OSDN Git Service
Glenn Kasten [Tue, 26 Feb 2013 17:20:22 +0000 (09:20 -0800)]
Remove tee sink debugging at compile time
Bug:
8223560
Change-Id: Iddbfb06c45d43d9f20bb428215dd4094931e19a7
Eino-Ville Talvala [Tue, 19 Feb 2013 18:40:14 +0000 (10:40 -0800)]
CameraService and Stagefright: Support AppOps
Camera:
- Signal to AppOpsService when camera usage starts and stops
- Listen to permissions revocations and act on them
- Currently just kill camera connection when permissions lost
Stagefright:
- Pass on client name, UID to camera as needed
Bug:
8181262
Change-Id: I9e33c9d05e9daa77dbb2d795045d08eb887ec8f0
Glenn Kasten [Sat, 23 Feb 2013 01:19:43 +0000 (01:19 +0000)]
Merge "Update tee sink"
Glenn Kasten [Thu, 10 Jan 2013 20:31:01 +0000 (12:31 -0800)]
Update tee sink
Implement rotation to reduce long-term storage use.
Implement optional per-track tee.
Dynamically enable at runtime based on property, instead of at compile-time.
Dynamic frame count not yet implemented.
Bug:
8223560
Change-Id: I3706443c6ec0cb0c6656dc288715a02ad5fea63a
Igor Murashkin [Wed, 20 Feb 2013 21:36:17 +0000 (13:36 -0800)]
ProCamera: Fix waitForFrameBuffer not handling multiple outstanding frames
If the CpuConsumer triggered multiple onFrameAvailable callbacks in between
a single waitForFrameBuffer call, the old code would only handle 1 callback.
This meant on two subsequent waitForFrameBuffer calls the second would always
timeout when two buffers were already available to be unlocked.
Bug:
8238112
Change-Id: Ibefca35005ac5c408e5ada97ec4a4344a9e3e497
Igor Murashkin [Thu, 21 Feb 2013 22:45:03 +0000 (14:45 -0800)]
ProCamera: add waitForFrameBuffer/waitForFrameResult blocking calls
Change-Id: I851d41aeecaa15245d5b9d622132e8706d6e292c
Igor Murashkin [Thu, 21 Feb 2013 22:18:10 +0000 (14:18 -0800)]
Camera: ProCameraTests: clean up a bit, still needs more cleanup
Change-Id: Ia7e18c81002c8d98176f21553cf13cbd13cfdf5d
Igor Murashkin [Thu, 21 Feb 2013 21:49:26 +0000 (13:49 -0800)]
ProCamera: Add getInfo for getting static metadata (and a test)
Change-Id: I3b87dfbc621d98bf41f37a892a1012baf85b5166
Igor Murashkin [Thu, 21 Feb 2013 20:02:29 +0000 (12:02 -0800)]
Camera: ProCamera - implement onResultReceived callback for metadata callbacks
Change-Id: I46775402b007244bc383d6343a620eebbd492aad
Igor Murashkin [Thu, 21 Feb 2013 19:43:14 +0000 (11:43 -0800)]
Camera: ProCameraTests - add CpuConsumer tests for dual streams
Change-Id: Ia010450dc79ac2841645b8c2e1553c9ec74ee5e3
Igor Murashkin [Thu, 21 Feb 2013 03:29:53 +0000 (19:29 -0800)]
Camera: ProCamera - add createStreamCpu and unit test
Change-Id: I468172dbfdd78510b273bf9d119c950cbeda7ea3
Igor Murashkin [Thu, 21 Feb 2013 03:15:15 +0000 (19:15 -0800)]
Camera: Change ProCamera to take IGraphicBufferProducer
Change-Id: Iec62eead6d179aa5486f7719143340976bb76e7d
Igor Murashkin [Thu, 21 Feb 2013 03:02:36 +0000 (19:02 -0800)]
Camera: ProCamera2Client implement createStream,createDefaultRequest
The StreamingImage unit test should now work when TEST_DISPLAY_SECS is set.
Change-Id: I9cf3ad5bd5a76096ab21bd39e631d636c4ff3d3f
Igor Murashkin [Thu, 21 Feb 2013 02:24:43 +0000 (18:24 -0800)]
Camera: ProCamera2Client - add createStream (service is stubbed) and unit test
Change-Id: I1ae7ba9b24f5883c214c19a7ed0eaf0c802d69c1
Igor Murashkin [Thu, 21 Feb 2013 01:57:31 +0000 (17:57 -0800)]
Camera: ProCamera - add createStream stub and unit test for it
Change-Id: Ic05130e63f4f2c0c3278ba348b192992169f105f
Igor Murashkin [Wed, 30 Jan 2013 18:14:24 +0000 (10:14 -0800)]
Camera: ProCameraTests - add asynchronous locking unit test
Change-Id: Ib79eb84046c9ed898bfb086a6600265fc351924c
Igor Murashkin [Thu, 21 Feb 2013 01:41:57 +0000 (17:41 -0800)]
Camera: ProClient add asynchronous locks and such
Change-Id: I551e5e5e76d9be733fab5224beaa7309268c0f38
Igor Murashkin [Thu, 21 Feb 2013 01:15:11 +0000 (17:15 -0800)]
Camera: Add ProCamera private binder interface for an API2-light functionality
Change-Id: I2af7a807c99df75ea659e6e6acc9c4fca6a56274
Igor Murashkin [Thu, 21 Feb 2013 00:50:13 +0000 (16:50 -0800)]
Camera: Add unit test infrastructure for libcameraclient
Change-Id: I957538663ae8332d26f3640c63a75efc79f4ac5c
Igor Murashkin [Thu, 14 Feb 2013 02:23:39 +0000 (18:23 -0800)]
Camera: Add exists function to CameraMetadata
Change-Id: I081f61417fca7c8408e9e767129647ddce5364ab
Igor Murashkin [Wed, 13 Feb 2013 23:53:56 +0000 (15:53 -0800)]
Camera: Move CameraMetadata.h from service to client library
Change-Id: I940ce86f318f37ae5b73f912a6e589415150125f
Mathias Agopian [Thu, 21 Feb 2013 23:21:20 +0000 (23:21 +0000)]
Merge "replace getISurfaceTexture by getIGraphicBufferProducer"
Glenn Kasten [Thu, 21 Feb 2013 20:57:15 +0000 (20:57 +0000)]
Merge "Permit conditional registry of mediaserver extensions"
Glenn Kasten [Thu, 21 Feb 2013 00:47:28 +0000 (16:47 -0800)]
Permit conditional registry of mediaserver extensions
Change-Id: I94dc0d038e702dfe1779a50a1de0fae9bad15057
Alex Ray [Wed, 20 Feb 2013 23:52:12 +0000 (23:52 +0000)]
Merge "CameraService: Log camera module name"
Alex Ray [Wed, 20 Feb 2013 21:39:37 +0000 (13:39 -0800)]
CameraService: Log camera module name
Explicitly logs camera module loaded. This is useful on systems with
multiple camera modules installed. Also this make the camera service
match audioflinger, which logs hardware module names as well.
Change-Id: I280bf12df2d517f936916b8ba3cb662e671d0216
Ed Heyl [Wed, 20 Feb 2013 21:47:28 +0000 (13:47 -0800)]
Merge remote-tracking branch 'goog/jb-mr2-release'
Mathias Agopian [Wed, 20 Feb 2013 02:25:33 +0000 (18:25 -0800)]
replace getISurfaceTexture by getIGraphicBufferProducer
Change-Id: I723a3863558126f2635a544a39bb3ec88d499c60
Glenn Kasten [Mon, 18 Feb 2013 20:47:49 +0000 (12:47 -0800)]
A reference cannot be re-bound
Bug:
8213067
Bug:
6490974
Change-Id: If546215641fbc9a0fb1e8af8b85cc01f0a4b26ab
Glenn Kasten [Mon, 18 Feb 2013 20:47:49 +0000 (12:47 -0800)]
A reference cannot be re-bound
Bug:
8213067
Bug:
6490974
Change-Id: If546215641fbc9a0fb1e8af8b85cc01f0a4b26ab
The Android Automerger [Tue, 19 Feb 2013 17:51:09 +0000 (09:51 -0800)]
merge in jb-mr2-release history after reset to master
Glenn Kasten [Fri, 15 Feb 2013 23:55:26 +0000 (23:55 +0000)]
Merge "Revert "Temporary additional logging to investigate bug""
Glenn Kasten [Fri, 15 Feb 2013 23:55:04 +0000 (23:55 +0000)]
Revert "Temporary additional logging to investigate bug"
This reverts commit
32584a7d672864b20ab8b83a3cb23c1858e908b7
Change-Id: I9dc680578b955b1af462eeb7a49d61a0d45eb81b
Glenn Kasten [Fri, 15 Feb 2013 23:54:53 +0000 (23:54 +0000)]
Merge "Temporary additional logging to investigate bug"
Glenn Kasten [Wed, 13 Feb 2013 22:46:45 +0000 (14:46 -0800)]
Temporary additional logging to investigate bug
The bug appears related to continuing to use an invalid buffer provider
in fast mixer after track destruction, so focus the added logs in that area.
Also includes a bug fix: was calling log in an unsafe place
near Threads.cpp AudioFlinger::PlaybackThread::createTrack_l line 1250.
Details:
- include caller pid or client pid where appropriate
- increase log buffer size
- log mFastIndex when AudioMixer sees an invalid bufferProvider.
- log both potentially modified and actually modified tracks in FastMixer.
- fix benign bug where sq->end() was called more than once.
- log StateQueue push() call and return.
- increase StateQueue size from 4 to 8 entries
- log mixer->enable(), bufferProvider, and currentTrackMask
- log buffer provider addresses
- increase fast mixer log buffer again
- check logf format vs. argument list compatibility
- add logging to AudioMixer
- add checking of magic field in AudioMixer to detect overwrites
- add bool AudioMixer::enabled()
- increase log buffer sizes yet again
- enable assertion checking without ALOGV
- improve a few log messages
- check for corruption in more places
- log in all the process hooks
- add new mixer APIs so we can check for corruption of mixer state
- fix a build warning
Bug:
6490974
Change-Id: Ib0c4a73dcf606ef9bd898313b3b40ef61ab42f51
Mathias Agopian [Fri, 15 Feb 2013 22:05:59 +0000 (22:05 +0000)]
Merge "Refactoring: Rename SurfaceTextureClient to Surface"
Mathias Agopian [Fri, 15 Feb 2013 01:11:27 +0000 (17:11 -0800)]
Refactoring: Rename SurfaceTextureClient to Surface
Change-Id: I4e8a8b20914cb64edc37abe68233fbc9f2b5d830
Marco Nelissen [Fri, 15 Feb 2013 16:26:24 +0000 (08:26 -0800)]
Remove FragmentedMP4Extractor
MPEG4Extractor now supports fragmented mp4 files.
Change-Id: I5659a51f4e5e4407a12535e69238fe3abffda7dc
Marco Nelissen [Wed, 13 Feb 2013 17:27:28 +0000 (09:27 -0800)]
Add support for fragmented mp4 to MPEG4Extractor
This makes FragmentedMP4Extractor obsolete. It will be removed in a
separate change.
Change-Id: Ida74c07ccf84983e20a1320ee24ffc7a5c083859
Glenn Kasten [Wed, 13 Feb 2013 22:46:45 +0000 (14:46 -0800)]
Temporary additional logging to investigate bug
The bug appears related to continuing to use an invalid buffer provider
in fast mixer after track destruction, so focus the added logs in that area.
Also includes a bug fix: was calling log in an unsafe place
near Threads.cpp AudioFlinger::PlaybackThread::createTrack_l line 1250.
Details:
- include caller pid or client pid where appropriate
- increase log buffer size
- log mFastIndex when AudioMixer sees an invalid bufferProvider.
- log both potentially modified and actually modified tracks in FastMixer.
- fix benign bug where sq->end() was called more than once.
- log StateQueue push() call and return.
- increase StateQueue size from 4 to 8 entries
- log mixer->enable(), bufferProvider, and currentTrackMask
- log buffer provider addresses
- increase fast mixer log buffer again
- check logf format vs. argument list compatibility
- add logging to AudioMixer
- add checking of magic field in AudioMixer to detect overwrites
- add bool AudioMixer::enabled()
Bug:
6490974
Change-Id: I1f3f18aa62d9fbd35bc32285b669f5ba40efe28e
Glenn Kasten [Thu, 14 Feb 2013 23:57:13 +0000 (23:57 +0000)]
Merge "Revert "Temporary additional logging to investigate bug""
Glenn Kasten [Thu, 14 Feb 2013 23:57:02 +0000 (23:57 +0000)]
Revert "Temporary additional logging to investigate bug"
This reverts commit
639482c24c911b125398b31883ba6d55faebe28b
Change-Id: I11f2829072ab11e18b0663024f27bf31192f1d39
Glenn Kasten [Thu, 14 Feb 2013 23:56:52 +0000 (23:56 +0000)]
Merge "Temporary additional logging to investigate bug"
Glenn Kasten [Wed, 13 Feb 2013 22:46:45 +0000 (14:46 -0800)]
Temporary additional logging to investigate bug
The bug appears related to continuing to use an invalid buffer provider
in fast mixer after track destruction, so focus the added logs in that area.
Also includes a bug fix: was calling log in an unsafe place
near Threads.cpp AudioFlinger::PlaybackThread::createTrack_l line 1250.
Details:
- include caller pid or client pid where appropriate
- increase log buffer size
- log mFastIndex when AudioMixer sees an invalid bufferProvider.
- log both potentially modified and actually modified tracks in FastMixer.
- fix benign bug where sq->end() was called more than once.
- log StateQueue push() call and return.
- increase StateQueue size from 4 to 8 entries
- log mixer->enable(), bufferProvider, and currentTrackMask
- log buffer provider addresses
- increase fast mixer log buffer again
- check logf format vs. argument list compatibility
- add logging to AudioMixer
- add checking of magic field in AudioMixer to detect overwrites
- add bool AudioMixer::enabled()
Bug:
6490974
Change-Id: I1f3f18aa62d9fbd35bc32285b669f5ba40efe28e
ztenghui [Thu, 14 Feb 2013 20:42:53 +0000 (20:42 +0000)]
Merge "Add a test utility for MediaMuxer."
The Android Automerger [Thu, 14 Feb 2013 13:55:39 +0000 (05:55 -0800)]
merge in jb-mr2-release history after reset to master
Glenn Kasten [Wed, 13 Feb 2013 22:46:45 +0000 (14:46 -0800)]
Temporary additional logging to investigate bug
The bug appears related to continuing to use an invalid buffer provider
in fast mixer after track destruction, so focus the added logs in that area.
Also includes a bug fix: was calling log in an unsafe place
near Threads.cpp AudioFlinger::PlaybackThread::createTrack_l line 1250.
Details:
- include caller pid or client pid where appropriate
- increase log buffer size
- log mFastIndex when AudioMixer sees an invalid bufferProvider.
- log both potentially modified and actually modified tracks in FastMixer.
- fix benign bug where sq->end() was called more than once.
- log StateQueue push() call and return.
Bug:
6490974
Change-Id: Iee7c8f40e20b6000cd8286c0ec6a14fff4a37af1
Glenn Kasten [Thu, 14 Feb 2013 00:02:53 +0000 (00:02 +0000)]
Merge "Revert "Temporary additional logging to investigate bug""
Glenn Kasten [Thu, 14 Feb 2013 00:02:36 +0000 (00:02 +0000)]
Revert "Temporary additional logging to investigate bug"
This reverts commit
0ddd56316262ac74a95e9edb595697c163136d6d
Change-Id: I180a928af6f5a38d15a5efe44cd1fe927b5d961c
Glenn Kasten [Thu, 14 Feb 2013 00:02:22 +0000 (00:02 +0000)]
Merge "Temporary additional logging to investigate bug"
Glenn Kasten [Wed, 13 Feb 2013 22:46:45 +0000 (14:46 -0800)]
Temporary additional logging to investigate bug
The bug appears related to continuing to use an invalid buffer provider
in fast mixer after track destruction, so focus the added logs in that area.
Also includes a bug fix: was calling log in an unsafe place
near Threads.cpp AudioFlinger::PlaybackThread::createTrack_l line 1250.
Details:
- include caller pid or client pid where appropriate
- increase log buffer size
- log mFastIndex when AudioMixer sees an invalid bufferProvider.
- log both potentially modified and actually modified tracks in FastMixer.
- fix benign bug where sq->end() was called more than once.
- log StateQueue push() call and return.
Bug:
6490974
Change-Id: Iee7c8f40e20b6000cd8286c0ec6a14fff4a37af1
ztenghui [Thu, 7 Feb 2013 23:12:10 +0000 (15:12 -0800)]
Add a test utility for MediaMuxer.
This test/utility copy samples from one video and mux into another video.
It support trimming, cutting audio or video track.
It can run simply as command line like:
adb shell muxer -a -v -s 1000 -e 8000 "/sdcard/DCIM/Camera/VID_*.mp4"
bug:
7991013
Change-Id: I8a2eeff3cabd001b6b2a7062d991dd076edbf22e
James Dong [Wed, 13 Feb 2013 18:18:22 +0000 (18:18 +0000)]
Merge "Add OMX.google.vpx.encoder"
The Android Automerger [Wed, 13 Feb 2013 13:55:41 +0000 (05:55 -0800)]
merge in jb-mr2-release history after reset to master
James Dong [Tue, 12 Feb 2013 18:36:01 +0000 (10:36 -0800)]
Add OMX.google.vpx.encoder
Change-Id: I843e6e542533884a94e105e1cb56f16f0440af61
Kunter Gultekin [Fri, 1 Feb 2013 15:01:15 +0000 (17:01 +0200)]
Adds VPX encoding support for stagefright.
Only following encoder settings are available
- target bitrate
- rate control (constant / variable)
- frame rate
- token partitioning
- error resilience
- reconstruction & loop filters
Only following color formats are recognized
- YUV420Planar
- YUV420SemiPlanar
- AndroidOpaque
Following settings are not configurable by the client
- encoding deadline is realtime
- the algorithm interface for encoder is vp8
- fractional bits of frame rate is discarded
- timebase is fixed to 1/
1000000
Requires libvpx to be built with encoder support enabled.
Requires openmax 1.1.2 extension headers.
Relevant tests exist in cts repo.
Change-Id: I650f1aca83e7dc93f79d7e6cba7ac24f26e66d40
Signed-off-by: Kunter Gultekin <kuntergultekin@google.com>
Glenn Kasten [Tue, 12 Feb 2013 20:12:42 +0000 (12:12 -0800)]
Temporary additional logging to investigate bug
The bug appears related to continuing to use an invalid buffer provider
in fast mixer after track destruction, so focus the added logs in that area.
Also includes a bug fix: was calling log in an unsafe place
near Threads.cpp AudioFlinger::PlaybackThread::createTrack_l line 1250.
Details:
- include caller pid or client pid where appropriate
- increase log buffer size
Bug:
6490974
Change-Id: I4c030f171343fe4b483eae0ddea4427118d8d4b1
Glenn Kasten [Tue, 12 Feb 2013 20:44:09 +0000 (20:44 +0000)]
Merge "Revert "Temporary additional logging to investigate bug""
Glenn Kasten [Tue, 12 Feb 2013 20:43:58 +0000 (20:43 +0000)]
Revert "Temporary additional logging to investigate bug"
This reverts commit
3051df27261e9952c0e642dec548515250e85f6a
Change-Id: I8bf5c3e91b65bd20de26f480c367c2854b62373c
Glenn Kasten [Tue, 12 Feb 2013 20:43:45 +0000 (20:43 +0000)]
Merge "Temporary additional logging to investigate bug"
Glenn Kasten [Tue, 12 Feb 2013 20:12:42 +0000 (12:12 -0800)]
Temporary additional logging to investigate bug
The bug appears related to continuing to use an invalid buffer provider
in fast mixer after track destruction, so focus the added logs in that area.
Also includes a bug fix: was calling log in an unsafe place
near Threads.cpp AudioFlinger::PlaybackThread::createTrack_l line 1250.
Details:
- include caller pid or client pid where appropriate
- increase log buffer size
Bug:
6490974
Change-Id: I4c030f171343fe4b483eae0ddea4427118d8d4b1
ztenghui [Mon, 4 Feb 2013 23:59:38 +0000 (15:59 -0800)]
Add the native MediaMuxer support.
MediaAdapter: a helper class to convert the push model to pull model.
MediaMuxer: the real muxer.
bug:
7991013
Change-Id: If3b79551bc6332bc81f5c2740885e579a5c4abf9
The Android Automerger [Tue, 12 Feb 2013 15:10:36 +0000 (07:10 -0800)]
merge in jb-mr2-release history after reset to master
Jean-Michel Trivi [Tue, 5 Feb 2013 00:26:02 +0000 (16:26 -0800)]
Add support for querying if a stream is active remotely
Bug
7485803
Change-Id: I0744374f130fd2dd0714102354cffed2fa915361
The Android Automerger [Mon, 11 Feb 2013 15:10:54 +0000 (07:10 -0800)]
merge in jb-mr2-release history after reset to master
James Dong [Sat, 9 Feb 2013 02:16:02 +0000 (18:16 -0800)]
Revert "Adds VPX encoding support for stagefright."
This lib was not part of the build, but it is built anyway for userdebug image (not for eng though).
let me revert it for now...
This reverts commit
6c6bb9873f55853fe74d8f45ad3ae116636d8be7.
James Dong [Sat, 9 Feb 2013 00:47:03 +0000 (00:47 +0000)]
Merge "Adds VPX encoding support for stagefright."
Mike Lockwoood [Fri, 8 Feb 2013 21:51:43 +0000 (21:51 +0000)]
Merge "MTP: Write initial data to correct file offset in SendPartialObject"
Mike Lockwoood [Fri, 8 Feb 2013 21:25:01 +0000 (13:25 -0800)]
MTP: Write initial data to correct file offset in SendPartialObject
Change-Id: I84288aeda3e65e6e6487f11d32a72910cd16cff2
Andreas Huber [Thu, 7 Feb 2013 23:25:59 +0000 (23:25 +0000)]
Merge "Allow for dynamic reconfiguration of the video bitrate used"
Andreas Huber [Thu, 7 Feb 2013 19:38:08 +0000 (11:38 -0800)]
A few more patches to fix wfd tcp unicast transport.
Change-Id: Ie2f1b1e56c487ac4c3ef19d9e79022a35084e042
Andreas Huber [Thu, 7 Feb 2013 18:56:14 +0000 (10:56 -0800)]
Allow for dynamic reconfiguration of the video bitrate used
to encode video while running as a wfd source.
Change-Id: I44f7b2350c88fc5807047c61bfe594ef8fa79275
Kunter Gultekin [Fri, 1 Feb 2013 15:01:15 +0000 (17:01 +0200)]
Adds VPX encoding support for stagefright.
Only following encoder settings are available
- target bitrate
- rate control (constant / variable)
- frame rate
- token partitioning
- error resilience
- reconstruction & loop filters
Only following color formats are recognized
- YUV420Planar
- YUV420SemiPlanar
- AndroidOpaque
Following settings are not configurable by the client
- encoding deadline is realtime
- the algorithm interface for encoder is vp8
- fractional bits of frame rate is discarded
- timebase is fixed to 1/
1000000
Requires libvpx to be built with encoder support enabled.
Requires openmax 1.1.2 extension headers.
Relevant tests exist in cts repo.
Change-Id: If759edb8db36acbd24dcb53d159a54e942766020
Signed-off-by: Kunter Gultekin <kuntergultekin@google.com>
The Android Automerger [Thu, 7 Feb 2013 15:10:37 +0000 (07:10 -0800)]
merge in jb-mr2-release history after reset to master
Roger Jönsson [Wed, 23 Jan 2013 17:18:08 +0000 (18:18 +0100)]
Avoid rebuffering after RTSP pause
If pausing an RTSP stream, an RTSP Pause request is sent and then
if the stream is immediately resumed again, an RTSP Play request
will be sent to the server.
But the new data after the pause will not be buffered until
Sender Reports have arrived again on both channels.
Meanwhile the player will resume playback and continue consuming
the already existing buffer.
This means that there is a risk that the buffer is emptied while
waiting for sender reports.
This commit simply adds a delay before the RTSP pause request is
sent, allowing some additional RTSP buffering that might be needed
when the stream is resumed again.
Also, if the stream is resumed again before the RTSP pause request
is sent, there is no need for any RTSP pause request, hence it is
omitted.
Change-Id: I928c8bfb5e99a6a146dcda4e51e528973ecbe065
joakim johansson [Wed, 23 Jan 2013 16:18:56 +0000 (17:18 +0100)]
EOS fixes for RTSP streams
The fix takes care of several near end of stream use cases:
seek, pause and fake timestamps.
Change-Id: I5f5fa881b1f619dfd5e1afd2af957082345c59eb
Måns Zigher [Wed, 23 Jan 2013 13:48:57 +0000 (14:48 +0100)]
RTSP: Parse session level control attribute from SDP
If a=control: is present at session-level in the SDP response,
RFC2326:C.1.1 defines the URL to be used for aggregate commands.
This includes PLAY and PAUSE but not TEARDOWN.
Change-Id: Iaa1dc2271d00df39dc83477a99fda6fbeb73c5b4
Roger Jönsson [Mon, 21 Jan 2013 16:15:45 +0000 (17:15 +0100)]
Enable pause/resume for RTSP streaming
When a stream is paused, RTSP Pause is also sent to the server.
Otherwise the buffering might continue until the memory runs out.
When the stream is resumed, RTSP Play will be sent in order to
resume the buffering.
Change-Id: I5dc1761140827c532451638c3fd3f34271e5b9ab
Roger Jönsson [Mon, 21 Jan 2013 15:26:41 +0000 (16:26 +0100)]
RTSP buffering improvements
Added buffering start and end notifications for RTSP.
MEDIA_INFO_BUFFERING_START is sent when buffering is started
and MEDIA_INFO_BUFFERING_END is sent when the buffer has
filled up.
This patch also adds RTSP end of stream handling.
EOS is signalled when BYE is received OR when
detecting end of stream even if no actual EOS is received.
Change-Id: I5cccb6845060ae6afd66d9f735b89da81476cd13
Andreas Huber [Wed, 6 Feb 2013 18:44:39 +0000 (10:44 -0800)]
HLS now properly publishes its "seekable" flags after connection
has successfully completed and a sufficient amount of data fetched,
and only then signals that preparation is completed.
Change-Id: I7684a14238b826909f518f2af506966e522dfcfc
The Android Automerger [Wed, 6 Feb 2013 15:10:35 +0000 (07:10 -0800)]
merge in jb-mr2-release history after reset to master
Andreas Huber [Tue, 5 Feb 2013 22:47:13 +0000 (14:47 -0800)]
RTSP now properly publishes its "seekable" flags after connection
has successfully completed and only then signals that preparation is
complete.
Change-Id: I1a60f718e673fe1462c69369c40eafbed6a14326
Andreas Huber [Tue, 5 Feb 2013 22:46:45 +0000 (22:46 +0000)]
Merge "Support for a "preparation" state that can take care of lengthy"
Andreas Huber [Tue, 5 Feb 2013 21:59:56 +0000 (13:59 -0800)]
Support for a "preparation" state that can take care of lengthy
operations in NuPlayer and its sources. Sources also can publish their
flags now and the mediaplayer UI will be able to pick up on these.
Change-Id: I4f2b7e5d105dcb4b6c9132cd0e8799efa0c6a14b
Andreas Huber [Tue, 5 Feb 2013 21:16:39 +0000 (13:16 -0800)]
Fix a typo in MediaPlayerFactory that would prevent us from
opting in to nuplayer for general media playback.
Change-Id: I050f5178aadbb0b8bf422861ef885745c0b9006a
Roger Jönsson [Thu, 17 Jan 2013 12:22:31 +0000 (13:22 +0100)]
Detect live streams
The information is used to decide on visibility of pause button and
to handle the duration clock correctly.
Change-Id: I286ac992fd171c7fc313e429326d38b6fc80e3fb
Andreas Huber [Tue, 5 Feb 2013 18:29:37 +0000 (18:29 +0000)]
Merge "Added HTTP support for SDP files."
Oscar Rydhé [Mon, 20 Feb 2012 09:15:48 +0000 (10:15 +0100)]
Added HTTP support for SDP files.
Added support for playing SDP files from http links. Previously,
SDP files only worked when started from rtsp links
(rtsp://a.b.c/abc.sdp), but they are just as common in http links.
patch provided by "Oscar Rydhé <oscar.rydhe@sonyericsson.com>"
Change-Id: Ic73af3a9a002009dbe8b04c267a4621bf7fe2f46
Andreas Huber [Tue, 5 Feb 2013 18:26:03 +0000 (18:26 +0000)]
Merge "Groundwork to support bidirectional, asynchronous communication"
Andreas Huber [Tue, 5 Feb 2013 18:14:26 +0000 (10:14 -0800)]
Groundwork to support bidirectional, asynchronous communication
between NuPlayer and its sources.
Change-Id: I1989022d806206b926555add3aa5c1fcf37aa78d
Andreas Huber [Tue, 5 Feb 2013 17:55:48 +0000 (17:55 +0000)]
Merge "Revive the code to support TCP interleaved transport"
Marco Nelissen [Tue, 5 Feb 2013 17:39:04 +0000 (17:39 +0000)]
Merge "Return error code if setting scaling mode fails"
Marco Nelissen [Tue, 5 Feb 2013 17:03:22 +0000 (09:03 -0800)]
Return error code if setting scaling mode fails
The return code from native_window_set_scaling_mode() was ignored.
Looking at the code review comments that introduced this code, it
seems like the intention was to return the error code in that case.
Change-Id: I9592cc378f0a0b960d37178aa0525fc17e8734ba
The Android Automerger [Tue, 5 Feb 2013 15:10:41 +0000 (07:10 -0800)]
merge in jb-mr2-release history after reset to master
James Dong [Mon, 4 Feb 2013 18:18:48 +0000 (18:18 +0000)]
Merge "Remove a few unused header includes from PlaybackSession.cpp"
Eric Laurent [Mon, 4 Feb 2013 17:05:30 +0000 (17:05 +0000)]
Merge "AudioFlinger: fix RecordThread initial device"
Glenn Kasten [Mon, 4 Feb 2013 16:15:34 +0000 (08:15 -0800)]
Fix crash when destroying a DuplicatingThread
Bug:
8121343
Change-Id: I9b1fc56c753e465e8a08916022e52db6c3a36906
The Android Automerger [Mon, 4 Feb 2013 15:10:31 +0000 (07:10 -0800)]
merge in jb-mr2-release history after reset to master
Eric Laurent [Sat, 2 Feb 2013 01:57:04 +0000 (17:57 -0800)]
AudioFlinger: fix RecordThread initial device
A regression was introduced when the audio device enums where
modified for a 32 bit representation: the device passed when
constructing a RecordThread was still the concatenation of input
device and output device bit fields on one 32 bit value which
is not possible anymore.
The fix consists in modifying the RecordThread constructor to accept
separate values for input and output devices.
Change-Id: I81fb5f4718428b54251e65d74b86e198ce15193e
James Dong [Wed, 30 Jan 2013 21:31:59 +0000 (13:31 -0800)]
Remove a few unused header includes from PlaybackSession.cpp
Change-Id: Id3147e2f2d5d75ccc172e67802639cd43ed68870