platform/upstream/efl.git
9 years agoeolian_cxx: Use C++ wrappers instead of native types as parameters
Vitor Sousa [Wed, 24 Jun 2015 14:23:01 +0000 (15:23 +0100)]
eolian_cxx: Use C++ wrappers instead of native types as parameters

Summary:
Using C++ wrapper types in the signature of methods of the generated
classes.
Now, when the type is an Eo type defined in Eolian, eolian_cxx will detect
it and use the equivalent C++ wrapper.

Types defined in Eolian no longer need to be specified in the lookup table,
so removed them from there.

Disable wrapping of non-const char*. The current wrapper (unique_ptr) is
not suited for some required operations like ownership acquisition from
C code.

Fix constructing properties that only have the "set" method.

Fix translation of non-const parameters.

Modified Makefiles to allow proper inclusion of evas_cxx headers required
by some tests and examples.

Move one auxiliary function to another reader to make it more accessible.

@feature

Reviewers: felipealmeida, q66, cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2752

9 years agoecore-buffer: fix that installing pkgconfig when ecore_buffer is enabled.
Seunghun Lee [Wed, 24 Jun 2015 13:18:11 +0000 (15:18 +0200)]
ecore-buffer: fix that installing pkgconfig when ecore_buffer is enabled.

Summary: ecore-buffer.pc file should be installed when ecore_buffer is enabled.

Reviewers: gwanglim, devilhorns

Reviewed By: devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2761

9 years agoexamples: Fix build after ecore_buffer commit.
Stefan Schmidt [Wed, 24 Jun 2015 11:46:43 +0000 (13:46 +0200)]
examples: Fix build after ecore_buffer commit.

Since 8b621775619b9959fe952b095b3baaa7aaa99572 make examples fails to build
as the examples have been included without checking if ecore_buffer is
actually enabled to build.

make[2]: *** No rule to make target '../../../src/lib/ecore_buffer/libecore_buffer.la', needed by 'ecore_buffer_example'.  Stop.

I'm running make examples during the nightlies now in the hope to catch those.

9 years agoefl - efreetd service move from dbus session bus to ecore ipc
Carsten Haitzler (Rasterman) [Wed, 24 Jun 2015 10:24:21 +0000 (19:24 +0900)]
efl - efreetd service move from dbus session bus to ecore ipc

this fixes warnings about no efreet dbus session bus in non session
environments as brought up on the mailing lists with:

Subject: Re: [E-devel] [EGIT] [core/efl] master 01/01: edje: unset
efreet cache update flag to prevent dbus connections

this moves all of efreetd client and server to ecore ipc, with client
auto-launching efreetd if not found as a service and trying for up to
500ms to connect. efreetd times out on last connection or no
connections after 10sec so it wont hang around forever if not in use.
it seems to work in my testing, so let me know if there is an issue.

@fix

9 years agoeo - silence ERR logs on constructor fail - this is valid behavior
Carsten Haitzler (Rasterman) [Wed, 24 Jun 2015 09:35:46 +0000 (18:35 +0900)]
eo - silence ERR logs on constructor fail - this is valid behavior

@fix

XXX: Given EFL usage of objects, construction is a perfectly valid thing
to do. we shouldn't complain about it as handling a NULL obj creation is
the job of the caller. a perfect example here is ecore_con and ecore_ipc
where you create a con or ipc obj then set up type/destination/port and
the finalize of the constructor does the actual connect and thus this
fails or succeeds based on if service is there.

until there is a better solution - don't complain here.

9 years agomore evas render 2 work - should not affect normal behavior or code
Carsten Haitzler (Rasterman) [Thu, 18 Jun 2015 11:48:37 +0000 (20:48 +0900)]
more evas render 2 work - should not affect normal behavior or code
paths unless turned on at runtime

9 years agoecore-drm: Quiet down output from libinput
Chris Michael [Tue, 23 Jun 2015 17:26:44 +0000 (13:26 -0400)]
ecore-drm: Quiet down output from libinput

Summary: As libinput has matured enough now, we can reduce the noise
in outputs logs by adjusting libinput log level priority

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeolian: relax doc parsing a little (no need to check if @since is allowed)
Daniel Kolesa [Tue, 23 Jun 2015 15:44:17 +0000 (16:44 +0100)]
eolian: relax doc parsing a little (no need to check if @since is allowed)

9 years agoeolian: simplify doc parsing logic a bit
Daniel Kolesa [Tue, 23 Jun 2015 15:34:45 +0000 (16:34 +0100)]
eolian: simplify doc parsing logic a bit

9 years agoecore-drm: Fix setting proper output subpixel value
Chris Michael [Tue, 23 Jun 2015 15:10:42 +0000 (11:10 -0400)]
ecore-drm: Fix setting proper output subpixel value

Summary: This adds a minor internal function to convert
DRM_MODE_SUBPIXEL values into something that the wayland protocol
supports.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Fix issue of outputs not getting registered with the wayland registry
Chris Michael [Tue, 23 Jun 2015 15:01:22 +0000 (11:01 -0400)]
ecore-drm: Fix issue of outputs not getting registered with the wayland registry

Summary: This fixes T2465: QT5 apps don't work. The issue here is that
when the randr code runs in E we make calls to
ecore_drm_output_enable/disable which is supposed to register the
outputs with the wayland registry. This was not happening due to the
enable/disabled checks at the top of these functions.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: Fix issue of events not getting dispatched properly
Chris Michael [Tue, 23 Jun 2015 15:00:07 +0000 (11:00 -0400)]
ecore-wayland: Fix issue of events not getting dispatched properly

Summary: In order for wayland events to be processed properly, we
should be dispatching any pending events off of the queue First before
we make the call to flush.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeolian: new doc token lexer/parser
Daniel Kolesa [Tue, 23 Jun 2015 14:28:46 +0000 (15:28 +0100)]
eolian: new doc token lexer/parser

This should allow us to more easily extend the format if desired
and overall makes the doc syntax parsing more readable and simpler.

9 years agoRevert "edje: unset efreet cache update flag to prevent dbus connections"
Daniel Kolesa [Tue, 23 Jun 2015 09:22:36 +0000 (10:22 +0100)]
Revert "edje: unset efreet cache update flag to prevent dbus connections"

This reverts commit 1edb35fff3fe54ac7eea1ba2c26e509284b4e470.

Accessing symbols from other libs like this should never be done. The
consequences of this are not significant enough to do this; better solution
can be found but definitely not like this.

9 years agoEcore_Con: Fix make distcheck
Jean-Philippe Andre [Tue, 23 Jun 2015 06:17:45 +0000 (15:17 +0900)]
Ecore_Con: Fix make distcheck

9 years agofb engine : Add NULL type handling in output free.
SangHyeon Lee [Mon, 22 Jun 2015 23:11:39 +0000 (08:11 +0900)]
fb engine : Add NULL type handling in output free.

Summary:
Add NULL type check in output_free of evas fb engine.
If engine setup is failed,
Render_Engine wil be NULL so output_free also need to
handling NULL check.

Test Plan:
It needs specific condition to reproduce,
           engine of ecore_evas is set to fb, and setup is failed,
           then Render_Engine is NULL, but ecore_evas_free will call
           output_free in fb engine's evas_engine.c

Reviewers: raster, cedric, Hermet

Reviewed By: Hermet

Subscribers: cedric, seoz, eagleeye, singh.amitesh

Differential Revision: https://phab.enlightenment.org/D2743

9 years agoedje: unset efreet cache update flag to prevent dbus connections
Mike Blumenkrantz [Mon, 22 Jun 2015 18:53:13 +0000 (14:53 -0400)]
edje: unset efreet cache update flag to prevent dbus connections

efreet_init began to be called with the addition of translatable strings,
but allowing dbus connections in edje (and thus edje-derived tools) will
result in errors when edje_init is called without a dbus connection available

9 years agoecore_con: eoifying ecore_con_eet.
Srivardhan Hebbar [Mon, 22 Jun 2015 16:55:54 +0000 (18:55 +0200)]
ecore_con: eoifying ecore_con_eet.

Summary:
This is still work in progress. I've added new file for temporary
purpose. Idea is to first eoify everything then change its namespace properly.

Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2602

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoedje: add possibility to speed up animation per object.
Cedric BAIL [Mon, 22 Jun 2015 16:46:12 +0000 (18:46 +0200)]
edje: add possibility to speed up animation per object.

This could come handy for previewing the effect of accelerating animation.

@feature

9 years agoedje: add support to control the transition time of edje animation globally.
jiin.moon [Mon, 22 Jun 2015 15:56:08 +0000 (17:56 +0200)]
edje: add support to control the transition time of edje animation globally.

Summary:
Add the transition scale flag and the transtion scale value.
This flag can be change using "transition_scale_enable" in edje program.
If flag is true, the transition scale value affect the transition speed of edje
The transition scale value can be change by elm_config.

@feature

Test Plan: elementary_config

Reviewers: Hermet, cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2287

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore-buffer: Add ecore_buffer library to EFL.
Seunghun Lee [Mon, 22 Jun 2015 14:23:54 +0000 (10:23 -0400)]
ecore-buffer: Add ecore_buffer library to EFL.

Summary:
Ecore_Buffer is abstraction of graphic buffer.
it supports backend of shm, x11_dri2 and x11_dri3 for now,
and this library also provides method to share buffers between processes.
Ecore_Buffer_Provider and Ecore_Buffer_Consumer is for this, sharing buffer.
provider draws something in to Ecore_Buffer, and consumer receives and displays it.
the binary, bq_mgr is a connection maker for buffer provider and consumer.
it can be included Enlightenment as a deamon later.

@feature

Test Plan:
1. Configure with --enable-ecore-buffer and --enable-always-build-examples to build examples.
2. Run bq_mgr, it connects consumer and provider.
3. Run ecore_buffer_provider_example and ecore_buffer_consumer_example

Reviewers: lsj119, gwanglim, cedric, zmike, jpeg, raster, devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2197

9 years agoedje: add of edje-3d to edje_calc.c
perepelits.m [Mon, 22 Jun 2015 09:17:26 +0000 (11:17 +0200)]
edje: add of edje-3d to edje_calc.c

Summary: Adding of Mesh_Node, Light and Camera to _edje_part_recalc.

Reviewers: raster, Hermet, cedric

Reviewed By: cedric

Subscribers: cedric, artem.popov

Differential Revision: https://phab.enlightenment.org/D2639

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: fix Evas.Canvas3D color pick
Oleksandr Shcherbina [Mon, 22 Jun 2015 09:11:29 +0000 (11:11 +0200)]
evas: fix Evas.Canvas3D color pick

Summary:
After this commit 684c51360f1591de4b8aee1f52d4edf8e8353822 it didn't work
Miss bind color pick texture to drawable data
@fix

Reviewers: Hermet, cedric

Reviewed By: cedric

Subscribers: cedric, Ievgen, se.osadchy

Differential Revision: https://phab.enlightenment.org/D2724

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_con: convert docs
Daniel Kolesa [Fri, 19 Jun 2015 14:44:22 +0000 (15:44 +0100)]
ecore_con: convert docs

9 years agoecore_con_base: converted docs
Daniel Kolesa [Fri, 19 Jun 2015 14:18:38 +0000 (15:18 +0100)]
ecore_con_base: converted docs

9 years agoefl_gfx_view: convert docs
Daniel Kolesa [Fri, 19 Jun 2015 14:01:31 +0000 (15:01 +0100)]
efl_gfx_view: convert docs

9 years agoevas canvas3d: Correct Eolian namespace
Oleksandr Shcherbina [Fri, 19 Jun 2015 09:34:48 +0000 (10:34 +0100)]
evas canvas3d: Correct Eolian namespace

Summary: Proceed fixing T2479

Reviewers: cedric, tasn

Subscribers: cedric, tasn

Maniphest Tasks: T2479

Differential Revision: https://phab.enlightenment.org/D2723

9 years agoecore_audio: convert docs
Daniel Kolesa [Thu, 18 Jun 2015 15:37:01 +0000 (16:37 +0100)]
ecore_audio: convert docs

9 years agoefl_gfx_gradient: convert docs
Daniel Kolesa [Thu, 18 Jun 2015 15:28:33 +0000 (16:28 +0100)]
efl_gfx_gradient: convert docs

9 years agoecore_audio_in: convert docs
Daniel Kolesa [Thu, 18 Jun 2015 14:55:17 +0000 (15:55 +0100)]
ecore_audio_in: convert docs

9 years agoedje: Fix Coverity CID1299026 and CID1299027: Explicit null dereference
Chris Michael [Thu, 18 Jun 2015 14:29:06 +0000 (10:29 -0400)]
edje: Fix Coverity CID1299026 and CID1299027: Explicit null dereference

Summary: Coverity reports that this potentially is passing a NULL
pointer to strcmp. As the GETSTR macro Can possible return NULL we
should be checking that return before passing it to strcmp.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-render2: Fix Coverity CID1306141: Dereference null return value
Chris Michael [Thu, 18 Jun 2015 14:24:40 +0000 (10:24 -0400)]
evas-render2: Fix Coverity CID1306141: Dereference null return value

Summary: Coverity reports that eina_thread_queue_wait was returning
NULL and thus msg variable being used here causes dereferencing a null
pointer.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-input-evas: Fix Coverity CID1306803 (logically dead code)
Chris Michael [Thu, 18 Jun 2015 14:20:07 +0000 (10:20 -0400)]
ecore-input-evas: Fix Coverity CID1306803 (logically dead code)

Summary: This cleans up the _ecore_event_evas_mouse_button function.
Coverity was reporting that the state tests in the 'if' function could
not be be reached due to eel variable being null. This moves the if
(!eel) check to run before the state tests.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-gl-drm: Cleanup compiler warnings when building gl_drm engine
Chris Michael [Thu, 18 Jun 2015 13:59:29 +0000 (09:59 -0400)]
evas-gl-drm: Cleanup compiler warnings when building gl_drm engine

Summary: This cleans up nasty compiler warnings when building the
gl_drm engine. If we do not define EGL_EGLEXT_PROTOTYPES then we end
up with implicit declaration warnings

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoEo: Fix windows support.
Tom Hacohen [Thu, 18 Jun 2015 11:30:13 +0000 (12:30 +0100)]
Eo: Fix windows support.

This is heavily based on a patch by Vincent Torri. I just refactored it
a bit so it doesn't break ABI on Linux, only on Windows (where it was
broken anyway).

This patch changes things so on Windows, functions are looked up only
based on their name. Because of the indirection (and export/import
tables) windows does, this is the only reasonable way to make it work.

9 years agoEo: Fix a potentially dangerous lack of {}.
Tom Hacohen [Thu, 18 Jun 2015 12:44:51 +0000 (13:44 +0100)]
Eo: Fix a potentially dangerous lack of {}.

You should always use curly brackets. Especially when the inside statement
has its own curlys. This can be confusing and has already lead to bugs in
many projects.

9 years agoeina/simple_xml_parser: don't parse the <, > in the attribute string.
ChunEon Park [Thu, 18 Jun 2015 10:51:06 +0000 (19:51 +0900)]
eina/simple_xml_parser: don't parse the <, > in the attribute string.

The tag identifier should not be inside of the string "~".
Previous logic doesn't care that scenario,
the parser context can be corrupted and it will be no more useful.

@fix

9 years agoeina: add eina_simple_xml example.
ChunEon Park [Thu, 18 Jun 2015 10:50:00 +0000 (19:50 +0900)]
eina: add eina_simple_xml example.

This example just prints tag and it's attributes in xml file.
But this example intended for the scenario,
if the attributes have the "<" or" >" characters inside of the string.

9 years agoefl_control,efl_text: convert docs
Daniel Kolesa [Thu, 18 Jun 2015 10:21:43 +0000 (11:21 +0100)]
efl_control,efl_text: convert docs

9 years agoefl_model_base: convert docs
Daniel Kolesa [Thu, 18 Jun 2015 10:01:12 +0000 (11:01 +0100)]
efl_model_base: convert docs

9 years agoExamples: Fix emotion basic example
Daniel Zaoui [Thu, 18 Jun 2015 02:41:45 +0000 (05:41 +0300)]
Examples: Fix emotion basic example

9 years agoExamples: fix compilation due to inline
Daniel Zaoui [Thu, 18 Jun 2015 02:32:39 +0000 (05:32 +0300)]
Examples: fix compilation due to inline

Undefined references were occurring on inline (not static) functions.
Adding static to these functions solves the issue.

9 years agoecore-x: move grabbed_is check to multi_handler
Lukasz Stanislawski [Wed, 17 Jun 2015 17:45:35 +0000 (13:45 -0400)]
ecore-x: move grabbed_is check to multi_handler

Summary:
Previous implementation used mouse_handler to create mouse events from touch
events if the device was grabbed and no emulated mouse events were emitted by X server.
However the same functionality could be achieved by using multi_handler with
additional condition on grabbed devices.

Test Plan: elementary_test -> Multitouch

Reviewers: cedric, raster, devilhorns

Subscribers: seoz, cedric

Differential Revision: https://phab.enlightenment.org/D2700

9 years agoefl_gfx_base: convert docs
Daniel Kolesa [Wed, 17 Jun 2015 16:32:34 +0000 (17:32 +0100)]
efl_gfx_base: convert docs

9 years agoevas_test_mesh: 3d -> canvas3d (fixes broken test compile)
Daniel Kolesa [Wed, 17 Jun 2015 16:29:51 +0000 (17:29 +0100)]
evas_test_mesh: 3d -> canvas3d (fixes broken test compile)

9 years agoevas: change interface evas_3d to evas_canvas3d
Oleksandr Shcherbina [Wed, 17 Jun 2015 14:31:14 +0000 (16:31 +0200)]
evas: change interface evas_3d to evas_canvas3d

Summary:
Regard to https://phab.enlightenment.org/T2479 it is need for right generation
Eolian files

Reviewers: cedric

Subscribers: cedric, tasn

Differential Revision: https://phab.enlightenment.org/D2711

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeina: add eina_quaternion_set
perepelits.m [Wed, 17 Jun 2015 14:15:36 +0000 (16:15 +0200)]
eina: add eina_quaternion_set

Reviewers: cedric, raster, Hermet

Subscribers: cedric, artem.popov

Differential Revision: https://phab.enlightenment.org/D2717

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas_canvas3d: cube chaotic was not initialized and relied on luck to be zeroed by...
Ivan Furs [Wed, 17 Jun 2015 14:09:24 +0000 (16:09 +0200)]
evas_canvas3d: cube chaotic was not initialized and relied on luck to be zeroed by default.

Summary: Cube chaotic was rotated, because angle in structure of "SceneData" have random number.

Reviewers: cedric, Hermet

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2643

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore-evas-drm: Default to GBM_FORMAT_ARGB8888
Chris Michael [Wed, 17 Jun 2015 14:18:28 +0000 (10:18 -0400)]
ecore-evas-drm: Default to GBM_FORMAT_ARGB8888

Summary: Seems some platforms (my setup here at least) don't have EGL
configs which suppose GBM_FORMAT_XRGB8888, so default to ARGB

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-gl-drm: Add support for EGL_MESA_platform_gbm
Chris Michael [Wed, 17 Jun 2015 14:16:50 +0000 (10:16 -0400)]
evas-gl-drm: Add support for EGL_MESA_platform_gbm

Summary: This change adds support for using EGL_MESA_platform_gbm to
get the display directly from gbm and to create a window surface
(usable by egl) directly from the gbm surface.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore_input_evas: remove tab and unusable code
Ji-Youn Park [Wed, 17 Jun 2015 13:52:16 +0000 (22:52 +0900)]
ecore_input_evas: remove tab and unusable code

remove tab and code which is not used

9 years agoeolian: remove the outdated and ugly database printer (needs rework)
Daniel Kolesa [Wed, 17 Jun 2015 13:26:07 +0000 (14:26 +0100)]
eolian: remove the outdated and ugly database printer (needs rework)

9 years agoeolian: check for C type keyword before trying to append as C type keyword
Daniel Kolesa [Wed, 17 Jun 2015 09:29:16 +0000 (10:29 +0100)]
eolian: check for C type keyword before trying to append as C type keyword

This fixes a segv when non-type Eolian keyword is used.

@fix

9 years agoeina_matrix: Added test cases for fixed point matrix functions
Vivek Ellur [Tue, 16 Jun 2015 15:47:29 +0000 (17:47 +0200)]
eina_matrix: Added test cases for fixed point matrix functions

Summary:
Added test cases for all the functions related to fixed point matrix

Signed-off-by: Vivek Ellur <vivek.ellur@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2701

9 years agoevas: use the right function with freetype backend.
Cedric BAIL [Tue, 16 Jun 2015 15:22:42 +0000 (17:22 +0200)]
evas: use the right function with freetype backend.

9 years agoevas: properly handle cleanup case.
Cedric BAIL [Tue, 16 Jun 2015 15:21:50 +0000 (17:21 +0200)]
evas: properly handle cleanup case.

9 years agoefl_image: convert docs
Daniel Kolesa [Tue, 16 Jun 2015 14:41:48 +0000 (15:41 +0100)]
efl_image: convert docs

9 years agoefl_player: convert docs
Daniel Kolesa [Tue, 16 Jun 2015 14:36:33 +0000 (15:36 +0100)]
efl_player: convert docs

9 years agoefl_gfx_gradient_radial: convert docs
Daniel Kolesa [Tue, 16 Jun 2015 14:31:02 +0000 (15:31 +0100)]
efl_gfx_gradient_radial: convert docs

9 years agoefl_text_properties: convert docs
Daniel Kolesa [Tue, 16 Jun 2015 14:28:24 +0000 (15:28 +0100)]
efl_text_properties: convert docs

9 years agoefl_gfx_shape: convert docs
Daniel Kolesa [Tue, 16 Jun 2015 14:23:06 +0000 (15:23 +0100)]
efl_gfx_shape: convert docs

9 years agoefl_player: Add playable property
Vitor Sousa [Tue, 16 Jun 2015 12:53:06 +0000 (13:53 +0100)]
efl_player: Add playable property

Summary:
Add "playable" property to Efl.Player.
It should be used to check if the currently loaded data can be
played.

Reviewers: tasn, felipealmeida, raster

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2704

9 years agoecore_imf/wayland: Add NULL check before accessing
Jihoon Kim [Tue, 16 Jun 2015 07:27:29 +0000 (16:27 +0900)]
ecore_imf/wayland: Add NULL check before accessing

9 years ago Ecore_input: Add "ECORE_EVENT_MOUSE_BUTTON_CANCEL" event
Ji-Youn Park [Tue, 16 Jun 2015 06:33:43 +0000 (15:33 +0900)]
 Ecore_input: Add "ECORE_EVENT_MOUSE_BUTTON_CANCEL" event

If ecore get the mouse cancel event(ex: display off, or window stack change) from below, we need to deal with.
Currently, evas also generates up event using evas_event_feed_mouse_cancel function when window hide is called.
if cancel event occurs, ecore also call evas_event_mouse_feed_cancel like window hide situation.
cancel event means all button or touch event should be canceled, but in the future, if we need to deal each cancel event according to touch point,
we need to add api like evas_event_feed_multi_cancel

@feature

9 years agoEdje: Add doc in edcref for description.clip_to
Jean-Philippe Andre [Tue, 16 Jun 2015 04:56:53 +0000 (13:56 +0900)]
Edje: Add doc in edcref for description.clip_to

Thanks Dave for pointing this out.

9 years agoevas/gl: Add null checks
Dongyeon Kim [Tue, 16 Jun 2015 04:47:16 +0000 (13:47 +0900)]
evas/gl: Add null checks

Summary: Add null check for glGetString version check and evgl_engine

Test Plan: Local tests

Reviewers: jpeg

Reviewed By: jpeg

Subscribers: wonsik, cedric

Differential Revision: https://phab.enlightenment.org/D2699

9 years agoEvas: Remove shader_3d .x generated file from BUILT_SOURCES
Jean-Philippe Andre [Tue, 16 Jun 2015 03:03:51 +0000 (12:03 +0900)]
Evas: Remove shader_3d .x generated file from BUILT_SOURCES

This should fix the build after make clean.
Even though those .x files can be re-generated at compile time, they
are checked in in Git and shipped in tarballs. They must not be removed
during make clean.

Thanks Ross Vandegrift for the report.

@fix

9 years agoEvas GL: Fix error message (ERR --> INF)
Jean-Philippe Andre [Mon, 15 Jun 2015 08:51:18 +0000 (17:51 +0900)]
Evas GL: Fix error message (ERR --> INF)

9 years agoevas/gl_x11, gl_cocoa: remove static variable in eng_font_draw
Dongyeon Kim [Mon, 15 Jun 2015 08:48:10 +0000 (17:48 +0900)]
evas/gl_x11, gl_cocoa: remove static variable in eng_font_draw

Summary:
When evas is recreated, this static variable value remains,
so it references already freed memory.

Test Plan: Local tests

Reviewers: cedric, jpeg

Reviewed By: jpeg

Subscribers: wonsik, mer.kim, cedric

Differential Revision: https://phab.enlightenment.org/D2678

9 years agoEdje: Add test case for "no_render" and "clip_to" state flag
Jean-Philippe Andre [Thu, 11 Jun 2015 08:46:50 +0000 (17:46 +0900)]
Edje: Add test case for "no_render" and "clip_to" state flag

9 years agoEvas: Add "no-render" flag for proxy sources & clippers
Jean-Philippe Andre [Tue, 12 May 2015 08:37:01 +0000 (17:37 +0900)]
Evas: Add "no-render" flag for proxy sources & clippers

Those objects should never be rendered on the canvas, even if they
are visible. On the other hand, they need to be rendered in mask or
proxy surfaces.

note: this patch includes some extra whitespaces changes :(

@feature

9 years agoEdje: Add clip as a property of a part description
Jean-Philippe Andre [Tue, 28 Apr 2015 09:56:02 +0000 (18:56 +0900)]
Edje: Add clip as a property of a part description

Since masking, for performance and themeing reasons, it becomes
interesting to be able to switch clippers on the fly. In particular,
switching from an IMAGE mask to a standard RECT clipper can save a
lot of power when masking is not required.

This new flag "description.clip_to" will behave a bit like a mix of
rel.to and visible:

- It points to a part by name, just like part.clip_to. This will
  override the clipper set by the part, or override the default clipper.

- Like "visible", it can not be interpolated between two values, so
  it will switch only at the end of a transition.

- By default there is no clip override, which means Edje will fallback
  to the part's clipper, if any, or the base (group's) clipper.

NOTE:
- Since a clipper that does not clip anything becomes a standard object,
  it is visible and rendered. This will in 99.999% cases not be the
  intended behaviour. Currently we can simply use a transparent RECT
  in order to always have something clipped by the clipper, but this is
  a hack and this will trigger rendering of masks in their surfaces even
  when they are not actually used.
  Ideally, there should be a flag indicating to Edje & Evas that an object
  should be considered a clipper in all situations, and never be rendered
  on screen.

TODO:
- Support Edje Edit
- Add Embryo & Lua functions
- Add support in edje_convert
- Add Edje/Evas flag to mark objects as "no_render"

@feature

9 years agoeina: add test cases for various eina_matrix operations
Vivek Ellur [Mon, 15 Jun 2015 07:06:31 +0000 (09:06 +0200)]
eina: add test cases for various eina_matrix operations

Summary:
Added test cases for various operations of 3*3 matrix

Signed-off-by: Vivek Ellur <vivek.ellur@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2692

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas-gl-drm: Cleanup gl_drm engine code a bit more
Chris Michael [Fri, 12 Jun 2015 15:48:30 +0000 (11:48 -0400)]
evas-gl-drm: Cleanup gl_drm engine code a bit more

Summary: This cleans up the gl-drm engine code a bit and avoids the
need to constantly destroy and recreate the gbm device itself. This
also adds checks during Outbuf create to handle the swap_mode as per
previously added enum.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-gl-drm: Move gbm_surface field to Outbuf structure
Chris Michael [Fri, 12 Jun 2015 15:47:47 +0000 (11:47 -0400)]
evas-gl-drm: Move gbm_surface field to Outbuf structure

Summary: This moves the Engine_Info's gbm_surface field to the Outbuf
structure

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-gl-drm: Add swap_mode enum
Chris Michael [Fri, 12 Jun 2015 15:44:29 +0000 (11:44 -0400)]
evas-gl-drm: Add swap_mode enum

Summary: This adds an enum we can use for setting/determining the
swap_mode to use for the gl_drm engine

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas_event: evas_event_feed_mouse_cancel set EVAS_EVENT_FLAG_ON_HOLD
Ji-Youn Park [Fri, 12 Jun 2015 14:10:07 +0000 (23:10 +0900)]
evas_event: evas_event_feed_mouse_cancel set EVAS_EVENT_FLAG_ON_HOLD

Evas_event_feed_mouse_cancel generates mouse up event if it meets situation witch mouse should be canceled.
but app also have to know this information. I add cancel information into evas up event generated by cancel function.
@fix

9 years agoevas/vg : Fixed issue regarding geometry change of vg object.
Subhransu Mohanty [Fri, 12 Jun 2015 08:24:55 +0000 (17:24 +0900)]
evas/vg : Fixed issue regarding geometry change of vg object.

9 years agoevas/engines : renamed ector native engine to "default" from "freetype"
Subhransu Mohanty [Wed, 15 Apr 2015 02:31:29 +0000 (11:31 +0900)]
evas/engines : renamed ector native engine to "default" from "freetype"

9 years agoevas_smart: Add a check for zero size before malloc
Vaibhav Gupta [Fri, 12 Jun 2015 08:47:13 +0000 (09:47 +0100)]
evas_smart: Add a check for zero size before malloc

Summary:
Size may be zero while allocating interfaces private data,
when NULL is passed in EVAS_SMART_SUBCLASS_IFACE_NEW for ifaces argument.

If size is 0, then malloc() returns either NULL, or a unique pointer
value that can later be successfully passed to free(). It is implementation
dependent. This case would be likely to occur in case of smart classes with
no interfaces.

@fix

Signed-off-by: Vaibhav Gupta <g.vaibhav1@samsung.com>
Reviewers: raster, Hermet, tasn, cedric

Subscribers: SanghyeonLee, sachin.dev, singh.amitesh, cedric

Differential Revision: https://phab.enlightenment.org/D2679

9 years agoedje entry - also filter 0x7f (DEL) char on input as it's invalid entry
Carsten Haitzler (Rasterman) [Fri, 12 Jun 2015 08:39:51 +0000 (17:39 +0900)]
edje entry - also filter 0x7f (DEL) char on input as it's invalid entry

@fix

9 years agoEvas filters: Remove unnecessary copy of buffers
Jean-Philippe Andre [Fri, 12 Jun 2015 07:44:43 +0000 (16:44 +0900)]
Evas filters: Remove unnecessary copy of buffers

When doing blur from alpha to rgba, there was an extra copy
step added, that was not required.

This should improve the performance a little bit in this situation.

9 years agoEvas filters: Fix blend with color with rgba buffers
Jean-Philippe Andre [Fri, 12 Jun 2015 06:44:31 +0000 (15:44 +0900)]
Evas filters: Fix blend with color with rgba buffers

- blend { color = red }
- blur { color = red }
were not working with RGBA input. This patch fixes that.

@fix

9 years agoedje: Remove unused variable
Chris Michael [Thu, 11 Jun 2015 16:59:21 +0000 (12:59 -0400)]
edje: Remove unused variable

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeolian: remove support for old event doc syntax
Daniel Kolesa [Thu, 11 Jun 2015 16:14:51 +0000 (17:14 +0100)]
eolian: remove support for old event doc syntax

9 years agoeo: move all event doc comments to new syntax
Daniel Kolesa [Thu, 11 Jun 2015 16:09:02 +0000 (17:09 +0100)]
eo: move all event doc comments to new syntax

9 years agoeolian: remove support for old doc style on types and vars
Daniel Kolesa [Thu, 11 Jun 2015 15:52:45 +0000 (16:52 +0100)]
eolian: remove support for old doc style on types and vars

9 years agoevas,ecore_con: more struct/enum doc updates
Daniel Kolesa [Thu, 11 Jun 2015 15:51:43 +0000 (16:51 +0100)]
evas,ecore_con: more struct/enum doc updates

9 years agoeolian: fix a bug in expr serializer with wrong sign insertion
Daniel Kolesa [Thu, 11 Jun 2015 15:41:32 +0000 (16:41 +0100)]
eolian: fix a bug in expr serializer with wrong sign insertion

9 years agoedje,evas: convert type docs to new format
Daniel Kolesa [Thu, 11 Jun 2015 15:14:37 +0000 (16:14 +0100)]
edje,evas: convert type docs to new format

9 years agoeolian/generator: fix enum field doc generator bug
Daniel Kolesa [Thu, 11 Jun 2015 15:13:22 +0000 (16:13 +0100)]
eolian/generator: fix enum field doc generator bug

This resulted in a field using the next field's docs rather
than its own, because it incremented the member before
even trying to retrieve the docs.

@fix

9 years agoEvas: Remove unused variables.
Tom Hacohen [Thu, 11 Jun 2015 08:49:29 +0000 (09:49 +0100)]
Evas: Remove unused variables.

These were introduced in ed01a32c136aaba8fb23a12afaf334b4c47fc62e.

9 years agoEo: Add Null check
Thiep Ha [Thu, 11 Jun 2015 08:46:36 +0000 (09:46 +0100)]
Eo: Add Null check

Summary: Add Null checking when allocate memory.

Reviewers: cedric, tasn

Reviewed By: tasn

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2664

9 years agoedje: add po file support in edc translation.
katpavalli [Thu, 11 Jun 2015 08:40:32 +0000 (10:40 +0200)]
edje: add po file support in edc translation.

Summary:
   Implementation to support .po files in edc for translation

Test Plan:
   Test Code to test this implementation is done as part of efl/src/examples/edje/edje-text.c and efl/src/examples/edje/text.edc
   edje_cc -md . text.edc && gcc -o edje-text edje-text.c `pkg-config --libs --cflags ecore-evas edje evas ecore eo`
   ./edje-text

   1) Click On the text "Click here"
      The language gets changed.

Reviewers: shilpasingh, cedric

Reviewed By: shilpasingh, cedric

Subscribers: cedric, rajeshps, govi, poornima.srinivasan

Differential Revision: https://phab.enlightenment.org/D2573

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: move evas_object_inject to super's ctor
Daniel Hirt [Thu, 11 Jun 2015 07:58:09 +0000 (09:58 +0200)]
evas: move evas_object_inject to super's ctor

Summary:
As we always call evas_object_inject in every Evas Object's ctcor,
it seems sensible to move this repeated bit of code to the super
(Evas.Object).

Test Plan: Expedite, Elementary_Test and pretty much everything

Reviewers: cedric, raster

Subscribers: JackDanielZ, cedric

Differential Revision: https://phab.enlightenment.org/D2665

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas - eo canvas events - move viewport to last to match enums in legacy
Carsten Haitzler (Rasterman) [Thu, 11 Jun 2015 01:51:47 +0000 (10:51 +0900)]
evas - eo canvas events - move viewport to last to match enums in legacy

9 years agoEvas GL: Make INF message more explicit
Jean-Philippe Andre [Thu, 21 May 2015 00:54:19 +0000 (09:54 +0900)]
Evas GL: Make INF message more explicit

Direct rendering is actually quite hard to enable, so make the
message very explicit for elm_glview users. Thanks zmike for
reporting this msg.

9 years agoEvas GL: Fix coverity CID 1304559, 1304560
Jean-Philippe Andre [Wed, 10 Jun 2015 23:39:43 +0000 (08:39 +0900)]
Evas GL: Fix coverity CID 1304559, 1304560

Summary:
<CID 1304559: Logically dead code>
The dead code is only valid for GLES backend, so move if statement
to be used for GLES backend only.
<CID 1304560: Bad bit shift operation>
When calculating depth bit, bit shifting could be done with negative values.
@fix

Test Plan: Local tests

Reviewers: jpeg

Reviewed By: jpeg

Subscribers: wonsik, cedric

Differential Revision: https://phab.enlightenment.org/D2654

9 years agoelua: unexpose "old" documentation api from eolian bindings
Daniel Kolesa [Wed, 10 Jun 2015 16:35:31 +0000 (17:35 +0100)]
elua: unexpose "old" documentation api from eolian bindings

9 years agoelua: remove unused code from lualian
Daniel Kolesa [Wed, 10 Jun 2015 16:33:14 +0000 (17:33 +0100)]
elua: remove unused code from lualian