platform/upstream/efl.git
6 years agoevas: Fix sign mismatch in evas between.c and .h definitions
Ross Vandegrift [Mon, 27 Nov 2017 06:38:16 +0000 (15:38 +0900)]
evas: Fix sign mismatch in evas between.c and .h definitions

Summary:
Parameters w and h are declared as int for evas_common_rgba_image_from_data
and evas_common_rgba_image_from_copied_data in evas_image_data.c.  This
does not match the prototypes in evas_image_private.h which declares them
unsigned.

Original report: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=748026

Reviewers: jpeg

Reviewed By: jpeg

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5540

6 years agoelementary: fixed single typos in few modules
Pawel Aksiutowicz [Mon, 27 Nov 2017 06:35:58 +0000 (15:35 +0900)]
elementary: fixed single typos in few modules

Reviewers: stanluk, lukasz.stanislawski

Reviewed By: lukasz.stanislawski

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5534

6 years agoEfl.Ui.Format: remove the dupicalte code
Amitesh Singh [Mon, 27 Nov 2017 07:55:04 +0000 (16:55 +0900)]
Efl.Ui.Format: remove the dupicalte code

struct tm check is already done at the beginning of the function.

6 years agoefl_ui_bg: add Efl.Gfx.color_get.
Sungtaek Hong [Mon, 27 Nov 2017 02:55:24 +0000 (11:55 +0900)]
efl_ui_bg: add Efl.Gfx.color_get.

Summary:
since legacy elm_bg_color_set/get is changed to efl_gfx_color_set/get,
efl_gfx_color_get also should get color of inner rect.

@fix

Test Plan:
1) run elementary test.
           2) run Bg EO API(Efl.Ui.Bg)
           3) set/unset "Bg color" check.
           4) observe printed color values.

Reviewers: jpeg, woohyun, cedric

Differential Revision: https://phab.enlightenment.org/D5429

6 years agoelm: fix legacy code
Marcel Hollerbach [Sun, 26 Nov 2017 18:29:36 +0000 (19:29 +0100)]
elm: fix legacy code

earlier elm_object_focused_object_get returned gengrid genlist toolbar
objects if a item of them is focused. This should fix that.

thank you @davemds for the report!

6 years agoecore-wl2: Fix resource leak
Chris Michael [Sun, 26 Nov 2017 17:12:53 +0000 (12:12 -0500)]
ecore-wl2: Fix resource leak

Coverity detected a resource leak here because we were not freeing the
malloc'd 'obo' variable.

Fixes Coverity CID1382907

Signed-off-by: Chris Michael <cp.michael@samsung.com>
6 years agoelm_code: set the color scheme from the default edc.
Al Poole [Sun, 26 Nov 2017 12:36:42 +0000 (12:36 +0000)]
elm_code: set the color scheme from the default edc.

This allows us to override in the application also. Work from
myself and @ajwillia.ms.

6 years agoecore ipc/con: fix nasty ... they dont mutually exclude
Carsten Haitzler (Rasterman) [Sun, 26 Nov 2017 05:01:14 +0000 (14:01 +0900)]
ecore ipc/con: fix nasty ... they dont mutually exclude

by deleting the socket they wont mutually exclude at bind which means
eny new server just nides the pror one and you can get lots of copies
of the same server. this wasnt the case before. it shouldnt have been.
i think ther was an rm path for stale sockets when connecting failed
or something. anyway... this here was causing multilpe efreetd's and
all sorts of nastiness. this is the root cause. so... fix it.

@fix

6 years agoelm ifrace scrollable - fix uninitialized values on scroll asjust
Carsten Haitzler (Rasterman) [Fri, 24 Nov 2017 14:52:08 +0000 (23:52 +0900)]
elm ifrace scrollable - fix uninitialized values on scroll asjust

dragable values were invalid (not fetched) so vx/y were junk and this
was making decisions based on that. guarantee it to be 0.

@fix

6 years agoefl ui slider - fix crashes with fmt callback data and class data
Carsten Haitzler (Rasterman) [Fri, 24 Nov 2017 14:45:14 +0000 (23:45 +0900)]
efl ui slider - fix crashes with fmt callback data and class data

the format free cb was passed the class data (sd) as the callback data
but this was freed by the time _indi_default_format_free_cb was called
causing an invalid access. fix this by looking it up from the obj
which is probably invalid at this point...

6 years agoelua: correctly wrap gettext funcs
Daniel Kolesa [Fri, 24 Nov 2017 09:53:27 +0000 (10:53 +0100)]
elua: correctly wrap gettext funcs

Now, we cannot directly register funcs defined by a different
signature than the lua standard (int (*)(lua_State *)) so we
have to correctly wrap those with proper conversions etc.

6 years agoelua - use safeptr with lightuserdata to work around luajit 64bit issue
Carsten Haitzler (Rasterman) [Thu, 23 Nov 2017 09:07:01 +0000 (18:07 +0900)]
elua - use safeptr with lightuserdata to work around luajit 64bit issue

luajit only supports 47bits on 64bit for lightuser data (bad! don't
mess with opaque pointers... really unexpected and bad and probably
should be fixed in luajit by doing things like on 32bit where
pointers are not messed with etc.)

6 years agoeina safepointer - mimic 47 bit eoid patch
Carsten Haitzler (Rasterman) [Thu, 23 Nov 2017 09:06:12 +0000 (18:06 +0900)]
eina safepointer - mimic 47 bit eoid patch

same thing as b209b9a92286209281c12e81fd9dd72dc28c7495 but for safeptr

6 years agoeo - by default on 64bit only use 47 bits because of luajit
Carsten Haitzler (Rasterman) [Thu, 23 Nov 2017 04:27:52 +0000 (13:27 +0900)]
eo - by default on 64bit only use 47 bits because of luajit

luajit wants to check the upper 17 bits of pointers and assume they
are all 0 if on a 64bit architecture. it will panic and barf if they
are used. we have been using them in our eoid's for a long time.

my take on this is that this is unportable. assuming how many bits are
or are not valid in an opaque pointer (void *) that is simply passed
to you and you cannot dereference or use (no size information even for
what amount of data it points to etc.), is just wrong. it's not
portable and it's trying to be too smart and creating such issues. my
take is that luajit needs a fix for this in the longer term. but for
now let's have a 47 bit mode and go with that. it does mean i have to
drop our generation counter to 10 bits on 64bit (from 27 bits) which
increases likelihood of eoid re-use being falsely detected as valid
(before on 64bit it was 1 in 130 million or so chance, with this 47
bit change it's 1 in 1000. but to be fair on 32bit it's 7 bits for gen
count so 1 in 127 ... so still more than 10x "safer" than on 32bit...
but still...). the relevant check in luajit is:

  (((uint64_t)(p) >> 47) ? (lj_err_msg(L, LJ_ERR_BADLU), NULL) : (p))

it ONLY does this on 64bit. on 32bit pointers are not checked for
validity at all.

as an aside, armv8.2 seemingly will bring 52bit addresses so luajit is
going to fall over flat on a newer rev of armv8. it may be x86 also
uses more bits. last i knew it was 48bits so the 47 bit check luajit
does i think is even wrong for x86-64. to be detailed i read:

amd64 arch == 48 bits (so luajit is wrong). even better In addition,
the AMD specification requires that the most significant 16 bits of
any virtual address, bits 48 through 63, must be copies of bit 47 (in
a manner akin to sign extension). so if the upper bit of 48 is set
THEN all the 16 upper bits must be 1... breaking luajit, even if it
were 47bit and this rule applied. I read the architecture allows for
up to 52bits of actual addresses so architecture-wise this is even
wrong...

So I smell a core bug here in luajit. Certainly in the number of bits
it insists must be 0 (the upper 17 bits where on amd64/x86-64 it
should be the upper 16 bits... and even then these may NOT be 0 if bit
47 (the upper bit of the lower 48 is 1).... so the whole check is
invalid... :(

at least the above is at a theoretical level. i believe that the
addresses divide the 48 bits into 2 chunks (thus 47)... but at the
PHYSICAL level with no mmu and virtual memory. arm64 has this:

https://www.kernel.org/doc/Documentation/arm64/memory.txt

note in all cases the 2nd chunk of memory has at leats some upper bits
of physical addresses beign 1 ... which makes luajit invalid tyo use
without virtual memory remapping these away from high bits.

@fix

6 years agombe: Fix invalid part handling
Jean-Philippe Andre [Fri, 24 Nov 2017 08:22:31 +0000 (17:22 +0900)]
mbe: Fix invalid part handling

Only the text parts need special handling.
Ping @CHAN :)

6 years agowin: Fix window background EO APIs
Jean-Philippe Andre [Fri, 24 Nov 2017 08:10:30 +0000 (17:10 +0900)]
win: Fix window background EO APIs

Broken in a previous commit

6 years agoelm: add custom access state for Elm.Hover
Lukasz Stanislawski [Fri, 24 Nov 2017 07:50:01 +0000 (16:50 +0900)]
elm: add custom access state for Elm.Hover

Reviewers: jpeg

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5523

6 years agolayout: Fix erroneous warnings (from earlier patch)
Jean-Philippe Andre [Fri, 24 Nov 2017 07:23:48 +0000 (16:23 +0900)]
layout: Fix erroneous warnings (from earlier patch)

Oops. Bad check :)

6 years agoelm_part: Mark as auto-unref before finalize
Jean-Philippe Andre [Fri, 24 Nov 2017 07:19:21 +0000 (16:19 +0900)]
elm_part: Mark as auto-unref before finalize

This simplifies the IMPLEMENT macro, which can now be a proper function.
Also, do not set a parent for the part object, as it's not useful and
could mess up the refcount.

6 years agoeo: Make sure auto_unref only applies after finalize
Jean-Philippe Andre [Fri, 24 Nov 2017 07:15:03 +0000 (16:15 +0900)]
eo: Make sure auto_unref only applies after finalize

It would be pretty stupid to create an object that dies before it even
gets created. Auto-unref is for parts and should probably not be used
anywhere else :)

6 years agopart bg: Fix up some unnecessary complexity
Jean-Philippe Andre [Fri, 24 Nov 2017 06:17:05 +0000 (15:17 +0900)]
part bg: Fix up some unnecessary complexity

See previous commit :)

Note: right now the background part has a small 3d indent which comes
from the legacy theme being used. This will be fixed soon.

6 years agoefl_ui_layout: create efl_ui_bg when efl_part for background is called.
Sungtaek Hong [Fri, 24 Nov 2017 05:41:31 +0000 (14:41 +0900)]
efl_ui_layout: create efl_ui_bg when efl_part for background is called.

Reviewers: woohyun, jpeg, cedric

Subscribers: felipealmeida, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5151

6 years agoEfl.Ui.Format: add string format checking for user input
Amitesh Singh [Fri, 24 Nov 2017 05:56:10 +0000 (14:56 +0900)]
Efl.Ui.Format: add string format checking for user input

6 years agowidget: Add EINA_UNUSED
Jean-Philippe Andre [Fri, 24 Nov 2017 05:36:01 +0000 (14:36 +0900)]
widget: Add EINA_UNUSED

6 years agoefl_ui_widget: secure elm_legacy_add flag
Sungtaek Hong [Fri, 24 Nov 2017 05:32:36 +0000 (14:32 +0900)]
efl_ui_widget: secure elm_legacy_add flag

Summary:
_elm_legacy_add goes back to EINA_FALSE after setting sd->legacy.
if constructor get called again after going back to EINA_FALSE,
sd->legacy should remain EINA_TRUE.
also, elm_legacy_add() should not be called non-elm_widget.

Test Plan:
Run elementary test->Efl.Ui.Text.Label.
Check legacy flag in _elm_theme_object_set() for efl_ui_win.
Check legacy flag for efl_ui_text after scrollable text is added.

Reviewers: jpeg, woohyun

Reviewed By: jpeg

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5529

6 years agoevas: remove warning from image_loaders/png/evas_image_load_png
Pawel Aksiutowicz [Thu, 23 Nov 2017 23:43:36 +0000 (15:43 -0800)]
evas: remove warning from image_loaders/png/evas_image_load_png

Reviewers: stanluk, lukasz.stanislawski, cedric

Reviewed By: cedric

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5501

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoefl_ui_bg: fix unclipped image on EFL_UI_IMAGE_SCALE_TYPE_NONE
Sungtaek Hong [Thu, 23 Nov 2017 23:43:23 +0000 (15:43 -0800)]
efl_ui_bg: fix unclipped image on EFL_UI_IMAGE_SCALE_TYPE_NONE

@fix

Summary:
efl_ui_bg has internal image and image may be larger than
         resize_obj when EFL.UI.IMAGE.SCALE_TYPE is none.
         Now resize_obj clips the internal image.

Test Plan:
1.Run elementary_test, part background(will be added soon)
2.Create a bg in a box with other widgets.
  Set bg as EFL_UI_IMAGE_SCALYE_TYPE_NONE.

Reviewers: jpeg, cedric

Reviewed By: cedric

Subscribers: woohyun

Differential Revision: https://phab.enlightenment.org/D5509

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoelementary: fix typos in elm_win_legacy
Pawel Aksiutowicz [Thu, 23 Nov 2017 23:41:41 +0000 (15:41 -0800)]
elementary: fix typos in elm_win_legacy

Reviewers: stanluk, lukasz.stanislawski, cedric

Reviewed By: cedric

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5511

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoelementary: fix typos in elm_gesture_layer
Pawel Aksiutowicz [Thu, 23 Nov 2017 23:40:58 +0000 (15:40 -0800)]
elementary: fix typos in elm_gesture_layer

Reviewers: stanluk, lukasz.stanislawski

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5512

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoelementary: fix typos in elm_atspi_bridge
Pawel Aksiutowicz [Thu, 23 Nov 2017 23:40:12 +0000 (15:40 -0800)]
elementary: fix typos in elm_atspi_bridge

Reviewers: stanluk, lukasz.stanislawski

Reviewed By: lukasz.stanislawski

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5513

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoecore_con: quiet 2 "clobbered" warnings in ecore_con
Pawel Aksiutowicz [Thu, 23 Nov 2017 23:38:48 +0000 (15:38 -0800)]
ecore_con: quiet 2 "clobbered" warnings in ecore_con

Reviewers: stanluk, lukasz.stanislawski

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5514

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoevas: fix typos for canvas
Pawel Aksiutowicz [Thu, 23 Nov 2017 23:37:56 +0000 (15:37 -0800)]
evas: fix typos for canvas

Reviewers: stanluk, lukasz.stanislawski

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5516

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoelementary: fix compilation of cxx examples
Lauro Moura [Thu, 23 Nov 2017 23:34:50 +0000 (15:34 -0800)]
elementary: fix compilation of cxx examples

Summary: Missing ecore_con library

Reviewers: felipealmeida, jpeg

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5517

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoelementary: Windows requires EAPI for the legacy .eo.h
Lauro Moura [Thu, 23 Nov 2017 23:32:08 +0000 (15:32 -0800)]
elementary: Windows requires EAPI for the legacy .eo.h

Summary:
After the reorganization of elm eos, some sources include the .eo
headers directly instead of through Elementary.h. This causes problems
on windows as the declarations won't be decorated with the dllexport
attributes.

Reviewers: cedric, felipealmeida, jpeg, vtorri

Subscribers: jenkins

Tags: #windows, #efl

Differential Revision: https://phab.enlightenment.org/D5519

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoecore_evas: use access() instead of stat()
Lauro Moura [Thu, 23 Nov 2017 23:30:58 +0000 (15:30 -0800)]
ecore_evas: use access() instead of stat()

Summary: stat() wasn't working on Windows due to some msys2/mingw quirk.

Reviewers: felipealmeida, cedric, vtorri

Subscribers: jenkins, jpeg

Tags: #windows

Differential Revision: https://phab.enlightenment.org/D5520

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoelementary: fix compilation of elm_code test on windows
Lauro Moura [Thu, 23 Nov 2017 23:30:29 +0000 (15:30 -0800)]
elementary: fix compilation of elm_code test on windows

Summary:
Added some guards to avoid redefinition of functions.

Partially fixes T5866, as there is still the question whether we should
test internal functions or not, as stated by jpeg in the comments.

Reviewers: vtorri, felipealmeida, jpeg, cedric

Reviewed By: cedric

Subscribers: jenkins, cedric

Maniphest Tasks: T5866

Differential Revision: https://phab.enlightenment.org/D5521

Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoecore: during fd cleanup we didn't remove freed fd handler from the ready fd list.
Cedric Bail [Thu, 23 Nov 2017 23:12:30 +0000 (15:12 -0800)]
ecore: during fd cleanup we didn't remove freed fd handler from the ready fd list.

6 years agoelm: Add hack to make test case work
Jean-Philippe Andre [Thu, 23 Nov 2017 06:20:00 +0000 (15:20 +0900)]
elm: Add hack to make test case work

Because of the way elm_code test case is written, by directly including
the C file, we end up with two symbols for the internal _elm_legacy_add
flag. This makes some test case fail (after applying a pending patch).

Solution found by Sungtaek Hong.

elm_code test case needs to be fixed. Don't include the C files
directly, testing static inlines should be done through a common .x or
something, but not by including the C file itself. This has led and will
lead to many issues.

6 years agoeolian: Add API's for part enumeration
Jean-Philippe Andre [Thu, 23 Nov 2017 05:49:49 +0000 (14:49 +0900)]
eolian: Add API's for part enumeration

@feature

6 years agobg: Remove trailing \n in ERR message
Jean-Philippe Andre [Thu, 23 Nov 2017 05:31:47 +0000 (14:31 +0900)]
bg: Remove trailing \n in ERR message

6 years agoefl_access: Add reading_info_type set/get APIs
Shilpa Singh [Thu, 23 Nov 2017 03:07:52 +0000 (12:07 +0900)]
efl_access: Add reading_info_type set/get APIs

Summary:
Information obtained by atspi client is name, role, description, state of an object.
reading_info_type_set/get APIs give control to application to decide on the information
that can be exposed.

Test Plan:
The reading info is added as an attribute of an accessible object, on query of
attribute, reading_info_type and corresponding character buffer should be given to ATSPI clients.

Reviewers: kimcinoo

Subscribers: cedric, govi, rajeshps, jpeg

Differential Revision: https://phab.enlightenment.org/D5524

6 years agoelm_code: don't crash on NULL scope
Andy Williams [Wed, 22 Nov 2017 19:31:19 +0000 (19:31 +0000)]
elm_code: don't crash on NULL scope

This could be triggerd with the cursor was not in the file

6 years agoefl_ui_focus_manager: remove unnecessary debugging printf.
Al Poole [Wed, 22 Nov 2017 16:11:12 +0000 (16:11 +0000)]
efl_ui_focus_manager: remove unnecessary debugging printf.

6 years agoefl_ui_focus_manager: fixup previous commit
Marcel Hollerbach [Wed, 22 Nov 2017 13:57:57 +0000 (14:57 +0100)]
efl_ui_focus_manager: fixup previous commit

i forgot to git commit --amend ... i am sorry :[

6 years agoefl_ui_focus_manager: try to error out instead of a infinit recusrion
Marcel Hollerbach [Wed, 22 Nov 2017 13:45:47 +0000 (14:45 +0100)]
efl_ui_focus_manager: try to error out instead of a infinit recusrion

6 years agoelm_gengrid: jump back to the last selected item on gengrid focus jump
Marcel Hollerbach [Wed, 22 Nov 2017 11:18:30 +0000 (12:18 +0100)]
elm_gengrid: jump back to the last selected item on gengrid focus jump

this moves the item into selection and focus that was last focused or
selected.

6 years agoefl_ui_win: use new api
Marcel Hollerbach [Wed, 22 Nov 2017 09:45:58 +0000 (10:45 +0100)]
efl_ui_win: use new api

we can just call the manager to prepare itself in the direction we have,
this means we dont need this loop here.

6 years agoefl_ui_focus_manager: move preparing of the manager to its own api
Marcel Hollerbach [Tue, 21 Nov 2017 22:41:17 +0000 (23:41 +0100)]
efl_ui_focus_manager: move preparing of the manager to its own api

with this some imlpementations can decide if they want to fallback to
some older focused item once they need to prepare themself for this
call.

6 years agoelm_gengrid: respect item config
Marcel Hollerbach [Tue, 21 Nov 2017 21:14:26 +0000 (22:14 +0100)]
elm_gengrid: respect item config

6 years agoelm_gengrid: remove on_focus_update
Marcel Hollerbach [Tue, 21 Nov 2017 21:02:40 +0000 (22:02 +0100)]
elm_gengrid: remove on_focus_update

this hook is not executed anymore, code for remembering the last focused
item will be added in the following commits.

6 years agodocgen: Correct title plugin usage
Andy Williams [Wed, 22 Nov 2017 09:59:52 +0000 (09:59 +0000)]
docgen: Correct title plugin usage

6 years agoEfl.Ui.Calendar: fix wrong define - again!
Amitesh Singh [Wed, 22 Nov 2017 08:14:07 +0000 (17:14 +0900)]
Efl.Ui.Calendar: fix wrong define - again!

6 years agoefl_ui_calendar: fix wrong define
WooHyun Jung [Wed, 22 Nov 2017 07:54:30 +0000 (16:54 +0900)]
efl_ui_calendar: fix wrong define

6 years agoelm_genlist: Add signals to communicate drag start and stop to atspi client.
Prasoon Singh [Wed, 22 Nov 2017 05:17:41 +0000 (14:17 +0900)]
elm_genlist: Add signals to communicate drag start and stop to atspi client.

Summary:
Send signal EFL_ACCESS_STATE_ANIMATED when dragging starts and stops to atspi clients and also set EFL_ACCESS_STATE_ANIMATED
when reorder mode is enabled.

Test Plan: When reorder happens atspi client should receive object:state-changed:animated signal.

Reviewers: kimcinoo, shilpasingh

Reviewed By: shilpasingh

Subscribers: cedric, govi, rajeshps, jpeg

Differential Revision: https://phab.enlightenment.org/D5515

6 years agoelm: efl access: fix the warnings.
Amitesh Singh [Wed, 22 Nov 2017 04:45:21 +0000 (13:45 +0900)]
elm: efl access: fix the warnings.

There were many warnings while building elm.

--->
./src/lib/elementary/efl_access.eo.h:433:55: warning: duplicate â€˜const’ declaration specifier
 -Wduplicate-decl-specifier]
 EOAPI void efl_access_attribute_append(Eo *obj, const const char *key, const const char *value);
<---

6 years agoefl_access: Add attribute_append, attributes_clear APIs
Shilpa Singh [Wed, 22 Nov 2017 02:54:44 +0000 (11:54 +0900)]
efl_access: Add attribute_append, attributes_clear APIs

Summary:
Add attribute append and attributes clear API, attributes of widget/widget_item helps in adding additional information
about the widget/widget item in the form of key-value pair.

Test Plan:
Query the attributes using atspi_accessible_get_attributes in atspi_client and an hash table consisting
of updates attributes should be returned.

Signed-Off By: Shilpa Singh <shilpa.singh@samsung.com>
Signed-Off By: Lukasz Wlazly <l.wlazly@partner.samsung.com>

Reviewers: kimcinoo, lukasz.stanislawski

Subscribers: cedric, govi, rajeshps, jpeg

Differential Revision: https://phab.enlightenment.org/D5510

6 years agowayland_shm: Fix partial redraws
Derek Foreman [Tue, 21 Nov 2017 22:48:30 +0000 (16:48 -0600)]
wayland_shm: Fix partial redraws

I accidentally the age tracking.

6 years agoecore_wl2: Add dmabuf allocations for vc4
Derek Foreman [Tue, 21 Nov 2017 22:29:55 +0000 (16:29 -0600)]
ecore_wl2: Add dmabuf allocations for vc4

This should theoretically be better for software rendering on rpi3 since
it allows the compositor to use our sw rendered buffers as a texture.

6 years agoelm_widget: remove focused unfocused events
Marcel Hollerbach [Mon, 20 Nov 2017 21:56:43 +0000 (22:56 +0100)]
elm_widget: remove focused unfocused events

the ones form efl.ui.focus.object are now taken.

6 years agoprefs elm_entry: change to the new event
Marcel Hollerbach [Mon, 20 Nov 2017 21:54:49 +0000 (22:54 +0100)]
prefs elm_entry: change to the new event

6 years agoefl_ui_text: remove the event emits here
Marcel Hollerbach [Mon, 20 Nov 2017 21:38:08 +0000 (22:38 +0100)]
efl_ui_text: remove the event emits here

with the previous commit we emit the legacy events for each interface
event, and thus this event is not needed.

However, in elm_entry this means that changing editable will _not_
reemit the focus event (where i am not that sure if that is correct or
not).

6 years agoelm_fileselector: forward focus changed from entry to fileselector
Marcel Hollerbach [Mon, 20 Nov 2017 21:25:46 +0000 (22:25 +0100)]
elm_fileselector: forward focus changed from entry to fileselector

the old event is then emitted from elm_widget.c

6 years agoelm_widget: move event emission to a callback
Marcel Hollerbach [Mon, 20 Nov 2017 21:22:02 +0000 (22:22 +0100)]
elm_widget: move event emission to a callback

the callback gets triggered if the new event got triggered.

6 years agoefl_ui_multibuttonentry: migrate to efl_ui_focus_object events
Marcel Hollerbach [Mon, 20 Nov 2017 20:34:33 +0000 (21:34 +0100)]
efl_ui_multibuttonentry: migrate to efl_ui_focus_object events

this merges the two two events callbacks into one

6 years agoelm_layout: call the super for calling the events and methods
Marcel Hollerbach [Mon, 20 Nov 2017 20:12:34 +0000 (21:12 +0100)]
elm_layout: call the super for calling the events and methods

this simplifies the code

6 years agoelm_multibuttonentry: on_focus_update is not needed anymore
Marcel Hollerbach [Mon, 20 Nov 2017 20:07:35 +0000 (21:07 +0100)]
elm_multibuttonentry: on_focus_update is not needed anymore

the widget itself never gets focus, the elements of the button are
getting them. MBE is basically just a box. The text input for the entry
is setted by elm_entry itself when it gets focus.

The legacy "focused" event is emitted somewhere seperatly.

6 years agoelm_spinner: lets use the new event
Marcel Hollerbach [Thu, 16 Nov 2017 22:30:30 +0000 (23:30 +0100)]
elm_spinner: lets use the new event

6 years agoelm_popup: lets use the new event
Marcel Hollerbach [Thu, 16 Nov 2017 22:24:15 +0000 (23:24 +0100)]
elm_popup: lets use the new event

6 years agoelm_hoversel: lets use the new event
Marcel Hollerbach [Thu, 16 Nov 2017 22:23:25 +0000 (23:23 +0100)]
elm_hoversel: lets use the new event

6 years agoelm_fileselector: lets use the new event
Marcel Hollerbach [Thu, 16 Nov 2017 22:13:42 +0000 (23:13 +0100)]
elm_fileselector: lets use the new event

6 years agogl_drm: Use the same gbm_device for multiple instances
Derek Foreman [Tue, 21 Nov 2017 18:26:17 +0000 (12:26 -0600)]
gl_drm: Use the same gbm_device for multiple instances

If we want to share a gl context (we do) between multiple instances of
gl_drm, we need to make sure they all use the same gbm_device.

This resolves a blocker for multi-output on the gl_drm backend.

6 years agoecore: reformat ecore_exe_posix.c
Mike Blumenkrantz [Tue, 21 Nov 2017 17:24:51 +0000 (12:24 -0500)]
ecore: reformat ecore_exe_posix.c

6 years agoecore_wl2: Fix buffer manager refcounting
Derek Foreman [Tue, 21 Nov 2017 16:09:15 +0000 (10:09 -0600)]
ecore_wl2: Fix buffer manager refcounting

commit 0cf806005e32827f174f42cba27192c47ed9c4a7 correctly fixed a
leaked buffer.  However, other code was already accounting for the
leaked reference to the buffer manager, so an extra deref happened
and broke the universe - but only on hardware that no developer
has access to for testing.

6 years agoefl_ui_focus_manager_calc: keep focus history arround
Marcel Hollerbach [Tue, 21 Nov 2017 15:15:46 +0000 (16:15 +0100)]
efl_ui_focus_manager_calc: keep focus history arround

this now keeps items arround even if a explicit other widget was
focused. This is usefull if we have a few logical items on the focus
stack and you remove them.

6 years agoefl_ui_focus_manager_calc: remove a logical item that caused a redirect
Marcel Hollerbach [Tue, 21 Nov 2017 15:15:02 +0000 (16:15 +0100)]
efl_ui_focus_manager_calc: remove a logical item that caused a redirect

so we dont try to focus a logical item.

6 years agoefl_ui_focus_layer: rely on _manager_in_chain_set
Marcel Hollerbach [Tue, 21 Nov 2017 15:14:18 +0000 (16:14 +0100)]
efl_ui_focus_layer: rely on _manager_in_chain_set

so we dont have to set the direct by hand anymore

6 years agoelm_widget: dont use the inner var for searching the top widget
Marcel Hollerbach [Tue, 21 Nov 2017 15:12:12 +0000 (16:12 +0100)]
elm_widget: dont use the inner var for searching the top widget

you think there is only one elm_parent behind the elementary property
parent? Oh no! Elm_Notify and Elm_Popup have theire own parent fields,
and those fields are not returned when calleing elm_widget_parent_get,
they are also not given to the elm_widget implementation of
elm_widget_top_get, so a call to elm_widget_top_get stalls at elm.notify
or elm.popup.

fix T6389

6 years agoedje_cc: Not reset inherit_script during group inherit
Jeeyong Um [Tue, 21 Nov 2017 14:51:21 +0000 (23:51 +0900)]
edje_cc: Not reset inherit_script during group inherit

If group inherits after setting "inherit_script: 1;", inherit_script
is overwritten by the value of parent group. However, inherit_script
indicates user wants to inherit script in this context, it should not
be initialized as false.

6 years agoedje_cc: Fix successive script inheritance
Jeeyong Um [Tue, 21 Nov 2017 14:48:28 +0000 (23:48 +0900)]
edje_cc: Fix successive script inheritance

6 years agoeina: warning removed from eina_debug_timer
Pawel Aksiutowicz [Tue, 21 Nov 2017 02:17:51 +0000 (11:17 +0900)]
eina: warning removed from eina_debug_timer

Reviewers: stanluk, lukasz.stanislawski

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5500

6 years agoeet: removed warning from eet_image
Pawel Aksiutowicz [Tue, 21 Nov 2017 02:12:56 +0000 (11:12 +0900)]
eet: removed warning from eet_image

Reviewers: stanluk, lukasz.stanislawski

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5505

6 years agoelementary: removed warning from efl_ui_win
Pawel Aksiutowicz [Tue, 21 Nov 2017 02:12:45 +0000 (11:12 +0900)]
elementary: removed warning from efl_ui_win

Reviewers: stanluk, lukasz.stanislawski

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5503

6 years agoeolian: Removed warning of unused var from eo_parser
Pawel Aksiutowicz [Tue, 21 Nov 2017 02:12:27 +0000 (11:12 +0900)]
eolian: Removed warning of unused var from eo_parser

Reviewers: stanluk, lukasz.stanislawski

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5502

6 years agoeina: Fix matrix header guard name
Bryce Harrington [Tue, 21 Nov 2017 02:00:53 +0000 (11:00 +0900)]
eina: Fix matrix header guard name

Summary:
eina_matrix.h includes matrix2, matrix3, and matrix4, so the header
guards shouldn't specify MATRIX3.

Reviewers: cedric

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D5508

6 years agodrm ee: Allow more than 4 buffers
Derek Foreman [Mon, 20 Nov 2017 22:53:31 +0000 (16:53 -0600)]
drm ee: Allow more than 4 buffers

Multi-head is hitting corner cases where there are lots of locked buffers
and it looks like right now 5 is the magic number that makes the problem
go away.

Make it possible to set 5 or more (via env var) for testing, make a macro
for MAX_BUFFERS instead of just a number.

6 years agowin: make _elm_win_opaque_update check that it has a wayland window
Derek Foreman [Mon, 20 Nov 2017 18:51:35 +0000 (12:51 -0600)]
win: make _elm_win_opaque_update check that it has a wayland window

This function will destroy non-wayland engine data, so it should
make sure it's actually operating on a wayland window.

Originally the sd->wl.win test was sufficient, but now wl.win is
present on non-wl windows to facilitate cut and paste, so we need
to check more thoroughly.

6 years agoecore-wl2: Destroy buffer before we exit the buffer_test function
Chris Michael [Mon, 20 Nov 2017 16:27:15 +0000 (11:27 -0500)]
ecore-wl2: Destroy buffer before we exit the buffer_test function

Small patch to destroy our test buffer before we exit the
_ecore_wl2_buffer_test function so that we do not leak here.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
6 years agoelementary: removed warning from elm_widget
Pawel Aksiutowicz [Mon, 20 Nov 2017 15:13:46 +0000 (16:13 +0100)]
elementary: removed warning from elm_widget

Reviewers: stanluk, lukasz.stanislawski, bu5hm4n

Reviewed By: bu5hm4n

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D5506

6 years agoelput: Remove unreachable code
Chris Michael [Mon, 20 Nov 2017 12:51:50 +0000 (07:51 -0500)]
elput: Remove unreachable code

As we are returning the evdev above, this block is unreachable so
remove it.

Fixes Coverity CID1382857

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
6 years agoeolian: include all test eo files in build (for distcheck)
Daniel Kolesa [Mon, 20 Nov 2017 12:03:32 +0000 (13:03 +0100)]
eolian: include all test eo files in build (for distcheck)

6 years agocxx: Fix slider example (bad downcast)
Jean-Philippe Andre [Mon, 20 Nov 2017 11:39:54 +0000 (20:39 +0900)]
cxx: Fix slider example (bad downcast)

6 years agoslider: Implement Efl.Content
Jean-Philippe Andre [Mon, 20 Nov 2017 11:39:44 +0000 (20:39 +0900)]
slider: Implement Efl.Content

6 years agoradio: Implement Efl.Content
Jean-Philippe Andre [Mon, 20 Nov 2017 11:39:34 +0000 (20:39 +0900)]
radio: Implement Efl.Content

6 years agoelm: Fix minor warning (return void)
Jean-Philippe Andre [Mon, 20 Nov 2017 11:34:17 +0000 (20:34 +0900)]
elm: Fix minor warning (return void)

6 years agocxx: Disable spinner example (broken)
Jean-Philippe Andre [Mon, 20 Nov 2017 11:31:07 +0000 (20:31 +0900)]
cxx: Disable spinner example (broken)

A new API for spinner is coming soon. This example code is now broken
since my change on Efl.Content vs. Efl.Container.

6 years agobutton: Implement Efl.Content set/get/unset
Jean-Philippe Andre [Mon, 20 Nov 2017 11:22:48 +0000 (20:22 +0900)]
button: Implement Efl.Content set/get/unset

This fixes make check (cxx compilation).

6 years agotests: Add EO check box in elementary_test
Jean-Philippe Andre [Mon, 20 Nov 2017 11:07:33 +0000 (20:07 +0900)]
tests: Add EO check box in elementary_test

If checked, it will list only the EO test cases.
Also add some stupid icons, and try to make the test case names more
uniform.

Test cases marked as EO are not necessarily 100% EO code, but it would
be good to make them so.

6 years agoimage: Return file+key even for edje objects
Jean-Philippe Andre [Mon, 20 Nov 2017 11:06:01 +0000 (20:06 +0900)]
image: Return file+key even for edje objects

6 years agoslider: Remove bad defines from EO file
Jean-Philippe Andre [Mon, 20 Nov 2017 10:22:15 +0000 (19:22 +0900)]
slider: Remove bad defines from EO file

6 years agowin: Make wm_available_rotations simpler for EO
Jean-Philippe Andre [Mon, 20 Nov 2017 08:28:20 +0000 (17:28 +0900)]
win: Make wm_available_rotations simpler for EO

The problem with the API is that it uses an array of ints, which is not
well defined in EO. We could set an array of pointers to int, but that
would be super awkward to use.

I believe the original API has been slightly over-engineered as it was
passing an array of available rotations when in reality only 4 rotations
could be supported (0, 90, 180, 270). It seems to me that the day
arbitrary rotation needs to be allowed, another API would be required
(maybe with a range, or a single bool flag to allow anything).

I have not modified the internal code, which still uses an array (as
ecore evas uses that).

Mote: ec464939d9b8e4daa removed wm_rotation_supported_get(), as well as
the preferred rotation from EO, but they could easily be added back if
needed.

Note 2: I couldn't test as desktop E doesn't support WM rotations.

Ref T5322