platform/upstream/efl.git
8 years agoefl - fix lots of little init/shutdown pairs that are wrong
Carsten Haitzler (Rasterman) [Mon, 4 Jul 2016 11:07:30 +0000 (20:07 +0900)]
efl - fix lots of little init/shutdown pairs that are wrong

i've fixed almost all the eina init/shutdown pairs to do the right
thing now... except one (ecore_shutdown) with comment inline where
eo_shutdown is not called. if this is called we are in crash land.
this needs further inspection.

8 years agoeina btlog - add -n option to not use color escapes
Carsten Haitzler (Rasterman) [Mon, 4 Jul 2016 11:04:38 +0000 (20:04 +0900)]
eina btlog - add -n option to not use color escapes

8 years agoEvasGL: Add support for Evas GL 3.1
DaeKwang Ryu [Mon, 4 Jul 2016 11:30:50 +0000 (20:30 +0900)]
EvasGL: Add support for Evas GL 3.1

Summary:
2 wrapper functions are added.
(glFramebufferParameteri, glGetFramebufferParameteriv)

update gl2ext.h(generated 20160209) in evas_gl.h

Test Plan: OpenGL ES Conformance Test

Reviewers: jpeg

Subscribers: spacegrapher, wonsik, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3782

8 years agoedje: Remove shadow variable
Jean-Philippe Andre [Mon, 4 Jul 2016 11:25:17 +0000 (20:25 +0900)]
edje: Remove shadow variable

8 years agoelm_frame: updated the eo prefix and documentation
Subhransu Mohanty [Mon, 4 Jul 2016 11:10:29 +0000 (20:10 +0900)]
elm_frame: updated the eo prefix and documentation

Summary: No idea what else can be done with this widget.

Reviewers: jpeg, cedric, woohyun

Subscribers: raster, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3930

8 years agospinner: spinner entry reactivate when focus come again.
Woochan Lee [Mon, 4 Jul 2016 10:55:53 +0000 (19:55 +0900)]
spinner: spinner entry reactivate when focus come again.

Summary:
When user edit spinner value on entry.
The user want to back on entry to edit spinner value even swiching view(focus out and focus in).

I considered a couple of exception case ( case that entry should not reactivate.)
1. User click spinner button to give focus.
2. User give focus to spinner's object using key action.

Test Plan:
Add sample in elementary_test.
Edit spinner value.
Gives focus to other window.
Back to spinner view.
See the action.

Reviewers: Hermet, woohyun, cedric, jpeg

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D4120

8 years agotest: Fix strange black rectangle in a test case
Jean-Philippe Andre [Mon, 4 Jul 2016 08:22:11 +0000 (17:22 +0900)]
test: Fix strange black rectangle in a test case

elementary_test -to "Entry on Page Scroll"

Click on "Click this and close popup"

A weird rectangle would be temporarily visible while the
popup is fading in. The rectangle was the block_events
rectangle. By default, elm_popup blocks events, and its
parent must be the whole blocked input area. This fixes the
parent in the example.

I think the block_events region should be fading in and out
as well, because it looks terrible right now.

Note: this changes the position of the popup, but makes
the whole "block_events" thing actually work (avoiding
other render artifacts).

8 years agotheme: Add fade-in/out for block regions (popup & notify)
Jean-Philippe Andre [Mon, 4 Jul 2016 09:01:40 +0000 (18:01 +0900)]
theme: Add fade-in/out for block regions (popup & notify)

Those are a bit faster than the content transitions. I believe the
popup transition time should be reduced. 500ms is way too long for
such animations.

8 years agoelm: test_glview - allow window to close
Amitesh Singh [Mon, 4 Jul 2016 08:03:25 +0000 (13:33 +0530)]
elm: test_glview - allow window to close

8 years agoevas: Fix rare issue with recursive proxy src invisible
Jean-Philippe Andre [Mon, 4 Jul 2016 07:06:04 +0000 (16:06 +0900)]
evas: Fix rare issue with recursive proxy src invisible

Scenario:
  smart {
    text
    proxy -> text, src_invisible
  }
  proxy -> smart

What we should see:
  smart {
    (blank)
    proxy -> text
  }
  proxy -> {
    (blank)
    proxy -> text
  }

What we saw:
  smart {
    (blank)
    proxy -> text
  }
  proxy -> {
    text
    proxy -> text
  }

Solution:
Check in evas render, when we're inside a proxy render, and the
proxy src_invisible flag is on (evas_object_source_visible_set(0),
that we're rendering the object itself to its proxy surface. If not,
it means we're rendering another proxy surface, ie. a parent smart
object's proxy surface.

Still loving evas render.

Fixes T4006.

@fix

8 years agoecore: Fix obvious crash with legacy timers
Jean-Philippe Andre [Mon, 4 Jul 2016 06:10:17 +0000 (15:10 +0900)]
ecore: Fix obvious crash with legacy timers

ecore_timer_del() checks a flag "inside_call" that can be
set before calling the timer cb... but it was never reset
to 0. So, all legacy timers would keep on ticking forever
and ever, until they return CANCEL.

Anyway, I find the distinction between eo_del and
ecore_timer_del very troubling. eo_del() should work
on a legacy timer. Ping @cedric. Maybe override eo_del()?

Fixes T3898

8 years agoedje - reduce memory footprint by rearranging structs and types
Carsten Haitzler (Rasterman) [Mon, 4 Jul 2016 05:59:59 +0000 (14:59 +0900)]
edje - reduce memory footprint by rearranging structs and types

this should cut some memory used by edje by using smaller types like
shorts instead of ints where we just dont need a full int range and
short will do, and re-ordering in memory data soit packs better when
accoutning for alignment

8 years agoexamples: Fix typo in filename
Jean-Philippe Andre [Mon, 4 Jul 2016 04:45:48 +0000 (13:45 +0900)]
examples: Fix typo in filename

8 years agoElementary: cnp - check on NULL data in elm_drag_start
Artem Popov [Mon, 4 Jul 2016 01:49:55 +0000 (10:49 +0900)]
Elementary: cnp - check on NULL data in elm_drag_start

Summary:
check on NULL add in wayland elm_drag_start API
@fix

Test Plan: Call elm_grag_start with data = NULL under wayland

Reviewers: zmike, NikaWhite, jpeg

Reviewed By: NikaWhite, jpeg

Subscribers: bu5hm4n, myoungwoon, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D4119

8 years agoelm/gengrid: Fix null pointer dereferences.
Daniel Juyung Seo [Sun, 3 Jul 2016 22:29:53 +0000 (07:29 +0900)]
elm/gengrid: Fix null pointer dereferences.

CID: 1357210 and 1357209.

8 years agoevas/textblock: Mark unused variable.
Daniel Juyung Seo [Sun, 3 Jul 2016 22:24:21 +0000 (07:24 +0900)]
evas/textblock: Mark unused variable.

8 years agoecore-wl2: Fix error handling for bad fd
Chris Michael [Sun, 3 Jul 2016 22:23:45 +0000 (18:23 -0400)]
ecore-wl2: Fix error handling for bad fd

If we end up failing to get the fd from ecore_main_fd_handler_fd_get,
then we should just bail out of this function and try again later.

Fixes Coverity CID1357213

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoelementary: Grab all key presses in elm_code for now
Andy Williams [Sun, 3 Jul 2016 19:48:14 +0000 (20:48 +0100)]
elementary: Grab all key presses in elm_code for now

Working around issues where new key bindings were breaking the editor

8 years agoelementary: add test for edje external combobox
Jean Guyomarc'h [Sun, 3 Jul 2016 10:00:33 +0000 (12:00 +0200)]
elementary: add test for edje external combobox

8 years agoelementary: release genlist item classes when done
Jean Guyomarc'h [Sun, 3 Jul 2016 09:31:09 +0000 (11:31 +0200)]
elementary: release genlist item classes when done

8 years agoedje_external: add combobox external
Jean Guyomarc'h [Sun, 3 Jul 2016 09:29:35 +0000 (11:29 +0200)]
edje_external: add combobox external

The combobox widget can now be used from edje externals.
It accepts the "guide" parameter, which is of type string.

8 years agoelementary: document elm_table parameters
Jean Guyomarc'h [Sat, 2 Jul 2016 18:41:28 +0000 (20:41 +0200)]
elementary: document elm_table parameters

8 years agoelementary: Update connman theme for all security
Andy Williams [Sun, 3 Jul 2016 16:25:37 +0000 (17:25 +0100)]
elementary: Update connman theme for all security

Provide the appropriate lock icon for less common security types.
@fix
also allow the icons to grow on scaled displays

8 years agoUi text: port changed_cb from entry
Daniel Hirt [Sun, 3 Jul 2016 16:57:23 +0000 (16:57 +0000)]
Ui text: port changed_cb from entry

8 years agoCanvas text: simplify efl_text_get
Daniel Hirt [Sun, 3 Jul 2016 16:40:34 +0000 (16:40 +0000)]
Canvas text: simplify efl_text_get

8 years agoUi text: move defer call to a function
Daniel Hirt [Sun, 3 Jul 2016 16:38:03 +0000 (16:38 +0000)]
Ui text: move defer call to a function

Make it easier to call/debug this.

8 years agoUi text: fix leak in selection_changed
Daniel Hirt [Thu, 30 Jun 2016 19:39:29 +0000 (19:39 +0000)]
Ui text: fix leak in selection_changed

8 years agoCanvas text: fix leak in annotation_insert
Daniel Hirt [Thu, 30 Jun 2016 19:37:02 +0000 (19:37 +0000)]
Canvas text: fix leak in annotation_insert

8 years agoCanvas text: fix leak in efl_text_set
Daniel Hirt [Thu, 30 Jun 2016 19:36:06 +0000 (19:36 +0000)]
Canvas text: fix leak in efl_text_set

8 years agoUi text: fix invalid call to image api
Daniel Hirt [Tue, 28 Jun 2016 19:01:01 +0000 (19:01 +0000)]
Ui text: fix invalid call to image api

8 years agoUi text: send signals to cursors explicitly
Daniel Hirt [Tue, 28 Jun 2016 12:24:35 +0000 (12:24 +0000)]
Ui text: send signals to cursors explicitly

The edje code used to register the cursor objects as "edje sub-objects" of the
edje entry object. This is not edje API, so will send these signals explicitly
to the objects.

8 years agoCanvas text cursor: fix bug after code port
Daniel Hirt [Sun, 3 Jul 2016 13:57:22 +0000 (13:57 +0000)]
Canvas text cursor: fix bug after code port

Fixes T4005.
Also, unified repeated code and added a test for multiple cursors.

8 years agoevas update buf - dont ref and unref outbug as an evas image - it's not
Carsten Haitzler (Rasterman) [Sun, 3 Jul 2016 06:43:04 +0000 (15:43 +0900)]
evas update buf - dont ref and unref outbug as an evas image - it's not

we need calls to ref/unref them from engines, but atm it's ok because
they dont get deleted until the flush is done and not used after that...

@fix

8 years agoelm layout - add and delete signal cbs from the same consistent obj
Carsten Haitzler (Rasterman) [Sun, 3 Jul 2016 05:04:45 +0000 (14:04 +0900)]
elm layout - add and delete signal cbs from the same consistent obj

elm layout signal handling was all over the place. using 3 different
ways of adding or deleteing signals from the object. it uses either
obj directly, eo_super(obj) or wd->resize_obj. come on. be consistent.
so using wd->resize_obj worked before and now works properly with
sgnal cbs PROPELY deleted unlike before.

@fix

8 years agoedje signal cb - dont delete already deleted signals
Carsten Haitzler (Rasterman) [Sun, 3 Jul 2016 05:01:04 +0000 (14:01 +0900)]
edje signal cb - dont delete already deleted signals

let's assume you registered a callback twice with the same ptr data
and same func and same sig/src... if you delete it once you're ok.
delete twice... and you re-delete the first one (just makging it for
delete_me). this fixes this corner case

@fix

8 years agoecore-wl2: Fix uninitialized variable
Chris Michael [Sat, 2 Jul 2016 11:17:23 +0000 (07:17 -0400)]
ecore-wl2: Fix uninitialized variable

Coverity reports that len is used un-init here, so let's give it a
value at declaration time

Fixes Coverity CID1357214

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoeolian: cleanup type serialization logic
Daniel Kolesa [Fri, 1 Jul 2016 13:34:24 +0000 (14:34 +0100)]
eolian: cleanup type serialization logic

8 years agogengrid: fix focus issue when prev/next item is disabled
Jehun Lim [Fri, 1 Jul 2016 02:52:39 +0000 (11:52 +0900)]
gengrid: fix focus issue when prev/next item is disabled

Summary:
After setting focus to disabled item, focus does not work properly.
It is necessary to consider disabled state of each item
when selecting next focused item.

Signed-off-by: Jehun Lim <jehun.lim@samsung.com>
Test Plan:
1. elementary_test -to 'gengrid disabled item focus'
2. check 'item select on focus disable'
3. move focus to disabled item
4. observe focus and compare before this patch

Reviewers: cedric

Subscribers: minkyu, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D4115

8 years agotests: Disable profile_exists as it fails on jenkins
Jean-Philippe Andre [Fri, 1 Jul 2016 07:52:24 +0000 (16:52 +0900)]
tests: Disable profile_exists as it fails on jenkins

8 years agotheme: return enum from elm_widget_style_set instead of bool
Jinyong Park [Fri, 1 Jul 2016 06:09:42 +0000 (15:09 +0900)]
theme: return enum from elm_widget_style_set instead of bool

Summary:
if trying to apply incorrect theme, widget apply default theme and return TRUE.
so there is no way to check it really apply correct theme.
To resolve this problem, _elm_theme_set return three type enum

* related history : 4ca3ef45146e05908a13d1010909abeba9693ad6
* elm_object_style_set is public api, so I didn't change it.
* typedef name [ Theme_Apply ] is temporarily, please suggest better one.

@fix

Reviewers: singh.amitesh, herb, Hermet, cedric, jpeg, raster

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D4073

8 years agoelm: fix compilation
Amitesh Singh [Fri, 1 Jul 2016 04:25:51 +0000 (09:55 +0530)]
elm: fix compilation

"
make[1]: *** No rule to make target 'efl_ui_text.h', needed by 'all-am'.  Stop.
"

8 years agoedje_pick: load just used images.
Jaehwan Kim [Fri, 1 Jul 2016 02:18:27 +0000 (11:18 +0900)]
edje_pick: load just used images.

If the edj is made by partial groups, the images that groups use
should be included.

@fix

8 years agowayland_shm: free dmabuf resources on fallback
Derek Foreman [Thu, 30 Jun 2016 18:44:51 +0000 (13:44 -0500)]
wayland_shm: free dmabuf resources on fallback

Add destructors for the buffer managers, and clean up properly when
falling back.

Fixes Coverity CID1357163 and CID1357162

8 years agoelc_hoversel: just little fix of expand/noitem behavior
Vitalii Vorobiov [Thu, 30 Jun 2016 18:12:52 +0000 (21:12 +0300)]
elc_hoversel: just little fix of expand/noitem behavior

when hoversel has no item at all and use clicks on it, then it goes into
state called "expanded".
and so then, no matter how many items user would try to add, hoversel won't work
anymore.

@fix

8 years agoexamples: update various gitignore files for new example binaries
Stefan Schmidt [Thu, 30 Jun 2016 16:14:38 +0000 (18:14 +0200)]
examples: update various gitignore files for new example binaries

These have been forgotten to be added recently.

8 years agoatspi_image: improve existing documentation
Stefan Schmidt [Thu, 30 Jun 2016 16:11:18 +0000 (18:11 +0200)]
atspi_image: improve existing documentation

Mixin and property descriptions.

8 years agoatspi_widget_action: add mixin description
Stefan Schmidt [Thu, 30 Jun 2016 16:10:29 +0000 (18:10 +0200)]
atspi_widget_action: add mixin description

8 years agoelm_scrollable: add mixin description
Stefan Schmidt [Thu, 30 Jun 2016 16:09:41 +0000 (18:09 +0200)]
elm_scrollable: add mixin description

8 years agoatspi_component: improve existing documentation
Stefan Schmidt [Thu, 30 Jun 2016 15:31:56 +0000 (17:31 +0200)]
atspi_component: improve existing documentation

Ad more descriptions, consolidate property descriptions and fix smaller
problems.

8 years agoatspi_action: improve existing documentation
Stefan Schmidt [Thu, 30 Jun 2016 15:31:10 +0000 (17:31 +0200)]
atspi_action: improve existing documentation

Fix typos, add parameter descriptions and consolidate property describtions.

8 years agoatspi_accessible: improve existing documentation
Stefan Schmidt [Thu, 30 Jun 2016 14:34:24 +0000 (16:34 +0200)]
atspi_accessible: improve existing documentation

Make sure we describe whole properties and not getters and setters only. Adding
some more docs while being there.

8 years agoeolian: add support for static and terminated arrays
Daniel Kolesa [Thu, 30 Jun 2016 13:04:03 +0000 (14:04 +0100)]
eolian: add support for static and terminated arrays

Adds two new type types, STATIC_ARRAY and TERMINATED_ARRAY. Static arrays are
only allowed as struct members right now - they translate to regular C static
arrays (allowing them elsewhere wouldn't be good, as C isn't very good at
working with the size information). Terminated arrays are basically sequences
of data terminated at the end. The base type of static arrays can be any type
that is not marked ref (explicit ref may get allowed later). The base type of
terminated arrays has the same restriction plus that it has to be either
implicitly reference type (i.e. translating to pointer in C), integer type
or a character. In case of ref types, the terminator is NULL. In case of
integer types, the terminator is a zero. In case of character types, the
terminator is also a zero (null terminator like C strings).

@feature

8 years agoecore-wl2: Fix issue where hash would not get destroyed on error
Chris Michael [Thu, 30 Jun 2016 15:14:40 +0000 (11:14 -0400)]
ecore-wl2: Fix issue where hash would not get destroyed on error

Coverity reports that we could not reach the eina_hash_free here on
erorr condition. This was due to an erroneous return NULL above the if
(hash_create) check.

Fixes Coverity CID1357171

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-drm2: Don't leak return from _output_name_get function
Chris Michael [Thu, 30 Jun 2016 15:11:22 +0000 (11:11 -0400)]
ecore-drm2: Don't leak return from _output_name_get function

Coverity reports that we end up leaking the return from
_output_name_get as the storage is never freed.

Fixes Coverity CID1357161

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-wl2: Fix issue of passing negative number to close() and read()
Chris Michael [Thu, 30 Jun 2016 15:07:09 +0000 (11:07 -0400)]
ecore-wl2: Fix issue of passing negative number to close() and read()

We should be checking the return value of ecore_main_fd_handler_fd_get
calls as they can return a negative number...which cannot be passed to
the close() or read() functions.

Fixes Coverity CID1357152 and CID1357153

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-wl2: Check the return value of ecore_main_fd_handler_fd_get
Chris Michael [Thu, 30 Jun 2016 15:05:14 +0000 (11:05 -0400)]
ecore-wl2: Check the return value of ecore_main_fd_handler_fd_get

This patch fixes an issue where ecore_main_fd_handler_fd_get could be
returning a negative number and passing that to close() which cannot
accept negative numbers.

Fixes Coverity CID1357152

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-evas-wayland: Fix issue of passing wrong values to resize
Chris Michael [Thu, 30 Jun 2016 15:03:07 +0000 (11:03 -0400)]
ecore-evas-wayland: Fix issue of passing wrong values to resize

Coverity detected that the same code was being passed to resize
regardless of canvas orientation. This patch fixes the issue by
passing the proper values to window_resize.

Fixes Coverity CID1357150

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-wl2: Fix derefence after null check
Chris Michael [Thu, 30 Jun 2016 15:01:22 +0000 (11:01 -0400)]
ecore-wl2: Fix derefence after null check

'window' here could potentially be null so we should be checking that
it is valid before trying to assign event window

Fixes Coverity CID1357148

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoelementary: Fix explicit null dereference
Chris Michael [Thu, 30 Jun 2016 14:59:28 +0000 (10:59 -0400)]
elementary: Fix explicit null dereference

This patch fixes an issue reported by coverity where 'type' variable
could be null and passing null to strcmp is not good ;)

Fixes Coverity CID1357147

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoevas-wayland-shm: Remove logically dead code
Chris Michael [Thu, 30 Jun 2016 14:56:01 +0000 (10:56 -0400)]
evas-wayland-shm: Remove logically dead code

Coverity reports that this bpp check is actuall dead code due to the
fact that bpp can never be < 0.

Fixes Coverity CID1357144

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoelput: Check return of read function call
Chris Michael [Thu, 30 Jun 2016 14:54:22 +0000 (10:54 -0400)]
elput: Check return of read function call

If we fail to read any bytes from the fd, then we should be returning
a failure here.

Fixes Coverity CID1357143

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-evas: Fix error handling issues in wayland_shm engine
Chris Michael [Thu, 30 Jun 2016 14:51:29 +0000 (10:51 -0400)]
ecore-evas: Fix error handling issues in wayland_shm engine

This patch fixes 2 Coverity issues where engine_info_set was being
called, but the return was never being checked.

Fixes Coverity CID1357141 and CID1357142

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-input: Check return fo ecore_main_fd_handler_fd_get
Chris Michael [Thu, 30 Jun 2016 14:47:56 +0000 (10:47 -0400)]
ecore-input: Check return fo ecore_main_fd_handler_fd_get

This patch fixes 2 coverity Error handling issues.
ecore_main_fd_handler_fd_get can return a negative fd on error so we
should be checking that return before trying to make use of the fd.

Fixes Coverity CID1356632 and CID1356624

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoedje: fix return type for part_text_cursor_content_get
Daniel Hirt [Thu, 30 Jun 2016 16:53:40 +0000 (16:53 +0000)]
edje: fix return type for part_text_cursor_content_get

Another spot that should've stayed "char *" and not "string". Also added the
missing 'own'.
Found due to a warning after ee36fdc2183a3de9569cac5ae74c0fb30d3d76db.

8 years agoector: add basic description to renderer mixins
Stefan Schmidt [Thu, 30 Jun 2016 11:14:50 +0000 (13:14 +0200)]
ector: add basic description to renderer mixins

8 years agoector: further document gl buffer base class
Stefan Schmidt [Thu, 30 Jun 2016 10:49:38 +0000 (12:49 +0200)]
ector: further document gl buffer base class

8 years agoefl: gfx_shape: add some more documentation and improve existing one
Stefan Schmidt [Thu, 30 Jun 2016 10:29:52 +0000 (12:29 +0200)]
efl: gfx_shape: add some more documentation and improve existing one

8 years agoefl: gfx_shape: document property itself and not the getters and setters
Stefan Schmidt [Thu, 30 Jun 2016 10:18:21 +0000 (12:18 +0200)]
efl: gfx_shape: document property itself and not the getters and setters

8 years agoedje_edit: remove useless pointer grab as it conflicts with elementary combobox
Vitalii Vorobiov [Thu, 30 Jun 2016 09:08:46 +0000 (12:08 +0300)]
edje_edit: remove useless pointer grab as it conflicts with elementary combobox

Summary:
So currently having an issue related to clip_to function from
EdjeEdit API. I am not sure about how and why combobox use pointer grabbing
(just started to use it) but when setting NULL into clip_to of edje edit it
suddenly block entire screen away from mouse. Only when point is moved out of
window and back, only then it will be able to access widgets again.

Take a look at video and maybe try example to reproduce problem

I am not sure if deleting those functions from edje_edit is actually right
decision. But it will fix the problem of course.

Test Plan: See attached video and examples in attached files

Reviewers: NikaWhite, reutskiy.v.v, Hermet, cedric, raster

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D4081

8 years agoRevert "elm: Fix wrong use of char instead of string and void out parameter"
Daniel Hirt [Thu, 30 Jun 2016 11:32:02 +0000 (11:32 +0000)]
Revert "elm: Fix wrong use of char instead of string and void out parameter"

This reverts commit 61f52a26fb5a512e77d3141e398182107b1cf752.

This fixes the reported break on T4000.

8 years agoslider: fix edje warning for vertical style
Jean-Philippe Andre [Thu, 30 Jun 2016 07:09:09 +0000 (16:09 +0900)]
slider: fix edje warning for vertical style

Follow-up of D4117

8 years agoslider: fix edje warning
Shuhrat Dehkanov [Thu, 30 Jun 2016 07:05:28 +0000 (16:05 +0900)]
slider: fix edje warning

Summary:
ERR<18878>:edje lib/edje/edje_calc.c:2767 _edje_part_recalc_single() file /usr/local/share/elementary/themes/default.edj, group elm/slider/horizontal/default has a non-fixed part 'elm.swallow.icon'. You should add 'fixed: 1 0'. But in order to optimize the edje calc, we add it automatically.
ERR<18878>:edje lib/edje/edje_calc.c:2767 _edje_part_recalc_single() file /usr/local/share/elementary/themes/default.edj, group elm/slider/horizontal/default has a non-fixed part 'elm.swallow.end'. You should add 'fixed: 1 0'. But in order to optimize the edje calc, we add it automatically.

Test Plan: Launch an app with slider, open the slider and observe the console output

Reviewers: cedric, singh.amitesh, jpeg

Subscribers: seoz, minkyu

Differential Revision: https://phab.enlightenment.org/D4117

8 years agoelm_widget: Fix mouse wheel event propagation
Jean-Philippe Andre [Thu, 30 Jun 2016 06:20:28 +0000 (15:20 +0900)]
elm_widget: Fix mouse wheel event propagation

The wrong type was propagated. Luckily it didn't crash.

Thanks @Hermet for the report.

8 years agoedje: Lower ERR to INF for message about non fixed parts
Jean-Philippe Andre [Thu, 30 Jun 2016 05:32:50 +0000 (14:32 +0900)]
edje: Lower ERR to INF for message about non fixed parts

This message really is just informational at best, since it's
not an error, and we even implemented an optimization.

It should be done at compile (edje_cc) time though.
See comments on D3801

8 years agoevas object: Move display_mode to Efl.Gfx.Size.Hint
Jean-Philippe Andre [Thu, 30 Jun 2016 05:23:42 +0000 (14:23 +0900)]
evas object: Move display_mode to Efl.Gfx.Size.Hint

8 years agowin: Remove font & image cache functions
Jean-Philippe Andre [Thu, 30 Jun 2016 02:47:38 +0000 (11:47 +0900)]
win: Remove font & image cache functions

Those actually belong to elm_config, or rather Efl.Config:
  efl_config_int_set("cache_image_cache_size", 42);
  efl_config_int_set("cache_font_cache_size", 1337);

8 years agolayout: Add NULL check in _elm_layout_text_set function
Jaeun Choi [Thu, 30 Jun 2016 05:20:44 +0000 (14:20 +0900)]
layout: Add NULL check in _elm_layout_text_set function

When the parameter 'text' is NULL in elm_layout_text_set function,
the sub object data with the same part name is removed
from the layout's sub object list and the function returns immediately.
However, if the text part doesn't exist in the list,
a new sub object data is appended to the list even though the text is NULL.
This patch adds NULL check so the function can return in such cases.

8 years agophotocam: add missing EOLIAN prefix
Amitesh Singh [Thu, 30 Jun 2016 05:11:40 +0000 (10:41 +0530)]
photocam: add missing EOLIAN prefix

8 years agoecore-wl2: Fix formatting
Chris Michael [Wed, 29 Jun 2016 17:08:26 +0000 (13:08 -0400)]
ecore-wl2: Fix formatting

NB: No functional changes

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoelementary: Fix erroneous usage of EINA_UNUSED
Chris Michael [Wed, 29 Jun 2016 15:48:53 +0000 (11:48 -0400)]
elementary: Fix erroneous usage of EINA_UNUSED

'data' parameter is actually used in this function, so remove mis-use
of EINA_UNUSED in function

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoelementary: Fix over-zealous optimization attempt
Chris Michael [Wed, 29 Jun 2016 14:52:34 +0000 (10:52 -0400)]
elementary: Fix over-zealous optimization attempt

This is a fix for some over-zealous optimization attempt. Basically,
we cannot optimize out multiple calls to ecore_wl2_window_input_get as
the window can change in the for loop.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoelementary: Fix bug in wl dnd where all subsequent mimetypes would be
Chris Michael [Wed, 29 Jun 2016 13:59:32 +0000 (09:59 -0400)]
elementary: Fix bug in wl dnd where all subsequent mimetypes would be
accepted

This fixes an issue where in wl dnd, if one mimetype was accepted by a
drop, then all subsequent types would have been accepted also..

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoelementary: Reformat elm_cnp to be readable
Chris Michael [Wed, 29 Jun 2016 13:27:17 +0000 (09:27 -0400)]
elementary: Reformat elm_cnp to be readable

NB: No functional changes, just formatting so I can read this hot mess

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoRevert "layout: Add NULL check in _elm_layout_text_set function"
Jaeun Choi [Wed, 29 Jun 2016 09:25:17 +0000 (18:25 +0900)]
Revert "layout: Add NULL check in _elm_layout_text_set function"

This reverts commit e3fd84d1de9885aa05021603a50c0287d21d2765.

NULL checking point is wrong.

8 years agolayout: Add NULL check in _elm_layout_text_set function
Jaeun Choi [Wed, 29 Jun 2016 07:15:46 +0000 (16:15 +0900)]
layout: Add NULL check in _elm_layout_text_set function

When the parameter 'text' is NULL in elm_layout_text_set function,
the sub object data with the same part name is removed
from the layout's sub object list and the function returns immediately.
However, if the text part doesn't exist in the list,
a new sub object data is appended to the list even though the text is NULL.
This patch adds NULL check so the function can return in such cases.

8 years agoecore: Add basic test case for efl version
Jean-Philippe Andre [Wed, 29 Jun 2016 05:56:26 +0000 (14:56 +0900)]
ecore: Add basic test case for efl version

Not very useful test but still good to make sure we
preserve the API.

8 years agoelm: Add app's EFL information to EFL_MAIN and ELM_MAIN
Jean-Philippe Andre [Wed, 29 Jun 2016 05:20:53 +0000 (14:20 +0900)]
elm: Add app's EFL information to EFL_MAIN and ELM_MAIN

Bindings should do the same.

8 years agoefl: Add Efl.Version struct and APIs
Jean-Philippe Andre [Wed, 29 Jun 2016 02:49:24 +0000 (11:49 +0900)]
efl: Add Efl.Version struct and APIs

The original idea behind knowing the app's version of EFL is not
a great story. It comes from the fact that some bugs exist in
earlier versions of EFL, and some things need to be fixed. But
those fixes may break behaviour for older apps. This patch is
opening the way to the slippery slope of bug compatibility.

Unfortunately this is a requirement if we want to be able to move
forward and not break apps when we fix bugs (behaviour or ABI).

I hope we will not need to implement too many (if any) workaround
such issues. For now, this will only be used as debugging info.

EFL_MAIN() and ELM_MAIN() will both set the app's EFL version
automatically at startup time. Some internal helpers can be added
later to check how the app build-time and run-time version of
EFL differ.

@feature

8 years agoloop: Add main class property to return the main loop
Jean-Philippe Andre [Wed, 29 Jun 2016 02:15:06 +0000 (11:15 +0900)]
loop: Add main class property to return the main loop

Note: this is both @class and @property. Hope that's ok for
all bindings.

This returns same as ecore_main_loop_get() (which now uses the eo
api instead).

Ping @cedric (so he can check this patch).

8 years agowin: Change type of composite interface for Efl.Config
Jean-Philippe Andre [Wed, 29 Jun 2016 02:08:51 +0000 (11:08 +0900)]
win: Change type of composite interface for Efl.Config

Use Efl.Config.Global instead, since it's what the window will
expose, and it provides access to the "profile" config options.

8 years agoevas: Add & fix legacy APIs for text filters
Jean-Philippe Andre [Wed, 29 Jun 2016 01:58:57 +0000 (10:58 +0900)]
evas: Add & fix legacy APIs for text filters

Those are marked as EINA_DEPRECATED since they are for legacy only,
and EDC is the preferred method.
Note: these are new in EFL 1.18.

I would rather have not added them as legacy APIs. This patch is
here to align to Tizen which added them by mistake. Also, remove
the "name" argument from filter_program_set as it's not required,
and also not there in Tizen's code base.

Sorry for the mess.

8 years agoevas: Add 2 EO examples for maps
Jean-Philippe Andre [Wed, 29 Jun 2016 01:24:39 +0000 (10:24 +0900)]
evas: Add 2 EO examples for maps

Those are copy & paste & eo-ified versions of the legacy
samples. I did not touch the other legacy APIs, only maps.

8 years agoevas: Implement evas map support in eo as a mixin
Jean-Philippe Andre [Tue, 28 Jun 2016 06:40:41 +0000 (15:40 +0900)]
evas: Implement evas map support in eo as a mixin

It relies a bit on evas legacy APIs and will only work on
evas objects (Efl.Canvas.Object) for now.

The main difference with Evas_Map is that there is no
separate map object, as the functions apply directly to
any canvas object.

For convenience, most functions will automatically populate
the map if there was no previous map info. While this may
be convenient, the object's size changes still need to
be tracked to update the map info.

8 years agoevas: Move evas_map APIs from common to legacy header
Jean-Philippe Andre [Tue, 28 Jun 2016 05:21:58 +0000 (14:21 +0900)]
evas: Move evas_map APIs from common to legacy header

+ define Evas_Map struct in Evas_Legacy.h (not eo)

Otherwise, just a simple cut & paste

8 years agoevas: Move map to a separate mixin
Jean-Philippe Andre [Tue, 28 Jun 2016 05:12:17 +0000 (14:12 +0900)]
evas: Move map to a separate mixin

Evas map is not translated to proper eo apis yet.
Efl.Canvas.Map will be a mixin implementing the evas map
API in a bindable API.

8 years agoecore_drm: Don't pass uninit garbage to drm ioctls
Derek Foreman [Tue, 28 Jun 2016 20:57:41 +0000 (15:57 -0500)]
ecore_drm: Don't pass uninit garbage to drm ioctls

Hasn't actually caused us any problems so far, but valgrind
whines about it, so fixing anyway.

8 years agoelementary: try to keep const type in config test.
Cedric BAIL [Tue, 28 Jun 2016 18:15:50 +0000 (11:15 -0700)]
elementary: try to keep const type in config test.

8 years agoecore: fix typo in test.
Cedric BAIL [Tue, 28 Jun 2016 18:10:51 +0000 (11:10 -0700)]
ecore: fix typo in test.

Thanks Stefan.

8 years agoecore_wl2: use ECORE_WL2_EVENT_DATA_SOURCE_END
Marcel Hollerbach [Tue, 28 Jun 2016 14:46:51 +0000 (16:46 +0200)]
ecore_wl2: use ECORE_WL2_EVENT_DATA_SOURCE_END

this event is emited each time the dragging is cancled. So the dragwin
should be deleted here.

8 years agoelementary/fileselector: Do not use deprecated APIs.
Daniel Juyung Seo [Tue, 28 Jun 2016 14:30:30 +0000 (23:30 +0900)]
elementary/fileselector: Do not use deprecated APIs.