platform/upstream/efl.git
6 years agoeio: move efl.io.manager.close to use the new future.
Cedric BAIL [Tue, 12 Sep 2017 21:08:40 +0000 (14:08 -0700)]
eio: move efl.io.manager.close to use the new future.

6 years agoeio: simplify promise use.
Cedric BAIL [Tue, 12 Sep 2017 17:22:21 +0000 (10:22 -0700)]
eio: simplify promise use.

6 years agoeina: cleanup documentation for promise.
Cedric BAIL [Tue, 12 Sep 2017 17:16:44 +0000 (10:16 -0700)]
eina: cleanup documentation for promise.

6 years agotests: fix build after elm atspi renaming
Stefan Schmidt [Tue, 12 Sep 2017 14:01:44 +0000 (16:01 +0200)]
tests: fix build after elm atspi renaming

Renamed but make check not taken into account.

6 years agogenlist: fix decorate mode dangling pointer in deletion
SangHyeon Jade Lee [Tue, 12 Sep 2017 09:38:44 +0000 (18:38 +0900)]
genlist: fix decorate mode dangling pointer in deletion

fix decorate mode crash issue reported by Jack Daniel in T6000
which is occured by dangling pointer in item deletion on decorate mode.

Signed-off-by: SangHyeon Jade Lee <dltkdgus1764@gmail.com>
6 years agoefl.ui.slider: implement efl.text intf
Amitesh Singh [Tue, 12 Sep 2017 09:25:41 +0000 (18:25 +0900)]
efl.ui.slider: implement efl.text intf

efl_text_set/get would work on slider text part.

6 years agoefl.ui.range: remove interval_enable/set/get APIs from interface
Amitesh Singh [Tue, 12 Sep 2017 08:29:18 +0000 (17:29 +0900)]
efl.ui.range: remove interval_enable/set/get APIs from interface

Remove interval related apis from interface since interval slider
implements these features now. Although legacy range APIs would still
work on slider widget.

6 years agointerval slider: Add new interval slider widget
Amitesh Singh [Tue, 12 Sep 2017 01:07:58 +0000 (10:07 +0900)]
interval slider: Add new interval slider widget

This widget is a slider with two indicators which allows
to have interval of value.

6 years agoedje: Add legacy compatibility code
Jean-Philippe Andre [Tue, 12 Sep 2017 08:21:32 +0000 (17:21 +0900)]
edje: Add legacy compatibility code

This affects:
 - edje_object_part_swallow
 - edje_object_part_swallow_get

Fixes T5944

@fix

6 years agoector: Copyedit Ector doxygen
Bryce Harrington [Tue, 12 Sep 2017 07:07:13 +0000 (16:07 +0900)]
ector: Copyedit Ector doxygen

Summary:
- Fix various spelling and grammar errors.
- Add a slight bit more explanation of what Ector is.
- Condense compilation directions similar to Evas.

Reviewers: cedric

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D5176

6 years agoelm: rename Elm.Interface.Atspi.Component => Efl.Access.Component
Lukasz Stanislawski [Tue, 12 Sep 2017 06:49:29 +0000 (15:49 +0900)]
elm: rename Elm.Interface.Atspi.Component => Efl.Access.Component

Reviewers: jpeg, cedric

Differential Revision: https://phab.enlightenment.org/D5164

6 years agoelm: rename Elm.Interface.Atspi.Image => Efl.Access.Image
Lukasz Stanislawski [Tue, 12 Sep 2017 06:48:39 +0000 (15:48 +0900)]
elm: rename Elm.Interface.Atspi.Image => Efl.Access.Image

Reviewers: jpeg, cedric

Differential Revision: https://phab.enlightenment.org/D5163

6 years agoelm: rename Elm.Interface.Atspi.Window => Efl.Access.Window
Lukasz Stanislawski [Tue, 12 Sep 2017 06:48:10 +0000 (15:48 +0900)]
elm: rename Elm.Interface.Atspi.Window => Efl.Access.Window

Reviewers: cedric, jpeg

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5162

6 years agoelm: rename Elm.Interface.Atspi.Value => Efl.Access.Value
Lukasz Stanislawski [Tue, 12 Sep 2017 06:36:24 +0000 (15:36 +0900)]
elm: rename Elm.Interface.Atspi.Value => Efl.Access.Value

Reviewers: jpeg, cedric, stanluk

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5161

6 years agoUi text: reduce calls to update functions
Daniel Hirt [Sat, 9 Sep 2017 10:21:06 +0000 (13:21 +0300)]
Ui text: reduce calls to update functions

Reduces calls to update functions when entering text.
Also, remaining calls to update the cursors were moved to the decoration
jobs.

6 years agoeio: first convertion to new future/promise.
Cedric BAIL [Mon, 11 Sep 2017 23:10:46 +0000 (16:10 -0700)]
eio: first convertion to new future/promise.

6 years agoeina: add eina_promise_data_set symetric to eina_promise_data_get.
Cedric BAIL [Mon, 11 Sep 2017 23:09:43 +0000 (16:09 -0700)]
eina: add eina_promise_data_set symetric to eina_promise_data_get.

6 years agoeina: fix typos in eina_promise_all_array documentation.
Cedric BAIL [Mon, 11 Sep 2017 23:08:23 +0000 (16:08 -0700)]
eina: fix typos in eina_promise_all_array documentation.

6 years agoevas: actually output might not be initialized until later.
Cedric BAIL [Mon, 11 Sep 2017 18:31:50 +0000 (11:31 -0700)]
evas: actually output might not be initialized until later.

6 years agoExamples: fix C++ examples
Daniel Zaoui [Mon, 11 Sep 2017 18:12:30 +0000 (21:12 +0300)]
Examples: fix C++ examples

The examples have not been updated after the last C++ changes.

6 years agoeina: fix build for big endian platforms
Stefan Schmidt [Mon, 11 Sep 2017 11:45:18 +0000 (13:45 +0200)]
eina: fix build for big endian platforms

Summary:
The code in eina_debug.c requires the eina_swapX() functions on big endian platforms, so include the required header.

Test Plan: Do build on big endian platform like ppc64 or s390x.

Subscribers: kubu, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5158

6 years agoevas textblock: support ZWJ, ZWNJ, LRM, RLM in escaped form
Youngbok Shin [Mon, 11 Sep 2017 05:53:04 +0000 (08:53 +0300)]
evas textblock: support ZWJ, ZWNJ, LRM, RLM in escaped form

Summary:
The patch supports the following escaped string from Evas Textblock.
These unicodes are frequently used.
ZWNJ U+200C - &zwnj;
ZWJ  U+200D - &zwj;
LRM  U+200E - &lrm;
RLM  U+200F - &rlm;
@feature

Test Plan: N/A

Reviewers: raster, cedric, herdsman, jpeg, woohyun

Differential Revision: https://phab.enlightenment.org/D5173

6 years agoevas: don't search other fonts only for INHERITED script unicodes
Youngbok Shin [Mon, 11 Sep 2017 05:52:03 +0000 (08:52 +0300)]
evas: don't search other fonts only for INHERITED script unicodes

Summary:
INHERITED script unicodes are only meaningful when it comes
after the previous unicode from same font.
Even if there is no glyph for the INHERTED script unicode from current font,
don't search other font for loading the unicdoe as first unicode
of next item. It will be meaningless.
@fix

Test Plan:
Check the following Emoji sequence with an emoji font
which does not have variation selector glyphs.
{ 0x1F3F3, 0xFE0F, 0x200D, 0x1F308 }

Reviewers: raster, cedric, herdsman, jpeg, woohyun

Reviewed By: herdsman

Differential Revision: https://phab.enlightenment.org/D5156

6 years agoefl ui layout - fix warning fro wrong type... its a list not a hash.
Carsten Haitzler (Rasterman) [Sun, 10 Sep 2017 04:21:50 +0000 (13:21 +0900)]
efl ui layout - fix warning fro wrong type... its a list not a hash.

6 years agoeo-cxx: Require instantiate keyword for constructors calling efl_add to avoid ambiguity
Felipe Magno de Almeida [Thu, 7 Sep 2017 09:02:43 +0000 (18:02 +0900)]
eo-cxx: Require instantiate keyword for constructors calling efl_add to avoid ambiguity

6 years agoedje: Better debug logs for invalid calls to swallow
Jean-Philippe Andre [Wed, 6 Sep 2017 05:39:22 +0000 (14:39 +0900)]
edje: Better debug logs for invalid calls to swallow

These error messages will be triggered when efl_part() is used, and the
part is not a swallow part.

6 years agoEvas filters: remove needless check for negative case.
Sungtaek Hong [Mon, 4 Sep 2017 08:47:31 +0000 (17:47 +0900)]
Evas filters: remove needless check for negative case.

Summary:
Becuse ox is set to 0,
it cannot be negative, so negative check will be needless.

Reviewers: jpeg

Reviewed By: jpeg

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5155

6 years agoCanvas text: emit missing events during text changes
Daniel Hirt [Thu, 7 Sep 2017 13:53:01 +0000 (16:53 +0300)]
Canvas text: emit missing events during text changes

Some text changes were not picked up by the Ui.Text widget, so no size
evaluation was triggered.

6 years agoelm: Fix multiple connections from model to properties in Layout
Felipe Magno de Almeida [Thu, 7 Sep 2017 03:07:08 +0000 (12:07 +0900)]
elm: Fix multiple connections from model to properties in Layout

6 years agoecore: Add efl_model_copmosite_selection model
Felipe Magno de Almeida [Thu, 7 Sep 2017 03:06:17 +0000 (12:06 +0900)]
ecore: Add efl_model_copmosite_selection model

6 years agoefl: Make models 0-based index, not 1-based
Felipe Magno de Almeida [Thu, 7 Sep 2017 03:05:24 +0000 (12:05 +0900)]
efl: Make models 0-based index, not 1-based

6 years agoeio: Fix eio model children count in initialization of children model
Felipe Magno de Almeida [Thu, 7 Sep 2017 03:04:40 +0000 (12:04 +0900)]
eio: Fix eio model children count in initialization of children model

6 years agoRevert "elm: rename elm spinner to Efl.Ui.Spinner"
Amitesh Singh [Thu, 7 Sep 2017 02:12:52 +0000 (11:12 +0900)]
Revert "elm: rename elm spinner to Efl.Ui.Spinner"

This reverts commit 9836116cab93388cf3d079d39180f675f86be78f.

This is based on discussion today i had.
There would be a separate minimal spinner class instead
which facilates ways to extend it.

6 years agocxx: generate a constructor with a parent and a lambda function
Daniel Zaoui [Wed, 6 Sep 2017 22:24:52 +0000 (01:24 +0300)]
cxx: generate a constructor with a parent and a lambda function

Until now, one could not invoke functions into the constructor of a
widget possessing a parent. It is needed for widgets such as Efl.Ui.Check
where style is needed during construction.

@fix T5980

6 years agonaviframe: Fix to finish transition for newly pushed item
Jaehyun Cho [Wed, 6 Sep 2017 11:04:41 +0000 (20:04 +0900)]
naviframe: Fix to finish transition for newly pushed item

Suppose naviframe top item is "A" and a new item "B" is newly pushed.
In this case, if "A" is deleted by elm_object_item_del() before
transition for "B" is started, then signal for "B" is also not sent and
"B" does not become visible.

The above issue happened if the transition effect was implemented by
using deferred signals (i.e. "pushed,deferred" and "popped,deferred").

To resolve the above issue, the signal only for the deleted item is not
sent.

Suppose naviframe top item is "A" and a new item "B" is newly pushed.
In this case, if "B" is deleted by elm_object_item_del() before
transition for "B" is started, then signal for "A" is still sent and "A"
becomes invisible.

To resolve the above issue, if the deleted item is the top item and it
is in the middle of item push, then all the signals related to the
deleted item are not sent.

6 years agoRevert "naviframe: Fix to finish transition for newly pushed item"
Jaehyun Cho [Wed, 6 Sep 2017 11:03:26 +0000 (20:03 +0900)]
Revert "naviframe: Fix to finish transition for newly pushed item"

This reverts commit 8574128b67d336e796ca0f0abc9b97e65f7799d4.

6 years agoelementary: atspi accessible name uses plain text
Shinwoo Kim [Wed, 6 Sep 2017 10:54:59 +0000 (19:54 +0900)]
elementary: atspi accessible name uses plain text

The markup information of accessible name is not necessary.

6 years agoeo-cxx: Fix ompilation error for uncomplete Eina_Future type
Felipe Magno de Almeida [Wed, 6 Sep 2017 08:03:03 +0000 (17:03 +0900)]
eo-cxx: Fix ompilation error for uncomplete Eina_Future type

6 years agoelm test: slider: Add limited interval slider example.
Amitesh Singh [Wed, 6 Sep 2017 05:04:45 +0000 (14:04 +0900)]
elm test: slider: Add limited interval slider example.

6 years agoevas_language_utils: Fix build error without NLS
Jaehyun Cho [Wed, 6 Sep 2017 02:19:28 +0000 (11:19 +0900)]
evas_language_utils: Fix build error without NLS

6 years agoelementary: enhance accessible name of elm_entry
Shinwoo Kim [Wed, 6 Sep 2017 00:58:07 +0000 (09:58 +0900)]
elementary: enhance accessible name of elm_entry

The accessible name of elm_entry was guide text.
The accessible name of elm_entry should be its text.
If there is not text information, then guide text is using for accessible name.

6 years agofilters: Fix GL support aka. cedric's bork
Jean-Philippe Andre [Tue, 5 Sep 2017 07:29:01 +0000 (16:29 +0900)]
filters: Fix GL support aka. cedric's bork

6 years agofilters: Safety++
Jean-Philippe Andre [Mon, 4 Sep 2017 10:08:02 +0000 (19:08 +0900)]
filters: Safety++

This is an attempt at fixing crashes in empc.
Test scenario:
  ELM_ACCEL="" elementary_test -to "Gfx Filters"
  And mouse scroll like crazy in the spinner.

@fix

6 years agoeina: properly mark memory for valgrind before accessing it in one_big mempool.
Cedric Bail [Wed, 6 Sep 2017 00:01:03 +0000 (17:01 -0700)]
eina: properly mark memory for valgrind before accessing it in one_big mempool.

6 years agoecore_drm2: Better handling of blank time checks
Derek Foreman [Tue, 5 Sep 2017 17:08:02 +0000 (12:08 -0500)]
ecore_drm2: Better handling of blank time checks

On some systems we'll successfully complete the vblank ioctl but get
a reply of 0.  When that happens we can't use that time for ticking
as it will break all of the entire world.

Fixes immediate screen blank on rpi3.

@ref T5977

6 years agoRevert "evas_thread: only join the thread if the thread is still alive"
Marcel Hollerbach [Tue, 5 Sep 2017 12:36:17 +0000 (14:36 +0200)]
Revert "evas_thread: only join the thread if the thread is still alive"

This reverts commit e41d46c635bdf769d4b93da609883347168c719a.

This seems to be wrong

6 years agoevas_thread: only join the thread if the thread is still alive
Marcel Hollerbach [Mon, 4 Sep 2017 17:44:03 +0000 (19:44 +0200)]
evas_thread: only join the thread if the thread is still alive

otherwise we might join a invalid thread id, that could lead to a
deadlock. Lets not do that.

ref T5245

@fix

6 years agoelm_panel: set content as a sub object of panel object
Jaeun Choi [Tue, 5 Sep 2017 08:09:37 +0000 (17:09 +0900)]
elm_panel: set content as a sub object of panel object

since sd->bx is not an elm object, it is excluded from parent-child tree.
in case of a non-scrollable panel, the content needs to be attached to the tree
as a child of panel object directly.

6 years agoeina: fix compilation after b0rking in c9a0237770a7fb
Amitesh Singh [Tue, 5 Sep 2017 07:20:43 +0000 (16:20 +0900)]
eina: fix compilation after b0rking in c9a0237770a7fb

Ref T5966

6 years agoeina_promise: remove shadowed declaration warning
Thiep Ha [Tue, 5 Sep 2017 06:56:00 +0000 (15:56 +0900)]
eina_promise: remove shadowed declaration warning

6 years agofocus_manager: remove unused parameter warning
Thiep Ha [Tue, 5 Sep 2017 05:44:21 +0000 (14:44 +0900)]
focus_manager: remove unused parameter warning

6 years agotextpath: correct map point number
Thiep Ha [Tue, 5 Sep 2017 05:19:59 +0000 (14:19 +0900)]
textpath: correct map point number

The last map point is not counted if seg is too small.
Test: change start angle to 3, the last segment is missed.

6 years agotextpath: remove legacy api support
Thiep Ha [Tue, 5 Sep 2017 03:07:38 +0000 (12:07 +0900)]
textpath: remove legacy api support

6 years agotextpath: do not process line if failed to allocate segment
Thiep Ha [Tue, 5 Sep 2017 03:02:01 +0000 (12:02 +0900)]
textpath: do not process line if failed to allocate segment

6 years agotextpath: correct data gets from path
Thiep Ha [Tue, 5 Sep 2017 02:55:07 +0000 (11:55 +0900)]
textpath: correct data gets from path

values of px0, py0 should be stored from previous command.

6 years agoeina: fix destruction of the osx semaphores
Jean Guyomarc'h [Mon, 4 Sep 2017 22:09:57 +0000 (00:09 +0200)]
eina: fix destruction of the osx semaphores

Well... that's just embarassing... semaphore_destroy() actually takes
the mach task as its first parameter, not the second. This core
amazingly worked very fine on macOS earlier than Sierra.

Fixes T5245

6 years agoecore: fix return value
Jean Guyomarc'h [Mon, 4 Sep 2017 17:59:20 +0000 (19:59 +0200)]
ecore: fix return value

6 years agoefl_io_copier: Use the new Eina_Future API.
Guilherme Iscaro [Wed, 30 Aug 2017 20:25:28 +0000 (17:25 -0300)]
efl_io_copier: Use the new Eina_Future API.

6 years agoefl_net: Use the new Eina_Future API.
Guilherme Iscaro [Wed, 30 Aug 2017 20:24:27 +0000 (17:24 -0300)]
efl_net: Use the new Eina_Future API.

6 years agoEfl_Loop: add job, timeout and idle based on Eina_Future.
Gustavo Sverzut Barbieri [Sat, 26 Aug 2017 22:59:39 +0000 (19:59 -0300)]
Efl_Loop: add job, timeout and idle based on Eina_Future.

Since some clash with old version, then add Eina_FutureXXX to their
name, later we'll sed.

6 years agoexport efl_future_then() for Eina_Future syntax sugar.
Gustavo Sverzut Barbieri [Sat, 26 Aug 2017 22:39:23 +0000 (19:39 -0300)]
export efl_future_then() for Eina_Future syntax sugar.

This is actually written as efl_future_Eina_FutureXXX_then() as the
old API clashes, after removing the old code we'll "sed" to fix those.

6 years agoEina_Future: add eina_future_resolved()
Gustavo Sverzut Barbieri [Sat, 26 Aug 2017 22:04:43 +0000 (19:04 -0300)]
Eina_Future: add eina_future_resolved()

This is a helper that creates a promise, then a future and immediately
resolves the promise, this will let the future be dispatched as usual,
from a clean main loop context.

6 years agoEina_Promise/Eina_Future: Add example and tests.
Guilherme Iscaro [Tue, 8 Aug 2017 21:11:23 +0000 (18:11 -0300)]
Eina_Promise/Eina_Future: Add example and tests.

6 years agoEfl_Object: Add integration with Eina_Future.
Guilherme Iscaro [Fri, 25 Aug 2017 22:53:15 +0000 (19:53 -0300)]
Efl_Object: Add integration with Eina_Future.

This commit adds the EO support for the new future infra.
From now on there's no need to efl_future_link()/efl_future_unlink()
object and futures since the new API already handles that internally.

6 years agoEina: Add Eina_Promise/Eina_Future.
Guilherme Iscaro [Tue, 8 Aug 2017 21:10:36 +0000 (18:10 -0300)]
Eina: Add Eina_Promise/Eina_Future.

This commit adds a new promise/future API which aims to replace
efl_future.

6 years agoeina: properly track valgrind use of the mempool.
Cedric Bail [Mon, 4 Sep 2017 10:08:33 +0000 (03:08 -0700)]
eina: properly track valgrind use of the mempool.

T5966

6 years agoevas: proceed glyph iterator to handle next index properly
Youngbok Shin [Mon, 4 Sep 2017 07:18:05 +0000 (10:18 +0300)]
evas: proceed glyph iterator to handle next index properly

Summary:
When harfbuzz is enabled, _content_create_ot() function will be used
for shaping. If evas_common_font_int_cache_glyph_get() failed in some reason,
it never proceed gl_itr until the end.
It can cause weird rendering result. Because, all of gl_itr after the failure
can't have proper x_bear, y_bear and width.
@fix

Test Plan: N/A

Reviewers: raster, cedric, herdsman, jpeg

Differential Revision: https://phab.enlightenment.org/D5154

6 years agonaviframe: Fix to finish transition for newly pushed item
Jaehyun Cho [Mon, 4 Sep 2017 05:58:39 +0000 (14:58 +0900)]
naviframe: Fix to finish transition for newly pushed item

If current item was deleted while new item was pushed, then the signals
for the newly pushed item was not sent.

The above issue happened if the transition effect was implemented by
using deferred signals (i.e. "pushed,deferred" and "popped,deferred").

To resolve the above issue, the signals only for the deleted item is not
sent.

6 years agoelm_widget: move the complete regsiter/unregister logic
Marcel Hollerbach [Sat, 2 Sep 2017 17:08:18 +0000 (19:08 +0200)]
elm_widget: move the complete regsiter/unregister logic

We had here a little problem, state focus_state_eval function handled
the unregisteration and consideration of the focus flags and then only
called a helper function (which was a widget function), that then did
the registeration in logical or regular mode.

Elm scroller for example took that function overwrote it and did onyl
permit logical registrations. Then again a evaluation of the focus state
and flags took place, and the function considered elm_scroller should be
registered as regular object, but found it to be logical. This lead to
the problem that we permantently unregistered Elm.Scroller and
registered it again as logical just to unregister it again. This was on
the one side a performance downside. But also a bug since all items from
within the Elm_Scrollers sub manager are getting reparent onto the
parent, which means not the root of the scroller (the scroller itself)
is the logical entrypoint to the widget but rather this reparented
widget, which led to unexpected focus  warps like described in T5923.

tldr: this fixes T5923

6 years agoefl_ui_focus_manager: present a slightly better error message
Marcel Hollerbach [Sat, 2 Sep 2017 17:41:30 +0000 (19:41 +0200)]
efl_ui_focus_manager: present a slightly better error message

with this and the backtrace you can probebly figure out if you just
registered the same widget twice for different types

6 years agoefl_ui_focus_manager_sub: not needed anymore, handled by calc
Marcel Hollerbach [Fri, 1 Sep 2017 16:12:13 +0000 (18:12 +0200)]
efl_ui_focus_manager_sub: not needed anymore, handled by calc

6 years agoelm_widget: eval everything when the parent changes
Marcel Hollerbach [Fri, 1 Sep 2017 15:58:41 +0000 (17:58 +0200)]
elm_widget: eval everything when the parent changes

6 years agoefl_ui_win: specify the correct manager
Marcel Hollerbach [Fri, 1 Sep 2017 15:32:06 +0000 (17:32 +0200)]
efl_ui_win: specify the correct manager

a widget in higher hirachies is not registered in the lower one.

6 years agoefl_ui_focus_manager_calc: do not swallow a focus event
Marcel Hollerbach [Fri, 1 Sep 2017 15:31:06 +0000 (17:31 +0200)]
efl_ui_focus_manager_calc: do not swallow a focus event

if we have been redirecting before, we would have lost that event due to
  the element here beeing at the top of that focus stack

6 years agoefl_ui_focus_manager_sub: make it a mixin
Marcel Hollerbach [Thu, 13 Jul 2017 08:12:50 +0000 (10:12 +0200)]
efl_ui_focus_manager_sub: make it a mixin

This just brings the functionality, the manager that is used as sub can
still be decided and be brought in via inheritance or compositition

6 years agoecore_exe - fix failed finalize fork to reutrn NULL not ecore_exe_free
Carsten Haitzler (Rasterman) [Sat, 2 Sep 2017 13:33:00 +0000 (22:33 +0900)]
ecore_exe - fix failed finalize fork to reutrn NULL not ecore_exe_free

dont delete the obj during finalize... just retyurn NULL to fail.
fork() failed for me so i found this... ask not why fork failed... but
it did... and thus found this error handling case.

@fix

6 years agowayland_shm - clean out ptrs after free - seeing invalid mem/ptr access
Carsten Haitzler (Rasterman) [Sat, 2 Sep 2017 12:47:18 +0000 (21:47 +0900)]
wayland_shm - clean out ptrs after free - seeing invalid mem/ptr access

enlightenment internal windows insta segv e on rpi. after much hunting
it seems a fallback is happening and bunk ptrs are being used. this at
least will make the problems more reliable with null ptrs.

6 years agowayland_shm: Try a little harder to verify exynos dmabuf capabilities
Derek Foreman [Fri, 1 Sep 2017 19:49:22 +0000 (14:49 -0500)]
wayland_shm: Try a little harder to verify exynos dmabuf capabilities

Turns out the "device_open" function pretty much just tests calloc
functionality, and doesn't open any device.  So let's allocate a
tiny bo and discard it to make sure we're actually on exynos.

6 years agoLuncher: Remove the engage feature.
Stephen 'Okra' Houston [Fri, 1 Sep 2017 18:45:09 +0000 (13:45 -0500)]
Luncher: Remove the engage feature.

The engage feature was mainly there as a demo of the capabilities of bryce.  Now that we are nearing release we need to clean up our gadgets.  The engage style for the luncher gadget is not complete, and does not work adequately, and quite honestly better belongs as a feature of bryce itself not luncher.

6 years agoelm_win: Stop hiding the mouse cusor on window out under wayland
Derek Foreman [Fri, 1 Sep 2017 18:28:36 +0000 (13:28 -0500)]
elm_win: Stop hiding the mouse cusor on window out under wayland

We needed to do this to prevent burning cpu with animated cursors
before we did frame callback based animating.

Now the compositor will stop our frame callbacks when our cursor
is implicitly visible when the pointer isn't over our window,
so we don't have to use window hides for it.

6 years agoevas: fix event grabber child stacking
Mike Blumenkrantz [Thu, 31 Aug 2017 20:13:01 +0000 (16:13 -0400)]
evas: fix event grabber child stacking

broken in 0516cdc0f9dc0969ac96df6984af2413680cad01

@fix

6 years agoevas: handle object names with a list hash
Mike Blumenkrantz [Thu, 31 Aug 2017 20:13:01 +0000 (16:13 -0400)]
evas: handle object names with a list hash

a canvas can have multiple objects with the same name. assuming that name:obj
is a 1:1 ratio means that name-finding functions are likely to return invalid
objects

@fix

6 years agoecore_wl2: Don't set up callbacks on initial commit
Derek Foreman [Fri, 1 Sep 2017 17:03:49 +0000 (12:03 -0500)]
ecore_wl2: Don't set up callbacks on initial commit

We don't want to do frame callbacks until after we can actually draw stuff

6 years agoeolian: fix float parsing with locales with ',' decpoint
Daniel Kolesa [Fri, 1 Sep 2017 14:07:45 +0000 (16:07 +0200)]
eolian: fix float parsing with locales with ',' decpoint

6 years agobuild: Drop wayland protocol requirement to 1.9
Derek Foreman [Fri, 1 Sep 2017 13:45:47 +0000 (08:45 -0500)]
build: Drop wayland protocol requirement to 1.9

Some distros are sitting at 1.9 for a long time now, and 1.10
didn't add any critical functionality for us.

6 years agoSoftware gdi: fix build on Windows.
Mykyta Biliavskyi [Fri, 1 Sep 2017 10:50:34 +0000 (13:50 +0300)]
Software gdi: fix build on Windows.

Summary:
Function argument was renamed, but in function body still uses old
variable name.

Test Plan: Build on Windows host

Reviewers: cedric, vtorri

Reviewed By: vtorri

Subscribers: jpeg

Tags: #windows, #efl

Differential Revision: https://phab.enlightenment.org/D5152

6 years agoelm: rename elm spinner to Efl.Ui.Spinner
Amitesh Singh [Mon, 28 Aug 2017 08:05:22 +0000 (17:05 +0900)]
elm: rename elm spinner to Efl.Ui.Spinner

Ref T5900

6 years agoefl_config: Fix type of returned value (EO)
Jean-Philippe Andre [Thu, 31 Aug 2017 11:31:54 +0000 (20:31 +0900)]
efl_config: Fix type of returned value (EO)

An ugly const char * is returned from strdup, freed later with a custom
function that just calls free(). This was probably intended for
stringshare but now free is used. We have to fix the free function for
the EO API so let's keep free(). I'd rather avoid declaring
elm_config_profile_dir_free in the EO files.

6 years agoelm: Drop unused class inheritance (EO)
Jean-Philippe Andre [Thu, 31 Aug 2017 11:18:04 +0000 (20:18 +0900)]
elm: Drop unused class inheritance (EO)

6 years agoelm_test: Scale down icon by 8x
Jean-Philippe Andre [Thu, 31 Aug 2017 10:41:12 +0000 (19:41 +0900)]
elm_test: Scale down icon by 8x

This makes the icon test actually work. Otherwise the icon data is too
big and basically seems ignored by the compositor or X.

Note: In E (X11) it seems that the window icon remains unchanged? xprop
shows the proper data, though. Ping @zmike.

6 years agoeina_array: Fix magic checks to return
Jean-Philippe Andre [Thu, 31 Aug 2017 10:25:26 +0000 (19:25 +0900)]
eina_array: Fix magic checks to return

The EINA_MAGIC check inside eina_array were just printing the error
message, without failing. This could lead to more crashes than
necessary.

@fix

6 years agowin: Fixup doc & code of icon_object
Jean-Philippe Andre [Thu, 31 Aug 2017 09:32:02 +0000 (18:32 +0900)]
win: Fixup doc & code of icon_object

Ref T5322

6 years agowin: Use array for available_profiles (EO)
Jean-Philippe Andre [Thu, 31 Aug 2017 08:57:40 +0000 (17:57 +0900)]
win: Use array for available_profiles (EO)

The legacy API wraps the EO API.
All tested with "Configuration" test case in elm_test.

Ref T5322

6 years agowin: Drop calls to lower when using the EO API
Jean-Philippe Andre [Thu, 31 Aug 2017 07:58:02 +0000 (16:58 +0900)]
win: Drop calls to lower when using the EO API

Note: The distinction is made on how the window was created, not on
which API is used (evas_object_lower or efl_gfx_stack_lower or
elm_win_lower).

Ref T5322

6 years agowin: Implement raise from stack (EO)
Jean-Philippe Andre [Thu, 31 Aug 2017 07:45:08 +0000 (16:45 +0900)]
win: Implement raise from stack (EO)

I think the intercept is not required anymore. All raise functions call
the ecore evas raise function.

Ref T5322

6 years agofocus: Fix shadow warning
Jean-Philippe Andre [Thu, 31 Aug 2017 07:28:42 +0000 (16:28 +0900)]
focus: Fix shadow warning

6 years agowidget: Forward focus_highlight_style to win
Jean-Philippe Andre [Thu, 31 Aug 2017 07:10:43 +0000 (16:10 +0900)]
widget: Forward focus_highlight_style to win

See the previous commits. All focus_highlight APIs are defined in the
Widget class but only implemented at the Window level. For consistency I
believe this call should also be forwarded to the window. The previous
logic had absolutely no effect at all, except storing a stringshare.

The day focus_highlight becomes widget-specific (i.e. each widget has
its own highlight style) then this can be changed.

Note: This will apply to legacy API as well.

Ref T5363

6 years agowin: Move focus_highlight_animate to widget (EO)
Jean-Philippe Andre [Thu, 31 Aug 2017 07:08:10 +0000 (16:08 +0900)]
win: Move focus_highlight_animate to widget (EO)

Similar to the previous commits. Moved for consistency.

Ref T5363
Ref T5322

6 years agowin: Move focus_highlight_enabled to widget (EO)
Jean-Philippe Andre [Thu, 31 Aug 2017 06:57:53 +0000 (15:57 +0900)]
win: Move focus_highlight_enabled to widget (EO)

This was actually declared in the internal legacy API in widget.
Forwards the calls to the window.

Ref T

6 years agowin: Merge focus_highlight_style with widget (EO)
Jean-Philippe Andre [Thu, 31 Aug 2017 06:42:52 +0000 (15:42 +0900)]
win: Merge focus_highlight_style with widget (EO)

Note: elm_test "Focus Style" can be used to test this API. The test case
is a bit broken (overly complex EDC?) but if you're patient you can see
the difference between "glow" and "glow_effect".

Ref T5363
Ref T5322