platform/upstream/efl.git
8 years agoEvas 3d: Fix compilation for Windows
Jean-Philippe Andre [Fri, 8 Apr 2016 01:38:15 +0000 (10:38 +0900)]
Evas 3d: Fix compilation for Windows

near and far are #defined in windows.h. Old legacy stuff.
To be sure the code compiles, rename the variables.

Fixes T3423

8 years agoeina tmp get - support other tmp dir env vars and dont use xdg runtime
Carsten Haitzler (Rasterman) [Fri, 8 Apr 2016 02:03:44 +0000 (11:03 +0900)]
eina tmp get - support other tmp dir env vars and dont use xdg runtime

xdg runtime dir is NOT a tmp dir in the normal sense. it's not world
writable nor world readable. only for the user.  using
eina_environment_tmp_get() would imply that it is a regular tmp dir,
not a per-user private only runtime dir. that is something else
entirely.

@fix

8 years agoefl vpath - fix fallback handling to use right buffer for stat
Carsten Haitzler (Rasterman) [Fri, 8 Apr 2016 01:50:48 +0000 (10:50 +0900)]
efl vpath - fix fallback handling to use right buffer for stat

8 years agowayland: hook engine data from elm_win and update rect with evas size
Mike Blumenkrantz [Thu, 7 Apr 2016 18:14:14 +0000 (14:14 -0400)]
wayland: hook engine data from elm_win and update rect with evas size

the only way to accurately calculate the "evas" size in the engine from
window geometry is to have the size of the frame available to subtract from
window geometry

window geometry is NOT framespace--framespace is the entire csd region, possibly
containing a shadow, and window geometry is explicitly the region occupied by the
window, ie. not the shadowed part.

not my ideal solution to the synchronization issue here, but I guess this is a
benefit of the unified tree

fix T3396

8 years agoecore-wl2: redo Ecore_Wl2_Event_Window_Configure entirely
Mike Blumenkrantz [Thu, 7 Apr 2016 18:09:51 +0000 (14:09 -0400)]
ecore-wl2: redo Ecore_Wl2_Event_Window_Configure entirely

this is an event representing the "new" state of the surface after a
configure event. it must contain the exact states which could potentially
have changed in the configure in order to ensure synchronization between
csd states and window size.

ecore events for xdg-shell configures must be sent only upon receiving a
configure event since states are set by the compositor and not by the client

@fix

 #hoorayforbeta

8 years agoelm_win: handle icon object lifetimes more accurately, use icon_name in csd icon
Mike Blumenkrantz [Thu, 7 Apr 2016 18:04:21 +0000 (14:04 -0400)]
elm_win: handle icon object lifetimes more accurately, use icon_name in csd icon

in the case where an icon existed upon having an icon object set, the previous icon
object would be orphaned while still being visible. the new icon would then never
be set into the csd.

@fix

8 years agoelm_win: move frame obj creation/deletion for fullscreen property to state change cb
Mike Blumenkrantz [Wed, 6 Apr 2016 20:44:44 +0000 (16:44 -0400)]
elm_win: move frame obj creation/deletion for fullscreen property to state change cb

this was broken if the compositor unset fullscreen without being prompted by
the client (eg. ctrl+alt+f). it also created timing differences for csd calcs based
on when the fullscreen call occurred in the render cycle

@fix

8 years agoecore-wl2: remove maximized/fullscreen/transient from window types enum
Mike Blumenkrantz [Wed, 6 Apr 2016 20:48:58 +0000 (16:48 -0400)]
ecore-wl2: remove maximized/fullscreen/transient from window types enum

these aren't real window types and only complicate internals/api

 #hoorayforbeta

8 years agoecore_x: fix emission of ECORE_X_EVENT_XKB_NEWKBD_NOTIFY
Marcel Hollerbach [Tue, 5 Apr 2016 13:11:29 +0000 (15:11 +0200)]
ecore_x: fix emission of ECORE_X_EVENT_XKB_NEWKBD_NOTIFY

Summary:
the event XkbNewKeyboardNotify was never selected with
XkbSelectEventDetails so the event type could never occur.

The event is now each time emitted when a new event to the keyboard
happens. To clarify a bit: A new keyboard is always detected in x when
the set of keymaps changes.

@fix

Reviewers: raster, devilhorns, zmike

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3867

8 years agoElm_image: remove elm_image_sizing_eval from eo to legacy.
Ji-Youn Park [Thu, 7 Apr 2016 10:38:41 +0000 (19:08 +0830)]
Elm_image: remove elm_image_sizing_eval from eo to legacy.

elm_image_object_get api will be deprecated, so moved sizing eval func to legacy

8 years agoefl.progress: Rename property from value to progress_value
Yeshwanth Reddivari [Thu, 7 Apr 2016 09:55:44 +0000 (18:55 +0900)]
efl.progress: Rename property from value to progress_value

Reviewers: jpeg, singh.amitesh

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3872

8 years agoelm_nstate: introduce nstate widget and inherit check from it
Amitesh Singh [Thu, 7 Apr 2016 09:25:59 +0000 (18:25 +0900)]
elm_nstate: introduce nstate widget and inherit check from it

Test Plan:
elementary_test -to "nstate"

@feature

Reviewers: yashu21985, tasn, Hermet, seoz, smohanty, felipealmeida, JackDanielZ, jypark, woohyun, herdsman, raster, cedric, jpeg

Subscribers: saurabhbunty, seoz

Differential Revision: https://phab.enlightenment.org/D3786

8 years agoEfl: Rename flip and orient enums
Jean-Philippe Andre [Thu, 7 Apr 2016 08:32:40 +0000 (17:32 +0900)]
Efl: Rename flip and orient enums

Efl.Flip is now the enum, the interface is Flipable.
We could even use names like Efl.IFlip a la Java.

eo_prefix is used to have pretty names in C. legacy: null
is removed from the enums. orient_x0 is removed and only
defined in C with #define

8 years agoInterface: Progress - Add common interface for Slider and Progressbar
Yeshwanth Reddivari [Thu, 7 Apr 2016 07:50:03 +0000 (16:50 +0900)]
Interface: Progress - Add common interface for Slider and Progressbar

Summary:
Implement common interface efl_ui_progress and inherit slider and progressbar from common interface.
Currently legacy APIs will also call interface functions and later it can be deprecated.
This interface will be moved to EFL in src/lib/efl/interfaces when elementary is merged into efl.

Test Plan:
elementary_test -to 'slider'
elementary_test -to 'progressbar'

Reviewers: singh.amitesh, raster, tasn, felipealmeida, woohyun, cedric, jpeg

Subscribers: saurabhbunty, alok25

Differential Revision: https://phab.enlightenment.org/D3759

8 years agoEo: Try to fix windows build
Jean-Philippe Andre [Thu, 7 Apr 2016 06:46:47 +0000 (15:46 +0900)]
Eo: Try to fix windows build

For windows, EAPI needs to be redefined as dllexport/dllimport.
Since we marked all EO APIs as weak, we had two different EAPI
macros: EAPI and EWAPI. Unfortunately, EWAPI was never redefined
(only declared inside Eo.h).

See also a1a506e13e276d8ef6e522ae9.

See T3423. Thanks @vtorri for the report.

8 years agoEo: Mark class_get() as weak APIs
Jean-Philippe Andre [Thu, 7 Apr 2016 06:36:20 +0000 (15:36 +0900)]
Eo: Mark class_get() as weak APIs

While eolian-gen was generating EWAPI (weak) class_get()
symbol declarations, they were implemented as EAPI (strong).
Not sure if this mismatch had any significant effect.

This patch tries to address T3423 (windows build).
The mismatch between EAPI and EWAPI might be the problem.

8 years agoEcore: Try to fix build for windows
Jean-Philippe Andre [Thu, 7 Apr 2016 06:25:42 +0000 (15:25 +0900)]
Ecore: Try to fix build for windows

Move efl.h above ecore.h to not mess with EAPI's dllimport
vs. dllexport definition. This addresses T3423.

8 years agoEfl: Add internal strong symbol to fix build on GCC < 5.3
Jean-Philippe Andre [Thu, 7 Apr 2016 05:40:10 +0000 (14:40 +0900)]
Efl: Add internal strong symbol to fix build on GCC < 5.3

This fixes a crash in ecore_init, calling a weak function from
libefl that was resolved to NULL.

So, here's a fun thing happening with GCC < 5.3. Since a1a506e13e2
all EOAPI and EO class_get() functions are weak symbols. This means
that all APIs inside libefl.so are weak.

As a result, gcc linker with --as-needed skipped linking to libefl
since not a single strong symbol from libefl was required by
libecore. This is actually a bug in gcc linker since we do in fact
use symbols from libefl, just weak ones.

GCC 5.3 seems to be fixed, so people with GCC 5.3+ will not
experience any build/runtime issue. The current patch is
a workaround that bug, by artifically creating a strong symbol
required by ecore.

Other libraries than ecore might also need to call
__efl_internal_init, if they end up not being linked to libefl.

8 years agoecore: attempt to fix windows build by putting header in a more logical place.
Cedric Bail [Wed, 6 Apr 2016 21:34:45 +0000 (14:34 -0700)]
ecore: attempt to fix windows build by putting header in a more logical place.

8 years agoeolian: add Eolian support for Eina Promises
Felipe Magno de Almeida [Sat, 2 Apr 2016 01:50:28 +0000 (22:50 -0300)]
eolian: add Eolian support for Eina Promises

Add a promise object to allows Eolian interface to include promises
as a way to have asynchronous value return and composibility.

The usage is like this in a .eo file:

class Foo {
   methods {
      bar {
         params {
            @inout promise: Promise<int>;
         }
      }
   }
}

Which will create the following API interface:

void foo_bar(Eo* obj, Eina_Promise** promise);

and a Eina_Promise_Owner for the implementation, like this:

void _foo_bar(Eo* obj, Private_Data* pdata, Eina_Promise_Owner* promise);

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
8 years agoeo: add before and after macro hooks for API generation functions
Felipe Magno de Almeida [Fri, 11 Mar 2016 20:22:59 +0000 (17:22 -0300)]
eo: add before and after macro hooks for API generation functions

Add two parameters for macros that generate API functions in Eo so
that the generation can be customized with macros used by Eolian.

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
8 years agoRevert "Eo: Fix rare crash after call_resolve"
Tom Hacohen [Wed, 6 Apr 2016 10:17:05 +0000 (11:17 +0100)]
Revert "Eo: Fix rare crash after call_resolve"

I'm reverting this because according to jpeg it was possibly fixed in
5284b62e930f0bef0ed3125b3a485e0599451ef8.
I reverted this patch after his fix and followed his reproduction cases
and it seems that his second patch does indeed fix this issue so this
patch is no longer needed.

This reverts commit 0862b9d08384bc1d862b90952130ec988f56b33b.

8 years agoEo tests: Also test function calls in reinit test.
Tom Hacohen [Wed, 6 Apr 2016 09:16:24 +0000 (10:16 +0100)]
Eo tests: Also test function calls in reinit test.

Since we cache ops we also need to check function calls work
when we reinit eo, not just class_get functions.

This commit essentially verifies that
5284b62e930f0bef0ed3125b3a485e0599451ef8 was done correctly.

8 years agoefl build - ecore - add efl link as a PUBLIC link as efl headers it in
Carsten Haitzler (Rasterman) [Wed, 6 Apr 2016 09:18:58 +0000 (18:18 +0900)]
efl build - ecore - add efl link as a PUBLIC link as efl headers it in

I added Efl.h to Ecore.h - because we really want to use efl
interfaces ... like everywhere. so add it to the link too.

8 years agoelm_test: Fix warning with clang
Jean-Philippe Andre [Wed, 6 Apr 2016 07:36:55 +0000 (16:36 +0900)]
elm_test: Fix warning with clang

Invalid enum type

8 years agoelm_glview: Fix warning with clang
Jean-Philippe Andre [Wed, 6 Apr 2016 07:34:53 +0000 (16:34 +0900)]
elm_glview: Fix warning with clang

I'm using the same gcc construct to initiliaze a complex
struct with {} instead of {0}.

8 years agoelm_config: Fix typo and warning
Jean-Philippe Andre [Wed, 6 Apr 2016 07:33:16 +0000 (16:33 +0900)]
elm_config: Fix typo and warning

A major typo (hard to find with the naked eye) was present in
elm_config's list of text & color classes. See D3487.

I'm pretty sure this feature has not been used at all.

8 years agoels_box: Fix warning with clang
Jean-Philippe Andre [Wed, 6 Apr 2016 06:16:37 +0000 (15:16 +0900)]
els_box: Fix warning with clang

warning: comparison of constant 100 with expression of type
         'Evas_Aspect_Control' is always true
         [-Wtautological-constant-out-of-range-compare]

8 years agoelm: Fix some warnings with clang
Jean-Philippe Andre [Wed, 6 Apr 2016 06:22:20 +0000 (15:22 +0900)]
elm: Fix some warnings with clang

warning: missing field 'desc' initializer
         [-Wmissing-field-initializers]

Solution: use gcc extension to init structs with {}.
This is a bit ugly, but having too many warnings leads to
ignoring them and not noticing valid ones.

The warning is triggered because the first member of Eo_Event
is not a primitive type (it's a struct _Eo_Opaque *).

8 years agoelm_web: Fix warning (missing initializer)
Jean-Philippe Andre [Wed, 6 Apr 2016 07:24:33 +0000 (16:24 +0900)]
elm_web: Fix warning (missing initializer)

8 years agoelm_image: Fix warning (wrong enum type)
Jean-Philippe Andre [Wed, 6 Apr 2016 05:10:15 +0000 (14:10 +0900)]
elm_image: Fix warning (wrong enum type)

8 years agoEo: Fix function cache after eo reinit
Jean-Philippe Andre [Wed, 6 Apr 2016 05:02:05 +0000 (14:02 +0900)]
Eo: Fix function cache after eo reinit

The function call resolve cache may be broken after an eo
shutdown + init cycle, leading to calls to invalid functions.
This adds an static uint for each and every single EO API
entry point, as well as an extra if() check.

Now I'm not sure if the previous commit 0862b9d08384bc1d8 is
still necessary.

8 years agoEfl vpath: Fix usage of @class function
Jean-Philippe Andre [Wed, 6 Apr 2016 04:27:09 +0000 (13:27 +0900)]
Efl vpath: Fix usage of @class function

Calling an @class function with a real object is not safe.
Missing check somewhere? I believe this should have failed safely.

8 years agoEo: Fix rare crash after call_resolve
Jean-Philippe Andre [Wed, 6 Apr 2016 02:25:29 +0000 (11:25 +0900)]
Eo: Fix rare crash after call_resolve

It seems that calling a @class function with an EO object
(that was not the required Eo_Class) lead to a situation
where func->func was NULL. And that meant a crash after
call_resolve.

The proper fix is to properly call a @class function with a
class object.

8 years agoInterface: Flip and orientation interface
Yeshwanth Reddivari [Wed, 6 Apr 2016 01:46:04 +0000 (10:46 +0900)]
Interface: Flip and orientation interface

Summary:
Added flip and orientation interface and used them in evas_image.
Removed efl_image_orientation_set API and used efl_orientation_set and efl_flip_set API.
In implementation part, converted enums back and forth in order to keep current implementation as it is.

Test Plan: src/examples/evas/evas-images5.c

Reviewers: singh.amitesh, raster, tasn, herdsman, woohyun, cedric, felipealmeida, jpeg

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3844

8 years agoelm_layout: Attach object to iterator
Jean-Philippe Andre [Mon, 4 Apr 2016 12:39:29 +0000 (21:39 +0900)]
elm_layout: Attach object to iterator

This is part of a new API, not a fix

This was missing in 3c40ebb99887ba4c74c40e4509c49cc675877886

8 years agoelm_box: Minor fixes
Jean-Philippe Andre [Mon, 4 Apr 2016 09:57:26 +0000 (18:57 +0900)]
elm_box: Minor fixes

8 years agoElementary: Don't send selected signal to a part with no selected state. Removes...
Stephen Houston [Wed, 6 Apr 2016 01:39:35 +0000 (20:39 -0500)]
Elementary: Don't send selected signal to a part with no selected state.  Removes warnings.

8 years agoelm_win: update wayland opaque region during pre-flush...always
Mike Blumenkrantz [Tue, 5 Apr 2016 19:10:57 +0000 (15:10 -0400)]
elm_win: update wayland opaque region during pre-flush...always

@fix

8 years agoeo-cxx: Fix warning on weak symbols with inline functions
Felipe Magno de Almeida [Tue, 5 Apr 2016 18:38:49 +0000 (15:38 -0300)]
eo-cxx: Fix warning on weak symbols with inline functions

Removed the EOAPI which is not necessary as the function is already inlined

8 years agoecore: add tests for promises with ecore_thread_promise_run
Felipe Magno de Almeida [Sat, 2 Apr 2016 01:24:19 +0000 (22:24 -0300)]
ecore: add tests for promises with ecore_thread_promise_run

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
8 years agoecore: add promise for Ecore_Thread
Felipe Magno de Almeida [Fri, 1 Apr 2016 18:37:42 +0000 (15:37 -0300)]
ecore: add promise for Ecore_Thread

Add ecore_thread_promise_run function that returns a Promise
and runs function in another thread which you can set the
value on a Eina_Promise_Owner.

Eina_Promise* promise;
Ecore_Thread* thread = ecore_thread_promise_run
( &function_heavy, &cancellation_function, private_data,
sizeof(ValueType), &promise);

This calls function_heavy on another thread and returns
the Ecore_Thread and a Eina_Promise as an out-parameter.

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
8 years agoeina: add tests for promises
Felipe Magno de Almeida [Fri, 1 Apr 2016 18:30:37 +0000 (15:30 -0300)]
eina: add tests for promises

8 years agoeina: add promise
Felipe Magno de Almeida [Fri, 1 Apr 2016 17:49:58 +0000 (14:49 -0300)]
eina: add promise

Add a promise object that will allows Eolian interface to include promises
as a way to have asynchronous value return and composibility.

To understand better, let see the coming usage in a .eo file:

class Foo {
   methods {
      bar {
         params {
            @inout promise: Promise<int>;
         }
      }
   }
}

Which will create the following API interface:

void foo_bar(Eo* obj, Eina_Promise** promise);

and the equivalent declaration for implementation.

However, the API function will instantiate the Promise for the user
and the implementer of the class automatically. So the user of this
function will treat it as a @out parameter, while the developer of the
function will treat it like a @inout parameter.

So, the user will use this function like this:

Eina_Promise* promise; // No need to instantiate
foo_bar(obj, &promise);
eina_promise_then(promise, callback);

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 15:43:14 +0000 (11:43 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing
ecore_x_randr_crtc_panning_area_set function

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 14:14:19 +0000 (10:14 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing
ecore_x_randr_edid_display_interface_type_get function

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 13:58:11 +0000 (09:58 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing function
ecore_x_randr_screen_backlight_level_set

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 13:46:57 +0000 (09:46 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing API function
ecore_x_randr_edid_dpms_off_available_get

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 13:44:29 +0000 (09:44 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing
ecore_x_randr_edid_dpms_standby_available_get function

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 13:41:40 +0000 (09:41 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing
ecore_x_randr_edid_display_serial_get function

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 13:39:33 +0000 (09:39 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing
ecore_x_randr_edid_display_ascii_get function

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 13:24:42 +0000 (09:24 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing API function
ecore_x_randr_edid_dpms_standby_available_get

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 13:23:22 +0000 (09:23 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing function
ecore_x_randr_edid_dpms_available_get

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 13:20:53 +0000 (09:20 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for missing API function
ecore_x_randr_edid_manufacturer_serial_number_get

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing Ecore_X API
Chris Michael [Tue, 5 Apr 2016 13:18:10 +0000 (09:18 -0400)]
ecore-xcb: Add implementation for missing Ecore_X API

This patch adds an xcb implementation for ecore_x_randr_edid_model_get
function which was missing from ecore-xcb.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add implementation for missing API function
Chris Michael [Tue, 5 Apr 2016 13:16:49 +0000 (09:16 -0400)]
ecore-xcb: Add implementation for missing API function

This patch adds an xcb implementation for
ecore_x_randr_edid_manufacturer_model_get function

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add missing API function
Chris Michael [Tue, 5 Apr 2016 13:11:41 +0000 (09:11 -0400)]
ecore-xcb: Add missing API function
ecore_x_randr_edid_manufacturer_name_get

This patch adds an implementation inside ecore-xcb for missing API
function ecore_x_randr_manufacturer_name_get

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add missing API function ecore_x_randr_edid_info_has_valid_checksum
Chris Michael [Tue, 5 Apr 2016 13:06:06 +0000 (09:06 -0400)]
ecore-xcb: Add missing API function ecore_x_randr_edid_info_has_valid_checksum

This patch adds an implementation for
ecore_x_randr_edid_info_has_valid_checksum which was missing in the
ecore-xcb codebase.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Add missing ecore_x_randr_output_crtc_set API function
Chris Michael [Tue, 5 Apr 2016 12:59:24 +0000 (08:59 -0400)]
ecore-xcb: Add missing ecore_x_randr_output_crtc_set API function

This patch adds definition of missing API function
ecore_x_randr_output_crtc_set.

NB: This function also has no code in the xlib implementation

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoevas-wayland-shm: Update copyright notice
Chris Michael [Tue, 5 Apr 2016 11:42:56 +0000 (07:42 -0400)]
evas-wayland-shm: Update copyright notice

As portions of this code have been derived from existing code in
Weston, we should also be including their copyright/license text to
give credit.

Fixes T3421

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoeolian: Fix test cases
Jean-Philippe Andre [Tue, 5 Apr 2016 11:44:57 +0000 (20:44 +0900)]
eolian: Fix test cases

So sorry I missed this in my weak API patch! Bad JP didn't run
make check.

8 years agoMakefile: Add efl as internal dependency for ecore
Jean-Philippe Andre [Tue, 5 Apr 2016 11:10:17 +0000 (20:10 +0900)]
Makefile: Add efl as internal dependency for ecore

This commit broke the build:
 - 7c8b2da2863bb734236f5758288a00eb0d8c047a

Each and every single library including ecore now must
also include efl as well (for Efl.h).

8 years agoEvas smart: Mark all smart clippers as no_render
Jean-Philippe Andre [Tue, 5 Apr 2016 08:19:55 +0000 (17:19 +0900)]
Evas smart: Mark all smart clippers as no_render

This allows doing ugly things such as setting a clipper/mask
on any smart object, without having to worry about seeing a
white rectangle cover the whole screen.

While the above scenario is not a great idea (bypass elementary in
its handling of the clipper), having a big white rect covering the
entire UI seems even worse.

I wonder if elm objects shouldn't simply allow user clippers, and
somehow manage them along with the smart object internal clipper.

8 years agoEvas: Fix GL shader selection for afill
Jean-Philippe Andre [Tue, 5 Apr 2016 05:54:29 +0000 (14:54 +0900)]
Evas: Fix GL shader selection for afill

AFILL should be used only for certain images, which are
actually native external or dynamic content images.

All normal image don't need the AFILL flag, since they should
have proper argb and alpha flag.

8 years agopopup : fix create scroller bug when list item add.
Jinyong Park [Tue, 5 Apr 2016 08:29:12 +0000 (13:59 +0530)]
popup : fix create scroller bug when list item add.

Summary:
data-> scr = scroller object
data-> scroll = scrollable flag.
in _list_add function, when scroller object is not created, do _create_scroller.
 if (!sd->scroll) _create_scroller(obj);
must change to
 if (!sd->scr) _create_scroller(obj);

@fix

Reviewers: jaehwan, id213sin, cedric, raster, singh.amitesh

Reviewed By: singh.amitesh

Subscribers: herb, jpeg

Differential Revision: https://phab.enlightenment.org/D3866

8 years agoevas: image obj - use vath to find images (sync not async)
Carsten Haitzler (Rasterman) [Mon, 4 Apr 2016 12:50:30 +0000 (21:50 +0900)]
evas: image obj - use vath to find images (sync not async)

8 years agoelm - add vpath paths for application resources
Carsten Haitzler (Rasterman) [Mon, 4 Apr 2016 08:24:29 +0000 (17:24 +0900)]
elm - add vpath paths for application resources

this now allows vpath files to be looked uop using app directory
resources like @app.data/file.jpg ... if the object uses vpath to do
its hunting.

@feature

8 years agoefl ecore - init vpath on ecore init so we have a base object working
Carsten Haitzler (Rasterman) [Mon, 4 Apr 2016 08:12:38 +0000 (17:12 +0900)]
efl ecore - init vpath on ecore init so we have a base object working

this inits a new vpath object and adds it at priority 0 to the vpath
manager so you can use the vpath manager to create vpath file objects
and look things up.

@feature

8 years agoefl: vpath subsystem
Carsten Haitzler (Rasterman) [Fri, 1 Apr 2016 01:45:07 +0000 (10:45 +0900)]
efl: vpath subsystem

this adds a core vpath subsystem to efl that allows paths like:

  ~/file.jpg
  ~user/file.jpg
  (:tmp/file.jpg
  (:config/file.jpg
  (:videos/file.mp4
  (:pictures/file.jpg
  (:app.config/mycfg.cfg

etc. to be translated/looked up. it is desitgned to be async and call
event callbacks when ready. the reason for this complexity is fo in
future also handle:

  file:///whatever/file.jpg
  http://blah.com/file.jpg
  https://blah.com/file.jpg
  ssh://blah.com:~/file.jpg

etc.

@feature

8 years agoPopup: don't set wrap type if ELM_WRAP_NONE is given
Daniel Zaoui [Tue, 5 Apr 2016 05:39:26 +0000 (08:39 +0300)]
Popup: don't set wrap type if ELM_WRAP_NONE is given

If ELM_WRAP_NONE is forbidden, the internal variable should not be set
if this value is given as parameter.

8 years agoelc_popup: support siganl emit for the main layout of popup
Taehyub Kim [Tue, 5 Apr 2016 04:46:38 +0000 (10:16 +0530)]
elc_popup: support siganl emit for the main layout of popup

Summary:
Popup has the main layout to show popup view, but it is not resize object of popup.
As the result, we cannot emit the signal to the main layout when we want to change the state of the popup.
So, popup inherited a layout signal emit and redirect the signal to the main layout.

Test Plan:
1. add the custom signal in elm/popup/base/default layout
2. send the custom siganl in application code

Reviewers: woohyun, raster, cedric, Hermet, singh.amitesh

Reviewed By: singh.amitesh

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3852

8 years agoelm_win: newest is always unfocused when window just gets focus
WooHyun Jung [Tue, 5 Apr 2016 02:23:04 +0000 (11:23 +0900)]
elm_win: newest is always unfocused when window just gets focus

8 years agowayland_shm: Fix resize optimization
Derek Foreman [Mon, 4 Apr 2016 21:57:30 +0000 (16:57 -0500)]
wayland_shm: Fix resize optimization

We're supposed to allocate a large pool at startup and use it for
resizing to save pool allocations.  However, this was broken and
we ended up allocating both a large pool and a proper sized pool
every resize.

This restores correct behaviour.

8 years agoelm_win: conditionally emit focus and maximize signals in frame creation
Mike Blumenkrantz [Mon, 4 Apr 2016 21:20:49 +0000 (17:20 -0400)]
elm_win: conditionally emit focus and maximize signals in frame creation

@fix

8 years agoelm_win: update opaque region while fullscreened
Mike Blumenkrantz [Mon, 4 Apr 2016 20:30:59 +0000 (16:30 -0400)]
elm_win: update opaque region while fullscreened

@fix

8 years agoelm_win: unset opaque region when alpha is set
Mike Blumenkrantz [Mon, 4 Apr 2016 20:30:33 +0000 (16:30 -0400)]
elm_win: unset opaque region when alpha is set

@fix

8 years agowayland-egl: Fix use after free
Derek Foreman [Mon, 4 Apr 2016 18:08:55 +0000 (13:08 -0500)]
wayland-egl: Fix use after free

eng_window_use() uses outbuf->redirect, so the clever code in
eng_outbuf_reconfigure that avoided setting it to NULL to check
if we needed to recreate the redirect was not very clever at all.

8 years agowayland_egl: Fix redirect to texture
Derek Foreman [Mon, 4 Apr 2016 16:12:06 +0000 (11:12 -0500)]
wayland_egl: Fix redirect to texture

Previous redirect to texture approach broke multiple window applications
by having only a single redirect per context.  This approach allows
multiple redirections.

8 years agoecore-xcb: Implement missing keyrouter and keygrab functions for xcb
Chris Michael [Mon, 4 Apr 2016 12:54:45 +0000 (08:54 -0400)]
ecore-xcb: Implement missing keyrouter and keygrab functions for xcb

This patch implements missing API functions (in ecore xcb) for
ecore_x_window_keygrab_set, ecore_x_window_keygrab_unset,
ecore_x_e_keyrouter_set, ecore_x_e_keyrouter_get. These
were missing from the initial commit of keygrab/keyrouter code.

Fixes T3377

ref 5c3a08433aa5797582f6cda12b3bb51b89f367d0

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-xcb: Fix wrong initialization
Chris Michael [Mon, 4 Apr 2016 12:41:27 +0000 (08:41 -0400)]
ecore-xcb: Fix wrong initialization

As Ecore_X_Window is an 'int' type, we should not be initializing it
to NULL. This causes compiler warnings about initialization makes
integer from pointer without a cast.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoEvas x11: Fix warnings and invalid calls in engine
Jean-Philippe Andre [Mon, 4 Apr 2016 05:26:20 +0000 (14:26 +0900)]
Evas x11: Fix warnings and invalid calls in engine

This follows commit f10672dd7429dd98dd3b7d88d9c3c63aac392a40
which apparently forgot some changes in xlib and eglfs.

I hope there are no remains of the old native bind/unbind
functions. Also... I can't build everything, so I'm just hoping
this works. Note: GL and SW native images are not using the
same internal api prototypes.

8 years agoElm_image: remove resizable_up and resizable_down Eo api
Ji-Youn Park [Mon, 4 Apr 2016 01:44:49 +0000 (10:14 +0830)]
Elm_image: remove resizable_up and resizable_down Eo api

risizable_up and resizable_down api will be deprecated
after creating the apis to deal with image content using scale type

8 years agoevas - gif loader - handle missing colormap
Carsten Haitzler (Rasterman) [Mon, 4 Apr 2016 00:51:12 +0000 (09:51 +0900)]
evas - gif loader - handle missing colormap

follow on from dd90b6afadf706aafec9e53a6b1efa8f899ab277 - this handled
if a gif file has no colormap (it'll decode as blank now as a colormap
of all 0's is used instead).

@fix

8 years agoevas_gl_cocoa: fix segmentation fault at program exit
Jean Guyomarc'h [Sat, 2 Apr 2016 20:19:29 +0000 (22:19 +0200)]
evas_gl_cocoa: fix segmentation fault at program exit

Programs crashed on a segmentation fault when the last window was
closed. The eng_output_idle_flush() function was removed... but
since gl_cocoa does not properly use *_generic modules, the
output_idle_flush() function called by the render engine was
garbage (hence the segfault).

Now nothing is done... but at least we don't crash anymore.

8 years agoautotools: on OSX warn about prefix being /usr
Jean Guyomarc'h [Sat, 2 Apr 2016 20:17:53 +0000 (22:17 +0200)]
autotools: on OSX warn about prefix being /usr

With Apple's SIP, nobody is allowed to modify the system
(excepted /usr/local). If SIP is enabled, the installation
will just fail.

8 years agominor spelling fixes
Mike Frysinger [Sat, 2 Apr 2016 20:19:52 +0000 (16:19 -0400)]
minor spelling fixes

8 years agotab--
Dave Andreoli [Sat, 2 Apr 2016 11:18:29 +0000 (13:18 +0200)]
tab--

8 years agoelm win - null out freed data on smart del
Carsten Haitzler (Rasterman) [Sat, 2 Apr 2016 07:02:07 +0000 (16:02 +0900)]
elm win - null  out freed data on smart del

this fixes double-frees of strings in oddball cases where a deleted
window that is not done deleting is accessed.

@fix

8 years agoevas - gif loader - be a little more optimal in pixel lookups on decode
Carsten Haitzler (Rasterman) [Sat, 2 Apr 2016 04:22:11 +0000 (13:22 +0900)]
evas - gif loader - be a little more optimal in pixel lookups on decode

8 years agoevas: gif loader - fix out of bounds access on cmap of invalid pixels
Carsten Haitzler (Rasterman) [Sat, 2 Apr 2016 03:25:52 +0000 (12:25 +0900)]
evas: gif loader - fix out of bounds access on cmap of invalid pixels

if gif has example 4 colors in colormap, pixels provided still can
hold values higher than 3 (4, 8, 255 etc.) ass a pixel is still a
byte. it should not, but it could. technically it'd be nice for gitlib
to pad its palette out to 256 entires to ensure this cant be a
problem, but it doesn't have to , so make a local copy of the cmap
when decoding pixels and pad out to 256 entires (using color 0 as any
value > pallette ize is invalid anyway so any color will do).

this fixes a possible security attack vector in reading memory out of
bounds of an allocated array. not very far out of bounds - but enough
to cause a crash - ie a dos attack, (not to inject code though).

@fix

8 years agoelementary: install elementary cxx pkgconfig file.
Cedric BAIL [Fri, 1 Apr 2016 20:54:40 +0000 (13:54 -0700)]
elementary: install elementary cxx pkgconfig file.

T3397

8 years agoelementary: install desktop files in the correct directory.
Cedric BAIL [Fri, 1 Apr 2016 18:50:20 +0000 (11:50 -0700)]
elementary: install desktop files in the correct directory.

Fix for T3397.

8 years agotheme: fix a bunch of the fixed: X Y edje errors
Mike Blumenkrantz [Fri, 1 Apr 2016 20:22:45 +0000 (16:22 -0400)]
theme: fix a bunch of the fixed: X Y edje errors

8 years agoecore-drm: do not attempt to destroy sprite outputs which have never repainted
Mike Blumenkrantz [Fri, 1 Apr 2016 19:54:37 +0000 (15:54 -0400)]
ecore-drm: do not attempt to destroy sprite outputs which have never repainted

silences a surprisingly large number of errors on shutdown

@fix

8 years agoecore-wl2: Enable releasing keyboard, pointer, and touch if supported
Chris Michael [Fri, 1 Apr 2016 16:22:02 +0000 (12:22 -0400)]
ecore-wl2: Enable releasing keyboard, pointer, and touch if supported

This removes a fixme where we were not using wl_(pointer, keyboard,
touch)_release functions as they rely on newer versions of the wayland
protocol. As we now rely on 1.10 of wayland, these can be enabled now.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore_timer: remove redundancy in property docs
Stefan Schmidt [Fri, 1 Apr 2016 13:40:01 +0000 (15:40 +0200)]
ecore_timer: remove redundancy in property docs

Streamline this a bit to document the property itself and only extra information
the the methods if needed.

8 years agoecore_timer: document properties
Stefan Schmidt [Fri, 1 Apr 2016 13:01:48 +0000 (15:01 +0200)]
ecore_timer: document properties

Make sure that we actually document the property itself and not only the
property methods.

8 years agoevas-software-generic: remove native.func.data variable and data argument of native...
Chris Michael [Fri, 1 Apr 2016 11:56:54 +0000 (07:56 -0400)]
evas-software-generic: remove native.func.data variable and data argument of native calblacks.

Evas Image should be independent of render engine.
So remove native.func.data member of RGBA_Image, Evas_GL_Image
struct. And remove data argument,too.

ref f10672dd7429dd98dd3b7d88d9c3c63aac392a40

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-wl2: Move structures above functions
Chris Michael [Fri, 1 Apr 2016 11:34:43 +0000 (07:34 -0400)]
ecore-wl2: Move structures above functions

NB: This is just a cosmetic issue, no functional changes

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoevas-wayland-egl: Add www protocol handling to wayland-egl engine
Derek Foreman [Mon, 28 Mar 2016 17:56:23 +0000 (12:56 -0500)]
evas-wayland-egl: Add www protocol handling to wayland-egl engine

Use the new post-processing API and www extension to implement
CSS effects for wayland-egl applications.

Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Signed-off-by: Mike Blumenkrantz <zmike@osg.samsung.com>