-24 13:35:05 +01:00
Gil Forcada
192920e351 [l10n] Updated Catalan translation
(cherry picked from commit 1a39fb66c1531ef53c8a700fc85bde1fb97bc624)
2012-09-24 13:34:53 +01:00
Ask H. Larsen
ed32a052e7 Updated Danish translation
(cherry picked from commit bb5220dbd98517544d0ff8676061024fd8a361f3)
2012-09-24 13:34:42 +01:00
Chun-wei Fan
b743d95e21 Bug 682071-cogl/cogl-sdl.h: MSVC: Link to SDL when apps are built
Link to SDL.lib and SDLmain.lib if Cogl was built with the SDL winsys.

Recent changes to the SDL winsys introduced a direct dependency to
SDLmain.lib (and hence SDL.lib) when programs are built, causing linker
errors to appear when any programs using cogl (with the SDL winsys built
in) are built.

Since we cannot determine whether a Cogl build is built with the SDL winsys
at build time easily, we could use #pragma comment (lib, ...) whenever
cogl-sdl.h is included by cogl.h so that SDLmain.lib and SDL.lib is linked
into the resulting binary, so that the program can link and run correctly.

This does not add any external dependencies as the Cogl DLL already depends
on SDL.dll when it is built with the SDL winsys.
2012-09-19 15:33:47 +08:00
Robert Bragg
f904f2099a Post-release version bump to 1.11.7 2012-09-17 23:47:25 +01:00
Robert Bragg
3aa7bdaf49 Release 1.11.6 (snapshot) 2012-09-17 23:11:34 +01:00
Robert Bragg
7429fb8df7 Updates NEWS for the 1.11.6 release 2012-09-17 23:08:45 +01:00
Neil Roberts
1b3a7ac0ab gles2-context: Don't split the wrapper snippet into two
We don't need to split the wrapper snippet into two separate parts
because it should be ok to declare the flip uniform in the middle of
the shader as long as it is somewhere in the global scope. Therefore
we can just declare it right before the definition for the replacement
main function. This is important because we don't want to put anything
at the top of the application's shader in case it is using a
'#version' directive. In that case moving it to anything other than
the first line would break things.

This patch also adds a marker in a comment around the wrapper snippet
so that we can easily locate the snippet when glGetShaderSource is
called and remove it.

The wrapper for glGetAttachedShaders has been removed because there
are no longer any additional shaders attached to the program so we can
just let GL handle it directly.

Reviewed-by: Robert Bragg <robert@linux.intel.com>

(cherry picked from commit dbd92e24ae61dcbe7ef26f61c9117c5516a7fa87)
2012-09-17 23:06:21 +01:00
Robert Bragg
5f3768f086 gles2-context: only insert prelude + wrapper into strv once
In our wrapper for glShaderSource we special case when a vertex shader
is being specified so we can sneak in a wrapper for the main function to
potentially flip all rendering upside down for better integration with
Cogl.

Previously we were appending the wrapper to all the sub-strings passed
via the vector of strings to glShaderSource but we now grow the vector
instead and insert the prelude and wrapper strings into the beginning
and end of the vector respectively so we should only have one copy for a
single shader.

Reviewed-by: Neil Roberts <neil@linux.intel.com>

(cherry picked from commit d2904d518718e3fbf4441abe2c2bcfd63edfd64b)
2012-09-17 23:06:21 +01:00
Daniel Stone
a491789637 cogl-gles2-context: Append wrapper shader to user shaders
The SGX GLSL compiler refuses to accept shaders of the form:
    void foo();
    void bar() {
        foo();
    }
where foo is undefined at glShaderSource() time, left for definition at
link time.  To work around this, simply append the wrapper shader to
user shaders, rather than building a separate shader that's always
linked with user shaders.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Robert Bragg <robert@linux.intel.com>

(cherry picked from commit 96f02c445f763ace088be71dc32b3374b2cdbab2)
2012-09-17 23:06:21 +01:00
Robert Bragg
50ba5fa130 gpu-info: Match latest Mesa strings
Mesa now reports a vendor string of "Mesa Project" instead of "VMWare,
Inc." and the software rasterizer renderer string is now "Software
Rasterizer". This update cogl-gpu-info.c to recognize these new strings.

Thanks to Alexander Larsson for the original patch.

https://bugzilla.gnome.org/show_bug.cgi?id=683818

(cherry picked from commit dfacbbd96f3fbadaffa4a76dfd71c47ece6ed6a3)
2012-09-17 23:06:20 +01:00
Robert Bragg
0f0ee4a909 texture: Add a context pointer to each texture
As part of our on-going goal to remove our dependence on a global Cogl
context this patch adds a pointer to the context to each CoglTexture
so that the various texture apis no longer need to use
_COGL_GET_CONTEXT.

Reviewed-by: Neil Roberts <neil@linux.intel.com>

(cherry picked from commit 83131072eea395f18ab0525ea2446f443a6033b1)
2012-09-17 23:06:20 +01:00
Robert Bragg
ea3d8eca91 Don't take internal references on the context
We want applications to fully control the lifetime of a CoglContext
without having to worry that internal resources (such as the default
2d,3d and rectangle textures, or any caches we maintain) could result in
circular references that keep the context alive. We also want to avoid
making CoglContext into a special kind of object that isn't ref-counted
or that can't be used with object apis such as
cogl_object_set_user_data. Being able to reliably destroy the context is
important on platforms such as Android where you may be required
bring-up and tear-down a CoglContext numerous times throughout the
applications lifetime. A dissadvantage of this policy is that it is now
possible to leave other object such as framebuffers in an inconsistent
state if the context is unreferenced and destroyed. The documentation
states that all objects that directly or indirectly depend on a context
that has been destroyed will be left in an inconsistent state and must
not be accessed thereafter. Applications (such as Android applications)
that need to cleanly destroy and re-create Cogl resources should make
sure to manually unref these dependant objects before destroying the
context.

Reviewed-by: Neil Roberts <neil@linux.intel.com>

(cherry picked from commit 23ce51beba1bb739a224e47614a59327dfbb65af)
2012-09-17 23:06:20 +01:00
Robert Bragg
2942fd362d Avoid referencing file scope context in _context_new()
cogl_context_new() had a mixture of references to the file scope context
variable (_context) and the local (context) variable. This renames the
file scope variable to _cogl_context to catch unnecessary references to
the old name and fixes the code accordingly to reference the local
variable instead.

Reviewed-by: Neil Roberts <neil@linux.intel.com>

(cherry picked from commit 33a9397ee1ae1729200be2e5084cf43cebb64289)
2012-09-17 23:06:20 +01:00
Robert Bragg
66169276f4 texture-2d: improve new_from_foreign error handling
There were lots of places where cogl_texture_2d_new_from_foreign would
simply return NULL without returning a corresponding error. We now
return an error wherever we are returning NULL except in cases where the
user provided invalid data.

Reviewed-by: Neil Roberts <neil@linux.intel.com>

(cherry picked from commit a1efc9405a13ac8aaf692c5f631a3b8f95d2f259)
2012-09-17 23:06:20 +01:00
Ihar Hrachyshka
a676f700a3 Updated Belarusian translation.
(cherry picked from commit ded3f51fd6f853ed179bcdf585350b64bae4f80c)
2012-09-17 23:06:20 +01:00
Marek Černocký
3c792b23d7 Updated Czech translation
(cherry picked from commit 0638b17fbde629e5dfaf7e28ecb6da43b408b3ba)
2012-09-17 23:06:20 +01:00
Robert Bragg
7b5b0bef83 framebuffer: drop _ALLOCATE_FLAG_DEPTH24_STENCIL8
There are two extensions, GL_OES_packed_depth_stencil and
GL_EXT_packed_depth_stencil, that inform us that the hardware supports
packing the depth and stencil values together into one format.

The OES extension is the GLES equivalent of the EXT extension and the
two extensions provide the same enums with basically the same semantics,
except that the EXT extension is a lot more wordy due to a larger number
of features in the full OpenGL api and the OES extension has some
asymmetric limitations on when the GL_DEPTH_STENCIL and
GL_DEPTH24_STENCIL8 enums can be used as internal formats.

GL_OES_packed_depth_stencil doesn't allow the GL_DEPTH_STENCIL enum
to be passed to glRenderbufferStorage (GL_DEPTH24_STENCIL8 should be
used instead) and GL_OES_packed_depth_stencil doesn't allow
GL_DEPTH24_STENCIL8 to be passed as an internal format to glTexImage2D.

We had been handling the two extensions differently in Cogl by calling
try_creating_fbo with different flags depending on whether the OES or
EXT extension was available and passing GL_DEPTH_STENCIL to
glRenderbufferStorage when we have the EXT extension or
GL_DEPTH24_STENCIL8 with the OES extension.

To localize the code that deals with the differences between the
extensions this patch does away with the need for separate flags
so we now just have COGL_OFFSCREEN_ALLOCATE_FLAG_DEPTH_DEPTH_STENCIL and
right before calling glRenderbufferStorage we check which extension we
are using to decide whether to use the GL_DEPTH_STENCIL or
GL_DEPTH24_STENCIL8 enums.

(cherry picked from commit 88a05fac6609f88c0f46d9df2611d9fbaf159939)
2012-09-17 23:06:19 +01:00
Alban Browaeys
8d09b93572 meta-texture: Fix textures[] index
textures[iter_y.index * n_y_spans + iter_x.index]
only works for vertical rectangles when n_x_spans > 0 (ie x != {0} )

is also wrong for horizontal rectangles ( x = {0, 1, 2, 3} , y = {0, 1}
-> second line will start at 2 = iter_y.index * n_y_spans + iter_x.index
-> iteration are 0, 1, 2, 3, \n 2, 3, 4, 5 instead of 0, 1, 2, 3 \n 4, 5, 6, 7

Reviewed-by: Robert Bragg <robert@linux.inte.com>

(cherry picked from commit bf0d187f1b5423b9ce1281aab1333fa2dfb9863f)
2012-09-17 11:47:04 +01:00
Bruce Cowan
2bc5c494a6 Updated British English translation
(cherry picked from commit 3899ad34c94bde8403f6c53706fcd41644bacb06)
2012-09-17 11:45:53 +01:00
Gabor Kelemen
4ffb423199 Updated Hungarian translation
(cherry picked from commit 228216123f9eb8d741a05d5fec3756eaf7c7110d)
2012-09-17 11:45:40 +01:00
Neil Roberts
cb834c95e9 pipeline: Fix the layer index used when pruning layers
When pruning a pipeline to a set number of layers it records the index
of the first layer after the given number of layers have been found.
This is stored in a variable called 'first_index_to_prune' implying
that this layer should be included in the layers to be pruned. However
the subsequent if-statement was only pruning layers with an index
greater than the recorded index so it would presumably only prune the
following layers. This patch fixes it to use '>=' instead.

https://bugzilla.gnome.org/show_bug.cgi?id=683414

Reviewed-by: Robert Bragg <robert@linux.intel.com>

(cherry picked from commit d3063e8dea92a8f668acef6435cc68e0c901dc8d)
2012-09-07 11:44:41 +01:00
Neil Roberts
a18c97798f pipeline: Ensure the pipeline layer cache is freed when pruning layers
When pruning layers from a pipeline the pipeline cache would once be
freed due to the call to pre_change_notify but it would immediately be
recreated again when foreach_layer_internal is called. When n_layers
is later set to 0 it would end up with an invalid cache lying around.
This patch changes the order so that it will iterate the layers first
before triggering the pre-change notify so that the cache will be
cleared correctly.

https://bugzilla.gnome.org/show_bug.cgi?id=683414

Reviewed-by: Robert Bragg <robert@linux.intel.com>

(cherry picked from commit 1c8efdc838cc5ace380365cb54e0741645856edf)
2012-09-07 11:44:41 +01:00
Neil Roberts
27d74f09df kms: Fix the version check for GBM
The check for whether to use ‘stride’ instead of ‘pitch’ from the GBM
API tries to check whether the GBM version is >= 8.1.0. However it was
comparing the major and micro components independently so any version
with the minor part set to 0 would fail. The GBM version in Mesa
master is now 9.0.0 which breaks it. This patch changes it to check
the version using the COGL_VERSION_ENCODE macro instead.

Reviewed-by: Robert Bragg <robert@linux.intel.com>

(cherry picked from commit 38f1dc58b35023f9e6bbc0db746b1554bd0377fc)
2012-09-06 18:30:30 +01:00