platform/upstream/efl.git
8 years agoEfl.Ui.Flip: Fix compilation outside EFL
Jean-Philippe Andre [Tue, 7 Jun 2016 04:06:03 +0000 (13:06 +0900)]
Efl.Ui.Flip: Fix compilation outside EFL

After a previous patch, compilation of apps would fail
as the include order was invalid. Efl_Ui_Flip_Mode was
defined after the typedef using it unless EO APIs were
enabled.

8 years agoelm: Don't install private headers
Jean-Philippe Andre [Tue, 7 Jun 2016 01:39:31 +0000 (10:39 +0900)]
elm: Don't install private headers

Thanks tom for the report.

8 years agoefl_ui_flip: renamed elm_flip to efl_ui_flip and updated the eo class
Subhransu Mohanty [Mon, 6 Jun 2016 23:50:26 +0000 (16:50 -0700)]
efl_ui_flip: renamed elm_flip to efl_ui_flip and updated the eo class

Reviewers: woohyun, raster, jpeg, cedric

Reviewed By: jpeg, cedric

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3933

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoelementary/colorselector: use layout instead of edje object for components.
Shilpa Singh [Mon, 6 Jun 2016 23:47:03 +0000 (16:47 -0700)]
elementary/colorselector: use layout instead of edje object for components.

Summary:
Issue: As edje object is used instead of layout for components mode,  the content widgets like buttons are added as
subobject of colorselector hence when we unset the edje object on mode change the buttons still remain subobject of colorselector causing issues when all subobjects of colorselector is queried(sub objects are queried during atspi mode).
Soln: Use layout instead of edje object.

Signed-off-by: Shilpa Singh <shilpa.singh@samsung.com>
@fix

Test Plan: Change the mode to Palette mode and access the children of colorselector.

Reviewers: Hermet, cedric

Subscribers: govi, rajeshps, kimcinoo, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3987

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoelementary/colorselector: colornames updated with spaces.
Shilpa Singh [Mon, 6 Jun 2016 23:42:27 +0000 (16:42 -0700)]
elementary/colorselector: colornames updated with spaces.

Summary:
Colornames updated with spaces so that any TTS engine
can read them out in an understandable manner.
Signed-off-by: Shilpa Singh <shilpa.singh@samsung.com>
@fix

Test Plan:
Enable ATSPI mode
Tap colors and listen to colorname's

Reviewers: cedric

Subscribers: rajeshps, jpeg

Differential Revision: https://phab.enlightenment.org/D3997

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoelementary/tooltip: adjust the coordinate if tooltip was located out of screen
Minkyu Kang [Mon, 6 Jun 2016 23:41:41 +0000 (16:41 -0700)]
elementary/tooltip: adjust the coordinate if tooltip was located out of screen

Summary:
When reconfigure the tooltip as orient, the tooltip can located out of screen.
This patch will adjust the coordinate so that can be seen whole tooltip.

Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Reviewers: cedric, Hermet

Subscribers: seoz, jpeg

Differential Revision: https://phab.enlightenment.org/D4007

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas: fix typo in description of the engine functions.
Oleksandr Shcherbina [Mon, 6 Jun 2016 23:40:23 +0000 (16:40 -0700)]
evas: fix typo in description of the engine functions.

Reviewers: Hermet, raster, cedric

Reviewed By: cedric

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D4010

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas: fix order binding attributes to gl.
Oleksandr Shcherbina [Mon, 6 Jun 2016 23:39:38 +0000 (16:39 -0700)]
evas: fix order binding attributes to gl.

Summary:
In same cases(devices) can lead to mix up attributes in shaders.
@fix

Reviewers: cedric, Hermet, raster

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D4012

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas: button_mask as unsigned, UB fixes.
Benjamin Jacobs [Mon, 6 Jun 2016 23:33:54 +0000 (16:33 -0700)]
evas: button_mask as unsigned, UB fixes.

Summary:
* pointer.button is DATA32 which is unsigned, so this changes the
  definition of pointer_button_down_mask accordingly.

* Avoids UB in mask generation:
lib/evas/canvas/evas_events.c:1348:37: runtime error: left shift of 1 by 31 places cannot be represented in type 'int'

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D4019

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore: make sure lifecycle of Efl.Loop.Fd is correct.
Cedric BAIL [Mon, 6 Jun 2016 23:28:52 +0000 (16:28 -0700)]
ecore: make sure lifecycle of Efl.Loop.Fd is correct.

8 years agoecore: properly reset fd handler when parent get deleted.
Cedric BAIL [Mon, 6 Jun 2016 23:19:55 +0000 (16:19 -0700)]
ecore: properly reset fd handler when parent get deleted.

8 years agoelementary: Fix elm_code theme lookup and use correct naming
Andy Williams [Mon, 6 Jun 2016 22:59:52 +0000 (23:59 +0100)]
elementary: Fix elm_code theme lookup and use correct naming

8 years agoecore: add a tests that ensure we do implement lifecycle correctly.
Cedric BAIL [Mon, 6 Jun 2016 21:59:36 +0000 (14:59 -0700)]
ecore: add a tests that ensure we do implement lifecycle correctly.

8 years agoecore: migrate ecore_timer lifecycle to rely cleanly on Eo lifecycle.
Cedric BAIL [Mon, 6 Jun 2016 21:59:06 +0000 (14:59 -0700)]
ecore: migrate ecore_timer lifecycle to rely cleanly on Eo lifecycle.

8 years agoeina: ensure resource destruction on failure
Jean Guyomarc'h [Mon, 6 Jun 2016 21:40:29 +0000 (23:40 +0200)]
eina: ensure resource destruction on failure

In case of failure within eina_lock_new() (posix), a
pthread_mutexattr_t would have been left in an initialized state,
without any deinitialization being called.

Consequences would have been implementation defined.

8 years agoecore-xcb: Fix XCB to be usable again
Benjamin Jacobs [Mon, 6 Jun 2016 19:46:00 +0000 (15:46 -0400)]
ecore-xcb: Fix XCB to be usable again

Summary:
Well mostly, it seems there is an issue with multi-key events and
enlightenment. Let's merge this first before opening a ticket.

Reviewers: devilhorns

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D4017

@fix

8 years agoelementary: export symbol used by tests suite.
Cedric BAIL [Mon, 6 Jun 2016 19:11:07 +0000 (12:11 -0700)]
elementary: export symbol used by tests suite.

8 years agoeina_thread.c: fix include guard.
Benjamin Jacobs [Mon, 6 Jun 2016 18:23:09 +0000 (11:23 -0700)]
eina_thread.c: fix include guard.

DragonFlyBSD has pthread_setname_np but no pthread_setaffinity_np;
we still need to include pthread_np though.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeio: fix pointer comparison.
Benjamin Jacobs [Mon, 6 Jun 2016 18:20:37 +0000 (11:20 -0700)]
eio: fix pointer comparison.

Fix a warning on dragonfly.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_audio: uncondionally include libgen.h.
Benjamin Jacobs [Mon, 6 Jun 2016 18:18:25 +0000 (11:18 -0700)]
ecore_audio: uncondionally include libgen.h.

Basically everyone should have it.  If not, maybe we should consider
to test the value of _POSIX_C_SOURCE before testing platform specific
defines, or adding a feature check.

Fixes the following warning on DragonFlyBSD:

lib/ecore_audio/ecore_audio_obj_out_pulse.c:282:60: warning: implicit declaration of function 'basename' [-Wimplicit-function-declaration]
        class_vars.context = pa_context_new(class_vars.api, basename(argv[0]));
                                                            ^
lib/ecore_audio/ecore_audio_obj_out_pulse.c:282:60: warning: passing argument 2 of 'pa_context_new' makes pointer from integer without a cast [-Wint-conversion]
In file included from /usr/local/include/pulse/pulseaudio.h:29:0,
                 from lib/ecore_audio/ecore_audio_private.h:17,
                 from lib/ecore_audio/ecore_audio_obj_out_pulse.c:18:
/usr/local/include/pulse/context.h:172:13: note: expected 'const char *' but argument is of type 'int'
 pa_context *pa_context_new(pa_mainloop_api *mainloop, const char *name);
             ^

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_time.c: do not mess with the representation of clockid_t.
Benjamin Jacobs [Mon, 6 Jun 2016 18:16:44 +0000 (11:16 -0700)]
ecore_time.c: do not mess with the representation of clockid_t.

Clockid_t should be used as an opaque type. Some platform might want
to (and even do, e.g. DragonFlyBSD) declare clockid_t as an unsigned.

On such platform, testing the sign of clockid_t is never false, and
assigning it a negative value is an UB, which makes this code unlikely to
work as intended. Fixes black window on dragonfly!

Thanks to gcc for spotting this.

  CC       lib/ecore/lib_ecore_libecore_la-ecore_time.lo
  In file included from ../src/lib/eina/Eina.h:215:0,
                   from lib/ecore/Ecore.h:304,
                   from lib/ecore/ecore_time.c:18:
                   lib/ecore/ecore_time.c: In function 'ecore_time_get':

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_signal.c: factoring out the platform test to improve readability.
Benjamin Jacobs [Mon, 6 Jun 2016 18:14:55 +0000 (11:14 -0700)]
ecore_signal.c: factoring out the platform test to improve readability.

Turns out there is no PRI?SIGATOMIC in the C99 standard. Work around
that by deducing the effective integer type by comparing the
SIG_ATOMIC_MAX with integers *MAX.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoelput: Fix crash whem sd_session_get_seat fails
Chris Michael [Mon, 6 Jun 2016 18:53:35 +0000 (14:53 -0400)]
elput: Fix crash whem sd_session_get_seat fails

If systemd fails to give us a valid session id, then any further calls
to get a seat from that session are going to fail. As such, check for
a valid seat return from systemd before calling any systemd functions
which require session id.

Fixes T3785

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoeio: remove eio_model that should already be part of Eio.h.
Cedric BAIL [Mon, 6 Jun 2016 18:00:54 +0000 (11:00 -0700)]
eio: remove eio_model that should already be part of Eio.h.

8 years agoedje_edit: fix map API (light, perspective, rotation)
Vitalii Vorobiov [Mon, 6 Jun 2016 17:46:14 +0000 (20:46 +0300)]
edje_edit: fix map API (light, perspective, rotation)

it should be possible to set light or perspective to NULL
and return -1 when nothing is set
also ability to send NULL point to store rotation

@fix

8 years agoelementary: fix definition of external symbols
Jean Guyomarc'h [Mon, 6 Jun 2016 17:30:07 +0000 (19:30 +0200)]
elementary: fix definition of external symbols

Introduced by elm_code. Broke the build (link) Mac OS X.

8 years agoeolian: we don't support unnamed structs/enums, so don't check
Daniel Kolesa [Mon, 6 Jun 2016 15:15:14 +0000 (16:15 +0100)]
eolian: we don't support unnamed structs/enums, so don't check

8 years agoeolian: remove the c_type_named_get APIs
Daniel Kolesa [Mon, 6 Jun 2016 14:50:15 +0000 (15:50 +0100)]
eolian: remove the c_type_named_get APIs

These have inconsistent behavior and are generally unnecessary, so I'm
removing them. Use the c_type_get funcs instead.

8 years agoeolian: add a new references system to help replace pointers
Daniel Kolesa [Mon, 6 Jun 2016 14:22:56 +0000 (15:22 +0100)]
eolian: add a new references system to help replace pointers

It's now possible to mark struct fields and function params as "references",
which causes them to become pointers in C (in bindings, they become whatever
is necessary). They're not a part of the type and are much more restricted
than pointers, allowing bindings to be easier. This system will be gradually
utilized and expanded as required.

@feature

8 years agoexamples: eio: remove unused variable
Stefan Schmidt [Mon, 6 Jun 2016 13:18:46 +0000 (15:18 +0200)]
examples: eio: remove unused variable

8 years agoexamples: eio: mark unused parameter as such
Stefan Schmidt [Mon, 6 Jun 2016 13:16:36 +0000 (15:16 +0200)]
examples: eio: mark unused parameter as such

Really shutting up warnings here to see what else the compiler has to say.

8 years agotests: evas: remove unused variable
Stefan Schmidt [Mon, 6 Jun 2016 12:49:33 +0000 (14:49 +0200)]
tests: evas: remove unused variable

We set it but never going to use it afterwards. We already tests the
efl_gfx_buffer_colorspace_get() elsewhere in this tests so we can ignore it
for now until we need it and bring it back together with a user.

8 years agoelementary_code: Fix test failures post-merge
Andy Williams [Mon, 6 Jun 2016 13:14:24 +0000 (14:14 +0100)]
elementary_code: Fix test failures post-merge

Looks like some were legitimate failures that were not being caught
in the previous setup

8 years agoelementary_code: Use correct theme file post-merge
Andy Williams [Sun, 5 Jun 2016 23:01:34 +0000 (00:01 +0100)]
elementary_code: Use correct theme file post-merge

8 years agoMerge elm code, a code editing widget into the efl
Tom Hacohen [Sun, 5 Jun 2016 10:12:13 +0000 (11:12 +0100)]
Merge elm code, a code editing widget into the efl

It was decided to merge this and mark the code as BETA for now.
It's currently being used by EDI and should also be used by enventor and etc.

8 years agobenchmark: remove unused variable
Stefan Schmidt [Mon, 6 Jun 2016 12:17:07 +0000 (14:17 +0200)]
benchmark: remove unused variable

Looks like this was forgotten during some re-factoring of the code.

8 years agobenchmark: mark parameter as unused to ignore compile warning
Stefan Schmidt [Mon, 6 Jun 2016 12:16:28 +0000 (14:16 +0200)]
benchmark: mark parameter as unused to ignore compile warning

8 years agoeolian_cxx: add missing libgen.h include for basename() to fix build on OS X
Stefan Schmidt [Mon, 6 Jun 2016 09:07:51 +0000 (11:07 +0200)]
eolian_cxx: add missing libgen.h include for basename() to fix build on OS X

On Linux this went through and has gone unnoticed. basename() needs libgen.h
though and eolian_cxx was not including it.

8 years agoedje_edit: return lost API *_span_row/col_set declaration
Vyacheslav Reutskiy [Mon, 6 Jun 2016 07:42:14 +0000 (10:42 +0300)]
edje_edit: return lost API *_span_row/col_set declaration

I'm lost this declaration while add API for manager part items by
index and mark as deprecated old funcs

8 years agoeio: Fix leaks in tests
Lauro Moura [Mon, 6 Jun 2016 07:26:02 +0000 (04:26 -0300)]
eio: Fix leaks in tests

8 years agoelm: Add support for Efl.Model in Elm.Fileselector
Vitor Sousa [Mon, 6 Jun 2016 07:16:57 +0000 (04:16 -0300)]
elm: Add support for Efl.Model in Elm.Fileselector

Elm.Interface.Fileselector now supports Efl.Model objects, allowing users
to work with paths from different data models.

Example of model attribution:

ELm_Fileselector *fs = eo_add(EFL_FILESELECTOR_CLASS, NULL);
Efl_Model *model = ...;
elm_interface_fileselector_model_set(fs, model, NULL);

8 years agoeo-cxx: Remove commented headers
Felipe Magno de Almeida [Mon, 6 Jun 2016 07:09:23 +0000 (04:09 -0300)]
eo-cxx: Remove commented headers

8 years agoeolian-cxx: Fix compilation error on clang with const UDT initialization
Felipe Magno de Almeida [Mon, 6 Jun 2016 06:49:16 +0000 (03:49 -0300)]
eolian-cxx: Fix compilation error on clang with const UDT initialization

8 years agoeolian-cxx: fix compilation errors an olders gcc
Larry Jr [Mon, 6 Jun 2016 06:28:33 +0000 (03:28 -0300)]
eolian-cxx: fix compilation errors an olders gcc

8 years agoeio: fix free eina_promises in error_callbacks
Larry Jr [Mon, 6 Jun 2016 06:10:24 +0000 (03:10 -0300)]
eio: fix free eina_promises in error_callbacks

8 years agoeina: eina_promise_all callback cast removed
Larry Jr [Mon, 6 Jun 2016 05:20:44 +0000 (02:20 -0300)]
eina: eina_promise_all callback cast removed

8 years agoeina: Avoid promise early deletion when cancelled.
Lauro Moura [Mon, 6 Jun 2016 05:54:18 +0000 (02:54 -0300)]
eina: Avoid promise early deletion when cancelled.

Cancelling a promise will fulfill it but won't actually free the memory. This
memory is under custody of the owner, who must either call value_set or
error_set to finish it.

8 years agoeolian-cxx: Rewrite to accomodate new features of the C++ binding
Felipe Magno de Almeida [Mon, 6 Jun 2016 05:54:36 +0000 (02:54 -0300)]
eolian-cxx: Rewrite to accomodate new features of the C++ binding

8 years agoeina-cxx: Add placeholder for futures in C++ binding
Felipe Magno de Almeida [Mon, 6 Jun 2016 05:50:36 +0000 (02:50 -0300)]
eina-cxx: Add placeholder for futures in C++ binding

8 years agoeina-cxx: Fix wrongful operator++ for iterator and docs
Felipe Magno de Almeida [Mon, 6 Jun 2016 05:50:13 +0000 (02:50 -0300)]
eina-cxx: Fix wrongful operator++ for iterator and docs

8 years agoeina-cxx: Add c_str() to string_view
Felipe Magno de Almeida [Mon, 6 Jun 2016 05:49:47 +0000 (02:49 -0300)]
eina-cxx: Add c_str() to string_view

8 years agoeina-cxx: Restrict templated constructor
Felipe Magno de Almeida [Mon, 6 Jun 2016 05:49:28 +0000 (02:49 -0300)]
eina-cxx: Restrict templated constructor

8 years agoeina-cxx: Add optional specialization for pointer types
Felipe Magno de Almeida [Mon, 6 Jun 2016 05:49:00 +0000 (02:49 -0300)]
eina-cxx: Add optional specialization for pointer types

8 years agoelm: Fix wrong use of char instead of string and void out parameter
Felipe Magno de Almeida [Mon, 6 Jun 2016 05:47:15 +0000 (02:47 -0300)]
elm: Fix wrong use of char instead of string and void out parameter

8 years agoedje: Fix wrongful uses of char instead of string in Eolian
Felipe Magno de Almeida [Mon, 6 Jun 2016 05:46:39 +0000 (02:46 -0300)]
edje: Fix wrongful uses of char instead of string in Eolian

8 years agoeio: treat Eina_Accessor NULL as success instead of failure in model
Larry Jr [Mon, 6 Jun 2016 01:24:52 +0000 (22:24 -0300)]
eio: treat Eina_Accessor NULL as success instead of failure in model

We treated as errors when Eina_Accessor was NULL, which meant
that we could never get an empty range os children. Treat it
as success and pass an empty Eina_Accessor

8 years agoeio: Add mime_type property in Eio_Model
Larry Jr [Mon, 6 Jun 2016 00:21:43 +0000 (21:21 -0300)]
eio: Add mime_type property in Eio_Model

8 years agoeio: Eio_Model parent_get return parent model
Larry Jr [Sun, 5 Jun 2016 23:52:52 +0000 (20:52 -0300)]
eio: Eio_Model parent_get return parent model

8 years agoeio: Fix Eio_Model promise accessor
Larry Jr [Sun, 5 Jun 2016 23:31:21 +0000 (20:31 -0300)]
eio: Fix Eio_Model promise accessor

8 years agoecore: Fix ecore thread promises
Lauro Moura [Sun, 5 Jun 2016 23:19:41 +0000 (20:19 -0300)]
ecore: Fix ecore thread promises

Make sure we free the promise memory, set MAGIC numbers and avoid
leaking the underlying eina promise

8 years agoeina: Fix promise tests and avoid early deletion
Lauro Moura [Sun, 5 Jun 2016 19:10:52 +0000 (16:10 -0300)]
eina: Fix promise tests and avoid early deletion

- Progress notify refcount fixes, avoiding early deletion
- Avoid leaking memory on tests

8 years agoeina: Fixes on promises benchmark
Lauro Moura [Sat, 4 Jun 2016 16:49:51 +0000 (13:49 -0300)]
eina: Fixes on promises benchmark

- Updated after pointer vs value change
- Updated callbacks after removal of promise argument

8 years agoeo: Fix compilation of benchmark
Lauro Moura [Sat, 4 Jun 2016 16:51:42 +0000 (13:51 -0300)]
eo: Fix compilation of benchmark

- Missing Eo* self in method declarations.
- Gcc complaining about variable-sized object being initialized.

8 years agoeolian: Fix generating pointer promises for class types
Felipe Magno de Almeida [Sun, 5 Jun 2016 19:50:54 +0000 (16:50 -0300)]
eolian: Fix generating pointer promises for class types

8 years agoEo: fix reference hack when calling composite objects.
Tom Hacohen [Sun, 5 Jun 2016 14:16:58 +0000 (15:16 +0100)]
Eo: fix reference hack when calling composite objects.

We used to keep a reference to the parent object and have it in the call
structure although we were actually calling the function on the embedded
object. This was needed because we wanted to unref the parent correctly.
This was incorrect (and marked as a hack) and now I finally gotten
around to implementing the (amazingly simple) fix to remove this
workaround.

Essentially we just ref the comp object, unref the parent, and let the
normal eo call flow to unref the comp object correctly later on, like it
would have unreffed the extra ref we had for the parent.

8 years agoFDO icons: and finally install the icons in fdo format
Dave Andreoli [Sun, 5 Jun 2016 14:12:01 +0000 (16:12 +0200)]
FDO icons: and finally install the icons in fdo format

After this commit you will find a new icon theme in your system,
it is called Enlightenment-X and is build using the same icons used
in the elm theme. Using this theme you can have non-efl
and efl apps to use the same icons.

#StupidAutotools that require me to list each icon and each folder
to be able to install them

8 years agoEeze scanner: Fix setting of SUID.
Tom Hacohen [Sun, 5 Jun 2016 12:51:15 +0000 (13:51 +0100)]
Eeze scanner: Fix setting of SUID.

I guess eeze scanner was not operational at all because the SUID bit was
never set. The reason for that was that where it was put in the makefile
made it not be a make rule (where @ would have worked) but used by a special
autofoo rule.

8 years agogitignore++
Dave Andreoli [Sun, 5 Jun 2016 10:37:08 +0000 (12:37 +0200)]
gitignore++

8 years agoecore_evas_cocoa: better handling of resize callback
Jean Guyomarc'h [Sun, 5 Jun 2016 09:47:59 +0000 (11:47 +0200)]
ecore_evas_cocoa: better handling of resize callback

8 years agoevas-gl_cocoa: resize the gl context before resizing the window
Jean Guyomarc'h [Sun, 5 Jun 2016 09:38:49 +0000 (11:38 +0200)]
evas-gl_cocoa: resize the gl context before resizing the window

This order seems more right. It does not have a visible
impact though...

8 years agoecore_cocoa: fix main loop handling
Jean Guyomarc'h [Sun, 5 Jun 2016 09:29:25 +0000 (11:29 +0200)]
ecore_cocoa: fix main loop handling

Pffff... another tricky one...

-windowDidResize is actually also called when the window is not
resize by the user, leading to multiple event posting, and of
course the infamous call of ecore_main_loop_iterate() when it
was not paused at all, leading to messy events handling... the
most visible being the initial resizing of the window going rogue.

We now ensure with -windowWillStartLiveResize and -windowDidEndLiveResize
that we only send an event when the user requested it. Since the main
loop is paused at this point, calling ecore_main_loop_iterate()
becomes safe.

Fixes T3648

8 years agoecore_audio: remove useless eo data in core audio
Jean Guyomarc'h [Sun, 5 Jun 2016 07:26:29 +0000 (09:26 +0200)]
ecore_audio: remove useless eo data in core audio

8 years agoEina promise: Fix broken return value.
Tom Hacohen [Sun, 5 Jun 2016 10:00:47 +0000 (11:00 +0100)]
Eina promise: Fix broken return value.

I broke it in dc73f6644e0ed8fdbd2c6c8187591fb7eb8027c7.
Excuse of the day: I'm not working from my usual spot and there's screen glare
here that made me not see the *. Good excuse?

8 years agoGeneric modules: Fix out of source build.
Tom Hacohen [Sun, 5 Jun 2016 09:40:02 +0000 (10:40 +0100)]
Generic modules: Fix out of source build.

Autotools really really really sucks.
Because of the above, we need to manually list both source dirs and
build dirs, because some of the include files are in the source dirs and
some are in the build dir (generated).

8 years agoEina promise: Simplify code and remove shadow variable.
Tom Hacohen [Sun, 5 Jun 2016 09:39:13 +0000 (10:39 +0100)]
Eina promise: Simplify code and remove shadow variable.

There was no need for the extra variable. This also fixes a shadow
warning.

8 years agoEo: remove strict class checking for composite objects.
Tom Hacohen [Sun, 5 Jun 2016 08:58:19 +0000 (09:58 +0100)]
Eo: remove strict class checking for composite objects.

When compositing objects, we were checking that class_of(B) is in A's
inheritance tree before allowing attaching B as a composite object of A.
This is wrong and breaks a few cases, for example: B extends a class that
is in A's inheritance tree or B implements an interface that is in A's
inheritance tree.

Thanks to Marcel for reporting.

8 years agoEnable evas loaders by default, if dependencies are found
Dave Andreoli [Sat, 4 Jun 2016 19:45:28 +0000 (21:45 +0200)]
Enable evas loaders by default, if dependencies are found

I really cannot see why we should require to use --enable-feature on those, they should be safe.

8 years agoFix report for image loaders
Dave Andreoli [Sat, 4 Jun 2016 18:19:54 +0000 (20:19 +0200)]
Fix report for image loaders

In the final report we want to show if the loader is enabled, not if it's wanted

8 years agoFix building of librsvg if requested
Dave Andreoli [Sat, 4 Jun 2016 18:10:04 +0000 (20:10 +0200)]
Fix building of librsvg if requested

This typo was preventing the svg loader to be built in every cases

8 years agoecore_cocoa: fix broken scrolling
Jean Guyomarc'h [Sat, 4 Jun 2016 14:34:30 +0000 (16:34 +0200)]
ecore_cocoa: fix broken scrolling

Evas events recently changed, and involved the use
of the 'multi' field within a mouse mouve event.
It was used afterwards for mouse scrolling, but since
it was never set by ecore_cocoa, scrolling went
broken...

Fixes T3789

8 years agoelementary: remove redundant code
Jean Guyomarc'h [Sat, 4 Jun 2016 14:19:00 +0000 (16:19 +0200)]
elementary: remove redundant code

The same checks are performed later on.
There is no use doing it twice.

Only affects Cocoa windows.

8 years agoecore_cocoa: fix window hiding logic
Jean Guyomarc'h [Sat, 4 Jun 2016 14:00:50 +0000 (16:00 +0200)]
ecore_cocoa: fix window hiding logic

8 years agoecore_cocoa: remove extra whitespaces
Jean Guyomarc'h [Sat, 4 Jun 2016 13:58:27 +0000 (15:58 +0200)]
ecore_cocoa: remove extra whitespaces

8 years agoedje: don't return a boolean when a pointer is expected
Jean Guyomarc'h [Sat, 4 Jun 2016 14:51:38 +0000 (16:51 +0200)]
edje: don't return a boolean when a pointer is expected

Fix a warning thrown by Apple-llvm.

8 years agoelm_config: force icons in preview to keep aspect ratio
Dave Andreoli [Sat, 4 Jun 2016 09:42:42 +0000 (11:42 +0200)]
elm_config: force icons in preview to keep aspect ratio

#SizeHintsMess... I really cannot understand why aspect_fixed is not respected :(

8 years agoConfigure: 2 small in-string quote changes
Dave Andreoli [Sat, 4 Jun 2016 08:54:40 +0000 (10:54 +0200)]
Configure: 2 small in-string quote changes

This should be harmless and just fix syntax hilighting is some editors (geany at least)

8 years agoeina: Fix magic check for eina_promise_owner_buffer_get
Felipe Magno de Almeida [Sat, 4 Jun 2016 00:53:29 +0000 (21:53 -0300)]
eina: Fix magic check for eina_promise_owner_buffer_get

8 years agoeina: Fix EINA_MAGIC and eina_promise_then NULL check
Felipe Magno de Almeida [Sat, 4 Jun 2016 00:42:23 +0000 (21:42 -0300)]
eina: Fix EINA_MAGIC and eina_promise_then NULL check

8 years agoecore: add a test for detecting the double free problem.
Cedric BAIL [Fri, 3 Jun 2016 23:59:55 +0000 (16:59 -0700)]
ecore: add a test for detecting the double free problem.

8 years agoecore: properly handle double free case of timeout promise.
Cedric BAIL [Fri, 3 Jun 2016 23:57:53 +0000 (16:57 -0700)]
ecore: properly handle double free case of timeout promise.

The bug came from the fact we need to handle the destruction of the
main loop which destroy the underlying timer. The event handler that
catch the destruction of the timer can not make the difference between
eo_del call from the timeout code and eo_del from the main loop
destruction. By removing the event handler, the double free is properly
avoided.

8 years agoecore: use EINA_ERROR_PROMISE_CANCEL instead of a custom Ecore one.
Cedric BAIL [Fri, 3 Jun 2016 23:57:30 +0000 (16:57 -0700)]
ecore: use EINA_ERROR_PROMISE_CANCEL instead of a custom Ecore one.

8 years agoecore: do not generate error while rescheduling the last timer.
Cedric BAIL [Fri, 3 Jun 2016 23:55:56 +0000 (16:55 -0700)]
ecore: do not generate error while rescheduling the last timer.

8 years agoeina: protect promise death during callbacks cancel.
Cedric BAIL [Fri, 3 Jun 2016 23:54:38 +0000 (16:54 -0700)]
eina: protect promise death during callbacks cancel.

8 years agoecore: Ecore_Cb is used by elm_box.eo.
Cedric BAIL [Fri, 3 Jun 2016 23:45:12 +0000 (16:45 -0700)]
ecore: Ecore_Cb is used by elm_box.eo.

8 years agoecore: Efl.Timer should be Efl.Loop.Timer for coherence.
Cedric BAIL [Fri, 3 Jun 2016 23:01:08 +0000 (16:01 -0700)]
ecore: Efl.Timer should be Efl.Loop.Timer for coherence.

As we add more object in the main loop, they can't live in the top
namespace as they make little sense there (Efl.Fd !). For coherence,
everyone should in the loop namespace, so move timer there.

8 years agoecore: remove Ecore_Animator Eo object.
Cedric BAIL [Fri, 3 Jun 2016 22:44:40 +0000 (15:44 -0700)]
ecore: remove Ecore_Animator Eo object.

This is only a legacy API.

8 years agoeina: Remove Eina_Promise* parameter to callbacks to promises in examples
Felipe Magno de Almeida [Fri, 3 Jun 2016 23:03:13 +0000 (20:03 -0300)]
eina: Remove Eina_Promise* parameter to callbacks to promises in examples

8 years agoeina: Add NULL checks and EINA_MAGIC checks to promises
Felipe Magno de Almeida [Fri, 3 Jun 2016 23:02:02 +0000 (20:02 -0300)]
eina: Add NULL checks and EINA_MAGIC checks to promises

Add NULL checks and EINA_MAGIC checks that avoid crashing when it
shouldn't.

8 years agoeina: Remove more Eina_Promise* parameter in promise callback
Felipe Magno de Almeida [Fri, 3 Jun 2016 21:34:42 +0000 (18:34 -0300)]
eina: Remove more Eina_Promise* parameter in promise callback

Remove not very useful parameter to eina_promise_then callbacks.

8 years agowayland_shm: Fix more dmabuf failure corner cases
Derek Foreman [Fri, 3 Jun 2016 21:27:14 +0000 (16:27 -0500)]
wayland_shm: Fix more dmabuf failure corner cases

With the exynos allocator a failure can happen during a resize, and the
resize path wasn't properly falling back to wl_shm buffering.