summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Richard Underhill [Tue, 13 May 2014 13:47:32 +0000 (14:47 +0100)]
TextParameters object now contains only the specific parameters set
[Issue#] N/A
[Problem] TextParameters contained Outline, Glow, DropShadow and Gradient
values even if some were not used.
[Cause]
[Solution] A container of Dali::Any objects describe the TextParameters
Change-Id: I73ff83b01b8e84df72f0fc3659e035c6a730d747
Signed-off-by: Richard Underhill <r.underhill@partner.samsung.com>
Kingsley Stephens [Tue, 13 May 2014 12:19:48 +0000 (13:19 +0100)]
(Image) Make Image GetWidth and GetHeight return natural size
[Issue#] N/A
[Problem] The size of the image should be available at all times
[Cause] N/A
[Solution] Make GetWidth and GetHeight return the natural size
Change-Id: I92f9374b8582c92a356abd62af277cf6b73af86c
Andrew Cox [Tue, 13 May 2014 17:14:31 +0000 (18:14 +0100)]
Revert "[SRUK] Bitmap core patch 2 of 4 - Replace all uses of the Bitmap class with new simpler type ImageData in core / adaptor communication"
This reverts commit
fffd4cca80c05dbc332a2b28ade580894a3915f3.
!! Note !! This code is good but we want to delay this series of Adaptor
changes until late May for client stability reasons.
Change-Id: I35e2c3d6940c0b531f9e47f87b15bcbc1c0abfbc
Signed-off-by: Andrew Cox <andrew.cox@partner.samsung.com>
Artem Dergachev [Thu, 17 Apr 2014 14:29:41 +0000 (18:29 +0400)]
Optimize HitTestWithinLayer (20% faster HitTest).
Whenever a temporary HitActor structure instance is created, constructor and destructor functions
are called for its .actor field.
In certain scenarios this results in creating and instantly deleting around 1000 dummy Actor
objects during a single call of HitTest().
HitTestWithinLayer is an important bottleneck in such scenarios, second only to
UpdateNodesAndAttachments (depending on the touchevent:framerate ratio).
This change makes the test around 20% faster and additionally reduces power consumption.
Change-Id: Iad9f59f39f6229516f036cc49025c2e0ef72ad3d
Paul Wisbey [Wed, 7 May 2014 12:54:28 +0000 (13:54 +0100)]
Stubbed-out Dynamics implementation to reduce binary size
[Issue#] N/A
[Problem] Binary size
[Cause] Lots of optional features built by default
[Solution] Don't build dynamics support by default. A stub API is
still provided; applications should anyway have been checking whether
DynamicsWorld was available i.e. whether bullet plugin was installed.
libdali-core.so sizes:
Before:
1.4 MB (1,433,052 bytes)
After:
1.3 MB (1,369,312 bytes)
Note - apps should check Stage::InitializeDynamics() return value, before
using any other Dynamics APIs.
/**
* @return A handle to the world object of the dynamics simulation, or an empty handle if Dynamics capable
* of supporting the requirement in config is not available on the platform.
*/
DynamicsWorld InitializeDynamics(DynamicsWorldConfig config);
David Steele [Tue, 13 May 2014 13:05:38 +0000 (14:05 +0100)]
Removing unnecessary dependencies
Change-Id: I50d11b88986777f29ca094e8934b920de136d402
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Francisco Santos [Thu, 8 May 2014 15:39:21 +0000 (16:39 +0100)]
Add missing uniforms.
Change-Id: I528d0108c4bd58adeac6cbf0d64b8ef6b96ca812
Signed-off-by: Ferran Sole <ferran.sole@samsung.com>
Andrew Cox [Wed, 26 Mar 2014 11:06:55 +0000 (11:06 +0000)]
Bitmap core patch 2 of 4 - Replace all uses of the Bitmap class with new simpler type ImageData in core / adaptor communication
[Issue#] N/A
[Problem]
Bitmap is a complicated abstract base class with several
interactions across core. Exposing it to the adaptor for the simple
job of moving some pixels and metadata from disk and application
memory buffers back to the core is unnecessary complication.
[Cause]
Legacy implementation.
[Solution]
Update all users of Bitmap at the core / adaptor boundary to employ the
simpler ImageData class instead and translate that to a Bitmap within
core.
A follow-up patch moves the code for Bitmap out of integration API altogether.
http://slp-info.sec.samsung.net/gerrit/#/c/455649/
A second follow-up patch refactors Bitmap:
http://slp-info.sec.samsung.net/gerrit/#/c/460583/
Will build fine alone but needs the following adaptor patch to
send back images in ImageData objects instead of Bitmap ones:
http://slp-info.sec.samsung.net/gerrit/#/c/460595/
Toolkit TCT tests are fixed by this patch:
http://slp-info.sec.samsung.net/gerrit/#/c/477139/
Signed-off-by: Andrew Cox <andrew.cox@partner.samsung.com>
Change-Id: I6150b29121adca9c7be84b7b0620d6406d6717ca
Kimmo Hoikka [Tue, 6 May 2014 12:50:23 +0000 (13:50 +0100)]
ContextObserver removal, part II: Its gone
[Issue#] N/A
[Problem] memory consumption
[Cause] observer pattern causes classes to be virtual unnecessarily
[Solution] remove observer pattern
Change-Id: Id0223b54cf989caf12a55ccadcaf9f48dff9a5b9
Signed-off-by: Ferran Sole <ferran.sole@samsung.com>
jonghyun.ho [Wed, 7 May 2014 05:11:03 +0000 (14:11 +0900)]
Upload package dali_0.9.15.
Related packages : dali_0.9.15, dali-adaptor_0.9.15, dali-toolkit_0.9.15
Change-Id: I182f2c5a4d9ae7ad2298a48c3162e8681eebcc77
Signed-off-by: Ferran Sole <ferran.sole@samsung.com>
Kimmo Hoikka [Wed, 7 May 2014 13:18:50 +0000 (14:18 +0100)]
Fix regression by previous optimization for mDirtyFlags
[Issue#] N/A
[Problem] TextInput demo working in a funny way
[Cause] regression in flag handling
[Solution] fix it
Change-Id: I4a1b5ddf5cfabdc67d6c2148ad925e7c14be3868
Signed-off-by: Ferran Sole <ferran.sole@samsung.com>
Kingsley Stephens [Tue, 6 May 2014 12:55:26 +0000 (13:55 +0100)]
(9 patch) Make ImageActor SetImage handle 9 patch images
[Issue#] N/A
[Problem] SetImage would not load the 9 patch image
[Cause] Only ImageActor New would load the 9 patch image
[Solution] Make SetImage load the 9 patch
Change-Id: I3b8dea9aefd520dedb253671345307fff536f309
Signed-off-by: Ferran Sole <ferran.sole@samsung.com>
Kimmo Hoikka [Thu, 1 May 2014 19:29:41 +0000 (20:29 +0100)]
ContextObserver removal, part I: Program no longer is context observer
[Issue#] N/A
[Problem] too many context observers
[Cause] observer pattern
[Solution] remove observer pattern
Change-Id: Ica8dd62613d689d97b7a9846154a589f7626450d
Signed-off-by: Ferran Sole <ferran.sole@samsung.com>
Paul Wisbey [Sat, 3 May 2014 11:27:56 +0000 (12:27 +0100)]
Fix regression with BitmapImage (width/height variety)
[Issue#] N/A
[Problem] Pointer returned by BitmapImage::GetBuffer() is invalid. This only
affects the variety of BitmapImage where the data is allocated by ResourceClient.
[Cause] Regression from "[SRUK] Compressed Texture Support for Dali Core"
[Solution] Partial rollback of BitmapExternal, Integration::Bitmap etc.
Change-Id: I3480e935ce908c1ced6cd57f0df87f6072292e3b
Signed-off-by: Ferran Sole <ferran.sole@samsung.com>
Francisco Santos [Fri, 2 May 2014 14:52:18 +0000 (15:52 +0100)]
Clamp text color values to valid range.
Signed-off-by: Francisco Santos <f1.santos@samsung.com>
Kimmo Hoikka [Wed, 30 Apr 2014 15:25:05 +0000 (16:25 +0100)]
Align data members, pointers first, classes then, next built in types and flags last. saves 8 bytes per actor on desktop
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Javon Prince [Fri, 2 May 2014 12:25:44 +0000 (13:25 +0100)]
Add 'ExclusiveArch: armv7l' limit build to arm architecture
Signed-off-by: Javon Prince <javon.prince@samsung.com>
Richard Underhill [Wed, 30 Apr 2014 16:53:49 +0000 (17:53 +0100)]
Build Optimization for size ( -Os )
Signed-off-by: Richard Underhill <r.underhill@partner.samsung.com>
Kimmo Hoikka [Thu, 1 May 2014 12:43:14 +0000 (13:43 +0100)]
GPUBuffer is unnecessarily large
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Javon Prince [Tue, 29 Apr 2014 18:16:09 +0000 (19:16 +0100)]
Add BakeFinal to Animation::EndAction
When an animation is stopped, it will save the property
values as if it's duration had normally elapsed.
Signed-off-by: Javon Prince <javon.prince@samsung.com>
Adeel Kazmi [Thu, 1 May 2014 12:52:17 +0000 (13:52 +0100)]
(AutomatedTests) Fixed build break.
Signed-off-by: Adeel Kazmi <adeel.kazmi@samsung.com>
Kimmo Hoikka [Thu, 1 May 2014 09:45:03 +0000 (10:45 +0100)]
Fix code to not include std set directly
[Issue#] N/A
[Problem] set exports its allocator causing extra symbols in symbol table
[Cause] including set directly
[Solution] include set-wrapper instead
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Change-Id: Ic2590137403d7d4d6d39971960098ffda9b5a9b2
Kimmo Hoikka [Wed, 30 Apr 2014 14:20:29 +0000 (15:20 +0100)]
Avoid creating unnecessary strings for built in shader uniforms
[Issue#] N/A
[Problem] 270 x 19 string objects created unnecessarily
[Cause] generic API
[Solution] specialize API for two cases
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Change-Id: I07975c3c8b7ab629084029c0e2012f98cb6401dd
Julien Heanley [Wed, 30 Apr 2014 12:33:40 +0000 (13:33 +0100)]
(Observers)Fix memory issues during observer iteration
[Issue#] N/A
[Problem] Crash could occur when adding a previously removed actor back into the scene
[Cause] Observer classes were using SceneObjectRemoved callback to remove themselves
from that object's observer list while the object is in process of iterating over the list
[Solution] Mark object as disconnected from scene and remove observer later
Change-Id: I552017a0d8284973f565cd08275f1270fc8e9668
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Adeel Kazmi [Thu, 3 Apr 2014 14:30:24 +0000 (15:30 +0100)]
(Properties) Ensure TypeInfo is set when creating Type via TypeInfo
[Problem] When created named types, the wrong type-info is used
when setting/getting properties.
[Cause] Gets the type-info related to the current type as it
does not know how it was created.
[Solution] Set the type-info when creating the object.
Change-Id: I3942ccf7c2c3ba36601dae1c4dd79f64a3c64e2c
Signed-off-by: Adeel Kazmi <adeel.kazmi@samsung.com>
jonghyun.ho [Mon, 28 Apr 2014 23:39:26 +0000 (08:39 +0900)]
Upload package dali_0.9.14.
Related packages : dali_0.9.14, dali-adaptor_0.9.14, dali-toolkit_0.9.14
Change-Id: I37c4cfff77f58046f0d195c4f2ad3067cccf4415
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Javon Prince [Thu, 10 Apr 2014 08:20:15 +0000 (09:20 +0100)]
Text: Quality improvement
Adjust amount of anti-aliased pixels dependent on the font pixel size.
This will only be accurate for text that is not scaled and not rotated
[Issue#] (N/A)
[Problem]
[Cause]
[Solution]
Change-Id: If83426348a970a93c250e31d7abbd5e7fb221a4c
Signed-off-by: Javon Prince <javon.prince@samsung.com>
Adeel Kazmi [Fri, 25 Apr 2014 14:36:06 +0000 (15:36 +0100)]
(Gestures) Ensure an update is done if gesture properties are updated.
[Problem] If gesture properties are updated and if there are no queued messaged from the event
thread, then update is still not run.
[Cause] Gesture properties are updated directly rather than via a message.
[Solution] Process gestures whenever we have an update, if the properties are updated, then we
should update the scene as well.
Change-Id: Ief1ee545fb7a3af289623cf100fe24e18f296569
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Adeel Kazmi [Fri, 25 Apr 2014 12:11:47 +0000 (13:11 +0100)]
MessageQueue now correctly reports that an update is not required
[Problem] The message queue always reports that it has messages even if it does not.
[Cause] The sceneUpdate counter never gets decremented to 0. This causes the update to run
every time.
[Solution] It was being cleared within a for loop where messages are being processed, it should be
outside this loop so is cleared if there are no messages.
Change-Id: I1d490c9c25f07813f2f904d659e5116e6b27a50c
Signed-off-by: Adeel Kazmi <adeel.kazmi@samsung.com>
Ferran Sole [Mon, 28 Apr 2014 07:52:44 +0000 (08:52 +0100)]
Fix bug when reusing renderlist on stencil renderers
[Issue#] N/A
[Problem] When reusing renderlists for stencil renderers, the stencil renderers
are rendered in the color buffer sometimes.
[Cause] When reusing a renderlist for stencil renderers the appropiate flags
where not seted on the renderlist which causes this renderers to be rendered in
the color buffer.
[Solution] Added a function to reset the flags correctly for stencil
renderlists as in the rest of renderlists (opaque, transparent and overlay)
Change-Id: Ic80a33af8ac8ac6bf606610fae5d6279cee8cf81
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Paul Wisbey [Mon, 14 Apr 2014 17:35:26 +0000 (18:35 +0100)]
Share a single weight property between constraints
[Issue#] (N/A)
[Problem] Not easy to blend between layouts using Constraints
[Cause] Each constraint weight has to be animated individually
[Solution] Provide an API for sharing a single weight property
between many Constraints
Change-Id: Icd81656930d85a8335ca8ee334843df6d2fdf179
Signed-off-by: David Steele <david.steele@partner.samsung.com>
jonghyun.ho [Thu, 24 Apr 2014 05:44:56 +0000 (14:44 +0900)]
Upload package dali_0.9.13.
Related packages : dali_0.9.13, dali-adaptor_0.9.13, dali-toolkit_0.9.13
Change-Id: I3845908ec388826f5c9eeb928f9d62c15f57a2c3
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Andrew Cox [Wed, 26 Mar 2014 11:06:55 +0000 (11:06 +0000)]
ImageData class to replace Bitmap class with new simpler type for core / adaptor communication
[Issue#] N/A
[Problem]
Bitmap is a complicated abstract base class with several
interactions across core. Exposing it to the adaptor for
the simple job of moving some pixels and metadata from
disk and application memory buffers back to the core is
unnecessary complication.
[Cause]
Legacy implementation.
[Solution]
Introduce a simpler and concrete ImageData class.
Change-Id: Ia3595f496326554d73dca24215f197c97fa86b6e
Signed-off-by: Andrew Cox <andrew.cox@partner.samsung.com>
Kingsley Stephens [Tue, 22 Apr 2014 15:08:58 +0000 (16:08 +0100)]
(prevnet) Fix prevnet issues
[Issue#] N/A
[Problem] Prevnet identified some issues needing fixing
[Cause] N/A
[Solution] Adjust code to satisfy prevnet
Change-Id: I501e0f51eaeee823cc428fce0edb6c0b2de27bea
Signed-off-by: Andrew Cox <andrew.cox@partner.samsung.com>
jonghyun.ho [Thu, 17 Apr 2014 01:35:01 +0000 (10:35 +0900)]
Upload package dali_0.9.12.
Related packages : dali_0.9.12, dali-adaptor_0.9.12, dali-toolkit_0.9.12
Signed-off-by: Andrew Cox <andrew.cox@partner.samsung.com>
Change-Id: I295fd2e3c3aafb97d1cf3d5a35fa4272db6a386d
Julien Heanley [Fri, 4 Apr 2014 10:24:37 +0000 (11:24 +0100)]
(PanGesture) Added environment variable "PAN_PREDICTION_MODE" for selecting gesture algorithm
Change-Id: I726d7567c67f7ecdec18c102d788d932500d1b43
Signed-off-by: Andrew Cox <andrew.cox@partner.samsung.com>
Javon Prince [Wed, 16 Apr 2014 12:30:22 +0000 (13:30 +0100)]
Move view-mode.h from unmanaged to managed api
[Issue#] (N/A)
[Problem] Causes build bream with core-api test
[Cause] view-mode.h was not in capi
[Solution] move view-mode.h to capi
Change-Id: I3acdc090a5d05c3e80ee5299233cce75c2c6b3e9
Signed-off-by: Javon Prince <javon.prince@samsung.com>
Adeel Kazmi [Wed, 16 Apr 2014 16:11:53 +0000 (17:11 +0100)]
Updated Tests style
Ferran Sole [Tue, 15 Apr 2014 09:01:28 +0000 (10:01 +0100)]
Added tag ( id ) to Constraints so they can be identified ( and removed ) by its tag value
[Issue#] N/A
[Problem] ItemView is not flexible enough
[Cause] ItemView removes all constraints from actors it manages
[Solution] Added tag to constraints so ItemView can tag its constraints and remove only those when needed
Change-Id: Id2e016f7b9f906bd4fa62d6fe249bd1ad60a1d97
Francisco Santos [Tue, 15 Apr 2014 13:11:01 +0000 (14:11 +0100)]
Fix wrong check for warning.
Change-Id: Ia43d8e6e9934a6826cb1d20d7bab0885a782c325
Signed-off-by: Francisco Santos <f1.santos@samsung.com>
Francisco Santos [Fri, 11 Apr 2014 16:47:02 +0000 (17:47 +0100)]
Klocwork fixes.
Change-Id: I8b5b701138565b76382993e8a9437c09d8644167
Signed-off-by: Francisco Santos <f1.santos@samsung.com>
Julien Heanley [Fri, 11 Apr 2014 09:27:52 +0000 (10:27 +0100)]
(PropertyNotification) Force a property notification from step conditions everytime they return true
[Issue#] N/A
[Problem]
[Cause]
[Solution]
Change-Id: Ie40c936e3ed1e3daf0b9f5fbba22ea5e4eeb9817
David Steele [Wed, 9 Apr 2014 14:52:27 +0000 (15:52 +0100)]
Resynced test cases and removed TET framework
[Issue#] N/A
[Problem] Test cases not quite up-to-date
[Cause] Time
[Solution] Migrated remaining test cases, removed TET framework
Change-Id: I2811ac5c197d75f0971ac0a20ebb2883bb79b9ee
[Verification] Build Repo, build and execute tests on desktop.
Signed-off-by: David Steele <david.steele@partner.samsung.com>
jonghyun.ho [Fri, 11 Apr 2014 00:56:13 +0000 (09:56 +0900)]
Upload package dali_0.9.11.
Related packages : dali_0.9.11, dali-adaptor_0.9.11, dali-toolkit_0.9.11
daemyung jang [Thu, 10 Apr 2014 04:33:15 +0000 (13:33 +0900)]
Update the spec configuration.
[Issue#] N/A
[Problem] Not build.
[Cause] When the system make the user binary,
the debug states are skipped.
[Solution] Update spec configuration.
David Steele [Thu, 10 Apr 2014 12:20:39 +0000 (13:20 +0100)]
Fixed 9 patch filename parser
[Issue#] N/A
[Problem] 9 patch filenames were not being matched.
[Cause] Left in decrement after switching to reverse iterator
[Solution] Removed initial decrement
[Verification] Build Repo
Change-Id: I046850b7e08ddd864fe7d5168e66cba3ee056db0
Signed-off-by: David Steele <david.steele@partner.samsung.com>
David Steele [Mon, 24 Mar 2014 20:06:55 +0000 (20:06 +0000)]
Migrating to new test framework
[Issue#] N/A
[Problem]
[Cause] Web-Tct & testkit-lite used.
[Solution] Migrate all local test cases to use testkit-lite on desktop and target
[Verification]
Desktop - run build.sh, then execute.sh.
Open summary.xml in your browser to see the results.
Target - run "sudo ./tcbuild build", then "./tcbuild install", then use tct-mgr to execute the tests on target.
Change-Id: I9d0dee70b0704503f9e977298ec665a1fcc99563
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Javon Prince [Wed, 19 Mar 2014 13:44:18 +0000 (13:44 +0000)]
Changes to support stereoscopic 3D
Add left/right cameras to stage, seperated by a stereo base
See Dali::Application::SetStereoBase() in dali-adaptor
Change-Id: If198d430704be6706ef0c81057a9e1950fd43de7
Signed-off-by: Javon Prince <javon.prince@samsung.com>
Victor Cebollada [Fri, 21 Mar 2014 16:00:17 +0000 (16:00 +0000)]
GlyphImage added. Retrieves a glyph image representing a given character.
[Issue#] N/A
[Problem] N/A
[Cause] N/A
[Solution] N/A
Change-Id: Ia23e3a786b74024b6feba2f6b499610e27eef24f
Signed-off-by: Victor Cebollada <v.cebollada@samsung.com>
Julien Heanley [Tue, 8 Apr 2014 13:01:52 +0000 (14:01 +0100)]
(PropertyNotification) Fix for adding notifications before ProxyObject's scene object has been created
[Issue#] N/A
[Problem] Calling PropertyNotification::SetNotifyMode could cause a crash in update thread
[Cause] The message would be sent off to update scene object value even when scene object had not been created
[Solution] Prevent notify mode message from being sent
Change-Id: I36dabb991873fa1efc5fcb8b7f7bf111847cba3e
Signed-off-by: Julien Heanley <j.heanley@partner.samsung.com>
David Steele [Tue, 8 Apr 2014 10:54:22 +0000 (11:54 +0100)]
Fixed issue with filename parser for 9 patch images
[Issue#] N/A
[Problem] .filename9.jpg or .filename#.jpg matched incorrectly.
[Cause] matcher not stopping if not detected . before 9 or hash.
[Solution] Ensured matcher stops after character before 9.suffix or #.suffix.
Note, the filename ".9.jpg" still matches - is this desired?
[Verification] Build Repo
Change-Id: I32cc2e8db4d35e5237c614a0617ddf8dff198518
Signed-off-by: David Steele <david.steele@partner.samsung.com>
David Steele [Thu, 3 Apr 2014 17:38:11 +0000 (18:38 +0100)]
(Image) Add support for 9-patch images
[Issue#] N/A
[Problem] There is no support for 9 patch images with the
stretch / child info information stored in a 1 pixel border within
the image
[Cause] N/A
[Solution] Added new class NinePatchImage, which is automatically
used by Image::New when a 9 patch filename is seen.
Adding a 9 patch image to an ImageActor via an Image handle will generate
an alternative BitmapImage with the border cropped out. It will set up the
ImageActor to use the 9 patch style and set the stretch areas up appropriately.
Adding a 9 patch image to an ImageActor via a NinePatchImage handle will
use the image as is (i.e. without the border cropped out).
[Verification] Build Repo
Change-Id: I8fa5cc34220ea9abcf085c600279ef92bc12da00
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Paul Wisbey [Fri, 4 Apr 2014 14:35:33 +0000 (15:35 +0100)]
(PanGesture) Logging enabled via environment variable
[Issue#] N/A
[Problem] Touch smootness etc. is hard to profile
[Cause] No logging features available
[Solution] Added new environment variable, usage:
DALI_LOG_PAN_GESTURE=1 your-app.exe
Change-Id: I96a78aca0838cd50d02365e952f10a0a74fd515f
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Richard Underhill [Mon, 31 Mar 2014 14:39:35 +0000 (15:39 +0100)]
Disabled automatic creation of TextStyle::Impl object on every TextStyle object.
[Issue#] N/A
[Problem] Unnecessary memory overhead on unused Implementation objects
[Cause]
[Solution] TextStyle::Impl is only created when a parameter contained
in that object is written to. Any query for a parameter on an
uninitialized Implementation object will return a default value
directly, without creating that Implementation object.
Signed-off-by: Richard Underhill <r.underhill@partner.samsung.com>
Paul Wisbey [Wed, 2 Apr 2014 12:44:53 +0000 (13:44 +0100)]
pan smoothing and prediction
Rebase of this change:
This patch keeps track of a history of the the last 60 touches.
The time stamp of the update (which is a pretty regular 16.7ms),
queries the history for the touch point at the update timestamp
minus a constant (e.g. 16.7ms). The history returns a touch
point which is the linearly interpolated value between the two
nearest touches in the sample. Thus reflecting where the touch
would've been at fixed intervals of a video frame.
-- Reverted to just blending the last two touch video-frames
Appears that most of the problem is noise in the touch positions
rather than their alignment with the frame interval.
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Richard Underhill [Wed, 19 Mar 2014 11:25:20 +0000 (11:25 +0000)]
Boost::any replacement in Dali Core functions
[Issue#] N/A
[Problem]
[Cause]
[Solution]
[Notes]
This needs to be verified with the corresponding adaptor and CAPI
patches:
http://slp-info.sec.samsung.net/gerrit/#/c/450053/
http://slp-info.sec.samsung.net/gerrit/#/c/450681/
[Updates]
Patch Set 2 Changes:
* Updated comments according to Kimmo feedback.
Patch Set 3:
* Synced with CAPI patch set 5.
Patch Set 4:
* Simple rebase.
Signed-off-by: Richard Underhill <r.underhill@partner.samsung.com>
Adeel Kazmi [Tue, 25 Mar 2014 16:07:08 +0000 (16:07 +0000)]
(Properties) Create a new actor using a Property::Map & event-thread-only default properties.
[Problem] 1) Cannot create an actor from a Property::Map.
2) No mechanism of having event-thread only properties within Core.
[Cause] 1) Not available.
2) Not available.
[Solution] 1) Create a method which creates an actor from a property map and vice versa.
2) Added mechanism which checks whether a property can be used as an input to a
constraint.
Signed-off-by: Adeel Kazmi <adeel.kazmi@samsung.com>
Julien Heanley [Tue, 25 Mar 2014 11:21:18 +0000 (11:21 +0000)]
(PanGestures)PropertyOwner.ResetDefaultProperties is now called on PanGestures
[Issue#] N/A
[Problem] PanGesture properties were never being reset and so would cause constraints to be updated every update frame
[Cause]
[Solution] Implemented ResetDefaultProperties
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Adeel Kazmi [Tue, 25 Mar 2014 17:49:04 +0000 (17:49 +0000)]
(Properties) Fixes to Actor/ShaderEffect/ImageActor incorrectly setting properties.
[Problem] 1) Actor world-scale type and name is out-of-sync.
2) Cannot create a shader effect if geometry is not specified in script.
3) ImageActor setting nine-patch property does not work.
[Cause] 1) Mistake in table initialisation. Table is clearer in a later patch.
2) Only creating a shader-effect if geometry is specified.
3) Not specifying that our values are in pixels.
[Solution] 1) Fix.
2) Create with default geometry type.
3) Specify that we have specified the 9-patch border in pixels.
Signed-off-by: Adeel Kazmi <adeel.kazmi@samsung.com>
David Steele [Mon, 24 Mar 2014 19:01:03 +0000 (19:01 +0000)]
(RenderTasks) Fixed test cases for chained framebuffers
[Issue#] N/A
[Problem] Test case failed after previous update
[Cause] Not enough care and attention
[Solution] Updated the test case to show that the framebuffer gets drawn
[Verification] Build Repo
Signed-off-by: David Steele <david.steele@partner.samsung.com>
David Steele [Fri, 14 Mar 2014 18:37:59 +0000 (18:37 +0000)]
(RenderTask) Make FB image attachments immediately ready
[Problem] Image attachments rendering Framebuffers are not
immediately marked as ready.
[Cause] Poor design
[Solution] Ensure attachments that render FBOs are marked as ready so they can
be rendered straight away.
Stops flickering of text with snapshot view.
[Verification] Build Repo
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Julien Heanley [Mon, 24 Mar 2014 14:11:11 +0000 (14:11 +0000)]
(Animation) Fix for issue in keyframed animation when only 1 key frame present
[Issue#] N/A
[Problem] Keyframe was not affecting property value
[Cause] key frames only affect property if current progress is between first and last key frame (if only 1 key frame then it uses same time for first and last)
[Solution] if progress is greater than last key frame time then set property to last key frame
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Richard Underhill [Fri, 21 Mar 2014 14:05:18 +0000 (14:05 +0000)]
Implementation of boost::any container substitute
[Issue#] N/A
[Problem]
[Cause]
[Solution]
Change-Id: I8a6ec38bc0c3932f90aff6b5debb972f2fb9fa9a
Signed-off-by: Richard Underhill <r.underhill@partner.samsung.com>
Julien Heanley [Wed, 19 Mar 2014 11:42:56 +0000 (11:42 +0000)]
(PropertyNotifications) Added incremental step notification
[Issue#] N/A
[Problem] No simple notification for properties changing over time
[Cause]
[Solution] Added incremental step notification
Change-Id: I9ad819fb3b19b2f45f722b034412b294af038915
Signed-off-by: Julien Heanley <j.heanley@partner.samsung.com>
David Steele [Mon, 17 Mar 2014 17:00:38 +0000 (17:00 +0000)]
Updated non-capi public API documentation to new style.
Updated coding style guide to reflect true dali style and new doxygen
style.
Change-Id: I30e697223143d3c096885d1b542e307cc8865970
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Francisco Santos [Tue, 18 Mar 2014 11:41:57 +0000 (11:41 +0000)]
Reimplement InvertYAxis API in CameraActor.
The InvertYAxis now has a slightly different meaning.
The Y invertion is done in the projection matrix.
Change-Id: I5af0b8fb0856b69787889c35ebfa7fceb07541cb
Signed-off-by: Francisco Santos <f1.santos@samsung.com>
Francisco Santos [Mon, 17 Mar 2014 14:58:52 +0000 (14:58 +0000)]
Make while loop a compound statement to fix code analysis warning.
Change-Id: I9aab6460d5428ad23fdc933f4600d083c5dd3efb
Signed-off-by: Francisco Santos <f1.santos@samsung.com>
David Steele [Fri, 14 Mar 2014 11:18:53 +0000 (11:18 +0000)]
Updated CAPI documentation style.
Change-Id: Idbf19cf1fc92558e10b2c77cd03fef395225041c
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Julien Heanley [Thu, 13 Mar 2014 11:50:05 +0000 (11:50 +0000)]
(Animation) Added component index to allow animation of separate components of a Vector
[Issue#] N/A
[Problem] Other than internally fixed components (e.g. POSITION_X), sub components of a Vector property could not be animated, creating a need for some properties to be split into multiple float values
[Cause]
[Solution] Added componentIndex parameter to Property() constructor
Change-Id: I82cd672301e5ac46e1a9a04a642553ca5e56d12a
Signed-off-by: Julien Heanley <j.heanley@partner.samsung.com>
Francisco Santos [Wed, 12 Mar 2014 18:10:03 +0000 (18:10 +0000)]
Print the value that we are trying to find when uniform test fails.
Change-Id: Iaaceb6ce9431443a271d8e48bd2d79adb6aba146
Signed-off-by: Francisco Santos <f1.santos@samsung.com>
Paul Wisbey [Fri, 7 Mar 2014 17:24:08 +0000 (17:24 +0000)]
Custom property owning objects
[Issue#] n/a
[Problem] There is no way of using custom properties without an actor.
[Cause] Can't instantiate a Constrainable object.
[Solution] Add method to create Constrainable objects.
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Change-Id: I0233ec5621ffdbe032d2e6e6bec8ec12d2a5463f
David Steele [Thu, 6 Mar 2014 15:50:50 +0000 (15:50 +0000)]
(Matrix) Fixing matrix multiply by quaternion
[Issue#] N/A
[Problem] Last row of matrix has random values
[Cause] It's not being initialized or written to
[Solution] Ensure the last row is set appropriately
Change-Id: I1b5d3da3d970b7c311ead4203eb197d11d52ccd7
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Richard Underhill [Mon, 10 Mar 2014 10:46:09 +0000 (10:46 +0000)]
Removed legacy resource tracking / logging
[Issue#] N/A
[Problem]
[Cause]
[Solution]
Change-Id: Ia957b8c40aff77b15ee53e681a09a3dcd4d17d04
Signed-off-by: Richard Underhill <r.underhill@partner.samsung.com>
jonghyun.ho [Wed, 19 Feb 2014 07:44:02 +0000 (16:44 +0900)]
Add DALI_COMPOSE_SHADER macro to write shader source conveniently.
[Issue] N/A
[Problem] We normally use double quotation marks to write a string such as "Hello World".
However many symbols are needed to add multiple lines of string.
[Cause] We don't want to write quotation marks at every line.
[Solution] Add DALI_COMPOSE_SHADER macro to write shader source conveniently.
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Change-Id: I4862e9723a9b16a591385d0f6dd25b5fbbe96b67
Paul Wisbey [Mon, 10 Mar 2014 17:19:21 +0000 (10:19 -0700)]
Merge "Print line numbers with shader source."
Francisco Santos [Thu, 27 Feb 2014 17:39:00 +0000 (17:39 +0000)]
Print line numbers with shader source.
Change-Id: I9729b9bf75d5082e0a3be5a5a194e4afe0ede8c6
Signed-off-by: Francisco Santos <f1.santos@samsung.com>
Francisco Santos [Mon, 10 Mar 2014 10:50:52 +0000 (10:50 +0000)]
Changed view matrix calculation to fix related problems.
Changed the view matrix to be the inverse of the Camera View matrix in the FREE_LOOK camera.
Updated the LOOK_AT view matrix to match an equivalent FREE_LOOK camera.
Updated sorting of renderables to work with the new view space.
Updated shader viewport uniform translation to match new view space.
Change-Id: Ideeacd1990ee4ebde715362e8006b0f2adebda91
Signed-off-by: Francisco Santos <f1.santos@samsung.com>
Javon Prince [Thu, 6 Mar 2014 11:05:15 +0000 (11:05 +0000)]
Clamp text shadow offset to limit clipping and/or bleeding of the shadow
[Issue#] P140304-04522
[Problem] If the shadow offset value is too large artifacts can be seen.
[Cause] Range of values is relative to the font point size.
[Solution] Clamp the shadow offset value internally.
Change-Id: Idb26ac962b08fc302988d1ea101962e637ed1e44
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Adeel Kazmi [Thu, 6 Mar 2014 15:53:52 +0000 (15:53 +0000)]
(PropertyRegistration) Moved enums outside of PropertyRegistration class.
[Issue#] N/A
[Problem] PropertyRegistration is not in CAPI but the enum is being used by Toolkit headers that
are in CAPI.
[Cause] N/A
[Solution] Move the enum outside of CAPI as we do not want to include type-registry in CAPI yet.
Change-Id: Iabc97d6257c1bac1870dab5084a1c9a53d3fad7e
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Kimmo Hoikka [Fri, 28 Feb 2014 15:12:59 +0000 (15:12 +0000)]
Inline default layer sort function to avoid unnecessary function call
[Issue#] N/A
[Problem]
[Cause]
[Solution] inline it and make layer aware when default function is used
Change-Id: Ic5376536a9a0925dcfab71aa02f9a0c0a337ab1e
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Paul Wisbey [Tue, 4 Mar 2014 11:14:28 +0000 (11:14 +0000)]
Remove redundant Core::SendEvent() & NotificationEvent
[Issue#] N/A
[Problem] Notification event is redundant
[Cause] Previously there was no way of telling Core to process events,
except by sending an event.
[Solution] Use the ProcessEvents() method, which was separated from Core::SendEvent()
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Change-Id: I3ae05b53931a685c59b5b0e2a449c633901d7ede
Lee Morgan [Mon, 3 Mar 2014 12:57:20 +0000 (12:57 +0000)]
Emscripten/llvm related issues
[Issue] N/A
[Problem] Emscripten/llvm patch (f33bf99) highlighted issues
[Cause] N/A
[Solution] N/A
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Change-Id: I625e1cd852bf199b5bea261581c71ae4d0ff83e2
Lee Morgan [Tue, 17 Dec 2013 18:10:26 +0000 (18:10 +0000)]
Emscripten workarounds and llvm syntax fixes
[Issue] N/A
[Problem] N/A
[Cause] N/A
[Solution] N/A
Adeel Kazmi [Tue, 18 Feb 2014 17:27:45 +0000 (17:27 +0000)]
(Properties) Added ability to add non-animatable event-thread only properties via type-registry
[Issue#] N/A
[Problem] 1) Toolkit in particular, but objects in general have several properties which do not
require any instance on the update-thread. This was partly why the OnPropertySet()
virtual method was added to CustomActor but this is intrinsically broken as the same
property could be set indirectly from the OnPropertySet() method which could cause
an unwanted loop.
2) Additionally, due to the change of property numbers where there are numerical gaps
between base and derived class properties, it has become quite difficult to retrieve
all properties for a certain class/object.
[Cause] N/A
[Solution] 1) A property registration system has been introduced which allows properties to be
dynamically added to a class on library load time (using the type-registry).
Upon registration, a getter and setter is specified.
Number ranges are used for property indices to distinguish between property types:
- Default Properties range from 0 to
10000000.
- Currently, event-only properties use
10000000 to
19999999.
- Custom properties are anything equal to or greater than
50000000.
2) Added a method to Object which retrieves a container of all the indices that an
object has. Only the indices are required as the name, type etc. can be retrieved by
using existing methods (if the index is known).
Change-Id: Icd477996465a4fb291f446feffa65ce2eb8c9120
Agnelo Vaz [Fri, 28 Feb 2014 10:34:02 +0000 (10:34 +0000)]
(KeyEvents) Removing deprecated and unused KeyEvents
[Issue#] N/A
[Problem] N/A
[Cause] N/A
[Solution] Removing Deprecated parts of KeyEvent
Removing cursorOffset and numberOfCharacters memebers variables as not used
[Verification] Build Repo
Julien Heanley [Fri, 28 Feb 2014 10:49:52 +0000 (10:49 +0000)]
(PropertyNotifications) Fixed TET issue
[Issue#] N/A
[Problem]
[Cause]
[Solution]
Change-Id: I6bb009706f17807873d6757c77c671a40f509474
Kimmo Hoikka [Thu, 27 Feb 2014 17:46:05 +0000 (17:46 +0000)]
Fix the renderable visibility checking and its relationship with render once
[Issue#] N/A
[Problem] incorrect algorithm
[Cause]
[Solution] fix the logic and clarify
Change-Id: I7114e35209f8a11d4f9e5c6aaa619b1744b20015
Francisco Santos [Thu, 27 Feb 2014 18:44:08 +0000 (18:44 +0000)]
Hide the sort function implementation and remove the unused functions.
[Issue#] (N/A)
[Problem]
[Cause]
[Solution]
Julien Heanley [Tue, 25 Feb 2014 10:33:56 +0000 (10:33 +0000)]
(PropertyNotification) Added conditions to allow checking of a specific vector components
[Issue#] N/A
[Problem] N/A
[Cause] N/A
[Solution] N/A
Change-Id: I468b57fb1deb46a711073d0d578c2da9f475f66a
David Steele [Tue, 25 Feb 2014 13:25:04 +0000 (13:25 +0000)]
(ImageSize) Add natural image size handling
[Issue#] N/A
[Problem] Need a way of switching actor between natural image size and
constrained, animated or set size.
[Cause] Natural image size not handled correctly
[Solution] Add mechanism to switch to natural size, (using existing
Actor::SetSize() will override this)
[Verification] Build Repo
Change-Id: Id4641d87bc8de68e613a0e10d8dddf26cddae9fb
Signed-off-by: David Steele <david.steele@partner.samsung.com>
David Steele [Wed, 19 Feb 2014 20:43:47 +0000 (20:43 +0000)]
(ImageLoader) Added synchronous method to get closest image size
[Issue#] N/A
[Problem] Can't synchronously determine what the loaded image size will be
[Cause] Not implemented
[Solution] Added a new method to the platform abstraction to get the closest
size for a given requested size and orientation correction.
Changed image implementation to set the correct image size on creation.
Change-Id: I7dab185ecc18e6f8605a3c3996d6586c47e5fb6e
[Verification] Build Repo
Signed-off-by: David Steele <david.steele@partner.samsung.com>
David Steele [Mon, 17 Feb 2014 20:16:48 +0000 (20:16 +0000)]
(Images) Reducing thread tennis
[Issue#] N/A
[Problem] Non-loaded images do not get size set immediately, and
rely on messages from update thread to update their size.
[Cause] N/A
[Solution] Store image size in image base class and in created ticket's image
attributes. Remove message notification of creation of bitmaps, native images and framebuffers
where the size is already known.
This will also prevent LoadingFinished signal from being sent for
bitmaps, native images and framebuffers that are created directly.
LoadingFinished will only be sent for images created from filenames.
On adding images to image-actors, only listens to LoadingFinished signal if the image
has a filename, otherwise it uses the known size.
Change-Id: Ie08a0c9070d903b87447841af8ddc1da0234a51f
Signed-off-by: David Steele <david.steele@partner.samsung.com>
Paul Wisbey [Mon, 3 Mar 2014 18:14:45 +0000 (18:14 +0000)]
[SRUK] Initial copy from Tizen 2.2 version
Signed-off-by: Paul Wisbey <p.wisbey@samsung.com>
Change-Id: I34519cb86d06e4edd581a5ff010224e96d702aab
Sehong Na [Sun, 16 Feb 2014 23:48:25 +0000 (15:48 -0800)]
Initial empty repository