platform/upstream/efl.git
6 years agoeio: switch Efl.Io.Manager to actually use Array instead of Accessor (to avoid uneces...
Cedric Bail [Fri, 19 Jan 2018 21:49:24 +0000 (13:49 -0800)]
eio: switch Efl.Io.Manager to actually use Array instead of Accessor (to avoid unecessary callback).

6 years agoeio: share _efl_io_manager_future_cancel with all of eio.
Cedric BAIL [Thu, 21 Dec 2017 00:23:18 +0000 (16:23 -0800)]
eio: share _efl_io_manager_future_cancel with all of eio.

6 years agoeio: allow access to Eina_Value Eina_Stat description from all part of the library.
Cedric BAIL [Wed, 20 Dec 2017 00:03:20 +0000 (16:03 -0800)]
eio: allow access to Eina_Value Eina_Stat description from all part of the library.

6 years agoecore: move efl_model_list_value_get to be an internal function.
Cedric BAIL [Wed, 20 Dec 2017 22:19:56 +0000 (14:19 -0800)]
ecore: move efl_model_list_value_get  to be an internal function.

6 years agoeldbus: migrate and refactor eldbus.model to new efl.model API.
Cedric Bail [Thu, 26 Oct 2017 01:05:41 +0000 (18:05 -0700)]
eldbus: migrate and refactor eldbus.model to new efl.model API.

6 years agoeldbus: refactor model code.
Cedric Bail [Thu, 26 Oct 2017 01:05:19 +0000 (18:05 -0700)]
eldbus: refactor model code.

6 years agoecore: don't artificially block fonctionnality for nested loop.
Cedric BAIL [Fri, 2 Feb 2018 22:53:57 +0000 (14:53 -0800)]
ecore: don't artificially block fonctionnality for nested loop.

We do have nested loop when manually calling iterate and was used
in legacy code that got broken with this change.

6 years agoecore: move efl_loop_message_process to ecore_internal.h to allow other EFL component...
Cedric BAIL [Fri, 2 Feb 2018 22:50:54 +0000 (14:50 -0800)]
ecore: move efl_loop_message_process to ecore_internal.h to allow other EFL component to use it.

6 years agoecore: refactor and migrate efl.model.container to the new efl.model API.
Cedric Bail [Thu, 26 Oct 2017 01:03:43 +0000 (18:03 -0700)]
ecore: refactor and migrate efl.model.container to the new efl.model API.

6 years agoecore: refactor and migrate efl.model.composite.selection to the new efl.model API.
Cedric Bail [Thu, 26 Oct 2017 01:03:00 +0000 (18:03 -0700)]
ecore: refactor and migrate efl.model.composite.selection to the new efl.model API.

6 years agoecore: refactor and migrate efl.model.composite.boolean to the new efl.model API.
Cedric Bail [Thu, 26 Oct 2017 01:02:01 +0000 (18:02 -0700)]
ecore: refactor and migrate efl.model.composite.boolean to the new efl.model API.

6 years agoecore: convert efl.model.item to the new efl.model API.
Cedric Bail [Thu, 26 Oct 2017 01:00:13 +0000 (18:00 -0700)]
ecore: convert efl.model.item to the new efl.model API.

6 years agoecore: add a class to factorize efl.model.composite.
Cedric Bail [Wed, 25 Oct 2017 22:17:28 +0000 (15:17 -0700)]
ecore: add a class to factorize efl.model.composite.

6 years agoefl: make efl.model use of eina_value and eina_future instead of efl_future.
Cedric Bail [Wed, 25 Oct 2017 22:13:32 +0000 (15:13 -0700)]
efl: make efl.model use of eina_value and eina_future instead of efl_future.

6 years agoelm_focus_legacy: set custom chain
Marcel Hollerbach [Mon, 30 Apr 2018 13:32:03 +0000 (15:32 +0200)]
elm_focus_legacy: set custom chain

fix T6182

6 years agoefl_ui_widget: do not register if the child is missing in a custom chain
Marcel Hollerbach [Sat, 21 Apr 2018 16:54:50 +0000 (18:54 +0200)]
efl_ui_widget: do not register if the child is missing in a custom chain

6 years agoefl_ui_focus_composition: legacy compatibility
Marcel Hollerbach [Sat, 21 Apr 2018 16:54:22 +0000 (18:54 +0200)]
efl_ui_focus_composition: legacy compatibility

6 years agoelm_widget: Introduce elm_focus_legacy
Marcel Hollerbach [Thu, 12 Apr 2018 19:28:49 +0000 (21:28 +0200)]
elm_widget: Introduce elm_focus_legacy

Code located here is meant for a legacy tree and legacy widgets.
This is checked on the entry of the function

6 years agoefl_ui_focus_util: add new function for complementing the direction
Marcel Hollerbach [Fri, 13 Apr 2018 15:07:30 +0000 (17:07 +0200)]
efl_ui_focus_util: add new function for complementing the direction

6 years agoefl_canvas_object: Add event info to animation events in eo
Jaehyun Cho [Mon, 30 Apr 2018 15:20:03 +0000 (00:20 +0900)]
efl_canvas_object: Add event info to animation events in eo

6 years agoefl_canvas_object: Change name of animation event info
Jaehyun Cho [Mon, 30 Apr 2018 15:07:27 +0000 (00:07 +0900)]
efl_canvas_object: Change name of animation event info

Change name of animation event info to synchronize with other event
info names.

Change data name in animation event info to synchronize with data type
name, Efl.Event.Description.

6 years agoefl_canvas_object: Remove unnecessary import
Jaehyun Cho [Mon, 30 Apr 2018 14:44:02 +0000 (23:44 +0900)]
efl_canvas_object: Remove unnecessary import

6 years agoelm_widget_item_static_focus: do not call prepare on all realized items
Marcel Hollerbach [Mon, 30 Apr 2018 10:36:29 +0000 (12:36 +0200)]
elm_widget_item_static_focus:  do not call prepare on all realized items

we should spare out the group and disabled items

6 years agoelm_widget_item_static_focus: a other element could also hide behind the
Marcel Hollerbach [Mon, 30 Apr 2018 09:17:36 +0000 (11:17 +0200)]
elm_widget_item_static_focus: a other element could also hide behind the
adapter

6 years agoefl_ui_focus_manager_calc: drop redirect if there is no candidate
Marcel Hollerbach [Mon, 30 Apr 2018 08:51:52 +0000 (10:51 +0200)]
efl_ui_focus_manager_calc: drop redirect if there is no candidate

6 years agoelm_interface_scrollable: rework filtering
Marcel Hollerbach [Wed, 25 Apr 2018 12:01:48 +0000 (14:01 +0200)]
elm_interface_scrollable: rework filtering

A item might be in the middle of the viewport and still miss relations.

6 years agoelm_****: do not make group items focusable
Marcel Hollerbach [Wed, 25 Apr 2018 12:00:28 +0000 (14:00 +0200)]
elm_****: do not make group items focusable

6 years agoefl_ui_focus_manager_calc: make the preparation call tree style
Marcel Hollerbach [Wed, 25 Apr 2018 11:29:26 +0000 (13:29 +0200)]
efl_ui_focus_manager_calc: make the preparation call tree style

So we make sure that all children are prepared.

6 years agoelm_widget_item_static_focus: prepare the item not the widget
Marcel Hollerbach [Wed, 25 Apr 2018 11:28:54 +0000 (13:28 +0200)]
elm_widget_item_static_focus: prepare the item not the widget

fix T6894

6 years agoelm_multibuttonentry: Fix make check fail
Jaehyun Cho [Mon, 30 Apr 2018 10:42:04 +0000 (19:42 +0900)]
elm_multibuttonentry: Fix make check fail

Fix make check fail of elm_multibuttonentry caused by
939690a8a99b04d9e67b6b90bc80de639a7e8941

6 years agoefl.ui.layout_factory : use interface layout theme set for now
SangHyeon Lee [Mon, 30 Apr 2018 10:27:09 +0000 (19:27 +0900)]
efl.ui.layout_factory : use interface layout theme set for now

interface object need to use efl.ui.layout_object theme set
instead of legacy one, for applying proper theme naming rules.

6 years agoefl.ui.view_list : update example text/content name on legacy to efl
SangHyeon Lee [Mon, 30 Apr 2018 09:54:32 +0000 (18:54 +0900)]
efl.ui.view_list : update example text/content name on legacy to efl

update examples and edc as our new naming rule changes.

6 years agoevas: fix default return value of image-related legacy functions
Jaeun Choi [Mon, 30 Apr 2018 07:52:21 +0000 (16:52 +0900)]
evas: fix default return value of image-related legacy functions

In commit 8367f497b8f1e9a7b365c815f12deeab9c92e1a5,
type checking code was added for legacy image functions
in order to reduce failing calls of efl_xxx function.
The default return values of some functions for none-legacy
objects (including NULL), however, were not equal to eo-based
default return value, which broke compatibility.
This commit fixes it.

@fix

6 years agoecore animator, x, vsync ecore evas restore x vsync to work again
Carsten Haitzler (Rasterman) [Wed, 25 Apr 2018 09:22:56 +0000 (18:22 +0900)]
ecore animator, x, vsync ecore evas restore x vsync to work again

so i found the work with wayland and having animator sources broke
that same source from ecore_x that was there from long ago, so i've
put in an exception if there are x based engines from restting to a
timer animator because ecore_x would have switched toa custom ticker
already, and this just resets it. also just set the source after
setting the tick callbacks and ensure tick cb's are null before going
to timer source as well. this cleans up this little but of animaatior
vsync modification to properly vsync in both x and wayland too now.

@fix

6 years agoGenlist: a sane test for genlist item iteration
Dave Andreoli [Sat, 28 Apr 2018 10:22:51 +0000 (12:22 +0200)]
Genlist: a sane test for genlist item iteration

This is much more extensive than before and we are now
testing both tree and group items.

This test has been run in 9 stable releases down to 1.12
the only one that fail is 1.18 that is know to be broken
wrt item iteration.

Hopefully definitly close T5938

6 years agoGenlist: fix doc for item_next/prev_get()
Dave Andreoli [Sat, 28 Apr 2018 08:31:23 +0000 (10:31 +0200)]
Genlist: fix doc for item_next/prev_get()

See T5938 for details

6 years agoRevert "genlist: fix "insane" order [BUG COMPATIBILITY]"
Dave Andreoli [Sat, 28 Apr 2018 08:21:57 +0000 (10:21 +0200)]
Revert "genlist: fix "insane" order [BUG COMPATIBILITY]"

This reverts commit fd82c2521ebb9a324db8fdebd2c9a62b76ee6dc9.

This was based on a wrong assumption, see T5938 fro details

6 years agoefl.ui.view_list : fix signal prefix elm to efl
SangHyeon Lee [Sat, 28 Apr 2018 05:03:50 +0000 (14:03 +0900)]
efl.ui.view_list : fix signal prefix elm to efl

by the patch 6e49db0739c38a378c48c6c880f5b41c986ee1e9
efl.ui interface change signal prefix legacy elm to efl.
this patch is fixing efl.ui.view.list scroller issue

6 years agoefl_ui_list : introduce new list widget for simple usage
SangHyeon Lee [Mon, 23 Apr 2018 06:13:59 +0000 (15:13 +0900)]
efl_ui_list : introduce new list widget for simple usage

Summary:
most usage of simple list, items are very limited and loading performance is not serious.
to support those requirement, this efl.ui.list will create scrollable box with efl.pack.

user can create list by packing an pre-loaded item object, Efl.Ui.List.Item class.

Test Plan: tested in efl_ui_list_example_1.c in examples.

Reviewers: cedric, felipealmeida

Subscribers: woohyun, Jaehyun_Cho

Differential Revision: https://phab.enlightenment.org/D5861

6 years agotests: disable ecore system modules in elm_suite
Mike Blumenkrantz [Fri, 27 Apr 2018 19:43:31 +0000 (21:43 +0200)]
tests: disable ecore system modules in elm_suite

Summary:
ref T6864

Depends on D5967

Reviewers: stefan_schmidt

Subscribers: cedric

Maniphest Tasks: T6864

Differential Revision: https://phab.enlightenment.org/D5968

6 years agotests: disable efreetd for elm tests
Mike Blumenkrantz [Fri, 27 Apr 2018 19:42:38 +0000 (21:42 +0200)]
tests: disable efreetd for elm tests

Summary:
this is especially problematic for elm tests as the fork reset callback
in efreet will cause a huge number of server connections to be created

ref T6864

Depends on D5966

Reviewers: stefan_schmidt

Reviewed By: stefan_schmidt

Subscribers: cedric

Maniphest Tasks: T6864

Differential Revision: https://phab.enlightenment.org/D5967

6 years agoeina: Provide actual len (after possible truncation) to eina_stringshare_add_length
Xavi Artigas [Fri, 27 Apr 2018 17:43:12 +0000 (13:43 -0400)]
eina: Provide actual len (after possible truncation) to eina_stringshare_add_length

Summary:
vsnprintf() returns "The number of characters that would have been written if n had been sufficiently large",
so if this is larger than the requested length, we pass the requested length to eina_stringshare_add_length,
to avoid reading uninitialised memory.

Fix T6903

Reviewers: zmike

Subscribers: cedric

Maniphest Tasks: T6903

Differential Revision: https://phab.enlightenment.org/D6011

6 years agotests: use a global win object in fork mode when using buffer engine
Mike Blumenkrantz [Fri, 27 Apr 2018 12:59:04 +0000 (14:59 +0200)]
tests: use a global win object in fork mode when using buffer engine

Summary:
this avoids the creation of a new win object for each test

ref T6864

Reviewers: stefan_schmidt, cedric

Reviewed By: cedric

Subscribers: cedric

Maniphest Tasks: T6864

Differential Revision: https://phab.enlightenment.org/D5965

6 years agoefreet: add method for disabling efreetd connection
Mike Blumenkrantz [Fri, 27 Apr 2018 12:35:13 +0000 (14:35 +0200)]
efreet: add method for disabling efreetd connection

Summary:
some minimal applications, such as test suites, may want to
disable this if they are not in need of any of the
functionality that is provided

@feature

Depends on D5965

Reviewers: cedric, stefan_schmidt

Reviewed By: cedric

Subscribers: stefan_schmidt, cedric

Differential Revision: https://phab.enlightenment.org/D5966

6 years agoeolian_mono: remove non existing file references for distcheck
Stefan Schmidt [Fri, 27 Apr 2018 09:22:19 +0000 (11:22 +0200)]
eolian_mono: remove non existing file references for distcheck

These three files do no longer exist in tree and are failing distcheck
when preparing the dist tarball.

6 years agoci: enable many more options in misc build
Stefan Schmidt [Thu, 26 Apr 2018 16:15:30 +0000 (18:15 +0200)]
ci: enable many more options in misc build

6 years agoci: enable gl-drm and egl in wayland build
Stefan Schmidt [Thu, 26 Apr 2018 16:14:19 +0000 (18:14 +0200)]
ci: enable gl-drm and egl in wayland build

6 years agoci: travis: create misc-disabled build
Stefan Schmidt [Thu, 26 Apr 2018 16:05:03 +0000 (18:05 +0200)]
ci: travis: create misc-disabled build

This build should cover all kind of disable configure flags we support
in efl. Making sure we still work with all of these when a commit comes
in.

6 years agoci: travis: enable pulseaudio and openssl in macos build
Stefan Schmidt [Wed, 25 Apr 2018 16:17:24 +0000 (18:17 +0200)]
ci: travis: enable pulseaudio and openssl in macos build

These had been disabled due to former problems. They should work, so
time to re-enable them.

6 years agoeina: provide full string length to vsnprintf to allow for null character
Mike Blumenkrantz [Thu, 26 Apr 2018 16:24:16 +0000 (12:24 -0400)]
eina: provide full string length to vsnprintf to allow for null character

Summary: fix T6903

Reviewers: segfaultxavi

Reviewed By: segfaultxavi

Subscribers: cedric

Maniphest Tasks: T6903

Differential Revision: https://phab.enlightenment.org/D6009

6 years agoeolian: check event vs function/part conflicts as well
Daniel Kolesa [Thu, 26 Apr 2018 14:41:32 +0000 (16:41 +0200)]
eolian: check event vs function/part conflicts as well

Still behind the env var, because there would be a lot of failures
otherwise.

6 years agoeolian: use pointer hashes in validation for performance
Daniel Kolesa [Thu, 26 Apr 2018 14:23:13 +0000 (16:23 +0200)]
eolian: use pointer hashes in validation for performance

We can use pointer hashes because the keys will have the same
addresses during validation. This should be faster, so do that.

6 years agoeolian: stricter function/part redef checks
Daniel Kolesa [Thu, 26 Apr 2018 14:14:13 +0000 (16:14 +0200)]
eolian: stricter function/part redef checks

Now we also check part redefinitions and function redefinitions
are checked against parts as well.

6 years agoeolian: add event redefinition checks
Daniel Kolesa [Thu, 26 Apr 2018 14:07:28 +0000 (16:07 +0200)]
eolian: add event redefinition checks

As EFL fails build with this enabled, hide these behind the
EOLIAN_EVENT_REDEF_WARN environment variable for the time being.

6 years agoeolian: clear chash on each iteration during validation
Daniel Kolesa [Thu, 26 Apr 2018 13:45:12 +0000 (15:45 +0200)]
eolian: clear chash on each iteration during validation

This is necessary because inheritance trees need to be validated
individually for proper name conflict checks. It also prevents
eina_hash_add related errors.

6 years agoefl_mono: Names fixes after the big rename
Lauro Moura [Wed, 25 Apr 2018 01:46:04 +0000 (22:46 -0300)]
efl_mono: Names fixes after the big rename

Summary: Depends on D5997

Reviewers: felipealmeida

Reviewed By: felipealmeida

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5998

6 years agoefl_mono: Fix support for event arguments.
Lauro Moura [Thu, 19 Apr 2018 02:07:22 +0000 (23:07 -0300)]
efl_mono: Fix support for event arguments.

Summary:
Using a type visitor scheme similar to type and marshall_type, to ease
increasing coverage for other types if needed.

Also, changed the filter functions for those selections to check for the
full name of the regular_type_def, allowing a better specificity of the
name selections. This fixes things like "Eina.Error", which was
conflicting with Efl.Image.Load.Error when we compared only the last
name. It didn't appear before as Load.Error only appears in a
Efl.Image.Load event.
Depends on D5996

Reviewers: felipealmeida

Reviewed By: felipealmeida

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5997

6 years agoefl_mono: Actually test events by passing things by value and structs
Lauro Moura [Wed, 11 Apr 2018 23:01:38 +0000 (20:01 -0300)]
efl_mono: Actually test events by passing things by value and structs

Summary:
While the test events were declared as by value, the support code passed
by reference, which was - wrongly - the only implementation supported in
event code.

Also added test with Eina_Bool and structs (passed by pointer to events,
while not having the ptr modifier explicitly.
Depends on D5995

Reviewers: felipealmeida

Reviewed By: felipealmeida

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5996

6 years agoefl_mono: Move event generators to its own header.
Lauro Moura [Tue, 17 Apr 2018 00:44:27 +0000 (21:44 -0300)]
efl_mono: Move event generators to its own header.

Summary:
Making it easier to share code between self and inherited events.

During this move, the namespace and keyword headers were merged into the
name_helpers header.

Also added the first seed of a generic namespace reducer function,
to be used by other functions in later commits.
Depends on D5994

Reviewers: felipealmeida

Reviewed By: felipealmeida

Subscribers: segfaultxavi, cedric

Differential Revision: https://phab.enlightenment.org/D5995

6 years agoefl_mono: Instantiate parts in constructor in test
Lauro Moura [Mon, 16 Apr 2018 22:21:18 +0000 (19:21 -0300)]
efl_mono: Instantiate parts in constructor in test

Summary:
Don't try to lazily create the parts when getting them, as it would
modify the owner inside its @const Efl.Part.part.
Depends on D5993

Reviewers: felipealmeida

Reviewed By: felipealmeida

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5994

6 years agoefl_mono: Reorganize utility functions.
Lauro Moura [Mon, 16 Apr 2018 19:26:13 +0000 (16:26 -0300)]
efl_mono: Reorganize utility functions.

Summary:
Blacklist functions are all in the header/namespace blacklist.

Helper functions returning strings (names) are in the name_helpers
header. They act somewhat like "mini-generators".

Helpers.hh was left with other kind of helper functions (checks, etc)
that do not return strings.
Depends on D5992

Reviewers: felipealmeida

Reviewed By: felipealmeida

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5993

6 years agoefl_mono: Use PascalCase in events
Lauro Moura [Mon, 16 Apr 2018 15:36:51 +0000 (12:36 -0300)]
efl_mono: Use PascalCase in events

Summary:
To deal with events with the same name as some methods (Del, Invalidate,
etc), the suffix Evt was added.

Thus, now we use

obj.ButtonClickedEvt += callback;

Instead of

obj.BUTTON_CLICKED += cal

The argument classes use the same scheme, being called <Evt name>_Args.
Depends on D5991

Reviewers: felipealmeida

Reviewed By: felipealmeida

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5992

6 years agoeolian_mono: Register is not a kw in C#
Lauro Moura [Fri, 13 Apr 2018 03:24:57 +0000 (00:24 -0300)]
eolian_mono: Register is not a kw in C#

Summary: It was a reminiscing bit from the c++ generator.

Reviewers: felipealmeida

Reviewed By: felipealmeida

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5991

6 years agoefl theme: Replace legacy signal names with efl ui signal names
Jaehyun Cho [Thu, 26 Apr 2018 11:24:09 +0000 (20:24 +0900)]
efl theme: Replace legacy signal names with efl ui signal names

Replace legacy signal/source names with efl ui signal/source names.
e.g. From "elm,state,content,set" "elm" to "efl,state,content,set" "efl"

6 years agoefl_ui_win: fix the wrong bracket
Taehyub Kim [Thu, 26 Apr 2018 12:00:35 +0000 (21:00 +0900)]
efl_ui_win: fix the wrong bracket

Summary: efl_ui_win: fix the wrong bracket

Reviewers: Jaehyun_Cho

Reviewed By: Jaehyun_Cho

Subscribers: cedric, stefan

Differential Revision: https://phab.enlightenment.org/D6007

6 years agoefl_ui_timepicker: Fix to inherit efl ui theme
Jaehyun Cho [Thu, 26 Apr 2018 11:21:36 +0000 (20:21 +0900)]
efl_ui_timepicker: Fix to inherit efl ui theme

6 years agoefl_ui_datepicker: Fix to inherit efl ui theme
Jaehyun Cho [Thu, 26 Apr 2018 11:21:04 +0000 (20:21 +0900)]
efl_ui_datepicker: Fix to inherit efl ui theme

6 years agoefl theme: remove the elm legacy name of efl ui theme
Taehyub Kim [Thu, 26 Apr 2018 10:27:43 +0000 (19:27 +0900)]
efl theme: remove the elm legacy name of efl ui theme

Summary: remove the elm legacy name of efl ui theme

Test Plan: run elementary_test and test efl ui widget cases

Reviewers: Jaehyun_Cho, woohyun, cedric, raster, jpeg

Reviewed By: Jaehyun_Cho

Differential Revision: https://phab.enlightenment.org/D5934

6 years agoefl_ui_win: remove references to removed functions
Mike Blumenkrantz [Wed, 25 Apr 2018 16:22:40 +0000 (12:22 -0400)]
efl_ui_win: remove references to removed functions

Reviewers: ManMower

Reviewed By: ManMower

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D6004

6 years agoefl_ui_win: simplify checking theme version when adding frame
Mike Blumenkrantz [Wed, 25 Apr 2018 16:09:25 +0000 (12:09 -0400)]
efl_ui_win: simplify checking theme version when adding frame

Summary:
the win has no theme set, so this would always return null
Depends on D5956

Reviewers: cedric

Reviewed By: cedric

Differential Revision: https://phab.enlightenment.org/D5957

6 years agoelm_theme: use themes and extensions lists for _elm_theme_data_find
Mike Blumenkrantz [Tue, 17 Apr 2018 15:46:37 +0000 (11:46 -0400)]
elm_theme: use themes and extensions lists for _elm_theme_data_find

Differential Revision: https://phab.enlightenment.org/D5956

6 years agoelm_win: remove useless function
Mike Blumenkrantz [Wed, 25 Apr 2018 16:06:10 +0000 (12:06 -0400)]
elm_win: remove useless function

Summary: Depends on D5961

Reviewers: ManMower, cedric

Reviewed By: ManMower, cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5945

6 years agoefl_ui_win: resolve race conditions with csd sizing on first show
Mike Blumenkrantz [Tue, 17 Apr 2018 18:55:35 +0000 (14:55 -0400)]
efl_ui_win: resolve race conditions with csd sizing on first show

there were some cases where frame geometry was being calculated/applied
strangely as a result of moving all of the layout calcs to pre-render.
enforcing a frame calc+resize resolves these issues

Differential Revision: https://phab.enlightenment.org/D5961

6 years agoefl_ui_win: remove show intercept function
Mike Blumenkrantz [Tue, 17 Apr 2018 18:52:19 +0000 (14:52 -0400)]
efl_ui_win: remove show intercept function

this should just be handled in the pre-render callback where the rest
of the calc for the window is done

also removes an unnecessary smart calc

Differential Revision: https://phab.enlightenment.org/D5960

6 years agoelm_win: freeze frame object edje until first pre-render
Mike Blumenkrantz [Mon, 16 Apr 2018 17:17:42 +0000 (13:17 -0400)]
elm_win: freeze frame object edje until first pre-render

this avoids a substantial number of unnecessary recalcs and halves the calltime
for elm_win_add

ref T6884

Differential Revision: https://phab.enlightenment.org/D5944

6 years agoeet: add counter for number of frees needed in each directory
Mike Blumenkrantz [Wed, 25 Apr 2018 15:59:12 +0000 (11:59 -0400)]
eet: add counter for number of frees needed in each directory

Summary:
further optimize the eet_shutdown case of closing files and avoid
endlessly looping over empty nodes
Depends on D5950

Reviewers: cedric

Reviewed By: cedric

Differential Revision: https://phab.enlightenment.org/D5951

6 years agoeet: simplify shutdown file closing
Mike Blumenkrantz [Mon, 16 Apr 2018 22:48:05 +0000 (18:48 -0400)]
eet: simplify shutdown file closing

no need to free mempool memory when the mempool is about to be freed,
and eet_clearcache duplicates existing shutdown functionality

Differential Revision: https://phab.enlightenment.org/D5950

6 years agoelm_win: set manual render and freeze frame object during deferred free
Mike Blumenkrantz [Mon, 16 Apr 2018 20:29:46 +0000 (16:29 -0400)]
elm_win: set manual render and freeze frame object during deferred free

ensure no recalcs of any kind occur on a deleted window

Differential Revision: https://phab.enlightenment.org/D5949

6 years agoels_cursor: avoid forcing a recursive canvas layout recalc when adding cursor
Mike Blumenkrantz [Mon, 16 Apr 2018 20:07:14 +0000 (16:07 -0400)]
els_cursor: avoid forcing a recursive canvas layout recalc when adding cursor

use deferred size hints to avoid unnecessary recalcs

Differential Revision: https://phab.enlightenment.org/D5948

6 years agoedje: remove recalc from content_get and text_get
Mike Blumenkrantz [Mon, 16 Apr 2018 20:05:40 +0000 (16:05 -0400)]
edje: remove recalc from content_get and text_get

neither of these functions should force a recalc under any circumstance
as they are simply returning pointers

@fix

Differential Revision: https://phab.enlightenment.org/D5947

6 years agoedje: make most edje_util.c functions obey freeze
Mike Blumenkrantz [Mon, 16 Apr 2018 18:55:12 +0000 (14:55 -0400)]
edje: make most edje_util.c functions obey freeze

a number of these functions have implicit recalcs, which is bad because
it's a pretty significant perf bottleneck, but it can't be improved
without breaking existing behavior expectations so this is probably the
best that can be done

ref T6884

Differential Revision: https://phab.enlightenment.org/D5946

6 years agoci: travis: allow for some failures in the matrix build
Stefan Schmidt [Wed, 25 Apr 2018 09:51:33 +0000 (11:51 +0200)]
ci: travis: allow for some failures in the matrix build

We want to make sure to have a stable and reliable subset of builds that
define if the build passed or not.
We also want to have builds which are more experimental. They give us a
good insight, but we are not yet ready to have them supported officially
as need-to-pass build. Namely the macOS build.

Another side effect of this change is that we reduce the critical build
matrix to 5 builds. The exact number of parallel ones we are allowed. With
fast_finish Travis will not wait for the other ones to finish before
setting the build status. This will allow us to have all builds in
parallel and not waiting for build #6 to be finished.

6 years agotests: eio: fix filename naming in two tests
Stefan Schmidt [Mon, 23 Apr 2018 13:50:51 +0000 (15:50 +0200)]
tests: eio: fix filename naming in two tests

The file names are derived from the test case names. These two seem to
have been overlooked when copying one test case to another.
I hoped this would fix my problems I have with the eio test suite
recently, but sadly it does not. Time to go on a git bisect tour.

6 years agoIntrudoce Efl Ui Date, Time interface and picker.
Woochan Lee [Wed, 25 Apr 2018 07:43:53 +0000 (16:43 +0900)]
Intrudoce Efl Ui Date, Time interface and picker.

Summary: Create datepicker, timepicker with one manager class(efl_datetime).

Test Plan: datepicker, timepicker sample.

Reviewers: herb, Jaehyun_Cho, woohyun, SanghyeonLee

Reviewed By: Jaehyun_Cho

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5871

6 years agoRevert "ecore_wl2: Add dma_buf lock/unlock ioctls"
Wonki Kim [Wed, 30 May 2018 07:23:37 +0000 (16:23 +0900)]
Revert "ecore_wl2: Add dma_buf lock/unlock ioctls"

This reverts commit e94479afc903a2edb92873c3e92d3ddb66e80f0d.

6 years agoefl_ui_focus_composition: permit anything but adapters and widgets
Marcel Hollerbach [Tue, 24 Apr 2018 22:02:55 +0000 (00:02 +0200)]
efl_ui_focus_composition: permit anything but adapters and widgets

a wildcard gfx object could bring focus to a evas object that has no
smart parent, and thus keyboard events are left at a place where they
cannot come away from.

clock was using this to give focus to the edit buttons. The edit buttons
are now wrapped into theire own adapter objects which makes them
focusable like before. However, the no keybindings or anything are taken
to react on the value.

6 years agogen****: just set focus to genlist / gengrid if no item is available
Marcel Hollerbach [Tue, 24 Apr 2018 20:41:39 +0000 (22:41 +0200)]
gen****: just set focus to genlist / gengrid if no item is available

Additional move requests are still redirected to the gen**** widget.

ref T6805

6 years agoelm_genlist: do not not right left top down in action_move
Marcel Hollerbach [Tue, 24 Apr 2018 20:18:35 +0000 (22:18 +0200)]
elm_genlist: do not not right left top down in action_move

the internal movement logic was resulting in wrong movements skipping
children items.

6 years agoefl_ui_focus_composition: add fields for manager and parent
Marcel Hollerbach [Mon, 23 Apr 2018 13:23:17 +0000 (15:23 +0200)]
efl_ui_focus_composition: add fields for manager and parent

the manager objects are build on the assertion that registered elements
are returning the manager they are registered on if
efl_ui_focus_object_manager_get is called.

6 years agoecore_wl2: reduce wayland display flushes
Derek Foreman [Fri, 6 Apr 2018 17:13:45 +0000 (12:13 -0500)]
ecore_wl2: reduce wayland display flushes

Clean up various places where we do flushes that we don't need to
because some immediately following action is going to cause a flush.
Also fix places where we flush without actually doing anything.

6 years agoecore_wl2: Share same region between input and opaque when possible
Derek Foreman [Thu, 5 Apr 2018 19:59:26 +0000 (14:59 -0500)]
ecore_wl2: Share same region between input and opaque when possible

If input and opaque region are the same (they usually are) we can use
the same region for both.

6 years agoecore_wl2: Refactor region create
Derek Foreman [Thu, 5 Apr 2018 18:01:22 +0000 (13:01 -0500)]
ecore_wl2: Refactor region create

Trivial refactor

6 years agoecore_wl2: Make ecore_wl2_window_false_commit more robust
Derek Foreman [Wed, 28 Mar 2018 15:07:46 +0000 (10:07 -0500)]
ecore_wl2: Make ecore_wl2_window_false_commit more robust

False commit when a commit is already pending is an error, but for safety
it should be a nop.

Currently it would overwrite the existing frame callback which could
cause problems on window destruction.

6 years agoee_wayland: Add some NULL checks
Derek Foreman [Mon, 26 Mar 2018 17:29:41 +0000 (12:29 -0500)]
ee_wayland: Add some NULL checks

These just protect against caller stupidity.  Double register or double
unregister shouldn't happen anyway.

6 years agoecore_wl2: Don't unmap buffers at unlock
Derek Foreman [Fri, 9 Mar 2018 22:12:44 +0000 (16:12 -0600)]
ecore_wl2: Don't unmap buffers at unlock

We should be using dmabuf sync ioctls instead of mmap/munmap every draw,
this makes that happen.  The surface code continues to do what its always
done, and map/unlock.

6 years agoecore_wl2: Add ERR for buffer lock errors
Derek Foreman [Fri, 9 Mar 2018 22:11:33 +0000 (16:11 -0600)]
ecore_wl2: Add ERR for buffer lock errors

Should never double lock or double unlock, this isn't refcounted, and
probably doesn't need to be.

6 years agoecore_wl2: Add dma_buf lock/unlock ioctls
Derek Foreman [Fri, 9 Mar 2018 20:33:13 +0000 (14:33 -0600)]
ecore_wl2: Add dma_buf lock/unlock ioctls

These should help with coherency later when we don't unmap for every
surface post.

6 years agoecore_wl2: Add lock/unlock function pointers to buffers
Derek Foreman [Fri, 9 Mar 2018 20:32:20 +0000 (14:32 -0600)]
ecore_wl2: Add lock/unlock function pointers to buffers

For backend specific lock/unlock behaviour

6 years agoecore_wl2: Add ecore_wl2_buffer_lock
Derek Foreman [Fri, 9 Mar 2018 20:31:25 +0000 (14:31 -0600)]
ecore_wl2: Add ecore_wl2_buffer_lock

Currently we're making mapping implicitly take a lock.

6 years agoefl_ui_win: Remove _elm_win_opaque_update
Derek Foreman [Mon, 12 Feb 2018 21:09:14 +0000 (15:09 -0600)]
efl_ui_win: Remove _elm_win_opaque_update

This stuff is all in the wayland ecore_evas now, so elm no longer needs
to handle it.