TPEPlatform/2014-05-08: Difference between revisions

 
(21 intermediate revisions by 9 users not shown)
Line 75: Line 75:
** OOAD
** OOAD
** Discuss with Benjamin & Randy on re-factory design of codec/muxer object creation
** Discuss with Benjamin & Randy on re-factory design of codec/muxer object creation
*{{Bug|975346}} - [LayerScope] LayerScope is not work while HWC is chosen
** feedback
** Define clear heirarchy among Sender functions


==Media==
==Media==
Line 84: Line 87:


*{{Bug|969372}} Intermittent | test_mediarecorder_record_no_timeslice.html | Test timed out.
*{{Bug|969372}} Intermittent | test_mediarecorder_record_no_timeslice.html | Test timed out.
** debug
** Change the media recording object into global scope, can avoid the event handler gone issue.
** Change the media recording object into global scope, can avoid the event handler gone issue.
** Land testing patch.
** Land testing patch.


====Alfredo Yang====
====Alfredo Yang====
*{{Bug|1003749}} - Local preview can't display after few calls. (H264 OMX)
** debug
** Microphone doesn't output pcm data in driver level. It needs to check the latest vendor libraries on PVT device.
*{{Bug|984215}} - VideoConduit reads the rotation flag and decides the rotation degree
** WIP
** Try to reduce memory copy.
====John Lin====
====John Lin====
* {{Bug|989944}} {{Bug|989945}}
* {{Bug|989944}} - [B2G][WebRTC] Serious frame dropping when enabling HW H.264 video in real-time mode
* {{Bug|989945}} - [B2G][WebRTC] long video lag when using H.264 codec.
** debug
** Build latest video benchmarking code from Ekr
** Build latest video benchmarking code from Ekr
** Experiment with QCOM suggestions to shorten decoder latency: still > 200ms
** Experiment with QCOM suggestions to shorten decoder latency: still > 200ms
Line 95: Line 108:


====JW Wang====
====JW Wang====
*{{Bug|1000813}} - WARNING: NS_ENSURE_TRUE(mDecoderStateMachine) failed
** debug
** testing patch
*{{Bug|1000841}} - MOZ_Assert: Assertion failure: mIsActive
** debug
** refactor MediaDecoderStateMachine, move reader idle/active code to MediaDecoderReader
** testing on try server
*{{Bug|981153}} - Intermittent test_seek.html | Test timed out.
** debug
** timeout again when HTTP Cache v2 is on (ref: http://www.janbambas.cz/new-firefox-http-cache-backend-implementation/)
** OnStartRequest not called from http channel, looks like a bug of http cache instead of media cache
====Benjamin Chen====
====Benjamin Chen====
* {{Bug|1000195}} - RTSP audio does not auto-play
** r+|land
*{{Bug|951043}} - [MediaEncoder::GTest] Implement unit test of Vorbis encoding
** r+
** fixing build error
*{{Bug|993598}} - POST of MediaRecorder Blob drops first byte with long video slices
** debug
** Cowork with vchang to figure out this.
*{{Bug|987568}} Refine the encoder codec selection in MediaEncoder::CreateEncoder
** WIP
*{{Bug|1006530}} - [RTSP][V2.0] Closing a audio RTSP streaming via tab page causes system
** debug
** Help to find the root cause, threading issue


==Platform==
==Platform==
====Steven Lee====
====Steven Lee====
* {{Bug|981780}} - Build error in --disable-webrtc builds: "VP8TrackEncoder.cpp:319: error: undefined reference to 'NV12ToI420'" and more
** Have a patch and test passed but it's stolen by jesup
* Help Wayne to disable webrtc on tarako
** Done
* Bugzilla introduction
** smb://fileserver1/public/StevenLee/introductionToBugzilla.pdf
====Phoebe Chang====
====Phoebe Chang====
*{{Bug|924692}} - [Touch Caret] Display a touch caret according to caret position in input element
** R?|R+
** Add flag in FrameMetrics to solve pan/zoom delay, R+.
** Refine coordinate transformation; fixing boundary check bug.
** Try server submitted and debugging.
====Vincent Chen====
====Vincent Chen====
* {{bug|983489}} - 983489: Crash on nsIAppStartup.quit()
** D
** Backout due to Bug 963113
* {{bug|1007011}} Cannot run mochitest-remote
** F/D
** Solved by Bug 1002545 Comment 1
====Chiahung Tai====
====Chiahung Tai====
*{{Bug|970725}} - Adapt gUM capture resolution based on encoding resolution
*{{Bug|970725}} - Adapt gUM capture resolution based on encoding resolution
Line 122: Line 183:
* {{Bug|879917}} - the WebIDL parser should more gracefully handle dictionaries being defined twice.
* {{Bug|879917}} - the WebIDL parser should more gracefully handle dictionaries being defined twice.
** land
** land
* {{Bug|1006986}} - flame debug build failed on MOZ_ASSERT_IF() in dom/bluetooth/bluez/BluetoothDBusService.cpp
* {{Bug|1006986}} - flame debug build failed on MOZ_ASSERT_IF() in dom/bluetooth/bluez/BluetoothDBusService.cpp
** file|land
** file | land


==Graphics==
==Graphics==
Line 135: Line 196:


* {{Bug|1006336}} - Remove nsThread stuff from GLContext
* {{Bug|1006336}} - Remove nsThread stuff from GLContext
** r?
** r? | f+
** got f+


* Study how efficient of screenshot capture/card view mode
* Study how efficient of screenshot capture/card view mode
** study
** study
** found three issues, will discuss with gaia and fire bugs
*** Open app receives "screenshot take event" twice with size 3x3
*** Screenshot(card view) is done by sw rendering, consider hw acceleration
*** Take screenshot with keyboard(home+power key), keyboard will go hidden soon, but you still see keyboard from screenshot


====Morris Tseng====
====Morris Tseng====
Line 157: Line 221:
** The root cause might be that all app use css property "width:100% height:100%". When we rotate the phone, we resize the system app, all apps within system app will also resize. Gaia team will takes this task.
** The root cause might be that all app use css property "width:100% height:100%". When we rotate the phone, we resize the system app, all apps within system app will also resize. Gaia team will takes this task.


* {{Bug|1007349}} - Do not do reflow for non-visible app on phone rotation
* {{Bug|1007349}} - Fix invalid timestamp in orientation angle calculation
** review
** review


====Chiajung Hung====
====Chiajung Hung====
* {{Bug|814524}} - Make WebGLContext::TexImage2D avoid readback for video elements
** WIP
** Strange behaviour on new codebase
** Still debugging
* {{Bug|959089}} - Separate GrallocBuffer allocations from Compositor thread
** land
* {{Bug|1001069}} - Slow getImageData() and putImageData() performance as canvas size increases
** debug
====Boris Chiou====
====Boris Chiou====
* {{Bug|975346}} - [LayerScope] LayerScope is not work while HWC is chosen
* {{Bug|975346}} - [LayerScope] LayerScope is not work while HWC is chosen
** WIP|F?
** WIP | F?
** Part I:  General functions for EffectChains (feedback?)
** Part I:  General functions for EffectChains (feedback?)
** part II: Support LayerScope for HwcComposer2D (WIP and feedback?)
** part II: Support LayerScope for HwcComposer2D (WIP and feedback?)
Confirmed users
346

edits