This reverts commit 51698aefd4925f2dfa0310a321f836d433fa9258.
Reason for revert: Breaks builds because badly formatted deps
Original change's description:
> Put internal video codec factories into separate target
>
> The purpose is to start splitting out the dependencies to the built-in
> SW video codecs, so that clients can decide to not depend on them and
> get a reduction in binary size.
>
> Replaces https://webrtc-review.googlesource.com/c/src/+/29101
>
> Bug: webrtc:7925
> Change-Id: I46b95aaf42ead70ba78776de60600b8a66a1fe0c
> Reviewed-on: https://webrtc-review.googlesource.com/33420
> Commit-Queue: Anders Carlsson <andersc@webrtc.org>
> Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#21381}
TBR=magjed@webrtc.org,andersc@webrtc.org
Change-Id: Ib85f77fea756f4beb6a95b45cb132cbdc424ef00
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:7925
Reviewed-on: https://webrtc-review.googlesource.com/35260
Reviewed-by: Anders Carlsson <andersc@webrtc.org>
Commit-Queue: Anders Carlsson <andersc@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21383}
The H264 encoder defaults to using the kCVPixelFormatType_420YpCbCr8BiPlanarFullRange
pixel format. If the frames coming into the encoder is RTCCVPixelBuffer frames,
we check the pixel format in the contained CVPixelBuffer and send the same format
to the encoder when possible, by switching the encoder's pixel format. When we
receive frames with buffers conforming to the RTCI420Buffer protocol, we copy
the frame contents to the target pixel buffer, hardcoded to be the default NV12.
This works except when switching incoming frames from RTCCVPixelBuffer frames to
I420 frames during runtime. If we received RTCCVPixelBuffers wrapping e.g. an
RGB CVPixelBuffer, the encoder's pixel format have been changed to RGB. If we
now get incoming frames in I420, we must convert these to RGB instead of NV12
to match the encoder's format.
This bug can be triggered by calling `[_localVideoTrack setIsEnabled:NO]` in
`ARDAppClient.m`. This will make the stream start sending black i420 frames to
the encoder.
This CL fixes this by resetting the compression session with the default NV12
format if the input frame type changes from native to I420.
Bug: webrtc:8638
Change-Id: I5d784d204b7b1d09313a0f4cea6302ea72e9ed94
Reviewed-on: https://webrtc-review.googlesource.com/33260
Reviewed-by: Peter Hanspers <peterhanspers@webrtc.org>
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Anders Carlsson <andersc@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21382}
The purpose is to start splitting out the dependencies to the built-in
SW video codecs, so that clients can decide to not depend on them and
get a reduction in binary size.
Replaces https://webrtc-review.googlesource.com/c/src/+/29101
Bug: webrtc:7925
Change-Id: I46b95aaf42ead70ba78776de60600b8a66a1fe0c
Reviewed-on: https://webrtc-review.googlesource.com/33420
Commit-Queue: Anders Carlsson <andersc@webrtc.org>
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21381}
This CL does the following:
* Split out MediaStream JNI code from peerconnection.cc to mediastream.h/mediastream.cc.
* Split out RtpSender JNI code from peerconnection.cc to rtpsender.h/rtpsender.cc.
* Split out TurnCustomizer JNI code from peerconnection.cc to turncustomizer.h/turncustomizer.cc.
* Add missing instanceof function to WrappedNativeVideoDecoder.java.
* Move some PeerConnectionFactory JNI declarations from pc/video.cc to peerconnectionfactory.cc.
* Add declaration to video.h for the JNI functions that depend on EglBase14_jni.h.
* Use a scoped object to store the global Java MediaStream objects that also call dispose.
Bug: webrtc:8278
Change-Id: I3c56a599b8bcbc8f34e5c5a7b9c9fe1d192ff3f3
Reviewed-on: https://webrtc-review.googlesource.com/34645
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21380}
This allows application to construct PeerConnection object in JNI and
pass that to Android API. API for wrapping Java PeerConnection Observers
is exposed for convenience.
Bug: webrtc:8662
Change-Id: Id110b92e6bb5ab00661cd50616d05c3e18a1697d
Reviewed-on: https://webrtc-review.googlesource.com/34520
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Reviewed-by: Jonas Oreland <jonaso@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21379}
C++ API allows passing all configuration through RTCConfiguration
object. This adds all values previously passed through PC constraints
to Java RTCConfiguration object and deprecates API that takes PC
contraints.
Using the deprecated API overrides the values in RTCConfigration
object.
Bug: webrtc:8663, webrtc:8662
Change-Id: I128432c3caba74403513fb1347ff58830c643885
Reviewed-on: https://webrtc-review.googlesource.com/33460
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Taylor Brandstetter <deadbeef@webrtc.org>
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21357}
On some devices `android.hardware.Camera.open` returns null
instead of raising exception. It causes `NPE` inside
`Camera1Session.create` when method `setPreviewTexture` is
invoked on local variable `camera`, which is `null`.
Bug: webrtc:8658
Change-Id: Ic65b4aef2c0b8b65735a9db02433b536bfe92ddd
Reviewed-on: https://webrtc-review.googlesource.com/33620
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21352}
One reason for the circular deps is that common_types.h is a
historical dumping ground for various structs and defines that
are believed to be generally useful. I tried moving things out
that did not appear to be used downstream (StreamCounters,
RtpCounters etc) and moved the things that seemed used
(RtpHeader + supporting structs) to a new file api/rtp_headers.h.
This makes their place in the api more clear while moving out
the things that don't belong in the API in the first place.
I had to extract out typedefs.h from webrtc_common to resolve
another circular dependency. I believe checks includes typedefs,
but common depends on checks.
Bug: webrtc:7745
Change-Id: I725d49616b1ec0cdc8b74be7c078f7a4d46f084b
Reviewed-on: https://webrtc-review.googlesource.com/33001
Commit-Queue: Patrik Höglund <phoglund@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21295}
This CL removes the following GN variables: rtc_build_libyuv,
rtc_libyuv_dir (as requested in webrtc:7906).
It also removes some unneeded dependencies on //third_party/libyuv.
WebRTC targets were using public_deps to depend on //third_party/libyuv
and this created a build graph where targets that were depending on
//third_party/libyuv were not declaring the dependency to GN because
they were somehow getting it from another target that was exposing
//third_party/libyuv header files even if it wasn't directly depending
on it.
Bug: webrtc:8605, webrtc:7906
Change-Id: If71f7988fd80421dc2ad887cf94c2ac66366c3fb
Reviewed-on: https://webrtc-review.googlesource.com/32201
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21275}
This splits things out of rtc_base and makes dependencies explicit.
Bug: webrtc:6828
Change-Id: Id521896c3c43595349021c857bec216e429a0c8d
Reviewed-on: https://webrtc-review.googlesource.com/32780
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Patrik Höglund <phoglund@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21264}
This splits things out of rtc_base and makes dependencies explicit.
Bug: webrtc:6828
Change-Id: Ib813c7bd9e4de7ab015acb917bc09ee7204ba7bd
Reviewed-on: https://webrtc-review.googlesource.com/31940
Commit-Queue: Patrik Höglund <phoglund@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21245}
The problem was that the encoder was feeded with frames that had 0 as
a timestamp. This confused the encoder. H264 high profile support
clause was also wrong and is corrected.
Bug: webrtc:8601
Change-Id: Ic5a893b4b7573e694f865b63620843b2c9aa489f
Reviewed-on: https://webrtc-review.googlesource.com/32300
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21234}
Using fully qualified paths to include libyuv headers allows WebRTC to
avoid to rely on the //third_party/libyuv:libyuv_config target to
set the -I compiler flag.
Today some WebRTC targets depend on //third_party/libyuv only to
include //third_party/libyuv:libyuv_config but with fully qualified
paths this should not be needed anymore.
A follow-up CL will remove //third_party/libyuv from some targets that
don't need it because they are not including libyuv headers.
Bug: webrtc:8605
Change-Id: Icec707ca761aaf2ea8088e7f7a05ddde0de2619a
No-Try: True
Reviewed-on: https://webrtc-review.googlesource.com/28220
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Magnus Flodman <mflodman@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21209}
The generated JNI files are actually implementation files that can only
be included from one place. For generating JNI code for
PeerConnection.java, this means we need to collect all JNI
PeerConnection code into one place. This CL moves
peerconnectionobserver_jni.h to peerconnection.h, and merges
peerconnectionobserver_jni.cc and peerconnection_jni.cc. It also moves
converter functions JavaToNativeIceServers and
JavaToNativeRTCConfiguration to peerconnection.h since they are also
related to PeerConnection.java.
This CL does not contain anything else than a pure code move.
TBR=sakal@webrtc.org
Bug: webrtc:8278
Change-Id: I31c696c1aab495e09f95e5ba5d79f95df7c34a0e
Reviewed-on: https://webrtc-review.googlesource.com/31483
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21200}
This moves all WebRTC internal code from using
SessionDescriptionInterface::type() which returns a string and
from using CreateSessionDescription with a string type parameter.
Bug: webrtc:8613
Change-Id: I1cdd93dc4b26dec157e22476fdac569d5da2810a
Reviewed-on: https://webrtc-review.googlesource.com/29500
Commit-Queue: Steve Anton <steveanton@webrtc.org>
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Zhi Huang <zhihuang@webrtc.org>
Reviewed-by: Peter Thatcher <pthatcher@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21147}
A lot of WebRTC targets were depending on //third_party/libyuv using
public_deps instead of deps. This causes issues because a the
inclusion of libyuv headers is not declared to the build system and
this creates hidden dependencies that put the modularity of the project
at risk.
Bug: webrtc:8603
Change-Id: Ide0ceb84eb5640ae664dc782f3a722b55c3b601a
No-Try: True
Reviewed-on: https://webrtc-review.googlesource.com/28120
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Magnus Flodman <mflodman@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21039}
This means we will properly request a new keyframe if decoding fails.
Bug: webrtc:8600
Change-Id: Id213686f016c5418bf04b2ee68bd19dbbe1ea954
Reviewed-on: https://webrtc-review.googlesource.com/28101
Reviewed-by: Rasmus Brandt <brandtr@webrtc.org>
Commit-Queue: Kári Helgason <kthelgason@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#21036}
This fixes a bug where AppRTCMobile would crash at runtime when
built without VP9 support.
Bug: webrtc:8602
Change-Id: Id2db79c3ff8136f06dc049afcc5197e9356fd25b
Reviewed-on: https://webrtc-review.googlesource.com/27983
Reviewed-by: Oleh Prypin <oprypin@webrtc.org>
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Oleh Prypin <oprypin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20982}
Changes places where we explicitly construct an Optional to instead use
nullopt or the requisite value type only.
This CL was uploaded by git cl split.
Bug: None
Change-Id: Idf275495910f651ec35f641611926e62414daa9a
Reviewed-on: https://webrtc-review.googlesource.com/23610
Commit-Queue: Oskar Sundbom <ossu@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20972}
The data-channel only target "peerconnectionfactory_no_media_objc"
transitively depends on libvpx via "peerconnectionfactory_base_objc".
This CL breaks the dependency between "objc_peeerconnectionfactory_base"
and libvpx moving RTCVideoCodecVP8.mm and RTCVideoCodecVP9.mm to
"peerconnectionfactory_objc" (together with RTCVideoCodecH264.mm).
Bug: webrtc:8594
Change-Id: Idfe3024163012925f017ad8c585b7ae21e86c319
Reviewed-on: https://webrtc-review.googlesource.com/27480
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20969}
Changes places where we explicitly construct an Optional to instead use
nullopt or the requisite value type only.
This CL was uploaded by git cl split.
R=magjed@webrtc.org
Bug: None
Change-Id: I78842b6bb8ae345bcb852feee3908fdaf955c664
Reviewed-on: https://webrtc-review.googlesource.com/23574
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Oskar Sundbom <ossu@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20956}
This reverts commit aede67a199ae0552074bfec4bb03cc9a6a5fba0f.
Reason for revert: Causes error:
JNI ERROR (app bug): local reference table overflow (max=512)'
Original change's description:
> Android: Generate JNI code for stats
>
> This CL also unifies the functions for converting from C++ to Java, and
> generates the boiler plate for converting C++ vectors to Java arrays.
>
> Bug: webrtc:8278
> Change-Id: I262e9162beae8a64ba0e8b6a27e1081207b03961
> Reviewed-on: https://webrtc-review.googlesource.com/26020
> Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
> Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#20918}
TBR=magjed@webrtc.org,sakal@webrtc.org
Change-Id: Ieb26ed8577bd489a4dd4f7542d16a7d0e11f409f
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webrtc:8278
Reviewed-on: https://webrtc-review.googlesource.com/26900
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20926}
This CL also unifies the functions for converting from C++ to Java, and
generates the boiler plate for converting C++ vectors to Java arrays.
Bug: webrtc:8278
Change-Id: I262e9162beae8a64ba0e8b6a27e1081207b03961
Reviewed-on: https://webrtc-review.googlesource.com/26020
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20918}
This reverts commit 2a9dbe6e7e722cf069b72ccff0051b1517706996.
Reason for revert: The fix does not help, it still fails at runtime
Original change's description:
> Adding -Wno-deprecated-declarations to declarations deprecated in iOS 10
>
> This CL will unblock the Chromium Roll while deprecated declarations
> will be removed from the WebRTC codebase.
>
> Bug: webrtc:8570
> Change-Id: I55cf78040758369ce45176cf0a00df50a87eb972
> No-Try: True
> Reviewed-on: https://webrtc-review.googlesource.com/25641
> Commit-Queue: Mirko Bonadei <mbonadei@webrtc.org>
> Reviewed-by: Patrik Höglund <phoglund@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#20858}
No-Try: True
Bug: webrtc:8570
Change-Id: Ib17ed05f8f239912661281329ecab175c23491d6
Reviewed-on: https://webrtc-review.googlesource.com/25964
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Oleh Prypin <oprypin@webrtc.org>
Commit-Queue: Oleh Prypin <oprypin@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20890}
For better consistency between the Objective-C API constant definitions
and the existing constants defined in the underlying core, re-use the
available video codec-name constants from cricket to define the peer
constants in the public API.
BUG=None
Change-Id: I8d5ddc2c1bd6670810fca1665aaf9a116620a34e
Reviewed-on: https://webrtc-review.googlesource.com/25360
Commit-Queue: Kári Helgason <kthelgason@webrtc.org>
Reviewed-by: Kári Helgason <kthelgason@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20883}
Also improves ownership model by using std::unique_ptr in a couple of
places instead of raw pointers.
Bug: webrtc:8278
Change-Id: I0429ec3c416b5baa1ffa21dad71e0d64b004c446
Reviewed-on: https://webrtc-review.googlesource.com/25020
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20863}
This is not the case for fallback codecs.
Bug: webrtc:7925
Change-Id: I5039d8879923a2db1e7c08bb5640763dd20ea8e7
Reviewed-on: https://webrtc-review.googlesource.com/24863
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20861}
It's now possible to generate JNI code for constructors and enums
correctly. This CL cleans that up.
Bug: webrtc:8278,webrtc:8551,webrtc:8556
Change-Id: I2284a30139cbb186c80713eb6113eda5659c16ad
Reviewed-on: https://webrtc-review.googlesource.com/25622
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20860}
Previously, wrapped native codec instances would leak the native object
if it was never used. This change fixes it by changing getNative method
to createNative.
Also fixes "Video codec hardware acceleration" setting in AppRTCMobile.
Bug: webrtc:7925
Change-Id: I53f6dc1dd5e37dea8d14278423122dede17719c5
Reviewed-on: https://webrtc-review.googlesource.com/24881
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#20859}
