We need to remove this so that adding images to the end of the list of images
for an ImageLayer doesn't force composition to happen even if nothing else
has changed.
--HG--
extra : commitid : 8ycbGC3r3R2
extra : rebase_source : 2db81491d697d45e2e2c4605a3a00db7aca8586b
We need this change so that when ImageHost has a next image to display
more than one composition-interval in the future, we skip the actual
compositing work in those intermediate composition(s) if nothing else
has changed.
This change is a little bit scary since it breaks any code that was
previously assuming ScheduleComposition would actually update the screen.
However, that code was already broken for BasicCompositor.
--HG--
extra : commitid : FUFFaJDSwfu
extra : rebase_source : eb2d5fadd5dbd417e7a25c05b148bd8964e8c2a3
For frame statistics to work properly, we have to notify an ImageContainer
when it has been composited. This requires a few changes, which have
been lumped together in this patch:
-- Create PImageContainer and ImageContainerParent/ImageContainerChild.
-- Add mFrameID and mProducerID everywhere we're passing around images.
-- Route composition notifications from the compositor back to
ImageContainerChild.
--HG--
extra : commitid : 7atVkOgdEhG
extra : rebase_source : caaba6a708ed267368df44609fb047abde9c3ca1
When ImageContainer and ImageClient are managing a list of images, the
individual Image serial numbers are no longer enough to detect whether the
state has changed.
--HG--
extra : commitid : 3EkV17zqHVt
extra : rebase_source : 92e7479b39bf499db0cc1db3534317c2fb8de68a
We'll need this later so ImageHost can select the correct image to use.
Adding a TimeStamp parameter to BeginFrame is a bit annoying since BeginFrame
is overridden by every subclass. It's a bit more convenient to just call a
separate non-virtual method just before we call BeginFrame.
--HG--
extra : commitid : 3QUviDozXL
extra : rebase_source : ba6bc3d846606f3f7598f4a3b687b68746067ce3
The picture rect logically belongs with the texture, and later patches will
make OpUseTexture take multiple textures, each of which needs its own
picture rect.
--HG--
extra : commitid : AF2YszcgNe1
extra : rebase_source : 826332fc5dfec25b712ff62e5812cb00aaac81a4
This makes normal playback consistent with the buffering state, which already
does this. We'll also need this when we handle multiple images, because then
we need to hande the entire queue of images to the ImageContainer without
pulling any of them off the queue.
--HG--
extra : commitid : B1zt05nhVn6
extra : rebase_source : af8a411f5f9be9d3a8c15633b2c4ad446be6d13f
This makes MediaDecoderStateMachine::GetVideoStreamPosition compute a
time that's more consistent with the audio clock.
--HG--
extra : commitid : JWQlELWFpTu
extra : rebase_source : 7c472cfd5188f5a948af16b734c25ddeed3651f8
There is some ambiguity about whether ScheduleComposite will necessarily
trigger a composite all the way to nsWindow::DrawWindowUnderlay. Android
robocop tests assume it will, because they rely on DrawWindowOverlay
being called so they can take a screenshot and make progress,
but this is a very fragile assumption. They also rely on the entire
window being painted, which is also a fragile assumption.
This patch improves the situation by explicitly invalidating the current
window area when Android Java code needs to trigger a composite. This avoids
regressions from future patches in this series which make composition bail
out when there is nothing invalid.
The resulting setup is still a bit fragile for my taste but I'm not sure
what the ideal solution would be.
--HG--
extra : commitid : 3t3xqRdZs24
extra : rebase_source : b23749613663ca805484776ccf5e36b4ff00e3fe
Otherwise we can get a crash with the following stack:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 14711]
0x5d99974e in mozilla::gl::GLContext::BeforeGLCall (this=0x6dbf0800,
funcName=0x60f251a4 <mozilla::gl::GLContext::raw_fDeleteProgram(unsigned int)::__PRETTY_FUNCTION__> "void mozilla::gl::GLContext::raw_fDeleteProgram(GLuint)") at /home/roc/mozilla-inbound/gfx/gl/GLContext.h:683
683 MOZ_ASSERT(IsCurrent());
(gdb) where
#0 0x5d99974e in mozilla::gl::GLContext::BeforeGLCall (this=0x6dbf0800,
funcName=0x60f251a4 <mozilla::gl::GLContext::raw_fDeleteProgram(unsigned int)::__PRETTY_FUNCTION__> "void mozilla::gl::GLContext::raw_fDeleteProgram(GLuint)") at /home/roc/mozilla-inbound/gfx/gl/GLContext.h:683
#1 0x5d99bed6 in mozilla::gl::GLContext::raw_fDeleteProgram (this=0x6dbf0800, program=210003)
at /home/roc/mozilla-inbound/gfx/gl/GLContext.h:2232
#2 0x5d99c10a in mozilla::gl::GLContext::fDeleteProgram (this=0x6dbf0800, program=210003)
at /home/roc/mozilla-inbound/gfx/gl/GLContext.h:2270
#3 0x5daa0ae6 in mozilla::layers::ShaderProgramOGL::~ShaderProgramOGL (this=0x6d7df000, __in_chrg=<optimized out>)
at /home/roc/mozilla-inbound/gfx/layers/opengl/OGLShaderProgram.cpp:491
#4 0x5da86bdc in mozilla::layers::CompositorOGL::CleanupResources (this=0x67ae4d70)
at /home/roc/mozilla-inbound/gfx/layers/opengl/CompositorOGL.cpp:177
--HG--
extra : commitid : LPnSogXNNio
extra : rebase_source : 0564dd5688916271c4a709ae6f15ba7ad493a761
On some Android versions, GraphicBuffer.h ends up including libui's
hardware.h, which #defines the symbols version_minor and version_major, which
are used as field names in Ogg Theora's th_info struct. Later patches will
require some files to include both Theora headers and LayerTypes.h.
--HG--
extra : commitid : HM7e5zfyoAz
extra : rebase_source : de8dd7e5bc6bfcfeb729cfab86fe44b8bb12ab80
Robert O'Callahan
Olli Pettay