This CL should be the last one in a series to finally unblock camera texture capture.
The SurfaceTexture.updateTexImage() calls are moved from the video renderers into MediaCodecVideoDecoder, and the destructor of the texture frames will signal MediaCodecVideoDecoder that the frame has returned. This CL also removes the SurfaceTexture from the native handle and only exposes the texture matrix instead, because only the video source should access the SurfaceTexture.
BUG=webrtc:4993
R=glaznev@webrtc.org, perkj@webrtc.org
Review URL: https://codereview.webrtc.org/1378033003 .
Cr-Commit-Position: refs/heads/master@{#10203}
The Java PeerConnection maintains a cached list of Java RtpSenders
and RtpReceivers so that the same objects are returned every time
getSenders() or getReceivers() is called. They are disposed of when
PeerConnection.dispose() is called, which will also dispose their
referenced MediaStreamTracks.
Review URL: https://codereview.webrtc.org/1368143002
Cr-Commit-Position: refs/heads/master@{#10189}
Every malformed URL should now produce an error message in JS, rather than
silently failing and possibly printing a warning message to the console (and
possibly crashing).
Also added some unit tests, and made "ParseIceServers" public.
BUG=445002
Review URL: https://codereview.webrtc.org/1344143002
Cr-Commit-Position: refs/heads/master@{#10186}
This allows to correctly report first frame event in applications which
use same remote video renderer for multiple calls.
R=wzh@webrtc.org
Review URL: https://codereview.webrtc.org/1384353002 .
Cr-Commit-Position: refs/heads/master@{#10176}
Fixed a problem where eglBase.makecurrent() could be called after the context had been released if SurfaceTextureHelper was first created and immedately disconnected.
Add the possibility to inject a thread to use instead of creating a new.
BUG= webrtc:4993
R=magjed@webrtc.org
Review URL: https://codereview.webrtc.org/1384923002 .
Cr-Commit-Position: refs/heads/master@{#10174}
This CL shouldn't make any functional changes. It adds a new VideoCapturerAndroid.nativeCreateVideoCapturer() instead of always using VideoCapturer.nativeCreateVideoCapturer(). The purpose is to simplify androidvideocapturer_jni and VideoCapturerAndroid.create(). This way, it is possible to use the ctor instead of VideoCapturerAndroid.init() to initialize variables, and they can be made final etc.
R=perkj@webrtc.org
Review URL: https://codereview.webrtc.org/1360173002 .
Cr-Commit-Position: refs/heads/master@{#10171}
VideoRendererGui may need to render incoming frames multiple times. We currently call VideoRenderer.renderFrameDone() while we still hold references to the OES texture. This CL makes a deep copy of the OES texture before calling renderFrameDone(). This will truly release the dependency to the incoming frame, so that video textures sources can rely on the renderFrameDone() callback.
This CL is a part of the plan in https://codereview.webrtc.org/1357923002/.
The texture copy doesn't cause any measurable performance difference on a Nexus 5 using VideoRendererGui in a AppRTCDemo loopback call.
BUG=webrtc:4993
TEST=Revert "Enable SurfaceViewRenderer for AppRTCDemo" https://codereview.webrtc.org/1356603004/ and try AppRTCDemo.
R=perkj@webrtc.org
Review URL: https://codereview.webrtc.org/1370113005 .
Cr-Commit-Position: refs/heads/master@{#10157}
This CL separates the types and code paths for textures vs byte buffers in MediaCodecVideoDecoder.dequeueOutputBuffer() and MediaCodecVideoDecoder::DeliverPendingOutputs(). The purpose is to prepare for lifetime management of textures received from the SurfaceTexture.
This CL is a part of the plan in https://codereview.webrtc.org/1357923002/.
BUG=webrtc:4993
Review URL: https://codereview.webrtc.org/1379383002
Cr-Commit-Position: refs/heads/master@{#10156}
This CL should not change the behaviour of the decoder. The purpose is to prepare for lifetime management of textures received from the SurfaceTexture. The main change is to only use exceptions for error signaling in MediaCodecVideoDecoder.dequeueOutputBuffer() and MediaCodecVideoDecoder.releaseOutputBuffer(), not both exceptions and error return values.
BUG=webrtc:4993
R=perkj@webrtc.org
Review URL: https://codereview.webrtc.org/1383983003 .
Cr-Commit-Position: refs/heads/master@{#10148}
This CL is a baby step towards consolidating the timestamps in cricket::VideoFrame and webrtc::VideoFrame, so that we can unify the frame classes in the future.
The elapsed time functionality is not really used. If a video sink wants to know the elapsed time since the first frame they can store the first timestamp themselves and calculate the time delta to later frames. This is already done in all video sinks that need the elapsed time. Having redundant timestamps in the frame classes is confusing and error prone.
TBR=pthatcher@webrtc.org
Review URL: https://codereview.webrtc.org/1324263004
Cr-Commit-Position: refs/heads/master@{#10131}
On some devices (confirmed Samsung) the focus mode is not configured correctly by default.
The fix explicitly set the focus mode to FOCUS_MODE_CONTINUOUS_VIDEO if this mode is supported.
BUG=webrtc:4991
Review URL: https://codereview.webrtc.org/1338773002
Cr-Commit-Position: refs/heads/master@{#10128}
Reason for revert:
This broke chromium.fyi bot.
Original issue's description:
> Change WebRTC SslCipher to be exposed as number only.
>
> This makes the SSL exposed as uint16_t which is the IANA value. GetRfcSslCipherName is introduced to handle the conversion to names from ID. IANA value will be used for UMA reporting. Names will still be used for WebRTC stats reporting.
>
> For SRTP, currently it's still string internally but is reported as IANA number.
>
> This is used by the ongoing CL https://codereview.chromium.org/1335023002.
>
> BUG=523033
>
> Committed: https://crrev.com/4fe3c9a77386598db9abd1f0d6983aefee9cc943
> Cr-Commit-Position: refs/heads/master@{#10124}
TBR=juberti@webrtc.org,rsleevi@chromium.org,pthatcher@webrtc.org,davidben@chromium.org,juberti@google.com,davidben@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=523033
Review URL: https://codereview.webrtc.org/1380603005
Cr-Commit-Position: refs/heads/master@{#10125}
This makes the SSL exposed as uint16_t which is the IANA value. GetRfcSslCipherName is introduced to handle the conversion to names from ID. IANA value will be used for UMA reporting. Names will still be used for WebRTC stats reporting.
For SRTP, currently it's still string internally but is reported as IANA number.
This is used by the ongoing CL https://codereview.chromium.org/1335023002.
BUG=523033
Review URL: https://codereview.webrtc.org/1337673002
Cr-Commit-Position: refs/heads/master@{#10124}
This CL refactors RendererCommon.getSamplingMatrix() so it does not have any dependecy to SurfaceTeture. The purpose is to prepare for a change in how texture frames are represented - only the texture matrix will be exposed, not the SurfaceTexture itself. This CL also adds an extra test for RendererCommon.rotateTextureMatrix().
R=hbos@webrtc.org
Review URL: https://codereview.webrtc.org/1375593002 .
Cr-Commit-Position: refs/heads/master@{#10118}
Reason for revert:
Relanding with SetConfiguration not pure virtual.
Original issue's description:
> Revert of Adding PeerConnectionInterface::SetConfiguration method. (patchset #4 id:60001 of https://codereview.webrtc.org/1317353005/ )
>
> Reason for revert:
> Broke FYI bots because SetConfiguration is pure virtual and MockPeerConnectionImpl doesn't implement it. Need to reland with SetConfiguration not pure virtual.
>
> Original issue's description:
> > Adding PeerConnectionInterface::SetConfiguration method.
> >
> > Also updated the JNI and Objective-C bindings. Later, will have a CL to
> > remove UpdateIce, which this method effectively replaces.
> >
> > BUG=webrtc:4945
> >
> > Committed: https://crrev.com/70702afbcb8418fe93747e7ed63bcbf5e56b90e9
> > Cr-Commit-Position: refs/heads/master@{#10040}
>
> TBR=guoweis@webrtc.org,pthatcher@webrtc.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=webrtc:4945
>
> Committed: https://crrev.com/7603c76ab077b1e2033bb179595129bd96797345
> Cr-Commit-Position: refs/heads/master@{#10041}
TBR=guoweis@webrtc.org,pthatcher@webrtc.org,tkchin@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:4945
Review URL: https://codereview.webrtc.org/1361273002
Cr-Commit-Position: refs/heads/master@{#10112}
The purpose of this CL is to use jlongFromPointer() for converting frame pointers to jlong instead of implicit casts which is not safe.
In order to respect constness, I had to make a small helper function for this.
BUG=webrtc:4993
R=perkj@webrtc.org
Review URL: https://codereview.webrtc.org/1373233002 .
Cr-Commit-Position: refs/heads/master@{#10108}
Reason for revert:
The top row in the video stream from the camera is messed up. The byte[] pointer is not the same as GetDirectBufferAddress() apparently.
Original issue's description:
> Android VideoCapturer: Send ByteBuffer instead of byte[]
>
> The purpose with this CL is to replace GetByteArrayElements() and ReleaseByteArrayElements() with GetDirectBufferAddress().
>
> R=hbos@webrtc.org
>
> Committed: https://crrev.com/cb3649b40b3fd6d5bbb0a92003b717e46ce90924
> Cr-Commit-Position: refs/heads/master@{#10091}
TBR=hbos@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.webrtc.org/1377783002
Cr-Commit-Position: refs/heads/master@{#10103}
This CL essentially converts [Local|Remote]TrackHandler to
Rtp[Sender|Receiver], and adds a "SetTrack" method for RtpSender.
It also gets rid of MediaStreamHandler and MediaStreamHandlerContainer,
since these classes weren't really anything more than containers.
PeerConnection now manages the RtpSenders and RtpReceivers directly.
Review URL: https://codereview.webrtc.org/1351803002
Cr-Commit-Position: refs/heads/master@{#10100}
This is the same sort of thing we do in C++ end-to-end PeerConnection
tests.
Review URL: https://codereview.webrtc.org/1361213002
Cr-Commit-Position: refs/heads/master@{#10098}
p2ptransportchannel. This CL does not use the new policy yet.
BUG=
Review URL: https://codereview.webrtc.org/1369773003
Cr-Commit-Position: refs/heads/master@{#10092}
The purpose with this CL is to replace GetByteArrayElements() and ReleaseByteArrayElements() with GetDirectBufferAddress().
R=hbos@webrtc.org
Review URL: https://codereview.webrtc.org/1372813002 .
Cr-Commit-Position: refs/heads/master@{#10091}
This cl moves back loading java SurfaceTextureHolder to the ClassReferenceHolder and use FindClass through ClassReferenceHolder. Without this, jni->FindClass returns nullptr in surfacetexturehelper_jni.cc.
BUG=webrtc:4993
R=magjed@webrtc.org
Review URL: https://codereview.webrtc.org/1370013002 .
Cr-Commit-Position: refs/heads/master@{#10086}
This CL also makes some small non-functional changes in ThreadUtils and EglBase to support SurfaceTextures and SurfaceTextureHelper.
BUG=webrtc:4993
R=hbos@webrtc.org
Review URL: https://codereview.webrtc.org/1368093003 .
Cr-Commit-Position: refs/heads/master@{#10085}
This cl adds a C++ counterpart of the Java SurfaceTextureHandler. It can be used for creating a webrtc::VideoFrames from a native handle and also guarantee that the Java SurfaceTexture is notified when the VideoFrame is no longer in use.
BUG=webrtc:4993
R=magjed@webrtc.org
Review URL: https://codereview.webrtc.org/1366413003 .
Cr-Commit-Position: refs/heads/master@{#10084}
Android hardware H.264 seems to keep a steady high-QP flow instead of
dropping frames, so framedrops aren't sufficient to detect a bad state
where downscaling would be beneficial.
BUG=webrtc:4968
R=magjed@webrtc.org, stefan@webrtc.org
Review URL: https://codereview.webrtc.org/1364253002 .
Cr-Commit-Position: refs/heads/master@{#10078}
Fix an issue where using setNeedsDisplay on a GLKView which has a frame
with size zero will make GLKView/iOS output the following error:
Failed to bind EAGLDrawable: <CAEAGLLayer: 0x1742282e0> to
GL_RENDERBUFFER 1 Failed to make complete framebuffer object 8cd6
(The error code 8cd6 corresponds to
GL_FRAMEBUFFER_INCOMPLETE_ATTACHMENT.)
GLKView will internally setup it's render buffer when the delegate is
about to draw into it. Previously when enableSetNeedsDisplay was set to
YES (default), then GLKView would still attempt to setup it's internal
buffer even if it's frame size is zero and that would cause
GL_FRAMEBUFFER_INCOMPLETE_ATTACHMENT.
By using enableSetNeedsDisplay = NO, RTCEAGLVideoView can guard against
calling -[GLKView display] if it's current frame size is empty.
Review URL: https://codereview.webrtc.org/1347013002
Cr-Commit-Position: refs/heads/master@{#10076}
This will allow Chromium's build files to be updated, so that when the
real RtpSender CL is submitted, it doesn't break the FYI bots.
Review URL: https://codereview.webrtc.org/1364813004
Cr-Commit-Position: refs/heads/master@{#10065}
Reason for revert:
Broke FYI bots because SetConfiguration is pure virtual and MockPeerConnectionImpl doesn't implement it. Need to reland with SetConfiguration not pure virtual.
Original issue's description:
> Adding PeerConnectionInterface::SetConfiguration method.
>
> Also updated the JNI and Objective-C bindings. Later, will have a CL to
> remove UpdateIce, which this method effectively replaces.
>
> BUG=webrtc:4945
>
> Committed: https://crrev.com/70702afbcb8418fe93747e7ed63bcbf5e56b90e9
> Cr-Commit-Position: refs/heads/master@{#10040}
TBR=guoweis@webrtc.org,pthatcher@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=webrtc:4945
Review URL: https://codereview.webrtc.org/1361263002
Cr-Commit-Position: refs/heads/master@{#10041}
Also updated the JNI and Objective-C bindings. Later, will have a CL to
remove UpdateIce, which this method effectively replaces.
BUG=webrtc:4945
Review URL: https://codereview.webrtc.org/1317353005
Cr-Commit-Position: refs/heads/master@{#10040}
Reason for revert:
This CL just landed: https://codereview.chromium.org/1323243006/
Which fixes the FYI bots for the original CL, and breaks them for this revert.
Original issue's description:
> Revert of TransportController refactoring. (patchset #6 id:100001 of https://codereview.webrtc.org/1350523003/ )
>
> Reason for revert:
> This CL causes problems with the WebRTC-in-Chromium FYI bots. Presumably it needs to be done in several steps, where removed files are emptied instead of removed in the first step.
>
> Original issue's description:
> > TransportController refactoring.
> >
> > Getting rid of TransportProxy, and in its place adding a
> > TransportController class which will facilitate access to and manage
> > the lifetimes of Transports. These Transports will now be accessed
> > solely from the worker thread, simplifying their implementation.
> >
> > This refactoring also pulls Transport-related code out of BaseSession.
> > Which means that BaseChannels will now rely on the TransportController
> > interface to create channels, rather than BaseSession.
> >
> > Committed: https://crrev.com/47ee2f3b9f33e8938948c482c921d4e13a3acd83
> > Cr-Commit-Position: refs/heads/master@{#10022}
>
> TBR=pthatcher@webrtc.org,deadbeef@webrtc.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Committed: https://crrev.com/a81a42f584baa0d93a4b93da9632415e8922450c
> Cr-Commit-Position: refs/heads/master@{#10024}
TBR=pthatcher@webrtc.org,torbjorng@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.webrtc.org/1361773005
Cr-Commit-Position: refs/heads/master@{#10036}
This CL makes the following changes:
* Instead of creating a new thread per startCapture()/stopCapture() cycle, VideoCapturerAndroid has a single thread that is initialized in the constructor and kept during the lifetime of the instance. This is more convenient because then it is always possible to post runnables without if-checks. This way, a lot of synchronize statements can be removed. Also, when the camera thread is preserved after stopCapture() it is possible to post late returnBuffer() calls to the correct thread.
* FramePool now enforces single thread use and returnBuffer() calls are posted to the camera thread. This is important because the camera should only be used from one thread, and we call camera.addCallbackBuffer() in returnBuffer().
* switchCamera() no longer returns false on failure, but instead signals the result via the callback.
* Update the test testCaptureAndAsyncRender() - it's not a valid use case to have outstanding frames when calling PeerConnectionFactory.dispose(). Instead, the renderer implementations should have release() functions that block until all frames are returned. The release() functions need to be called in the correct order with PeerConnectionFactory.dispose() last.
BUG=webrtc:4909
R=hbos@webrtc.org, perkj@webrtc.org
Review URL: https://codereview.webrtc.org/1350863002 .
Cr-Commit-Position: refs/heads/master@{#10025}
Reason for revert:
This CL causes problems with the WebRTC-in-Chromium FYI bots. Presumably it needs to be done in several steps, where removed files are emptied instead of removed in the first step.
Original issue's description:
> TransportController refactoring.
>
> Getting rid of TransportProxy, and in its place adding a
> TransportController class which will facilitate access to and manage
> the lifetimes of Transports. These Transports will now be accessed
> solely from the worker thread, simplifying their implementation.
>
> This refactoring also pulls Transport-related code out of BaseSession.
> Which means that BaseChannels will now rely on the TransportController
> interface to create channels, rather than BaseSession.
>
> Committed: https://crrev.com/47ee2f3b9f33e8938948c482c921d4e13a3acd83
> Cr-Commit-Position: refs/heads/master@{#10022}
TBR=pthatcher@webrtc.org,deadbeef@webrtc.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.webrtc.org/1358413003
Cr-Commit-Position: refs/heads/master@{#10024}
