* [android] Catches ‘NumberFormatException’ while invoking Integer.parseInt.
* [android] Catches ‘NullPointerException’ while converting Integer to int in Cocos2dxActivity.hideVirtualButton.
* Fixing issue where Scheduler::unscheduleAll() would not unschedule functions scheduled with Scheduler::performFunctionInCocosThread.
* Removing clear of _functionsToPerform from unscheduleAll, based on code review feedback. Adding a function removeAllFunctionsToBePerfomedInCocosThread, which will remove all pending functions scheduled with performFunctionOnCocosThread.
* Adding test case for Scheduler::removeAllFunctionsToBePerformedInCocosThread
* Adding proper cleanup code to Scheduler test for SchedulerRemoveAllFunctionsToBePerformedInCocosThread
* fixed#16938: Audio could not be played entirely on iOS/macOS. And refactored some code of decoding.
* Adds license header, removes unused macros.
* issue #16938: Corrects remaining frame count calculation.
* Updates include path.
* Adds comments for AudioDecoder class.
* Adds const for getter functions in AudioDecoder.
* issue #16938: Adds test case.
* Updates license years.
* Deletes trailing white spaces.
* Adds missing Scale9Sprite::initXXX overrided methods after refactoring Scale9 logic. Otherwise, `setupSlice9` will not be invoked.
Test case is: 1. Node: Scene3D -> Description -> Scale9Sprite display is wrong.
* Some fixes:
* Call Sprite::initWithFile(filename) in Scale9Sprite::initWithFile(filename) rather than Scale9Sprite::initWIthFile(filename, Rect::ZERO). Otherwise content size will be zero which is incorrect.
* setCapInsets in updateCapInset only when it’s in slice mode.
* setCapInsets should reset _insetLeft, _insetTop, _insetRight, _insetBottom member variables.
* Remove _previousCenterRectNormalized, just updateCapInset while switching to SLICE mode.
* Fixes JSTest -> ExtensionTest -> EditBoxTest: no background of editbox.
* Avoids an unused `setCapInsets` invocation in setupSlice9.
* fix: Scale9Sprite test case 18
Bug calculating rotated frames.
the width was being used instead of the height and vice-versa
* fixes test cases 8 and 9
* fix: Scale9Sprite test case 18
Bug calculating rotated frames.
the width was being used instead of the height and vice-versa
* fixes test cases 8 and 9
* initialize Director::_invalid on first init
* fix: Setting the GLView should be done after creating the RootViewController
* restore 2016 copyright on templates/cpp-template-default/proj.ios_mac/ios/AppController
* remove old iOS js and lua templates
* copy actual iOS templates from cpp to js and lua
* These variables can be accessed by reference
* Assign values in initialization list
* _range can also be in initialization list
* Prefixed ++ and added 'auto' to loops
Also removed extra whitespace
* Prefixed ++/--, and changed for-loop implementation to use 'auto' on some loops
* Use resolved file name (through dictionary lookup) in font atlas creation, to prevent creating multiple atlases for the same file.
Add font tests with and without font name replacement (through dictionary lookup).
* Fix cpp-tests xcode and cmake project
* Additional fix for cmake project
* Added -Wno-unused-parameter and removed all uses of the CC_UNUSED_PARAM macro
* Commented unused parameter names in .cpp files which previously used CC_UNUSED_PARAM
* Reverted -Wno-unused-param flag.
Moved deprecated touch methods definitions to .cpp file.
Commented more unused parameter names.
* Fixed some errors and warnings caused by the previous commit.
* Commented remaining unused parameter names in .cpp files.
* Fixed unused parameter warnings in headers.
* Fixed some more unused parameter warnings.
* Fixed some more unused parameters warnings.
* Fixed mistake in previous commit, missing ComAudioReader:: in method. Other warnings.
* Fixed build errors.
* Added missing file to CMakeLists
* Added method to switch between single and multitouch on Android.
* Moved variable definitions to the beginning of the class. Handle single touch for ACTION_MOVE and ACTION_CANCEL.
* Cleaner code by using if/else instead of break
* replaced some unordered_map::insert(std::make_pair(foo, bar)) with unordered_map::emplace(foo, bar)
* replaced some vector::push_back(std::make_pair(foo, bar)) with vector::emplace_back(foo, bar)
The old way will construct a std::pair first then call move constructor
when putting it into the container, while using emplace will construct
the pair in-place in the container. Also, the emplace way is shorter &
more concise.
* Add new methods to get the number of actions running in a
given node with a specific tag.
This is useful for cases that we want to know how many
animations with a **specific** tag is running in our target.
For example:
auto tag1Count = this->getNumberOfRunningActionsByTag(kMyTag1);
auto tag2Count = this->getNumberOfRunningActionsByTag(kMyTag2);
While this could be achieved (more or less) with callbacks
to increment the tagCount at start of action and decrement
at the end the proposed API is much more concise and less error
prone.
* Replace the old C casts to static_casts.
As discussed in #16789 issue thread @minggo asked
to change the old style of cast to newer, more safer
static_casts.
* Add the required parenthesis and normalize the spaces.
Normalize all the static_cast<Action *> to static_cast<Action*>
some places we have spaces, other places haven't.
Fix the missing parenthesis.
* SENSITIVE doesn't imply password, fixes crash on static cast in
* onTextChanged
* Fix type fontName instead of pFontName
* Set attributes that affect text content before setting initial text
In the case of a simple MTL file such as
newmtl cube
Ns 10.0000
Ni 1.5000
d 1.0000
Tr 0.0000
Tf 1.0000 1.0000 1.0000
illum 2
Ka 0.0000 0.0000 0.0000
Kd 0.5880 0.5880 0.5880
Ks 0.0000 0.0000 0.0000
Ke 0.0000 0.0000 0.0000
map_Ka cube.png
map_Kd cube.png
The entire istringstream would be consumed by LoadMTL and result in a 'not found' error due to the null check being in the wrong place
* Undo pull request #16742 while keeping fixed indentation on ccShader_UI_Gray.frag
* Removed unnecessary whitespace
* Switched implementation to C++11 string literal
Thanks to stevetranby for the suggestion
cc.Ray:intersects now returns two results, an indication of hit or
not, plus the distance along the ray. The function can still be called
expecting a single argument and hence is backward compatible.
Also correct the debug error message
* tileGid may overflow when use horizontal flip
kTMXTileHorizontalFlag = 0x80000000
when use horizontal flip, gid will bigger chan 0x7FFFFFFF
* use unsigned int to convert string to gid
* fixed#16735: [native] The behavior of (setRotation + setSkewX) is wrong.
Since I removed some logic, this patch may also improve a little bit performance.
* Adds test case for issue #16735
* fixed#16754: [android] Background music which is playing could not be paused while game enters background.
* Updates a comment of cocos/audio/android/AudioEngine-inl.h
* Support MultiView matrix array
Support MultiView matrix array and add new interface in GLProgram to
support shader header definitions.
* support getMatrixStackSize
* optimize code
* remove indents
* rename resetMatrixStack(unsigned int stackCount)
* Add document to interfaces
* Supplement document of render interface
* Use range-based for-loops and allocate std::vector size(), end(), cend(), rend(), crend() on the stack where favorable.
Other minor trivial changes were applied.
* Fixed Android compilation error
* Fixed windows-universal compilation error
* Adds slice9 support for Sprite.
how to use it:
// points coords
sprite->setCenterRect(Rect(x,y,w,h));
// normalized coords
sprite->setCenterRectNormalized(Rect(x,y,w,h));
starts scale9sprite in sprite
more slice 9 changes
sprite 9 slice works?
kind of works
correct anchor point
slice 9 works, at least with non-rotated atlases
streched works ok
better Y invert code.
cleaner, compatible with the previous code
yay, scaling workings...
need a better api now
sets scale correctly
yay! works as expected!
more fixes and tests
better test for box
setContentSize() changes size in non-9-slice mode sprites as well
setCenterRect() -> setCenterRectNormalized()
yet another test
adds setPositionNormalized()
adds setCenterRect() tests
remove devel team from xcode
tests: add one more tests
fun test!
improved test
yet another test for slice 9
* fixes anchorPoint issues
* adds documentation
* fix: using top-left coordinate for setRect
* sprite: fixes related to scale9 and tiled
* Sprite: slice 9 fixes
works Ok with rotated frames
uses `setCapInsets` instead of `setCenterRect` to be more familiar
with `UIScale9Sprite`
* fixes js and lua bindings for Autopolygon
* issue #16661: Replace c style uthash with std::unordered_map<K, V> in CCFontFNT.cpp
* issue #16661: Removes unused include "base/uthash.h" in CCGLProgram.cpp.
* Remove undrawn quads from the skybox mesh
CCSkybox had been implemented using a combination of two
inconsistent techniques. The rendering was being achieved via use of
the vertex shader's inherent support for cubemaps. That technique requires
only a single screen-covering quad, but the implemtation defined a cube.
Defining a cube mesh would be appropriate if one were simply mapping the
cubemap's 6 textures to faces, but is unnecessary if using the shader's
cubemap feature.
Not only was the use of a cube mesh unnecessary, but the particular way
the cube was defined and used meant that only one face would ever
contribute to the rendering. One of the other faces would always be culled
and the other four would be viewed edge on, mapping the the infinitesimally
thin lines defining the edges of the screen.
This commit simply removes the never-rendered faces, and adds comments
explaining the technique.
* Within test code, remove setScale calls applied to skyboxes.
A Skybox is defined in such a way that it's position, rotation and
scaling has no effect on it's rendering, so setScale has no effect.
The calls are removed from test code to avoid confusing anyone using
it as a template for their own programs.
* Make the Skybox correctly account for the camera's fov
The Skybox does not use the model/view and projection matricies. Instead
a single quad that maps exactly to the screen is rendered and the camera's
world matrix is passed into a shader that renders using cubemap lookups.
The way that works hardwires the fov to 90deg in both the horizontal and
vertical. That shows up particularly badly when the camera is pointed
directly downwards and rotated: the image deforms as it rotates.
This commit corrects the problem by using scaling factors from the
camera's projection matrix to prescale the matrix passed into the shader.
* Support audio interruption and resume callback when change earphone status.
* Refactor keyevent callback and add makecurrent for compatible with different binary.
* Enable Tizen indicator.
* Remove the unnecessary evasobject in the window and refactor the function for glview mode.
* Fix the compatible issue between Tizen 2.3 and 2.4.
* Fix the wrong directory of script resources for HellLua template project.
* Create performance-test project for Tizen platform.
* Use ++ prefix instead for non-primitive data types in CCAutoPolygon.cpp
* Use ++ prefix instead for non-primitive data types in SimpleAudioEngine.cpp
* Use ++ prefix instead for non-primitive data types in CCNodeLoaderLibrary.cpp
* Use ++ prefix instead for non-primitive data types in CCActionManagerEx.cpp
* Use ++ prefix instead for non-primitive data types in CCDataReaderHelper.cpp
* Use ++ prefix instead for non-primitive data types in HttpClient-android.cpp
* Use ++ prefix instead for non-primitive data types in CCDevice-linux.cpp
* Use ++ prefix instead for non-primitive data types in ScriptingCore.cpp
* Switched implementation to range-based for loops
* Added readability, fixed loop in CCDataReaderHelper.cpp
Thanks to Dimon4eg for the code review.
* Adds Spine binary-file format support
fixes issue #10438
* ignore SkeletonBinary.c from linter
* add another file to the exclude list
* lua manual fixes for spine
* SkeletonBinary.c and not .cpp
* missing files for win8.1
* don't compile it as crt
* new updates from spine
* Lua: remove unused 3rd argument from cc.mat4.transformVector
cc.mat4.transformVector takes a mat4 and a 4vector and returns the result
of applying the mat4 to the vector. Strangely it had been made to also
require a third argument called dst which it doesn't use but did check
was a vec4. This commit allows cc.mat4.transformVector to be called
without the unnecessary argument. We still permit calling with it for
backward compatibility.
Also changed is the wrapper in Cocos2d.lua so that a second form of
cc.mat4.transformVector can be used where the vector is supplied as
4 separate numbers.
Changed the test files to call without the unnecessary argument.
* Lua: remove unused argument from cc.mat4.createTranslation/Rotation
Both these functions have an extra unnecessary argument called dst,
although neither use it either to pass in or out a value. This
commit removes the extra argument and updates the test programs
correspondingly.
Also, within Cocos2d.lua, remove some definitions that are later
overwritten by new ones
* Removes insignificant log: `prevents overloading the autorelease pool`
* Removes unused retain -> autorelease stuff
* Checks the return value of `texture->initWithImage`, if it returns false, texture should be released and set to nullptr.
* More specifc log output in TextureCache::addImage(image, key)
* Checks arguments validation for VolatileTextureMgr::addImage
* fix: staggered hex maps returns correct tile
order was fixed in the array that contains the GIDs.
fixes Github issue #16512
* fix: add fix validation code for issue #16512
If playing a large audio and its loop flag is set to false, after the audio is played over, there will be 4 or 8 bytes memory leak by _rotateBufferThread variable.
Resolution:
The `isDestroy` flag should not be set to true in rotate buffer thread, to exiting the thread, we need another varible for this purpose.
Thanks @nicolechen819 for the feedback.
* getSlidBallRenderer methods to allow more customization of UISlider look and feel.
* Added const to declaration of getSlidBallRenderer methods for UISlider.
* fix: Label updateColor() and underline color work as expected (#16486)
fixes Github issue #15214 correctly. underline uses _displayColor
and not _textColorF. _textColorF is only for the text.
This also emulates the SystemFont underline behavior
fixes Github issue #16471
* fix conflict
* Issue 16471 (#16489)
* fix: Label updateColor() and underline color work as expected
fixes Github issue #15214 correctly. underline uses _displayColor
and not _textColorF. _textColorF is only for the text.
This also emulates the SystemFont underline behavior
fixes Github issue #16471
* docs: added more internal doc for issue_16471
* fixed#16473: OpenSLES error while play background music in writable path or sdcard. (#16475) (#16517)
* [ci skip][AUTO]: updating luabinding & jsbinding & cocos_file.json automatically
* ease action class macros to reduce code duplication
- update bindings for ease actions
- reorder clone method
- add back in create/etc methods to allow for EaseRateAction instances
- fix wrong clone return type
* Add note and minor cleanup
* use create in clone instead of new/init as in latest EaseBezierAction
* fix reverse for EaseIn/Out to match original code
* add assert for null
* refactor to match original code flow
* fix missing CC_DLL for windows builds
* fix win build issues due to abstract methods
* remove unused overrides
* fix: Label updateColor() and underline color work as expected
fixes Github issue #15214 correctly. underline uses _displayColor
and not _textColorF. _textColorF is only for the text.
This also emulates the SystemFont underline behavior
fixes Github issue #16471
* docs: added more internal doc for issue_16471
* Warning fixes in AudioDecoder.cpp
* AudioPauseResumeAfterPlay test case should start in onEnter.
And add a loop to pause & resume many times.
* AudioMixerController::initTrack should not setState for track. Otherwise, it may trigger Track state mess-up.
fixes Github issue #15214 correctly. underline uses _displayColor
and not _textColorF. _textColorF is only for the text.
This also emulates the SystemFont underline behavior
fixes Github issue #16471
* Update CCSAXParser.cpp
* #Improve XML parse performance with rapidxml pure SAX parser without compatible break.
* put rapidxml to external
* Update deps
Camera::_viewProjectionUpdated is not initialised in constructor. Should be initialised to false.
This can lead to redundant visibility checks if _viewProjectionUpdated happens to resolve true.
As Camera::visit() is never called, _viewProjectionUpdated is never actually updated, resulting in no visibility tests when the camera transform changes.
Fixes#16404
* Fix LNK4098 warnings
Fixes these two warnings:
```
warning: LNK4098: defaultlib 'MSVCRT' conflicts with use of other libs; use /NODEFAULTLIB:library
warning: LNK4098: defaultlib 'LIBCMT' conflicts with use of other libs; use /NODEFAULTLIB:library
```
* Fix LNK2001 error
The `version` library is not linked for MSVC compiler and it produces these errors when compiling with MSVC 19.
```
cocos2dInternal.lib(CCApplication-win32.cpp.obj) : error LNK2001: unresolved external symbol _GetFileVersionInfoSizeW@8
cocos2dInternal.lib(CCApplication-win32.cpp.obj) : error LNK2001: unresolved external symbol _GetFileVersionInfoW@16
cocos2dInternal.lib(CCApplication-win32.cpp.obj) : error LNK2001: unresolved external symbol _VerQueryValueW@16
```
This commit fixes that problem
* Update js auto binding settings with new ndk version
* Remove unnecessary property from Ref
* Merge part of Cocos2d-x-lite :
1. Improve network bindings
2. Use nullptr instead of NULL
3. Remove usage of some deprecated APIs
4. Add some defensive code and cleanup
* Fix spine track entry circle referencing issue
* cocos2d/cocos2d-x#16327: Fix global object leak issue in JSB
* Improve ScriptingCore::evalString, old implementation have been deprecated in Spidermonkey
* Update JS versions
* Add docs for web v3.13
* Improve ScriptingCore
* Manual merge cocos2d/cocos2d-x#16383 by DavidDeSimone
* Fix auto bindings configurations
* Update web ref
Adds Director::isValid()
removes DisplayLinkDirector()
removes virtual functions from Director()
Director is no longer an abstract class
github issue #14276
* new actions ResizeBy and ResizeTo
* FDP-5580-ResizeBy and ResizeTo fix and Test cpp : added a scale exemple in the test cpp to compare with our new resize + fix
* FDP-5580: actions ResizeBy and ResizeTo : changed some functions visibility
Currently, we couldn't find out the exact reason.
libwebsockets official said it's probably an issue of user code since 'libwebsockets' passed AutoBahn stressed Test.
* fix the return value when return null in java
fix the return value when return null in java
* fix getStringUTFCharsJNI
fix getStringUTFCharsJNI when srcjStr is null
* remove a word
remove a word added by mistake
* make a indention
line 233 utf8Str
* fix: TriangleCommands with custom uniforms can be batched
TriangleCommands with custom uniforms can be batched together.
This improves the performance when using custom uniforms without
adding any penalties when not using them
Github issue #16224
* better tests
* #improve node sort performance by std::sort + unsigned long long (_localZOrder)
* Use bit field for make _localZOrder storage more clearly.
* #Comment _localZOrder members for meaningful.
* fix issues
* #use explicit bits integer of localZOrder storage
* #Fix compile error, add include <cstdint> to CCNode.h
* use the optimization for 64bit and no optimization solution for 32bit
* Use the optimization for 64bit and no optimization solution for 32bit
* #Improve node sort performance 64-bit by std::sort + long long (_localZOrder)
* fix issues
* test case
* Revert "test case"
This reverts commit 85471d07a433f2b7f0265c3b7dd483fcc28e3e2f.
* fix issues
* use lambda, and any other sort issues.
* skip sotNodes
* fix compile error
* fix compile issue
* Support audio interruption and resume callback when change earphone status.
* Refactor keyevent callback and add makecurrent for compatible with different binary.
* Enable Tizen indicator.
* Remove the unnecessary evasobject in the window and refactor the function for glview mode.
* Fix the compatible issue between Tizen 2.3 and 2.4.
* Fix the wrong directory of script resources for HellLua template project.
* Create performance-test project for Tizen platform.
* Upgrades to Spine v3.4
Github issue #16115
* compiles and works on windows
tested on:
* win32
* win10
* win8.1
* include-linter: don't test spine files
we need them unmodified
* spine bindings fixes
also, doesn't use goblins-ffd. instead it uses goblins.
goblins-ffd is no longer present in spine v3.4
* fixed#16169: new android project crash in android 5.0.2 device (Nexus 7) when use 3.12
* fixed#16169: Adds comments for why we should not use `assignment operator` of `std::u16string`.
* Win32 move unicode utils global.
It need for easy converting utf-8 string in all code.
And use it api in user code also WinRT version.
* Move to CCUtils-win32
Also in WinRT version.
* Copyright
* #improve node sort performance by std::sort + unsigned long long (_localZOrder)
* Use bit field for make _localZOrder storage more clearly.
* #Comment _localZOrder members for meaningful.
* fix issues
* #use explicit bits integer of localZOrder storage
* #Fix compile error, add include <cstdint> to CCNode.h
* use the optimization for 64bit and no optimization solution for 32bit
translucent part of ETC format textures are rendered too bright due to
the fact that v3.12 sets PremultipliedAlpha to true by default. This
commit sets default value of _hasPremultipliedAlpha = false, deletes
unnecessary `_hasPremultipliedAlpha = false` code in multiple places,
making ETC format's _hasPremultipliedAlpha false while keeping other
formats unchanged.
* Pause instead of stopping music on resign
The [audioSource stop] causes the music to be stopped, and therefore it fails to resume later on a call to [audioSource resume], due to the addition of the if (!stopped) check in resume added in 26a04b38f2
* Don't re-pause music that has already been paused
In this situation:
1. Start game, music plays
2. Switch to Music app, game music paused, start other music
3. Switch to game, game music not resumed due to other music playing
4. Switch to Music app, stop other music
5. Switch back to game, game music should resume due to no other music playing
At step 5 the game music doesn't currently resume. This is because at step 4 when switching to the Music app, the game music gets re-paused (actually isPlaying is false, so systemPaused is set to NO, even though the music *is* still paused). This causes the music to not be resumed at step 5.
This change fixes this, by skipping the pause logic if systemPaused is already true.
Note that this is dependent on https://github.com/cocos2d/cocos2d-x/pull/16178
* Fix typo in previous fix.
* fixed#16170: Random crash in alGenBuffers(AudioCache::readDataTask) at startup.
* Minor fix: should -> may
* Minor fix: updates comments.
* Update comment.
* Comment fix again.
* [audio] Fixes some issues while stop/uncache right after play/preload on OSX/iOS platform.
* Better state control in AudioCache.
* Adds State enum class which contains INITIAL, LOADING, READY, FAILED states.
* AudioPlayer::destroy should check AudioCache::_isAudioLoaded flag, otherwise, app will freeze in cpp-tests/NewAudioEngineTest/6:Test invalid audio file.
* Should check whether state is INITIAL in the destructor of AudioCache.
* Adds test case for switching play state frequently.
* Skips invoking reading data callback if cache's state is INITIAL.
* Variables initialization for AudioCache class and fixes protential missing delete AL buffers.
* Audio latency fix for Android. Support preload effects on Android now.
Squashed commits:
[b6d80fe] log fix
[a0a918e] Fixes assetFd didn't be released while PcmData is returned from cache.
[4b956ba] Potential crash fix for PcmAudioPlayer while pause / resume.
[398ab8c] Updates LOG_TAG position in AudioEngine-inl.cpp
[e3634e7] include stdlib.h for posix_memalign
[9004074] fixes setVolume logical error.
[c96df46] Don't use another thread for mixing, enqueue is in a seperated thread, therefore doing mixing in another thread will waste more time.
[0a4c1a8] Adds setLoop, setVolume, setPostion support for Track
[c35fb20] Fixed include.
[cdd9d32] Do mixing by ourself. (TO BE POLISHED)
[6447025] µ -> u since µ could not be shown on some android devices.
[97be0c6] Don't send a silence clip.
[c1607ed] Make linter.py happy.
[0898b54] Puts enqueue & SetPlayState in PcmAudioPlayer::play to thread pool.
[b79fc01] Adds getDuration, getPosition support for PcmAudioPlayer
[80fa2ab] minor fix of the code position of resetting state to State::INITIALIZED
[d9c62f1] underrun fix for PcmAudioPlayer.
[9c2212a] UrlAudioPlayer, playOverMutex should be static, and should be used in update method.
[1519d2e] static variables
[19da936] _pcmAudioPlayer Null pointer check in AudioPlayerProvider.
[e6b0d14] Updates audio performance test.
[fc01dd4] Registers foreground & background event in AudioEngine-inl.cpp(android), the callback should invoke `provider`'s pause & resume method.
[e00a886] TBD: Pause & resume support for PcmAudioPlayerPool.
Since OpenSLES audio resources are expensive and device shared, we should delete all unused PcmAudioPlayers in pool while pause and re-create them while resume.
But this commit isn't finished yet, I don't find a better way to register pause&resume event in AudioEngine module.
[9e42ea3] Interleave mono audio to stereo audio. PcmAudioPlayerPool only contains PcmAudioPlayers with 2 channels.
[3f18d05] Adds a strategy for checking small size of different file formats.
[753ff49] Adds performance test for AudioEngine.
[09d3045] Releases an extra PcmAudioPlayer for UrlAudioPlayer while allocating PcmAudioPlayer fails.
[9dd4477] Using std::move for PcmData move constructor & move assignment.
[6ca3bcb] some fixes:
1) new -> new (std::nothrow)
2) break if allocate PcmAudioPlayer fails
3) renames 'initForPlayPcmData' to 'init'
4) PcmAudioPlayer destructor deadlock if 'init' failed
[54675b6] include path fix.
[a1903ca] More refactorings.
[19b9498] Makes linter.py happy. :)
[923c530] Fixes:
1) Avoid getFileInfo to be invoked twice
2) A critical bug fix for UrlAudioPlayer and adds detailed comments
3) __clang__ compiler option fix for AudioResamplerSinc.cpp.
[5ec4faf] minor fix.
[faaa0f3] output a log in the destructor of UrlAudioPlayer.
[9c20355] NewAudioEngineTest,TestControll crash fix.
[f114464] fixes an unused import.
[1dc5dab] Better algorithm for allocating PcmAudioPlayer.
[331a213] minor fix.
[e54084a] null -> nullptr
[f9a0389] Support uncache.
[89a364f] Removes unused update, and TODO uncache functionality.
[1732bf9] Supports AudioEngineImpl::setFinishCallback for android.
[43d1596] UrlAudioPlayer::stop fix.
[e2ee941] Test case fix in NewAudioEngineTest/AudioIssue11143Test
[5c5ba01] More fixes for making cpp-tests/New Audio Engine Test happy.
[8b554a3] Adds log while remove player from map.
[ed71322] If original file is larger than 30k bytes, consider it's a large audio file.
[fb1845a] Updates project.properties
[6f3839f] minor log output fix in AudioEngine-inl.cpp
[c68bc6c] Don't resample if the sample rate of the decoded pcm data matchs the device's.
[43ca45f] PcmAudioPlayers also need to be removed while they play over, but should not be deleted since their lifecycle is managed by PcmAudioPlayerPool.
[f5e63c9] Audio latency fix for Android. Support preload effects on Android now.
* Supports to loading audio files asynchronously.
* Crash fix for stop audio right after play2d.
* Minor fix for logic in AudioMixerController.cpp
* Adds missing files (CCThreadPool.h/.cpp).
* Minor fix for including.
* Minor fix for missing include <functional> in Track.h
* update license information in audio.h
* Don't use std::future/std::promise anymore since ndk counldn't support it well in armeabi arch.
* isSmallFile postion updated, fixes large audio file goto the checking logic of cache.
* std::atomic<int> isn't supported by ndk-r10e while compiling with `armeabi` arch, using a int with a mutex instead.
* fixes __isnanf & posix_memalign doesn't exist on low api (<=16) devices.
* namespace updated: cocos2d -> cocos2d::experimental
* Removes commented code in AudioMixerController.h/.cpp
* Removes unused code again, and fixes a memory leak of `Track` instance.
* Oops, namespace changed.
* Only outputs log in debug mode.
* Uses ALOGV for outputing logs in AudioEngine-inl.cpp
* const PcmData& -> PcmData for Track
* Fixes a protential crash in NewAudioEngineTest
* Adds `COCOS` prefix in header #ifndef COCOS_BALABALA #define COCOS_BALABALA
* Uses _ prefix for cocos code style instead of `m` prefix.
* Deletes AudioResamplerSinc related files.
* Bug fix from @minggo's reply on github.
* Don't need to invoke pause after in UrlAudioPlayer::prepare.
* Updates ThreadPool class, uses enum class and adds const keyword.
pandamicro